diff --git a/cleanrl/cleanrl/ppo_procgen.py b/cleanrl/cleanrl/ppo_procgen.py
new file mode 100644
index 0000000000000000000000000000000000000000..0a13317da85801a653b02427d27bda705f642fd3
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_procgen.py
@@ -0,0 +1,346 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/ppo/#ppo_procgenpy
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from procgen import ProcgenEnv
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "starpilot"
+    """the id of the environment"""
+    total_timesteps: int = int(25e6)
+    """total timesteps of the experiments"""
+    learning_rate: float = 5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 64
+    """the number of parallel game environments"""
+    num_steps: int = 256
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = False
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.999
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 8
+    """the number of mini-batches"""
+    update_epochs: int = 3
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.2
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.01
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+# taken from https://github.com/AIcrowd/neurips2020-procgen-starter-kit/blob/142d09586d2272a17f44481a115c4bd817cf6a94/models/impala_cnn_torch.py
+class ResidualBlock(nn.Module):
+    def __init__(self, channels):
+        super().__init__()
+        self.conv0 = nn.Conv2d(in_channels=channels, out_channels=channels, kernel_size=3, padding=1)
+        self.conv1 = nn.Conv2d(in_channels=channels, out_channels=channels, kernel_size=3, padding=1)
+
+    def forward(self, x):
+        inputs = x
+        x = nn.functional.relu(x)
+        x = self.conv0(x)
+        x = nn.functional.relu(x)
+        x = self.conv1(x)
+        return x + inputs
+
+
+class ConvSequence(nn.Module):
+    def __init__(self, input_shape, out_channels):
+        super().__init__()
+        self._input_shape = input_shape
+        self._out_channels = out_channels
+        self.conv = nn.Conv2d(in_channels=self._input_shape[0], out_channels=self._out_channels, kernel_size=3, padding=1)
+        self.res_block0 = ResidualBlock(self._out_channels)
+        self.res_block1 = ResidualBlock(self._out_channels)
+
+    def forward(self, x):
+        x = self.conv(x)
+        x = nn.functional.max_pool2d(x, kernel_size=3, stride=2, padding=1)
+        x = self.res_block0(x)
+        x = self.res_block1(x)
+        assert x.shape[1:] == self.get_output_shape()
+        return x
+
+    def get_output_shape(self):
+        _c, h, w = self._input_shape
+        return (self._out_channels, (h + 1) // 2, (w + 1) // 2)
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        h, w, c = envs.single_observation_space.shape
+        shape = (c, h, w)
+        conv_seqs = []
+        for out_channels in [16, 32, 32]:
+            conv_seq = ConvSequence(shape, out_channels)
+            shape = conv_seq.get_output_shape()
+            conv_seqs.append(conv_seq)
+        conv_seqs += [
+            nn.Flatten(),
+            nn.ReLU(),
+            nn.Linear(in_features=shape[0] * shape[1] * shape[2], out_features=256),
+            nn.ReLU(),
+        ]
+        self.network = nn.Sequential(*conv_seqs)
+        self.actor = layer_init(nn.Linear(256, envs.single_action_space.n), std=0.01)
+        self.critic = layer_init(nn.Linear(256, 1), std=1)
+
+    def get_value(self, x):
+        return self.critic(self.network(x.permute((0, 3, 1, 2)) / 255.0))  # "bhwc" -> "bchw"
+
+    def get_action_and_value(self, x, action=None):
+        hidden = self.network(x.permute((0, 3, 1, 2)) / 255.0)  # "bhwc" -> "bchw"
+        logits = self.actor(hidden)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(hidden)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = ProcgenEnv(num_envs=args.num_envs, env_name=args.env_id, num_levels=0, start_level=0, distribution_mode="easy")
+    envs = gym.wrappers.TransformObservation(envs, lambda obs: obs["rgb"])
+    envs.single_action_space = envs.action_space
+    envs.single_observation_space = envs.observation_space["rgb"]
+    envs.is_vector_env = True
+    envs = gym.wrappers.RecordEpisodeStatistics(envs)
+    if args.capture_video:
+        envs = gym.wrappers.RecordVideo(envs, f"videos/{run_name}")
+    envs = gym.wrappers.NormalizeReward(envs, gamma=args.gamma)
+    envs = gym.wrappers.TransformReward(envs, lambda reward: np.clip(reward, -10, 10))
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs = torch.Tensor(envs.reset()).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, next_done, info = envs.step(action.cpu().numpy())
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            for item in info:
+                if "episode" in item.keys():
+                    print(f"global_step={global_step}, episodic_return={item['episode']['r']}")
+                    writer.add_scalar("charts/episodic_return", item["episode"]["r"], global_step)
+                    writer.add_scalar("charts/episodic_length", item["episode"]["l"], global_step)
+                    break
+
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped)
+                    v_loss = 0.5 * v_loss_max.mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        print("SPS:", int(global_step / (time.time() - start_time)))
+        writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/ppo_rnd_envpool.py b/cleanrl/cleanrl/ppo_rnd_envpool.py
new file mode 100644
index 0000000000000000000000000000000000000000..0c1758274e6b7c47c94dd54818ddf6798506d6fb
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_rnd_envpool.py
@@ -0,0 +1,539 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/ppo-rnd/#ppo_rnd_envpoolpy
+import os
+import random
+import time
+from collections import deque
+from dataclasses import dataclass
+
+import envpool
+import gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from gym.wrappers.normalize import RunningMeanStd
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "MontezumaRevenge-v5"
+    """the id of the environment"""
+    total_timesteps: int = 2000000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 1e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 128
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.999
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.1
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.001
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+
+    # RND arguments
+    update_proportion: float = 0.25
+    """proportion of exp used for predictor update"""
+    int_coef: float = 1.0
+    """coefficient of extrinsic reward"""
+    ext_coef: float = 2.0
+    """coefficient of intrinsic reward"""
+    int_gamma: float = 0.99
+    """Intrinsic reward discount rate"""
+    num_iterations_obs_norm_init: int = 50
+    """number of iterations to initialize the observations normalization parameters"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+class RecordEpisodeStatistics(gym.Wrapper):
+    def __init__(self, env, deque_size=100):
+        super().__init__(env)
+        self.num_envs = getattr(env, "num_envs", 1)
+        self.episode_returns = None
+        self.episode_lengths = None
+
+    def reset(self, **kwargs):
+        observations = super().reset(**kwargs)
+        self.episode_returns = np.zeros(self.num_envs, dtype=np.float32)
+        self.episode_lengths = np.zeros(self.num_envs, dtype=np.int32)
+        self.lives = np.zeros(self.num_envs, dtype=np.int32)
+        self.returned_episode_returns = np.zeros(self.num_envs, dtype=np.float32)
+        self.returned_episode_lengths = np.zeros(self.num_envs, dtype=np.int32)
+        return observations
+
+    def step(self, action):
+        observations, rewards, dones, infos = super().step(action)
+        self.episode_returns += infos["reward"]
+        self.episode_lengths += 1
+        self.returned_episode_returns[:] = self.episode_returns
+        self.returned_episode_lengths[:] = self.episode_lengths
+        self.episode_returns *= 1 - infos["terminated"]
+        self.episode_lengths *= 1 - infos["terminated"]
+        infos["r"] = self.returned_episode_returns
+        infos["l"] = self.returned_episode_lengths
+        return (
+            observations,
+            rewards,
+            dones,
+            infos,
+        )
+
+
+# ALGO LOGIC: initialize agent here:
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        self.network = nn.Sequential(
+            layer_init(nn.Conv2d(4, 32, 8, stride=4)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(32, 64, 4, stride=2)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 64, 3, stride=1)),
+            nn.ReLU(),
+            nn.Flatten(),
+            layer_init(nn.Linear(64 * 7 * 7, 256)),
+            nn.ReLU(),
+            layer_init(nn.Linear(256, 448)),
+            nn.ReLU(),
+        )
+        self.extra_layer = nn.Sequential(layer_init(nn.Linear(448, 448), std=0.1), nn.ReLU())
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(448, 448), std=0.01),
+            nn.ReLU(),
+            layer_init(nn.Linear(448, envs.single_action_space.n), std=0.01),
+        )
+        self.critic_ext = layer_init(nn.Linear(448, 1), std=0.01)
+        self.critic_int = layer_init(nn.Linear(448, 1), std=0.01)
+
+    def get_action_and_value(self, x, action=None):
+        hidden = self.network(x / 255.0)
+        logits = self.actor(hidden)
+        probs = Categorical(logits=logits)
+        features = self.extra_layer(hidden)
+        if action is None:
+            action = probs.sample()
+        return (
+            action,
+            probs.log_prob(action),
+            probs.entropy(),
+            self.critic_ext(features + hidden),
+            self.critic_int(features + hidden),
+        )
+
+    def get_value(self, x):
+        hidden = self.network(x / 255.0)
+        features = self.extra_layer(hidden)
+        return self.critic_ext(features + hidden), self.critic_int(features + hidden)
+
+
+class RNDModel(nn.Module):
+    def __init__(self, input_size, output_size):
+        super().__init__()
+
+        self.input_size = input_size
+        self.output_size = output_size
+
+        feature_output = 7 * 7 * 64
+
+        # Prediction network
+        self.predictor = nn.Sequential(
+            layer_init(nn.Conv2d(in_channels=1, out_channels=32, kernel_size=8, stride=4)),
+            nn.LeakyReLU(),
+            layer_init(nn.Conv2d(in_channels=32, out_channels=64, kernel_size=4, stride=2)),
+            nn.LeakyReLU(),
+            layer_init(nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1)),
+            nn.LeakyReLU(),
+            nn.Flatten(),
+            layer_init(nn.Linear(feature_output, 512)),
+            nn.ReLU(),
+            layer_init(nn.Linear(512, 512)),
+            nn.ReLU(),
+            layer_init(nn.Linear(512, 512)),
+        )
+
+        # Target network
+        self.target = nn.Sequential(
+            layer_init(nn.Conv2d(in_channels=1, out_channels=32, kernel_size=8, stride=4)),
+            nn.LeakyReLU(),
+            layer_init(nn.Conv2d(in_channels=32, out_channels=64, kernel_size=4, stride=2)),
+            nn.LeakyReLU(),
+            layer_init(nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1)),
+            nn.LeakyReLU(),
+            nn.Flatten(),
+            layer_init(nn.Linear(feature_output, 512)),
+        )
+
+        # target network is not trainable
+        for param in self.target.parameters():
+            param.requires_grad = False
+
+    def forward(self, next_obs):
+        target_feature = self.target(next_obs)
+        predict_feature = self.predictor(next_obs)
+
+        return predict_feature, target_feature
+
+
+class RewardForwardFilter:
+    def __init__(self, gamma):
+        self.rewems = None
+        self.gamma = gamma
+
+    def update(self, rews):
+        if self.rewems is None:
+            self.rewems = rews
+        else:
+            self.rewems = self.rewems * self.gamma + rews
+        return self.rewems
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = envpool.make(
+        args.env_id,
+        env_type="gym",
+        num_envs=args.num_envs,
+        episodic_life=True,
+        reward_clip=True,
+        seed=args.seed,
+        repeat_action_probability=0.25,
+    )
+    envs.num_envs = args.num_envs
+    envs.single_action_space = envs.action_space
+    envs.single_observation_space = envs.observation_space
+    envs = RecordEpisodeStatistics(envs)
+    assert isinstance(envs.action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    rnd_model = RNDModel(4, envs.single_action_space.n).to(device)
+    combined_parameters = list(agent.parameters()) + list(rnd_model.predictor.parameters())
+    optimizer = optim.Adam(
+        combined_parameters,
+        lr=args.learning_rate,
+        eps=1e-5,
+    )
+
+    reward_rms = RunningMeanStd()
+    obs_rms = RunningMeanStd(shape=(1, 1, 84, 84))
+    discounted_reward = RewardForwardFilter(args.int_gamma)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    curiosity_rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    ext_values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    int_values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    avg_returns = deque(maxlen=20)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs = torch.Tensor(envs.reset()).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    num_updates = args.total_timesteps // args.batch_size
+
+    print("Start to initialize observation normalization parameter.....")
+    next_ob = []
+    for step in range(args.num_steps * args.num_iterations_obs_norm_init):
+        acs = np.random.randint(0, envs.single_action_space.n, size=(args.num_envs,))
+        s, r, d, _ = envs.step(acs)
+        next_ob += s[:, 3, :, :].reshape([-1, 1, 84, 84]).tolist()
+
+        if len(next_ob) % (args.num_steps * args.num_envs) == 0:
+            next_ob = np.stack(next_ob)
+            obs_rms.update(next_ob)
+            next_ob = []
+    print("End to initialize...")
+
+    for update in range(1, num_updates + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (update - 1.0) / num_updates
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += 1 * args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                value_ext, value_int = agent.get_value(obs[step])
+                ext_values[step], int_values[step] = (
+                    value_ext.flatten(),
+                    value_int.flatten(),
+                )
+                action, logprob, _, _, _ = agent.get_action_and_value(obs[step])
+
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, done, info = envs.step(action.cpu().numpy())
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(done).to(device)
+            rnd_next_obs = (
+                (
+                    (next_obs[:, 3, :, :].reshape(args.num_envs, 1, 84, 84) - torch.from_numpy(obs_rms.mean).to(device))
+                    / torch.sqrt(torch.from_numpy(obs_rms.var).to(device))
+                ).clip(-5, 5)
+            ).float()
+            target_next_feature = rnd_model.target(rnd_next_obs)
+            predict_next_feature = rnd_model.predictor(rnd_next_obs)
+            curiosity_rewards[step] = ((target_next_feature - predict_next_feature).pow(2).sum(1) / 2).data
+            for idx, d in enumerate(done):
+                if d and info["lives"][idx] == 0:
+                    avg_returns.append(info["r"][idx])
+                    epi_ret = np.average(avg_returns)
+                    print(
+                        f"global_step={global_step}, episodic_return={info['r'][idx]}, curiosity_reward={np.mean(curiosity_rewards[step].cpu().numpy())}"
+                    )
+                    writer.add_scalar("charts/avg_episodic_return", epi_ret, global_step)
+                    writer.add_scalar("charts/episodic_return", info["r"][idx], global_step)
+                    writer.add_scalar(
+                        "charts/episode_curiosity_reward",
+                        curiosity_rewards[step][idx],
+                        global_step,
+                    )
+                    writer.add_scalar("charts/episodic_length", info["l"][idx], global_step)
+
+        curiosity_reward_per_env = np.array(
+            [discounted_reward.update(reward_per_step) for reward_per_step in curiosity_rewards.cpu().data.numpy().T]
+        )
+        mean, std, count = (
+            np.mean(curiosity_reward_per_env),
+            np.std(curiosity_reward_per_env),
+            len(curiosity_reward_per_env),
+        )
+        reward_rms.update_from_moments(mean, std**2, count)
+
+        curiosity_rewards /= np.sqrt(reward_rms.var)
+
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value_ext, next_value_int = agent.get_value(next_obs)
+            next_value_ext, next_value_int = next_value_ext.reshape(1, -1), next_value_int.reshape(1, -1)
+            ext_advantages = torch.zeros_like(rewards, device=device)
+            int_advantages = torch.zeros_like(curiosity_rewards, device=device)
+            ext_lastgaelam = 0
+            int_lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    ext_nextnonterminal = 1.0 - next_done
+                    int_nextnonterminal = 1.0
+                    ext_nextvalues = next_value_ext
+                    int_nextvalues = next_value_int
+                else:
+                    ext_nextnonterminal = 1.0 - dones[t + 1]
+                    int_nextnonterminal = 1.0
+                    ext_nextvalues = ext_values[t + 1]
+                    int_nextvalues = int_values[t + 1]
+                ext_delta = rewards[t] + args.gamma * ext_nextvalues * ext_nextnonterminal - ext_values[t]
+                int_delta = curiosity_rewards[t] + args.int_gamma * int_nextvalues * int_nextnonterminal - int_values[t]
+                ext_advantages[t] = ext_lastgaelam = (
+                    ext_delta + args.gamma * args.gae_lambda * ext_nextnonterminal * ext_lastgaelam
+                )
+                int_advantages[t] = int_lastgaelam = (
+                    int_delta + args.int_gamma * args.gae_lambda * int_nextnonterminal * int_lastgaelam
+                )
+            ext_returns = ext_advantages + ext_values
+            int_returns = int_advantages + int_values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape(-1)
+        b_ext_advantages = ext_advantages.reshape(-1)
+        b_int_advantages = int_advantages.reshape(-1)
+        b_ext_returns = ext_returns.reshape(-1)
+        b_int_returns = int_returns.reshape(-1)
+        b_ext_values = ext_values.reshape(-1)
+
+        b_advantages = b_int_advantages * args.int_coef + b_ext_advantages * args.ext_coef
+
+        obs_rms.update(b_obs[:, 3, :, :].reshape(-1, 1, 84, 84).cpu().numpy())
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+
+        rnd_next_obs = (
+            (
+                (b_obs[:, 3, :, :].reshape(-1, 1, 84, 84) - torch.from_numpy(obs_rms.mean).to(device))
+                / torch.sqrt(torch.from_numpy(obs_rms.var).to(device))
+            ).clip(-5, 5)
+        ).float()
+
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                predict_next_state_feature, target_next_state_feature = rnd_model(rnd_next_obs[mb_inds])
+                forward_loss = F.mse_loss(
+                    predict_next_state_feature, target_next_state_feature.detach(), reduction="none"
+                ).mean(-1)
+
+                mask = torch.rand(len(forward_loss), device=device)
+                mask = (mask < args.update_proportion).type(torch.FloatTensor).to(device)
+                forward_loss = (forward_loss * mask).sum() / torch.max(
+                    mask.sum(), torch.tensor([1], device=device, dtype=torch.float32)
+                )
+                _, newlogprob, entropy, new_ext_values, new_int_values = agent.get_action_and_value(
+                    b_obs[mb_inds], b_actions.long()[mb_inds]
+                )
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                new_ext_values, new_int_values = new_ext_values.view(-1), new_int_values.view(-1)
+                if args.clip_vloss:
+                    ext_v_loss_unclipped = (new_ext_values - b_ext_returns[mb_inds]) ** 2
+                    ext_v_clipped = b_ext_values[mb_inds] + torch.clamp(
+                        new_ext_values - b_ext_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    ext_v_loss_clipped = (ext_v_clipped - b_ext_returns[mb_inds]) ** 2
+                    ext_v_loss_max = torch.max(ext_v_loss_unclipped, ext_v_loss_clipped)
+                    ext_v_loss = 0.5 * ext_v_loss_max.mean()
+                else:
+                    ext_v_loss = 0.5 * ((new_ext_values - b_ext_returns[mb_inds]) ** 2).mean()
+
+                int_v_loss = 0.5 * ((new_int_values - b_int_returns[mb_inds]) ** 2).mean()
+                v_loss = ext_v_loss + int_v_loss
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef + forward_loss
+
+                optimizer.zero_grad()
+                loss.backward()
+                if args.max_grad_norm:
+                    nn.utils.clip_grad_norm_(
+                        combined_parameters,
+                        args.max_grad_norm,
+                    )
+                optimizer.step()
+
+            if args.target_kl is not None:
+                if approx_kl > args.target_kl:
+                    break
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/fwd_loss", forward_loss.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        print("SPS:", int(global_step / (time.time() - start_time)))
+        writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/ppo_rubikscube_generalization.py b/cleanrl/cleanrl/ppo_rubikscube_generalization.py
new file mode 100644
index 0000000000000000000000000000000000000000..c82a368b88ca38644e83315d7b69e083dfe24aa6
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_rubikscube_generalization.py
@@ -0,0 +1,561 @@
+# PPO with small MLP for RAGEN Rubik's Cube 2x2 using the existing env (no env edits)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+import re
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.rubikscube.env import RubiksCube2x2Env
+from ragen.env.rubikscube.config import RubiksCube2x2Config
+
+
+class RubiksCubeWrapper(gym.Env):
+    """
+    Adapter to use ragen RubiksCube2x2Env with Gymnasium vector API.
+    - Converts text observation to one-hot vector of 24 stickers x 6 colors.
+    - Maps agent actions [0..11] to env actions [1..12].
+    - Exposes proper observation_space and action_space.
+    """
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: RubiksCube2x2Env):
+        super().__init__()
+        self._env = env
+        # 24 stickers, 6 colors -> one-hot size 144
+        self._colors = ['W', 'O', 'G', 'R', 'B', 'Y']
+        self._color_to_idx = {c: i for i, c in enumerate(self._colors)}
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(24 * len(self._colors),), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(12)
+        # precompile regex to extract 4 letters within [X, X]\n [X, X]
+        # Lines look like: "Up (U):    [W, W]\n       [W, W]"
+        self._face_pat = re.compile(r"\[(?:\s*([A-Z])\s*,\s*([A-Z])\s*\])\n\s*\[(?:\s*([A-Z])\s*,\s*([A-Z])\s*\])")
+
+    def _encode_obs(self, text_obs: str) -> np.ndarray:
+        # Extract the six face blocks in the order U, L, F, R, B, D as rendered
+        faces_order = ["Up (U):", "Left (L):", "Front (F):", "Right (R):", "Back (B):", "Down (D):"]
+        onehots: List[int] = []
+        # Build a map from header to its block text
+        lines = text_obs.splitlines()
+        # Collect blocks starting after header line and including the next line for second row
+        i = 0
+        blocks: List[str] = []
+        while i < len(lines):
+            line = lines[i]
+            for header in faces_order:
+                if line.startswith(header):
+                    # Join current line's bracketed pair and the next line (which contains the second pair)
+                    # Remove the "Header:" prefix to keep only the bracket portions
+                    content = line[len(header):].strip()
+                    next_line = lines[i + 1] if i + 1 < len(lines) else ""
+                    block = f"{content}\n{next_line}"
+                    blocks.append(block)
+                    break
+            i += 1
+        # Fallback: if regex fails, return zeros
+        if len(blocks) != 6:
+            return np.zeros(24 * len(self._colors), dtype=np.float32)
+        stickers: List[int] = []
+        for blk in blocks:
+            m = self._face_pat.search(blk)
+            if not m:
+                return np.zeros(24 * len(self._colors), dtype=np.float32)
+            # order: 0,1,2,3 per render() doc
+            c0, c1, c2, c3 = m.group(1), m.group(2), m.group(3), m.group(4)
+            stickers.extend([c0, c1, c2, c3])
+        # stickers now length 24; convert to one-hot
+        grid = np.zeros((24, len(self._colors)), dtype=np.float32)
+        for idx, ch in enumerate(stickers):
+            cidx = self._color_to_idx.get(ch, None)
+            if cidx is not None:
+                grid[idx, cidx] = 1.0
+        return grid.reshape(-1)
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        mapped = int(action) + 1  # 0..11 -> 1..12
+        text_obs, reward, done, info = self._env.step(mapped)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "RubiksCube2x2"
+    total_timesteps: int = 1000_000
+    learning_rate: float = 2.5e-4
+    num_envs: int = 8
+    num_steps: int = 128
+    anneal_lr: bool = True
+    gamma: float = 0.99
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    # Rubik specific（scramble_depth 即打乱步数 / 难度，可与 rotation 一一对应）
+    scramble_depth: int = 3
+    max_steps_env: int = 6
+    # 逗号分隔，如 "2,3,4,5"。在训练难度为 scramble_depth 下训练，仅在这些深度上做评估。
+    # 不设或留空则只在训练难度 scramble_depth 上评估（与旧行为一致）。
+    eval_scramble_depths: str | None = None
+
+    # runtime filled
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+    eval_splits: int = 20
+    eval_episodes: int = 10000
+
+
+def make_env(idx, run_name, seed, scramble_depth, max_steps_env, capture_video=False):
+    def thunk():
+        config = RubiksCube2x2Config(scramble_depth=scramble_depth, max_steps=max_steps_env, render_mode='text')
+        env = RubiksCube2x2Env(config)
+        env = RubiksCubeWrapper(env)
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps_env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = int(np.array(envs.single_observation_space.shape).prod())
+        hidden = 128
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None):
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # envs
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args.scramble_depth, args.max_steps_env, args.capture_video)
+        for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # storage
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # start
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    # eval helper similar to FrozenLake: collect greedy eval trajectories and write metrics.json
+    def collect_eval_trajectories(
+        agent_model,
+        make_env_fn,
+        n_episodes,
+        step_tag,
+        trajectory_subtag: str | None = None,
+        scramble_depth_eval: int | None = None,
+    ):
+        sub = f"step_{step_tag}" if not trajectory_subtag else f"step_{step_tag}_{trajectory_subtag}"
+        out_dir = Path(f"runs/{run_name}/trajectories/{sub}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env.reset(seed=args.seed + 100000 + collected)
+                traj_states = [state.tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env, '_max_episode_steps', None) or int(args.max_steps_env)
+                while not done:
+                    with torch.no_grad():
+                        logits = agent_model.actor(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(logits, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env.step(action)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(state.tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            if scramble_depth_eval is not None:
+                metrics["scramble_depth_eval"] = int(scramble_depth_eval)
+                metrics["scramble_depth_train"] = int(args.scramble_depth)
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    def parse_eval_scramble_depths(s: str | None, train_depth: int) -> List[int]:
+        if s is None or not str(s).strip():
+            return [train_depth]
+        return [int(x.strip()) for x in str(s).split(",") if x.strip()]
+
+    eval_depths_list = parse_eval_scramble_depths(args.eval_scramble_depths, args.scramble_depth)
+    print(
+        f"Train scramble_depth={args.scramble_depth} | "
+        f"Eval scramble_depths={eval_depths_list} (OOD 泛化评估)"
+    )
+
+    # training loop
+    eval_every_iters = max(1, args.num_iterations // args.eval_splits)
+    for iteration in range(1, args.num_iterations + 1):
+        # Anneal LR
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        # accumulate per-iteration episode successes
+        iter_successes = []
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            # Episode stats logging similar to FrozenLake
+            try:
+                mask = None
+                if isinstance(infos, dict):
+                    if "_episode" in infos:
+                        mask = np.asarray(infos["_episode"]).astype(bool)
+                    elif "episode" in infos and isinstance(infos["episode"], dict) and "_l" in infos["episode"]:
+                        mask = np.asarray(infos["episode"]["_l"]).astype(bool)
+                if mask is not None and np.any(mask):
+                    r_arr = np.asarray(infos.get("episode", {}).get("r", np.zeros_like(mask, dtype=float)))
+                    l_arr = np.asarray(infos.get("episode", {}).get("l", np.zeros_like(mask, dtype=int)))
+                    # prefer success from info; fallback to ep return > 0
+                    if "success" in infos:
+                        succ_arr = np.asarray(infos.get("success", np.zeros_like(mask, dtype=bool))).astype(float)
+                    else:
+                        try:
+                            succ_arr = (np.asarray(r_arr) > 0).astype(float)
+                        except Exception:
+                            succ_arr = np.zeros_like(mask, dtype=float)
+                    # collect iteration successes for training success rate
+                    try:
+                        for s in np.asarray(succ_arr)[mask]:
+                            iter_successes.append(float(s))
+                    except Exception:
+                        pass
+                    if args.track:
+                        try:
+                            import wandb
+                            log_dict = {
+                                "global_step": int(global_step),
+                                "rollout/ep_rew_mean": float(np.mean(r_arr[mask])) if np.any(mask) else None,
+                                "rollout/ep_len_mean": float(np.mean(l_arr[mask])) if np.any(mask) else None,
+                                "rollout/success_rate": float(np.mean(succ_arr[mask])) if np.any(mask) else None,
+                            }
+                            wandb.log(log_dict, step=global_step)
+                        except Exception:
+                            pass
+            except Exception:
+                pass
+
+        # GAE
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # update
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss = 0.5 * torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        # metrics
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        train_success_rate = float(np.mean(iter_successes)) if len(iter_successes) else 0.0
+        print(f"Iter {iteration:4d}/{args.num_iterations} | SPS: {sps:5d} | R: {rewards.mean().item():6.3f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "charts/progress": float(100.0 * iteration / max(1, args.num_iterations)),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "losses/value_loss": float(v_loss.item()),
+                    "losses/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "train/old_approx_kl": float(old_approx_kl.item()),
+                    "train/approx_kl": float(approx_kl.item()),
+                    "train/clipfrac": float(np.mean(clipfracs)) if len(clipfracs) else 0.0,
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(values.mean().item()),
+                    "perf/SPS": int(sps),
+                    "charts/SPS": int(sps),
+                    "train/success_rate": train_success_rate,
+                    "charts/train_success_rate": train_success_rate,
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                }, step=global_step)
+            except Exception:
+                pass
+
+        # periodic evaluation collection
+        if iteration % eval_every_iters == 0:
+            try:
+                wb_eval: Dict[str, Any] = {}
+                for ed in eval_depths_list:
+                    subtag = f"scramble{ed}"
+                    eval_thunk = make_env(0, run_name, args.seed + 9999, ed, args.max_steps_env, False)
+                    collect_eval_trajectories(
+                        agent,
+                        eval_thunk,
+                        n_episodes=args.eval_episodes,
+                        step_tag=global_step,
+                        trajectory_subtag=subtag,
+                        scramble_depth_eval=ed,
+                    )
+                    if args.track:
+                        try:
+                            import json as _json
+                            from pathlib import Path as _Path
+                            mpath = _Path(f"runs/{run_name}/trajectories/step_{global_step}_{subtag}/metrics.json")
+                            if mpath.exists():
+                                with mpath.open("r") as mf:
+                                    metrics = _json.load(mf)
+                                pfx = f"eval/scramble_{ed}"
+                                wb_eval[f"{pfx}/success_rate"] = metrics.get("success_rate")
+                                wb_eval[f"{pfx}/avg_return"] = metrics.get("avg_return")
+                                wb_eval[f"{pfx}/std_return"] = metrics.get("std_return")
+                                wb_eval[f"{pfx}/episodes"] = metrics.get("episodes")
+                        except Exception:
+                            pass
+                    print(
+                        f"Eval scramble_depth={ed}: collected {args.eval_episodes} trajectories "
+                        f"at global_step {global_step}"
+                    )
+                if args.track and wb_eval:
+                    try:
+                        import wandb
+                        wb_eval["global_step"] = int(global_step)
+                        wandb.log(wb_eval, step=global_step)
+                    except Exception:
+                        pass
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
diff --git a/cleanrl/cleanrl/ppo_sudoku_actionmask.py b/cleanrl/cleanrl/ppo_sudoku_actionmask.py
new file mode 100644
index 0000000000000000000000000000000000000000..bdb9f9d8c15488d72e5523cda91d94cc03dd1726
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_sudoku_actionmask.py
@@ -0,0 +1,588 @@
+# PPO with Action Masking for RAGEN Sudoku (4x4, max_step=20)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+
+import sys
+# 假设 ragen 库在两级目录之上，请根据实际情况调整
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sudoku.env import SudokuEnv
+from ragen.env.sudoku.config import SudokuEnvConfig
+
+
+class SudokuWrapper(gym.Env):
+    """
+    Adapter to use ragen SudokuEnv with Gymnasium vector API.
+    Improvements: Returns a Dict observation with 'action_mask' to prevent
+    the agent from modifying cells that are already filled.
+    """
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: SudokuEnv, grid_size: int):
+        super().__init__()
+        self._env = env
+        self._size = grid_size
+        # 0 denotes empty, 1..grid_size denote values
+        self._val_dim = self._size + 1
+        
+        # Actions: (row, col, num) -> Flattened
+        self._act_n = self._size * self._size * self._size
+        self.action_space = gym.spaces.Discrete(self._act_n)
+
+        # Observation: Dict with mask
+        self.observation_space = gym.spaces.Dict({
+            "observation": gym.spaces.Box(low=0.0, high=1.0, shape=(self._size * self._size * self._val_dim,), dtype=np.float32),
+            "action_mask": gym.spaces.Box(low=0.0, high=1.0, shape=(self._act_n,), dtype=np.float32)
+        })
+
+    def _encode_obs(self, text_obs: str) -> Dict[str, np.ndarray]:
+        # Parse the 'simple' grid format
+        vals: List[int] = []
+        for line in text_obs.splitlines():
+            ls = line.strip()
+            if len(ls) == 0: continue
+            if set(ls) <= {'-'}: continue
+            tokens = [t for t in ls.split() if t != '|']
+            if len(tokens) == 0: continue
+            for t in tokens:
+                if t == '.': vals.append(0)
+                else:
+                    try: v = int(t)
+                    except ValueError: v = 0
+                    vals.append(v)
+        
+        target = self._size * self._size
+        if len(vals) < target: vals.extend([0] * (target - len(vals)))
+        if len(vals) > target: vals = vals[:target]
+        
+        # One-hot encode grid
+        grid = np.zeros((target, self._val_dim), dtype=np.float32)
+        # Initialize mask (1.0 = valid, 0.0 = invalid)
+        mask = np.ones(self._act_n, dtype=np.float32)
+
+        for i, v in enumerate(vals):
+            v_clamped = int(v)
+            if v_clamped < 0 or v_clamped > self._size:
+                v_clamped = 0
+            grid[i, v_clamped] = 1.0
+
+            # If a cell is NOT empty (v_clamped != 0), mask all actions for this cell.
+            # Agent should not overwrite existing numbers.
+            if v_clamped != 0:
+                start_idx = i * self._size
+                end_idx = start_idx + self._size
+                mask[start_idx:end_idx] = 0.0
+
+        return {
+            "observation": grid.reshape(-1),
+            "action_mask": mask
+        }
+
+    @staticmethod
+    def _decode_action(action_id: int, grid_size: int) -> Tuple[int, int, int]:
+        g = grid_size
+        row = action_id // (g * g)
+        rem = action_id % (g * g)
+        col = rem // g
+        num = (rem % g) + 1
+        return row, col, num
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        row, col, num = self._decode_action(int(action), self._size)
+        act_str = f"{row+1},{col+1},{num}"
+        text_obs, reward, done, info = self._env.step(act_str)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "Sudoku"
+    total_timesteps: int = 2000_000
+    learning_rate: float = 3e-4
+    num_envs: int = 8
+    num_steps: int = 128
+    anneal_lr: bool = True
+    gamma: float = 0.99
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    # Sudoku specific
+    grid_size: int = 4
+    difficulty: str = "easy"
+
+    # runtime filled
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+
+    # eval
+    eval_splits: int = 40
+    eval_episodes: int = 10000
+
+
+def make_env(idx, run_name, seed, grid_size, difficulty, capture_video=False):
+    def thunk():
+        config = SudokuEnvConfig(
+            grid_size=grid_size,
+            difficulty=difficulty,
+            render_mode='text',
+            render_format='simple',
+        )
+        env = SudokuEnv(config)
+        env = SudokuWrapper(env, grid_size)
+        # Use env's own max_steps default if available, otherwise a sane cap
+        # Keeping your request for strict step limit logic, although wrapper enforces logic
+        max_steps = 81
+        # max_steps = int(grid_size * grid_size * 6)
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        # Accessing the shape from the Dict space
+        obs_shape = int(np.array(envs.single_observation_space['observation'].shape).prod())
+        hidden = 256 # Increased hidden size slightly for better capacity
+
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None, action_mask=None):
+        logits = self.actor(x)
+        
+        # Apply Action Masking
+        if action_mask is not None:
+            # Set logits of invalid actions to a very large negative number
+            logits = logits + (action_mask - 1.0) * 1e8
+
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # envs
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args.grid_size, args.difficulty, args.capture_video)
+        for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # storage
+    # Note: obs storage now only stores the flattened grid part
+    obs_shape = envs.single_observation_space['observation'].shape
+    mask_shape = envs.single_observation_space['action_mask'].shape
+    
+    obs = torch.zeros((args.num_steps, args.num_envs) + obs_shape).to(device)
+    masks = torch.zeros((args.num_steps, args.num_envs) + mask_shape).to(device) # Storage for masks
+    
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # start
+    global_step = 0
+    start_time = time.time()
+    
+    # envs.reset() returns a Dict of stacked arrays
+    next_obs_dict, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+    next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    
+    episode_returns = []
+    episode_steps = []
+    episode_successes = []
+
+    # Eval helper
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                obs_dict, _ = env.reset(seed=args.seed + collected)
+                # Handle single env dict unpacking
+                state = obs_dict['observation']
+                mask = obs_dict['action_mask']
+                
+                traj_states = [state.tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env, '_max_episode_steps', None) or int(args.grid_size * args.grid_size * 6)
+                
+                # Eval loop
+                current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+
+                while not done:
+                    with torch.no_grad():
+                        # Pass mask to actor during eval
+                        action, _, _, _ = agent_model.get_action_and_value(current_obs, action_mask=current_mask)
+                        action_item = int(action.item())
+                        
+                    next_obs_dict, reward, terminated, truncated, info = env.step(action_item)
+                    
+                    traj_actions.append(action_item)
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool(info.get('success', False)))
+                    
+                    state = next_obs_dict['observation']
+                    mask = next_obs_dict['action_mask']
+                    current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                    current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+                    
+                    traj_states.append(state.tolist())
+                    done = d
+                    
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    eval_every_iters = max(1, args.num_iterations // args.eval_splits)
+
+    # training loop
+    for iteration in range(1, args.num_iterations + 1):
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            optimizer.param_groups[0]["lr"] = frac * args.learning_rate
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            masks[step] = next_mask # Store mask
+            dones[step] = next_done
+
+            with torch.no_grad():
+                # PASS MASK HERE
+                action, logprob, _, value = agent.get_action_and_value(next_obs, action_mask=next_mask)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs_dict, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            
+            # Unpack dict again
+            next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+            next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+            next_done = torch.Tensor(next_done).to(device)
+
+            try:
+                mask = None
+                if isinstance(infos, dict):
+                    if "_episode" in infos:
+                        mask = np.asarray(infos["_episode"]).astype(bool)
+                    elif "episode" in infos and isinstance(infos["episode"], dict) and "_l" in infos["episode"]:
+                        mask = np.asarray(infos["episode"]["_l"]).astype(bool)
+                if mask is not None and np.any(mask):
+                    r_arr = np.asarray(infos.get("episode", {}).get("r", np.zeros_like(mask, dtype=float)))
+                    l_arr = np.asarray(infos.get("episode", {}).get("l", np.zeros_like(mask, dtype=int)))
+                    succ_arr = np.asarray(infos.get("success", np.zeros_like(mask, dtype=bool))).astype(float)
+                    for i in np.where(mask)[0]:
+                        episode_returns.append(float(r_arr[i]))
+                        episode_steps.append(global_step)
+                        episode_successes.append(float(succ_arr[i]))
+                    if args.track:
+                        try:
+                            import wandb
+                            log_dict = {
+                                "global_step": int(global_step),
+                                "rollout/ep_rew_mean": float(np.mean(r_arr[mask])) if np.any(mask) else None,
+                                "rollout/ep_len_mean": float(np.mean(l_arr[mask])) if np.any(mask) else None,
+                                "rollout/success_rate": float(np.mean(succ_arr[mask])) if np.any(mask) else None,
+                            }
+                            if np.any(mask):
+                                last_idx = np.where(mask)[0][-1]
+                                log_dict.update({
+                                    "train/episodic_return": float(r_arr[last_idx]),
+                                    "train/episodic_length": int(l_arr[last_idx]),
+                                    "train/success": float(succ_arr[last_idx]),
+                                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 100 else None,
+                                })
+                            wandb.log(log_dict, step=global_step)
+                        except Exception:
+                            pass
+            except Exception:
+                pass
+
+        # GAE
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten batch
+        b_obs = obs.reshape((-1,) + obs_shape)
+        b_masks = masks.reshape((-1,) + mask_shape) # Flatten masks
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # update
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                # PASS MASK HERE
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(
+                    b_obs[mb_inds], 
+                    action=b_actions.long()[mb_inds],
+                    action_mask=b_masks[mb_inds]
+                )
+                
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss = 0.5 * torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        # logging
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Val: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "train/old_approx_kl": float(old_approx_kl.item()),
+                    "train/approx_kl": float(approx_kl.item()),
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(values.mean().item()),
+                    "perf/SPS": int(sps),
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                }, step=global_step)
+            except Exception:
+                pass
+
+        if iteration % eval_every_iters == 0:
+            try:
+                eval_thunk = make_env(0, run_name, args.seed + 9999, args.grid_size, args.difficulty, False)
+                collect_eval_trajectories(agent, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import json as _json
+                        from pathlib import Path as _Path
+                        mpath = _Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = _json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at global_step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
\ No newline at end of file
diff --git a/cleanrl/cleanrl/ppo_sudoku_strongactionmask.py b/cleanrl/cleanrl/ppo_sudoku_strongactionmask.py
new file mode 100644
index 0000000000000000000000000000000000000000..ed8686a222dacc6e746a52276869bf76579c86bb
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_sudoku_strongactionmask.py
@@ -0,0 +1,616 @@
+# PPO with Action Masking for RAGEN Sudoku (4x4, max_step=20)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+
+import sys
+# 假设 ragen 库在两级目录之上，请根据实际情况调整
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sudoku.env import SudokuEnv
+from ragen.env.sudoku.config import SudokuEnvConfig
+
+
+class SudokuWrapper(gym.Env):
+    """
+    Adapter to use ragen SudokuEnv with Gymnasium vector API.
+    Improvements: Returns a Dict observation with 'action_mask' to prevent
+    the agent from modifying cells that are already filled.
+    """
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: SudokuEnv, grid_size: int):
+        super().__init__()
+        self._env = env
+        self._size = grid_size
+        # 0 denotes empty, 1..grid_size denote values
+        self._val_dim = self._size + 1
+        
+        # Actions: (row, col, num) -> Flattened
+        self._act_n = self._size * self._size * self._size
+        self.action_space = gym.spaces.Discrete(self._act_n)
+
+        # Observation: Dict with mask
+        self.observation_space = gym.spaces.Dict({
+            "observation": gym.spaces.Box(low=0.0, high=1.0, shape=(self._size * self._size * self._val_dim,), dtype=np.float32),
+            "action_mask": gym.spaces.Box(low=0.0, high=1.0, shape=(self._act_n,), dtype=np.float32)
+        })
+
+    def _encode_obs(self, text_obs: str) -> Dict[str, np.ndarray]:
+        # Parse the 'simple' grid format
+        vals: List[int] = []
+        for line in text_obs.splitlines():
+            ls = line.strip()
+            if len(ls) == 0: continue
+            if set(ls) <= {'-'}: continue
+            tokens = [t for t in ls.split() if t != '|']
+            if len(tokens) == 0: continue
+            for t in tokens:
+                if t == '.': vals.append(0)
+                else:
+                    try: v = int(t)
+                    except ValueError: v = 0
+                    vals.append(v)
+        
+        target = self._size * self._size
+        if len(vals) < target: vals.extend([0] * (target - len(vals)))
+        if len(vals) > target: vals = vals[:target]
+        
+        # One-hot encode grid
+        grid = np.zeros((target, self._val_dim), dtype=np.float32)
+        # Initialize mask (1.0 = valid, 0.0 = invalid)
+        mask = np.zeros(self._act_n, dtype=np.float32)
+
+        for i, v in enumerate(vals):
+            v_clamped = int(v)
+            if v_clamped < 0 or v_clamped > self._size:
+                v_clamped = 0
+            grid[i, v_clamped] = 1.0
+
+            # If a cell is NOT empty (v_clamped != 0), mask all actions for this cell.
+            # Agent should not overwrite existing numbers.
+            if v_clamped != 0:
+                start_idx = i * self._size
+                end_idx = start_idx + self._size
+                mask[start_idx:end_idx] = 0.0
+
+        # Build current grid for validity checks
+        cur = np.array(vals, dtype=np.int64).reshape(self._size, self._size)
+        box = int(np.sqrt(self._size))
+
+        # Set mask=1 only for Sudoku-rule-valid placements on empty cells
+        for i, v in enumerate(vals):
+            if int(v) == 0:
+                r = i // self._size
+                c = i % self._size
+                row_vals = set(cur[r, :].tolist())
+                col_vals = set(cur[:, c].tolist())
+                br = (r // box) * box
+                bc = (c // box) * box
+                box_vals = set(cur[br:br+box, bc:bc+box].reshape(-1).tolist())
+                start_idx = i * self._size
+                for num in range(1, self._size + 1):
+                    if (num not in row_vals) and (num not in col_vals) and (num not in box_vals):
+                        mask[start_idx + (num - 1)] = 1.0
+
+        # Safe fallback: if no valid actions found, allow all actions on empty cells; if still none, allow all
+        if mask.sum() == 0:
+            for i, v in enumerate(vals):
+                if int(v) == 0:
+                    start_idx = i * self._size
+                    end_idx = start_idx + self._size
+                    mask[start_idx:end_idx] = 1.0
+            if mask.sum() == 0:
+                mask[:] = 1.0
+
+        return {
+            "observation": grid.reshape(-1),
+            "action_mask": mask
+        }
+
+    @staticmethod
+    def _decode_action(action_id: int, grid_size: int) -> Tuple[int, int, int]:
+        g = grid_size
+        row = action_id // (g * g)
+        rem = action_id % (g * g)
+        col = rem // g
+        num = (rem % g) + 1
+        return row, col, num
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        row, col, num = self._decode_action(int(action), self._size)
+        act_str = f"{row+1},{col+1},{num}"
+        text_obs, reward, done, info = self._env.step(act_str)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "Sudoku"
+    total_timesteps: int = 2000_000
+    learning_rate: float = 3e-4
+    num_envs: int = 8
+    num_steps: int = 128
+    anneal_lr: bool = True
+    gamma: float = 0.99
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    # Sudoku specific
+    grid_size: int = 9
+    difficulty: str = "easy"
+
+    # runtime filled
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+
+    # eval
+    eval_splits: int = 2
+    eval_episodes: int = 4000
+
+
+def make_env(idx, run_name, seed, grid_size, difficulty, capture_video=False):
+    def thunk():
+        config = SudokuEnvConfig(
+            grid_size=grid_size,
+            difficulty=difficulty,
+            render_mode='text',
+            render_format='simple',
+        )
+        env = SudokuEnv(config)
+        env = SudokuWrapper(env, grid_size)
+        # Use env's own max_steps default if available, otherwise a sane cap
+        # Keeping your request for strict step limit logic, although wrapper enforces logic
+        max_steps = 81
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        # Accessing the shape from the Dict space
+        obs_shape = int(np.array(envs.single_observation_space['observation'].shape).prod())
+        hidden = 256 # Increased hidden size slightly for better capacity
+
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None, action_mask=None):
+        logits = self.actor(x)
+        
+        # Apply Action Masking
+        if action_mask is not None:
+            # Set logits of invalid actions to a very large negative number
+            logits = logits + (action_mask - 1.0) * 1e8
+
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # envs
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args.grid_size, args.difficulty, args.capture_video)
+        for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # storage
+    # Note: obs storage now only stores the flattened grid part
+    obs_shape = envs.single_observation_space['observation'].shape
+    mask_shape = envs.single_observation_space['action_mask'].shape
+    
+    obs = torch.zeros((args.num_steps, args.num_envs) + obs_shape).to(device)
+    masks = torch.zeros((args.num_steps, args.num_envs) + mask_shape).to(device) # Storage for masks
+    
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # start
+    global_step = 0
+    start_time = time.time()
+    
+    # envs.reset() returns a Dict of stacked arrays
+    next_obs_dict, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+    next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    
+    episode_returns = []
+    episode_steps = []
+    episode_successes = []
+
+    # Eval helper
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                obs_dict, _ = env.reset(seed=args.seed + collected)
+                # Handle single env dict unpacking
+                state = obs_dict['observation']
+                mask = obs_dict['action_mask']
+                
+                traj_states = [state.tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env, '_max_episode_steps', None) or int(args.grid_size * args.grid_size * 6)
+                
+                # Eval loop
+                current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+
+                while not done:
+                    with torch.no_grad():
+                        # Pass mask to actor during eval
+                        action, _, _, _ = agent_model.get_action_and_value(current_obs, action_mask=current_mask)
+                        action_item = int(action.item())
+                        
+                    next_obs_dict, reward, terminated, truncated, info = env.step(action_item)
+                    
+                    traj_actions.append(action_item)
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool(info.get('success', False)))
+                    
+                    state = next_obs_dict['observation']
+                    mask = next_obs_dict['action_mask']
+                    current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                    current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+                    
+                    traj_states.append(state.tolist())
+                    done = d
+                    
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    eval_every_iters = max(1, args.num_iterations // args.eval_splits)
+
+    # training loop
+    for iteration in range(1, args.num_iterations + 1):
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            optimizer.param_groups[0]["lr"] = frac * args.learning_rate
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            masks[step] = next_mask # Store mask
+            dones[step] = next_done
+
+            with torch.no_grad():
+                # PASS MASK HERE
+                action, logprob, _, value = agent.get_action_and_value(next_obs, action_mask=next_mask)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs_dict, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            
+            # Unpack dict again
+            next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+            next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+            next_done = torch.Tensor(next_done).to(device)
+
+            try:
+                mask = None
+                if isinstance(infos, dict):
+                    if "_episode" in infos:
+                        mask = np.asarray(infos["_episode"]).astype(bool)
+                    elif "episode" in infos and isinstance(infos["episode"], dict) and "_l" in infos["episode"]:
+                        mask = np.asarray(infos["episode"]["_l"]).astype(bool)
+                if mask is not None and np.any(mask):
+                    r_arr = np.asarray(infos.get("episode", {}).get("r", np.zeros_like(mask, dtype=float)))
+                    l_arr = np.asarray(infos.get("episode", {}).get("l", np.zeros_like(mask, dtype=int)))
+                    succ_arr = np.asarray(infos.get("success", np.zeros_like(mask, dtype=bool))).astype(float)
+                    for i in np.where(mask)[0]:
+                        episode_returns.append(float(r_arr[i]))
+                        episode_steps.append(global_step)
+                        episode_successes.append(float(succ_arr[i]))
+                    if args.track:
+                        try:
+                            import wandb
+                            log_dict = {
+                                "global_step": int(global_step),
+                                "rollout/ep_rew_mean": float(np.mean(r_arr[mask])) if np.any(mask) else None,
+                                "rollout/ep_len_mean": float(np.mean(l_arr[mask])) if np.any(mask) else None,
+                                "rollout/success_rate": float(np.mean(succ_arr[mask])) if np.any(mask) else None,
+                            }
+                            if np.any(mask):
+                                last_idx = np.where(mask)[0][-1]
+                                log_dict.update({
+                                    "train/episodic_return": float(r_arr[last_idx]),
+                                    "train/episodic_length": int(l_arr[last_idx]),
+                                    "train/success": float(succ_arr[last_idx]),
+                                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 100 else None,
+                                })
+                            wandb.log(log_dict, step=global_step)
+                        except Exception:
+                            pass
+            except Exception:
+                pass
+
+        # GAE
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten batch
+        b_obs = obs.reshape((-1,) + obs_shape)
+        b_masks = masks.reshape((-1,) + mask_shape) # Flatten masks
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # update
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                # PASS MASK HERE
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(
+                    b_obs[mb_inds], 
+                    action=b_actions.long()[mb_inds],
+                    action_mask=b_masks[mb_inds]
+                )
+                
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss = 0.5 * torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        # logging
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Val: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "train/old_approx_kl": float(old_approx_kl.item()),
+                    "train/approx_kl": float(approx_kl.item()),
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(values.mean().item()),
+                    "perf/SPS": int(sps),
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                }, step=global_step)
+            except Exception:
+                pass
+
+        if iteration % eval_every_iters == 0:
+            try:
+                eval_thunk = make_env(0, run_name, args.seed + 9999, args.grid_size, args.difficulty, False)
+                collect_eval_trajectories(agent, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import json as _json
+                        from pathlib import Path as _Path
+                        mpath = _Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = _json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at global_step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
\ No newline at end of file
diff --git a/cleanrl/cleanrl/ppo_trxl/pom_env.py b/cleanrl/cleanrl/ppo_trxl/pom_env.py
new file mode 100644
index 0000000000000000000000000000000000000000..9c807ce883c9faf3a75a2c89220600136e302cc9
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_trxl/pom_env.py
@@ -0,0 +1,186 @@
+import gymnasium as gym
+import numpy as np
+import pygame
+from gymnasium import spaces
+
+gym.register(
+    id="ProofofMemory-v0",
+    entry_point="pom_env:PoMEnv",
+    max_episode_steps=16,
+)
+
+
+class PoMEnv(gym.Env):
+    """
+    Proof of Concept Memory Environment
+
+    This environment is intended to assess whether the policy's memory is working.
+    The environment is based on a one dimensional grid where the agent can move left or right.
+    At both ends, a goal is spawned that is either punishing or rewarding.
+    During the very first two steps, the agent gets to know which goal leads to a positive or negative reward.
+    Afterwards, this information is hidden in the agent's observation.
+    The last value of the agent's observation is its current position inside the environment.
+    Optionally and to increase the difficulty of the task, the agent's position can be frozen until the goal information is hidden.
+    To further challenge the agent, the step_size can be decreased.
+    """
+
+    metadata = {"render_modes": ["human", "rgb_array", "debug_rgb_array"], "render_fps": 4}
+
+    def __init__(self, render_mode="human"):
+        self._freeze = True
+        self._step_size = 0.2
+        self._min_steps = int(1.0 / self._step_size) + 1
+        self._time_penalty = 0.1
+        self._num_show_steps = 2
+        self.render_mode = render_mode
+        glob = False
+
+        self.action_space = spaces.Discrete(2)
+        self.observation_space = spaces.Box(low=-1.0, high=1.0, shape=(3,), dtype=np.float32)
+
+        # Create an array with possible positions
+        num_steps = int(0.4 / self._step_size)
+        lower = min(-2.0 * self._step_size, -num_steps * self._step_size) if not glob else -1 + self._step_size
+        upper = max(3.0 * self._step_size, self._step_size, (num_steps + 1) * self._step_size) if not glob else 1
+        self.possible_positions = np.arange(lower, upper, self._step_size).clip(-1 + self._step_size, 1 - self._step_size)
+        self.possible_positions = list(map(lambda x: round(x, 2), self.possible_positions))  # fix floating point errors
+
+        # Pygame-related attributes for rendering
+        self.window = None
+        self.clock = None
+        self.width = 400
+        self.height = 80
+        self.cell_width = self.width / (2 * int(1 / self._step_size) + 1)
+
+    def step(self, action):
+        reward = 0.0
+        done = False
+
+        if self._num_show_steps > self._step_count:
+            self._position += self._step_size * (1 - self._freeze) if action == 1 else -self._step_size * (1 - self._freeze)
+            self._position = np.round(self._position, 2)
+
+            obs = np.asarray([self._goals[0], self._position, self._goals[1]], dtype=np.float32)
+
+            if self._freeze:
+                self._step_count += 1
+                return obs, reward, done, False, {}
+
+        else:
+            self._position += self._step_size if action == 1 else -self._step_size
+            self._position = np.round(self._position, 2)
+            obs = np.asarray([0.0, self._position, 0.0], dtype=np.float32)  # mask out goal information
+
+        # Determine reward and termination
+        if self._position == -1.0:
+            if self._goals[0] == 1.0:
+                reward += 1.0 + self._min_steps * self._time_penalty
+            else:
+                reward -= 1.0 + self._min_steps * self._time_penalty
+            done = True
+        elif self._position == 1.0:
+            if self._goals[1] == 1.0:
+                reward += 1.0 + self._min_steps * self._time_penalty
+            else:
+                reward -= 1.0 + self._min_steps * self._time_penalty
+            done = True
+        else:
+            reward -= self._time_penalty
+        self.rewards.append(reward)
+
+        if done:
+            info = {"reward": sum(self.rewards), "length": len(self.rewards)}
+        else:
+            info = {}
+
+        self._step_count += 1
+
+        return obs, reward, done, False, info
+
+    def reset(self, *, seed=None, options=None):
+        super().reset(seed=seed)
+        self.rewards = []
+        self._position = np.random.choice(self.possible_positions)
+        self._step_count = 0
+        goals = np.asarray([-1.0, 1.0])
+        self._goals = goals[np.random.permutation(2)]
+        obs = np.asarray([self._goals[0], self._position, self._goals[1]], dtype=np.float32)
+        return obs, {}
+
+    def render(self):
+        if self.render_mode not in self.metadata["render_modes"]:
+            return
+
+        # Initialize Pygame
+        if not pygame.get_init():
+            pygame.init()
+        if self.window is None and self.render_mode == "human":
+            pygame.display.init()
+            self.window = pygame.display.set_mode((self.width, self.height))
+            pygame.display.set_caption("Proof of Memory Environment")
+        if self.clock is None and self.render_mode == "human":
+            self.clock = pygame.time.Clock()
+
+        # Create surface
+        canvas = pygame.Surface((self.width, self.height))
+        canvas.fill((255, 255, 255))  # Fill the background with white
+
+        # Draw grid
+        num_cells = 2 * int(1 / self._step_size) + 1
+        for i in range(num_cells):
+            x = i * self.cell_width
+            pygame.draw.rect(canvas, (200, 200, 200), pygame.Rect(x, 0, self.cell_width, self.height), 1)
+
+        # Draw agent
+        agent_pos = int((self._position + 1) / self._step_size)
+        agent_x = agent_pos * self.cell_width + self.cell_width / 2
+        pygame.draw.circle(canvas, (0, 0, 255), (agent_x, self.height / 2), 15)
+
+        # Draw goals
+        show_goals = self._num_show_steps > self._step_count
+        if show_goals:
+            left_goal_color = (0, 255, 0) if self._goals[0] > 0 else (255, 0, 0)
+            pygame.draw.rect(canvas, left_goal_color, pygame.Rect(0, 0, self.cell_width, self.height))
+            right_goal_color = (0, 255, 0) if self._goals[1] > 0 else (255, 0, 0)
+            pygame.draw.rect(
+                canvas, right_goal_color, pygame.Rect(self.width - self.cell_width, 0, self.cell_width, self.height)
+            )
+        else:
+            pygame.draw.rect(canvas, (200, 200, 200), pygame.Rect(0, 0, self.cell_width, self.height))
+            pygame.draw.rect(
+                canvas, (200, 200, 200), pygame.Rect(self.width - self.cell_width, 0, self.cell_width, self.height)
+            )
+
+        # Render text information
+        font = pygame.font.SysFont(None, 24)
+        text = font.render(f"Goals are shown: {show_goals}", True, (0, 0, 0))
+        canvas.blit(text, (10, 10))
+
+        if self.render_mode == "human":
+            self.window.blit(canvas, (0, 0))
+            pygame.display.flip()
+            self.clock.tick(self.metadata["render_fps"])
+        elif self.render_mode in ["rgb_array", "debug_rgb_array"]:
+            return np.transpose(np.array(pygame.surfarray.pixels3d(canvas)), axes=(1, 0, 2))
+
+    def close(self):
+        if self.window is not None:
+            pygame.display.quit()
+            pygame.quit()
+            self.window = None
+            self.clock = None
+
+
+if __name__ == "__main__":
+    env = PoMEnv(render_mode="human")
+    o, _ = env.reset()
+    img = env.render()
+    done = False
+    rewards = []
+    const_action = 1
+    while not done:
+        o, r, done, _, _ = env.step(const_action)
+        rewards.append(r)
+        img = env.render()
+    print(f"Total reward: {sum(rewards)}, Steps: {len(rewards)}")
+    env.close()
diff --git a/cleanrl/cleanrl/ppo_trxl/ppo_trxl.py b/cleanrl/cleanrl/ppo_trxl/ppo_trxl.py
new file mode 100644
index 0000000000000000000000000000000000000000..7529546b0843e120d995b95dcc4069267dcc671f
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_trxl/ppo_trxl.py
@@ -0,0 +1,682 @@
+import os
+import random
+import time
+from collections import deque
+from dataclasses import dataclass
+
+import gymnasium as gym
+import memory_gym  # noqa
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from einops import rearrange
+from minigrid.wrappers import ImgObsWrapper, RGBImgPartialObsWrapper
+from pom_env import PoMEnv  # noqa
+from torch.distributions import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+
+    # Algorithm specific arguments
+    env_id: str = "MortarMayhem-Grid-v0"
+    """the id of the environment"""
+    total_timesteps: int = 200000000
+    """total timesteps of the experiments"""
+    init_lr: float = 2.75e-4
+    """the initial learning rate of the optimizer"""
+    final_lr: float = 1.0e-5
+    """the final learning rate of the optimizer after linearly annealing"""
+    num_envs: int = 32
+    """the number of parallel game environments"""
+    num_steps: int = 512
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_steps: int = 32 * 512 * 10000
+    """the number of steps to linearly anneal the learning rate and entropy coefficient from initial to final"""
+    gamma: float = 0.995
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 8
+    """the number of mini-batches"""
+    update_epochs: int = 3
+    """the K epochs to update the policy"""
+    norm_adv: bool = False
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.1
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    init_ent_coef: float = 0.0001
+    """initial coefficient of the entropy bonus"""
+    final_ent_coef: float = 0.000001
+    """final coefficient of the entropy bonus after linearly annealing"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.25
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+
+    # Transformer-XL specific arguments
+    trxl_num_layers: int = 3
+    """the number of transformer layers"""
+    trxl_num_heads: int = 4
+    """the number of heads used in multi-head attention"""
+    trxl_dim: int = 384
+    """the dimension of the transformer"""
+    trxl_memory_length: int = 119
+    """the length of TrXL's sliding memory window"""
+    trxl_positional_encoding: str = "absolute"
+    """the positional encoding type of the transformer, choices: "", "absolute", "learned" """
+    reconstruction_coef: float = 0.0
+    """the coefficient of the observation reconstruction loss, if set to 0.0 the reconstruction loss is not used"""
+
+    # To be filled on runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def make_env(env_id, idx, capture_video, run_name, render_mode="debug_rgb_array"):
+    if "MiniGrid" in env_id:
+        if render_mode == "debug_rgb_array":
+            render_mode = "rgb_array"
+
+    def thunk():
+        if "MiniGrid" in env_id:
+            env = gym.make(env_id, agent_view_size=3, tile_size=28, render_mode=render_mode)
+            env = ImgObsWrapper(RGBImgPartialObsWrapper(env, tile_size=28))
+            env = gym.wrappers.TimeLimit(env, 96)
+        else:
+            env = gym.make(env_id, render_mode=render_mode)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return gym.wrappers.RecordEpisodeStatistics(env)
+
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    # torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+def batched_index_select(input, dim, index):
+    for ii in range(1, len(input.shape)):
+        if ii != dim:
+            index = index.unsqueeze(ii)
+    expanse = list(input.shape)
+    expanse[0] = -1
+    expanse[dim] = -1
+    index = index.expand(expanse)
+    return torch.gather(input, dim, index)
+
+
+class PositionalEncoding(nn.Module):
+    def __init__(self, dim, min_timescale=2.0, max_timescale=1e4):
+        super().__init__()
+        freqs = torch.arange(0, dim, min_timescale)
+        inv_freqs = max_timescale ** (-freqs / dim)
+        self.register_buffer("inv_freqs", inv_freqs)
+
+    def forward(self, seq_len):
+        seq = torch.arange(seq_len - 1, -1, -1.0)
+        sinusoidal_inp = rearrange(seq, "n -> n ()") * rearrange(self.inv_freqs, "d -> () d")
+        pos_emb = torch.cat((sinusoidal_inp.sin(), sinusoidal_inp.cos()), dim=-1)
+        return pos_emb
+
+
+class MultiHeadAttention(nn.Module):
+    """Multi Head Attention without dropout inspired by https://github.com/aladdinpersson/Machine-Learning-Collection"""
+
+    def __init__(self, embed_dim, num_heads):
+        super().__init__()
+        self.embed_dim = embed_dim
+        self.num_heads = num_heads
+        self.head_size = embed_dim // num_heads
+
+        assert self.head_size * num_heads == embed_dim, "Embedding dimension needs to be divisible by the number of heads"
+
+        self.values = nn.Linear(self.head_size, self.head_size, bias=False)
+        self.keys = nn.Linear(self.head_size, self.head_size, bias=False)
+        self.queries = nn.Linear(self.head_size, self.head_size, bias=False)
+        self.fc_out = nn.Linear(self.num_heads * self.head_size, embed_dim)
+
+    def forward(self, values, keys, query, mask):
+        N = query.shape[0]
+        value_len, key_len, query_len = values.shape[1], keys.shape[1], query.shape[1]
+
+        values = values.reshape(N, value_len, self.num_heads, self.head_size)
+        keys = keys.reshape(N, key_len, self.num_heads, self.head_size)
+        query = query.reshape(N, query_len, self.num_heads, self.head_size)
+
+        values = self.values(values)  # (N, value_len, heads, head_dim)
+        keys = self.keys(keys)  # (N, key_len, heads, head_dim)
+        queries = self.queries(query)  # (N, query_len, heads, heads_dim)
+
+        # Dot-product
+        energy = torch.einsum("nqhd,nkhd->nhqk", [queries, keys])
+
+        # Mask padded indices so their attention weights become 0
+        if mask is not None:
+            energy = energy.masked_fill(mask.unsqueeze(1).unsqueeze(1) == 0, float("-1e20"))  # -inf causes NaN
+
+        # Normalize energy values and apply softmax to retrieve the attention scores
+        attention = torch.softmax(
+            energy / (self.embed_dim ** (1 / 2)), dim=3
+        )  # attention shape: (N, heads, query_len, key_len)
+
+        # Scale values by attention weights
+        out = torch.einsum("nhql,nlhd->nqhd", [attention, values]).reshape(N, query_len, self.num_heads * self.head_size)
+
+        return self.fc_out(out), attention
+
+
+class TransformerLayer(nn.Module):
+    def __init__(self, dim, num_heads):
+        super().__init__()
+        self.attention = MultiHeadAttention(dim, num_heads)
+        self.layer_norm_q = nn.LayerNorm(dim)
+        self.norm_kv = nn.LayerNorm(dim)
+        self.layer_norm_attn = nn.LayerNorm(dim)
+        self.fc_projection = nn.Sequential(nn.Linear(dim, dim), nn.ReLU())
+
+    def forward(self, value, key, query, mask):
+        # Pre-layer normalization (post-layer normalization is usually less effective)
+        query_ = self.layer_norm_q(query)
+        value = self.norm_kv(value)
+        key = value  # K = V -> self-attention
+        attention, attention_weights = self.attention(value, key, query_, mask)  # MHA
+        x = attention + query  # Skip connection
+        x_ = self.layer_norm_attn(x)  # Pre-layer normalization
+        forward = self.fc_projection(x_)  # Forward projection
+        out = forward + x  # Skip connection
+        return out, attention_weights
+
+
+class Transformer(nn.Module):
+    def __init__(self, num_layers, dim, num_heads, max_episode_steps, positional_encoding):
+        super().__init__()
+        self.max_episode_steps = max_episode_steps
+        self.positional_encoding = positional_encoding
+        if positional_encoding == "absolute":
+            self.pos_embedding = PositionalEncoding(dim)
+        elif positional_encoding == "learned":
+            self.pos_embedding = nn.Parameter(torch.randn(max_episode_steps, dim))
+        self.transformer_layers = nn.ModuleList([TransformerLayer(dim, num_heads) for _ in range(num_layers)])
+
+    def forward(self, x, memories, mask, memory_indices):
+        # Add positional encoding to every transformer layer input
+        if self.positional_encoding == "absolute":
+            pos_embedding = self.pos_embedding(self.max_episode_steps)[memory_indices]
+            memories = memories + pos_embedding.unsqueeze(2)
+        elif self.positional_encoding == "learned":
+            memories = memories + self.pos_embedding[memory_indices].unsqueeze(2)
+
+        # Forward transformer layers and return new memories (i.e. hidden states)
+        out_memories = []
+        for i, layer in enumerate(self.transformer_layers):
+            out_memories.append(x.detach())
+            x, attention_weights = layer(
+                memories[:, :, i], memories[:, :, i], x.unsqueeze(1), mask
+            )  # args: value, key, query, mask
+            x = x.squeeze()
+            if len(x.shape) == 1:
+                x = x.unsqueeze(0)
+        return x, torch.stack(out_memories, dim=1)
+
+
+class Agent(nn.Module):
+    def __init__(self, args, observation_space, action_space_shape, max_episode_steps):
+        super().__init__()
+        self.obs_shape = observation_space.shape
+        self.max_episode_steps = max_episode_steps
+
+        if len(self.obs_shape) > 1:
+            self.encoder = nn.Sequential(
+                layer_init(nn.Conv2d(3, 32, 8, stride=4)),
+                nn.ReLU(),
+                layer_init(nn.Conv2d(32, 64, 4, stride=2)),
+                nn.ReLU(),
+                layer_init(nn.Conv2d(64, 64, 3, stride=1)),
+                nn.ReLU(),
+                nn.Flatten(),
+                layer_init(nn.Linear(64 * 7 * 7, args.trxl_dim)),
+                nn.ReLU(),
+            )
+        else:
+            self.encoder = layer_init(nn.Linear(observation_space.shape[0], args.trxl_dim))
+
+        self.transformer = Transformer(
+            args.trxl_num_layers, args.trxl_dim, args.trxl_num_heads, self.max_episode_steps, args.trxl_positional_encoding
+        )
+
+        self.hidden_post_trxl = nn.Sequential(
+            layer_init(nn.Linear(args.trxl_dim, args.trxl_dim)),
+            nn.ReLU(),
+        )
+
+        self.actor_branches = nn.ModuleList(
+            [
+                layer_init(nn.Linear(args.trxl_dim, out_features=num_actions), np.sqrt(0.01))
+                for num_actions in action_space_shape
+            ]
+        )
+        self.critic = layer_init(nn.Linear(args.trxl_dim, 1), 1)
+
+        if args.reconstruction_coef > 0.0:
+            self.transposed_cnn = nn.Sequential(
+                layer_init(nn.Linear(args.trxl_dim, 64 * 7 * 7)),
+                nn.ReLU(),
+                nn.Unflatten(1, (64, 7, 7)),
+                layer_init(nn.ConvTranspose2d(64, 64, 3, stride=1)),
+                nn.ReLU(),
+                layer_init(nn.ConvTranspose2d(64, 32, 4, stride=2)),
+                nn.ReLU(),
+                layer_init(nn.ConvTranspose2d(32, 3, 8, stride=4)),
+                nn.Sigmoid(),
+            )
+
+    def get_value(self, x, memory, memory_mask, memory_indices):
+        if len(self.obs_shape) > 1:
+            x = self.encoder(x.permute((0, 3, 1, 2)) / 255.0)
+        else:
+            x = self.encoder(x)
+        x, _ = self.transformer(x, memory, memory_mask, memory_indices)
+        x = self.hidden_post_trxl(x)
+        return self.critic(x).flatten()
+
+    def get_action_and_value(self, x, memory, memory_mask, memory_indices, action=None):
+        if len(self.obs_shape) > 1:
+            x = self.encoder(x.permute((0, 3, 1, 2)) / 255.0)
+        else:
+            x = self.encoder(x)
+        x, memory = self.transformer(x, memory, memory_mask, memory_indices)
+        x = self.hidden_post_trxl(x)
+        self.x = x
+        probs = [Categorical(logits=branch(x)) for branch in self.actor_branches]
+        if action is None:
+            action = torch.stack([dist.sample() for dist in probs], dim=1)
+        log_probs = []
+        for i, dist in enumerate(probs):
+            log_probs.append(dist.log_prob(action[:, i]))
+        entropies = torch.stack([dist.entropy() for dist in probs], dim=1).sum(1).reshape(-1)
+        return action, torch.stack(log_probs, dim=1), entropies, self.critic(x).flatten(), memory
+
+    def reconstruct_observation(self):
+        x = self.transposed_cnn(self.x)
+        return x.permute((0, 2, 3, 1))
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    # Determine the device to be used for training and set the default tensor type
+    if args.cuda:
+        device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+        torch.set_default_device(device)
+    else:
+        device = torch.device("cpu")
+
+    # Environment setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name) for i in range(args.num_envs)],
+    )
+    observation_space = envs.single_observation_space
+    action_space_shape = (
+        (envs.single_action_space.n,)
+        if isinstance(envs.single_action_space, gym.spaces.Discrete)
+        else tuple(envs.single_action_space.nvec)
+    )
+    env_ids = range(args.num_envs)
+    env_current_episode_step = torch.zeros((args.num_envs,), dtype=torch.long)
+    # Determine maximum episode steps
+    max_episode_steps = envs.envs[0].spec.max_episode_steps
+    if not max_episode_steps:
+        envs.envs[0].reset()  # Memory Gym envs need to be reset before accessing max_episode_steps
+        max_episode_steps = envs.envs[0].max_episode_steps
+    if max_episode_steps <= 0:
+        max_episode_steps = 1024  # Memory Gym envs have max_episode_steps set to -1
+    # Set transformer memory length to max episode steps if greater than max episode steps
+    args.trxl_memory_length = min(args.trxl_memory_length, max_episode_steps)
+
+    agent = Agent(args, observation_space, action_space_shape, max_episode_steps).to(device)
+    optimizer = optim.AdamW(agent.parameters(), lr=args.init_lr)
+    bce_loss = nn.BCELoss()  # Binary cross entropy loss for observation reconstruction
+
+    # ALGO Logic: Storage setup
+    rewards = torch.zeros((args.num_steps, args.num_envs))
+    actions = torch.zeros((args.num_steps, args.num_envs, len(action_space_shape)), dtype=torch.long)
+    dones = torch.zeros((args.num_steps, args.num_envs))
+    obs = torch.zeros((args.num_steps, args.num_envs) + observation_space.shape)
+    log_probs = torch.zeros((args.num_steps, args.num_envs, len(action_space_shape)))
+    values = torch.zeros((args.num_steps, args.num_envs))
+    # The length of stored-memories is equal to the number of sampled episodes during training data sampling
+    # (num_episodes, max_episode_length, num_layers, embed_dim)
+    stored_memories = []
+    # Memory mask used during attention
+    stored_memory_masks = torch.zeros((args.num_steps, args.num_envs, args.trxl_memory_length), dtype=torch.bool)
+    # Index to select the correct episode memory from stored_memories
+    stored_memory_index = torch.zeros((args.num_steps, args.num_envs), dtype=torch.long)
+    # Indices to slice the episode memories into windows
+    stored_memory_indices = torch.zeros((args.num_steps, args.num_envs, args.trxl_memory_length), dtype=torch.long)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    episode_infos = deque(maxlen=100)  # Store episode results for monitoring statistics
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs)
+    # Setup placeholders for each environments's current episodic memory
+    next_memory = torch.zeros((args.num_envs, max_episode_steps, args.trxl_num_layers, args.trxl_dim), dtype=torch.float32)
+    # Generate episodic memory mask used in attention
+    memory_mask = torch.tril(torch.ones((args.trxl_memory_length, args.trxl_memory_length)), diagonal=-1)
+    """ e.g. memory mask tensor looks like this if memory_length = 6
+    0, 0, 0, 0, 0, 0
+    1, 0, 0, 0, 0, 0
+    1, 1, 0, 0, 0, 0
+    1, 1, 1, 0, 0, 0
+    1, 1, 1, 1, 0, 0
+    1, 1, 1, 1, 1, 0
+    """
+    # Setup memory window indices to support a sliding window over the episodic memory
+    repetitions = torch.repeat_interleave(
+        torch.arange(0, args.trxl_memory_length).unsqueeze(0), args.trxl_memory_length - 1, dim=0
+    ).long()
+    memory_indices = torch.stack(
+        [torch.arange(i, i + args.trxl_memory_length) for i in range(max_episode_steps - args.trxl_memory_length + 1)]
+    ).long()
+    memory_indices = torch.cat((repetitions, memory_indices))
+    """ e.g. the memory window indices tensor looks like this if memory_length = 4 and max_episode_length = 7:
+    0, 1, 2, 3
+    0, 1, 2, 3
+    0, 1, 2, 3
+    0, 1, 2, 3
+    1, 2, 3, 4
+    2, 3, 4, 5
+    3, 4, 5, 6
+    """
+
+    for iteration in range(1, args.num_iterations + 1):
+        sampled_episode_infos = []
+
+        # Annealing the learning rate and entropy coefficient if instructed to do so
+        do_anneal = args.anneal_steps > 0 and global_step < args.anneal_steps
+        frac = 1 - global_step / args.anneal_steps if do_anneal else 0
+        lr = (args.init_lr - args.final_lr) * frac + args.final_lr
+        for param_group in optimizer.param_groups:
+            param_group["lr"] = lr
+        ent_coef = (args.init_ent_coef - args.final_ent_coef) * frac + args.final_ent_coef
+
+        # Init episodic memory buffer using each environments' current episodic memory
+        stored_memories = [next_memory[e] for e in range(args.num_envs)]
+        for e in range(args.num_envs):
+            stored_memory_index[:, e] = e
+
+        for step in range(args.num_steps):
+            global_step += args.num_envs
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                obs[step] = next_obs
+                dones[step] = next_done
+                stored_memory_masks[step] = memory_mask[torch.clip(env_current_episode_step, 0, args.trxl_memory_length - 1)]
+                stored_memory_indices[step] = memory_indices[env_current_episode_step]
+                # Retrieve the memory window from the entire episodic memory
+                memory_window = batched_index_select(next_memory, 1, stored_memory_indices[step])
+                action, logprob, _, value, new_memory = agent.get_action_and_value(
+                    next_obs, memory_window, stored_memory_masks[step], stored_memory_indices[step]
+                )
+                next_memory[env_ids, env_current_episode_step] = new_memory
+                # Store the action, log_prob, and value in the buffer
+                actions[step], log_probs[step], values[step] = action, logprob, value
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            # Reset and process episodic memory if done
+            for id, done in enumerate(next_done):
+                if done:
+                    # Reset the environment's current timestep
+                    env_current_episode_step[id] = 0
+                    # Break the reference to the environment's episodic memory
+                    mem_index = stored_memory_index[step, id]
+                    stored_memories[mem_index] = stored_memories[mem_index].clone()
+                    # Reset episodic memory
+                    next_memory[id] = torch.zeros(
+                        (max_episode_steps, args.trxl_num_layers, args.trxl_dim), dtype=torch.float32
+                    )
+                    if step < args.num_steps - 1:
+                        # Store memory inside the buffer
+                        stored_memories.append(next_memory[id])
+                        # Store the reference of to the current episodic memory inside the buffer
+                        stored_memory_index[step + 1 :, id] = len(stored_memories) - 1
+                else:
+                    # Increment environment timestep if not done
+                    env_current_episode_step[id] += 1
+
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        sampled_episode_infos.append(info["episode"])
+
+        # Bootstrap value if not done
+        with torch.no_grad():
+            start = torch.clip(env_current_episode_step - args.trxl_memory_length, 0)
+            end = torch.clip(env_current_episode_step, args.trxl_memory_length)
+            indices = torch.stack([torch.arange(start[b], end[b]) for b in range(args.num_envs)]).long()
+            memory_window = batched_index_select(next_memory, 1, indices)  # Retrieve the memory window from the entire episode
+            next_value = agent.get_value(
+                next_obs,
+                memory_window,
+                memory_mask[torch.clip(env_current_episode_step, 0, args.trxl_memory_length - 1)],
+                stored_memory_indices[-1],
+            )
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # Flatten the batch
+        b_obs = obs.reshape(-1, *obs.shape[2:])
+        b_logprobs = log_probs.reshape(-1, *log_probs.shape[2:])
+        b_actions = actions.reshape(-1, *actions.shape[2:])
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+        b_memory_index = stored_memory_index.reshape(-1)
+        b_memory_indices = stored_memory_indices.reshape(-1, *stored_memory_indices.shape[2:])
+        b_memory_mask = stored_memory_masks.reshape(-1, *stored_memory_masks.shape[2:])
+        stored_memories = torch.stack(stored_memories, dim=0)
+
+        # Remove unnecessary padding from TrXL memory, if applicable
+        actual_max_episode_steps = (stored_memory_indices * stored_memory_masks).max().item() + 1
+        if actual_max_episode_steps < args.trxl_memory_length:
+            b_memory_indices = b_memory_indices[:, :actual_max_episode_steps]
+            b_memory_mask = b_memory_mask[:, :actual_max_episode_steps]
+            stored_memories = stored_memories[:, :actual_max_episode_steps]
+
+        # Optimizing the policy and value network
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            b_inds = torch.randperm(args.batch_size)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+                mb_memories = stored_memories[b_memory_index[mb_inds]]
+                mb_memory_windows = batched_index_select(mb_memories, 1, b_memory_indices[mb_inds])
+
+                _, newlogprob, entropy, newvalue, _ = agent.get_action_and_value(
+                    b_obs[mb_inds], mb_memory_windows, b_memory_mask[mb_inds], b_memory_indices[mb_inds], b_actions[mb_inds]
+                )
+
+                # Policy loss
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+                mb_advantages = mb_advantages.unsqueeze(1).repeat(
+                    1, len(action_space_shape)
+                )  # Repeat is necessary for multi-discrete action spaces
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = torch.exp(logratio)
+                pgloss1 = -mb_advantages * ratio
+                pgloss2 = -mb_advantages * torch.clamp(ratio, 1.0 - args.clip_coef, 1.0 + args.clip_coef)
+                pg_loss = torch.max(pgloss1, pgloss2).mean()
+
+                # Value loss
+                v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                if args.clip_vloss:
+                    v_loss_clipped = b_values[mb_inds] + (newvalue - b_values[mb_inds]).clamp(
+                        min=-args.clip_coef, max=args.clip_coef
+                    )
+                    v_loss = torch.max(v_loss_unclipped, (v_loss_clipped - b_returns[mb_inds]) ** 2).mean()
+                else:
+                    v_loss = v_loss_unclipped.mean()
+
+                # Entropy loss
+                entropy_loss = entropy.mean()
+
+                # Combined losses
+                loss = pg_loss - ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                # Add reconstruction loss if used
+                r_loss = torch.tensor(0.0)
+                if args.reconstruction_coef > 0.0:
+                    r_loss = bce_loss(agent.reconstruct_observation(), b_obs[mb_inds] / 255.0)
+                    loss += args.reconstruction_coef * r_loss
+
+                optimizer.zero_grad()
+                loss.backward()
+                torch.nn.utils.clip_grad_norm_(agent.parameters(), max_norm=args.max_grad_norm)
+                optimizer.step()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # Log and monitor training statistics
+        episode_infos.extend(sampled_episode_infos)
+        episode_result = {}
+        if len(episode_infos) > 0:
+            for key in episode_infos[0].keys():
+                episode_result[key + "_mean"] = np.mean([info[key] for info in episode_infos])
+
+        print(
+            "{:9} SPS={:4} return={:.2f} length={:.1f} pi_loss={:.3f} v_loss={:.3f} entropy={:.3f} r_loss={:.3f} value={:.3f} adv={:.3f}".format(
+                iteration,
+                int(global_step / (time.time() - start_time)),
+                episode_result["r_mean"],
+                episode_result["l_mean"],
+                pg_loss.item(),
+                v_loss.item(),
+                entropy_loss.item(),
+                r_loss.item(),
+                torch.mean(values),
+                torch.mean(advantages),
+            )
+        )
+
+        if episode_result:
+            for key in episode_result:
+                writer.add_scalar("episode/" + key, episode_result[key], global_step)
+        writer.add_scalar("episode/value_mean", torch.mean(values), global_step)
+        writer.add_scalar("episode/advantage_mean", torch.mean(advantages), global_step)
+        writer.add_scalar("charts/learning_rate", lr, global_step)
+        writer.add_scalar("charts/entropy_coefficient", ent_coef, global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/loss", loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/reconstruction_loss", r_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        model_data = {
+            "model_weights": agent.state_dict(),
+            "args": vars(args),
+        }
+        torch.save(model_data, model_path)
+        print(f"model saved to {model_path}")
+
+    writer.close()
+    envs.close()
diff --git a/cleanrl/cleanrl/ppo_trxl/pyproject.toml b/cleanrl/cleanrl/ppo_trxl/pyproject.toml
new file mode 100644
index 0000000000000000000000000000000000000000..22626cc491f23e400f34cc96c415b0ce29f32068
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_trxl/pyproject.toml
@@ -0,0 +1,34 @@
+[project]
+name = "ppo-trxl"
+version = "1.0.1"
+description = ""
+authors = [{ name = "Marco Pleines", email = "marco.pleines@tu-dortmund.de" }]
+requires-python = "~=3.10"
+license = "MIT"
+dependencies = [
+    "memory-gym>=1.0.2,<2",
+    "einops>=0.7.0,<0.8",
+    "minigrid>=2.3.1,<3",
+    "reprint>=0.6.0,<0.7",
+    "opencv-python>=4.9.0.80,<5",
+    "torch>=2.0.0,<3",
+    "torchaudio>=2.0.0,<3",
+    "wandb>=0.16.6,<0.17",
+    "tyro>=0.8.3,<0.9",
+    "tensorboard>=2.16.2,<3",
+]
+
+[tool.uv]
+
+[[tool.uv.index]]
+name = "pytorch"
+url = "https://download.pytorch.org/whl/cu118"
+explicit = true
+
+[tool.uv.sources]
+torch = { index = "pytorch" }
+torchaudio = { index = "pytorch" }
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
diff --git a/cleanrl/cleanrl/ppo_ultrahorizon.py b/cleanrl/cleanrl/ppo_ultrahorizon.py
new file mode 100644
index 0000000000000000000000000000000000000000..3901852dd6e8abbc363c764cc68df1398658351f
--- /dev/null
+++ b/cleanrl/cleanrl/ppo_ultrahorizon.py
@@ -0,0 +1,490 @@
+# PPO implementation for Ultrahorizon Grid Environment
+# Adapted from ppo_atari.py for text-based grid environment
+import os
+import random
+import time
+from dataclasses import dataclass
+import sys
+
+# Add parent directory to path to import the wrapper
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+from tqdm import tqdm
+from ultrahorizon_gym_wrapper import make_ultrahorizon_env, Difficulty
+import json
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = True
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "Ultrahorizon-v0"
+    """the id of the environment"""
+    difficulty: str = "EASY"
+    """difficulty level: EASY, MEDIUM, or HARD"""
+    total_timesteps: int = 10000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 4
+    """the number of parallel game environments"""
+    num_steps: int = 512
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.2
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.01
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    """
+    Neural network agent for Ultrahorizon environment.
+    Uses MLP instead of CNN since observations are feature vectors, not images.
+    """
+    def __init__(self, envs):
+        super().__init__()
+        # Get observation space dimension
+        obs_dim = np.array(envs.single_observation_space.shape).prod()
+        
+        # MLP network for processing feature vectors
+        self.network = nn.Sequential(
+            layer_init(nn.Linear(obs_dim, 256)),
+            nn.ReLU(),
+            layer_init(nn.Linear(256, 256)),
+            nn.ReLU(),
+            layer_init(nn.Linear(256, 128)),
+            nn.ReLU(),
+        )
+        self.actor = layer_init(nn.Linear(128, envs.single_action_space.n), std=0.01)
+        self.critic = layer_init(nn.Linear(128, 1), std=1)
+
+    def get_value(self, x):
+        return self.critic(self.network(x))
+
+    def get_action_and_value(self, x, action=None):
+        hidden = self.network(x)
+        logits = self.actor(hidden)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(hidden)
+
+
+def test_policy(agent, difficulty, device, num_episodes=10, seed=None):
+    """
+    Test the current policy and collect trajectories and rewards.
+    
+    Args:
+        agent: The trained agent
+        difficulty: Difficulty level for the environment
+        device: torch device
+        num_episodes: Number of episodes to test
+        seed: Random seed for testing
+    
+    Returns:
+        dict: Dictionary containing test results including trajectories and rewards
+    """
+    # Create a single test environment
+    test_env = make_ultrahorizon_env(difficulty=difficulty, free=True)()
+    
+    test_results = {
+        'episode_returns': [],
+        'episode_lengths': [],
+        'final_scores': [],
+        'trajectories': []  # Store trajectories for each episode
+    }
+    
+    agent.eval()  # Set agent to evaluation mode
+    
+    for episode in range(num_episodes):
+        obs, _ = test_env.reset(seed=seed + episode if seed is not None else None)
+        done = False
+        episode_return = 0
+        episode_length = 0
+        trajectory = {
+            'observations': [],
+            'actions': [],
+            'rewards': [],
+            'dones': []
+        }
+        
+        while not done:
+            # Store observation
+            trajectory['observations'].append(obs.tolist())
+            
+            # Get action from policy
+            with torch.no_grad():
+                obs_tensor = torch.Tensor(obs).unsqueeze(0).to(device)
+                action, _, _, _ = agent.get_action_and_value(obs_tensor)
+                action = action.cpu().numpy()[0]
+            
+            # Take action in environment
+            next_obs, reward, terminated, truncated, info = test_env.step(action)
+            done = terminated or truncated
+            
+            # Store trajectory data
+            trajectory['actions'].append(int(action))
+            trajectory['rewards'].append(float(reward))
+            trajectory['dones'].append(done)
+            
+            episode_return += reward
+            episode_length += 1
+            obs = next_obs
+        
+        # Store episode results
+        test_results['episode_returns'].append(float(episode_return))
+        test_results['episode_lengths'].append(episode_length)
+        test_results['final_scores'].append(float(info.get('final_score', 0.0)))
+        test_results['trajectories'].append(trajectory)
+    
+    agent.train()  # Set agent back to training mode
+    test_env.close()
+    
+    # Calculate statistics
+    test_results['mean_return'] = np.mean(test_results['episode_returns'])
+    test_results['std_return'] = np.std(test_results['episode_returns'])
+    test_results['mean_length'] = np.mean(test_results['episode_lengths'])
+    test_results['mean_final_score'] = np.mean(test_results['final_scores'])
+    test_results['std_final_score'] = np.std(test_results['final_scores'])
+    
+    return test_results
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}_{args.difficulty.upper()}__{int(time.time())}"
+    
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup - convert difficulty string to enum
+    difficulty_map = {
+        "EASY": Difficulty.EASY,
+        "MEDIUM": Difficulty.MEDIUM,
+        "HARD": Difficulty.HARD
+    }
+
+    difficulty = difficulty_map.get(args.difficulty.upper(), Difficulty.EASY)
+    
+    envs = gym.vector.SyncVectorEnv(
+        [make_ultrahorizon_env(difficulty=difficulty, free=True) for i in range(args.num_envs)],
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    
+    # Setup for testing at 10% intervals
+    test_interval = args.num_iterations // 10
+    test_results_history = []  # Store all test results
+    test_save_dir = f"runs/{run_name}/test_results"
+    os.makedirs(test_save_dir, exist_ok=True)
+
+    for iteration in tqdm(range(1, args.num_iterations + 1)):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            # Handle terminal observations correctly
+            # When an episode ends, next_obs is the initial state of the new episode
+            # We need to use the final observation from the terminated episode for bootstrapping
+            if "final_observation" in infos:
+                for idx, final_obs in enumerate(infos["final_observation"]):
+                    if final_obs is not None:
+                        # Replace the reset observation with the actual final observation
+                        next_obs[idx] = torch.Tensor(final_obs).to(device)
+
+            # import pdb;pdb.set_trace()
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        episode_return = info['episode']['r']
+                        episode_length = info['episode']['l']
+                        final_score = info.get('final_score', 0.0)  # Get cumulative score from environment
+                        # import pdb;pdb.set_trace()
+                        # print(f"global_step={global_step}, episodic_return={episode_return}, episodic_length={episode_length},final_score={final_score}")
+                        writer.add_scalar("charts/episodic_return", episode_return, global_step)
+                        writer.add_scalar("charts/episodic_length", episode_length, global_step)
+                        writer.add_scalar("charts/final_score", final_score, global_step)
+
+        # import pdb;pdb.set_trace()
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped)
+                    v_loss = 0.5 * v_loss_max.mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        
+        # Additional useful metrics
+        writer.add_scalar("charts/avg_reward", rewards.mean().item(), global_step)
+        writer.add_scalar("charts/avg_value", values.mean().item(), global_step)
+        writer.add_scalar("charts/max_reward", rewards.max().item(), global_step)
+        writer.add_scalar("charts/min_reward", rewards.min().item(), global_step)
+        
+        # Console output with key metrics
+        sps = int(global_step / (time.time() - start_time))
+        print(f"Iter {iteration}/{args.num_iterations} | SPS: {sps} | "
+              f"Avg Reward: {rewards.mean().item():.3f} | "
+              f"Value Loss: {v_loss.item():.4f} | "
+              f"Policy Loss: {pg_loss.item():.4f} | "
+              f"Entropy: {entropy_loss.item():.4f}")
+        print(f"global_step={global_step}, episodic_return={episode_return}, episodic_length={episode_length},final_score={final_score}")
+        writer.add_scalar("charts/SPS", sps, global_step)
+        
+        # Test policy at 10% intervals
+        if iteration % test_interval == 0 or iteration == args.num_iterations:
+            print(f"\n{'='*60}")
+            print(f"Testing policy at iteration {iteration}/{args.num_iterations} ({100*iteration//args.num_iterations}%)")
+            print(f"{'='*60}")
+            
+            test_results = test_policy(
+                agent=agent,
+                difficulty=difficulty,
+                device=device,
+                num_episodes=10,
+                seed=args.seed
+            )
+            
+            # Log test results to tensorboard
+            writer.add_scalar("test/mean_return", test_results['mean_return'], global_step)
+            writer.add_scalar("test/std_return", test_results['std_return'], global_step)
+            writer.add_scalar("test/mean_length", test_results['mean_length'], global_step)
+            writer.add_scalar("test/mean_final_score", test_results['mean_final_score'], global_step)
+            writer.add_scalar("test/std_final_score", test_results['std_final_score'], global_step)
+            
+            # Print test results
+            print(f"Test Results (10 episodes):")
+            print(f"  Mean Return: {test_results['mean_return']:.2f} ± {test_results['std_return']:.2f}")
+            print(f"  Mean Length: {test_results['mean_length']:.2f}")
+            print(f"  Mean Final Score: {test_results['mean_final_score']:.2f} ± {test_results['std_final_score']:.2f}")
+            print(f"{'='*60}\n")
+            
+            # Save test results to file
+            test_results_with_metadata = {
+                'iteration': iteration,
+                'global_step': global_step,
+                'training_progress': iteration / args.num_iterations,
+                'test_results': test_results
+            }
+            test_results_history.append(test_results_with_metadata)
+            
+            # Save individual test result
+            test_file = f"{test_save_dir}/test_iter_{iteration}.json"
+            with open(test_file, 'w') as f:
+                json.dump(test_results_with_metadata, f, indent=2)
+            print(f"Saved test results to {test_file}")
+
+    envs.close()
+    writer.close()
+    torch.save(agent.state_dict(), f"runs/{run_name}/agent_{global_step}.pt")
+    print(f"Saved model at global_step={global_step}")
+    
+    # Save all test results history
+    test_history_file = f"{test_save_dir}/test_history.json"
+    with open(test_history_file, 'w') as f:
+        json.dump(test_results_history, f, indent=2)
+    print(f"Saved complete test history to {test_history_file}")
diff --git a/cleanrl/cleanrl/pqn.py b/cleanrl/cleanrl/pqn.py
new file mode 100644
index 0000000000000000000000000000000000000000..fd0eb8e21620142b7066482eabe41b7f4fa1fb26
--- /dev/null
+++ b/cleanrl/cleanrl/pqn.py
@@ -0,0 +1,248 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/pqn/#pqnpy
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "CartPole-v1"
+    """the id of the environment"""
+    total_timesteps: int = 500000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 4
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run for each environment per update"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    start_e: float = 1
+    """the starting epsilon for exploration"""
+    end_e: float = 0.05
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.5
+    """the fraction of `total_timesteps` it takes from start_e to end_e"""
+    max_grad_norm: float = 10.0
+    """the maximum norm for the gradient clipping"""
+    q_lambda: float = 0.65
+    """the lambda for Q(lambda)"""
+
+
+def make_env(env_id, seed, idx, capture_video, run_name):
+    def thunk():
+        if capture_video and idx == 0:
+            env = gym.make(env_id, render_mode="rgb_array")
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        else:
+            env = gym.make(env_id)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        env.action_space.seed(seed)
+
+        return env
+
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    def __init__(self, env):
+        super().__init__()
+
+        self.network = nn.Sequential(
+            layer_init(nn.Linear(np.array(env.single_observation_space.shape).prod(), 120)),
+            nn.LayerNorm(120),
+            nn.ReLU(),
+            layer_init(nn.Linear(120, 84)),
+            nn.LayerNorm(84),
+            nn.ReLU(),
+            layer_init(nn.Linear(84, env.single_action_space.n)),
+        )
+
+    def forward(self, x):
+        return self.network(x)
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    # agent setup
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.RAdam(q_network.parameters(), lr=args.learning_rate)
+
+    # storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+            random_actions = torch.randint(0, envs.single_action_space.n, (args.num_envs,)).to(device)
+            with torch.no_grad():
+                q_values = q_network(next_obs)
+                max_actions = torch.argmax(q_values, dim=1)
+                values[step] = q_values[torch.arange(args.num_envs), max_actions].flatten()
+
+            explore = torch.rand((args.num_envs,)).to(device) < epsilon
+            action = torch.where(explore, random_actions, max_actions)
+            actions[step] = action
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                        writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                        writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+
+        # Compute Q(lambda) targets
+        with torch.no_grad():
+            returns = torch.zeros_like(rewards).to(device)
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    next_value, _ = torch.max(q_network(next_obs), dim=-1)
+                    nextnonterminal = 1.0 - next_done
+                    returns[t] = rewards[t] + args.gamma * next_value * nextnonterminal
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    next_value = values[t + 1]
+                    returns[t] = (
+                        rewards[t]
+                        + args.gamma * (args.q_lambda * returns[t + 1] + (1 - args.q_lambda) * next_value) * nextnonterminal
+                    )
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_returns = returns.reshape(-1)
+
+        # Optimizing the Q-network
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                old_val = q_network(b_obs[mb_inds]).gather(1, b_actions[mb_inds].unsqueeze(-1).long()).squeeze()
+                loss = F.mse_loss(b_returns[mb_inds], old_val)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(q_network.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+        writer.add_scalar("losses/td_loss", loss, global_step)
+        writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+        print("SPS:", int(global_step / (time.time() - start_time)))
+        writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/pqn_atari_envpool_lstm.py b/cleanrl/cleanrl/pqn_atari_envpool_lstm.py
new file mode 100644
index 0000000000000000000000000000000000000000..10d0de4e547981ed44b4784e4c2acf996a090bd5
--- /dev/null
+++ b/cleanrl/cleanrl/pqn_atari_envpool_lstm.py
@@ -0,0 +1,339 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/pqn/#pqn_atari_envpool_lstmpy
+import os
+import random
+import time
+from collections import deque
+from dataclasses import dataclass
+
+import envpool
+import gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "Breakout-v5"
+    """the id of the environment"""
+    total_timesteps: int = 10000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 8
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    start_e: float = 1
+    """the starting epsilon for exploration"""
+    end_e: float = 0.01
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.10
+    """the fraction of `total_timesteps` it takes from start_e to end_e"""
+    q_lambda: float = 0.65
+    """the lambda for the Q-Learning algorithm"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+class RecordEpisodeStatistics(gym.Wrapper):
+    def __init__(self, env, deque_size=100):
+        super().__init__(env)
+        self.num_envs = getattr(env, "num_envs", 1)
+        self.episode_returns = None
+        self.episode_lengths = None
+
+    def reset(self, **kwargs):
+        observations = super().reset(**kwargs)
+        self.episode_returns = np.zeros(self.num_envs, dtype=np.float32)
+        self.episode_lengths = np.zeros(self.num_envs, dtype=np.int32)
+        self.lives = np.zeros(self.num_envs, dtype=np.int32)
+        self.returned_episode_returns = np.zeros(self.num_envs, dtype=np.float32)
+        self.returned_episode_lengths = np.zeros(self.num_envs, dtype=np.int32)
+        return observations
+
+    def step(self, action):
+        observations, rewards, dones, infos = super().step(action)
+        self.episode_returns += infos["reward"]
+        self.episode_lengths += 1
+        self.returned_episode_returns[:] = self.episode_returns
+        self.returned_episode_lengths[:] = self.episode_lengths
+        self.episode_returns *= 1 - infos["terminated"]
+        self.episode_lengths *= 1 - infos["terminated"]
+        infos["r"] = self.returned_episode_returns
+        infos["l"] = self.returned_episode_lengths
+        return (
+            observations,
+            rewards,
+            dones,
+            infos,
+        )
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QNetwork(nn.Module):
+    def __init__(self, env):
+        super().__init__()
+        self.network = nn.Sequential(
+            layer_init(nn.Conv2d(1, 32, 8, stride=4)),
+            nn.LayerNorm([32, 20, 20]),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(32, 64, 4, stride=2)),
+            nn.LayerNorm([64, 9, 9]),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 64, 3, stride=1)),
+            nn.LayerNorm([64, 7, 7]),
+            nn.ReLU(),
+            nn.Flatten(),
+            layer_init(nn.Linear(3136, 512)),
+            nn.LayerNorm(512),
+            nn.ReLU(),
+        )
+        self.lstm = nn.LSTM(512, 128)
+        for name, param in self.lstm.named_parameters():
+            if "bias" in name:
+                nn.init.constant_(param, 0)
+            elif "weight" in name:
+                nn.init.orthogonal_(param, 1.0)
+        self.q_func = layer_init(nn.Linear(128, env.single_action_space.n))
+
+    def get_states(self, x, lstm_state, done):
+        hidden = self.network(x / 255.0)
+
+        # LSTM logic
+        batch_size = lstm_state[0].shape[1]
+        hidden = hidden.reshape((-1, batch_size, self.lstm.input_size))
+        done = done.reshape((-1, batch_size))
+        new_hidden = []
+        for h, d in zip(hidden, done):
+            h, lstm_state = self.lstm(
+                h.unsqueeze(0),
+                (
+                    (1.0 - d).view(1, -1, 1) * lstm_state[0],
+                    (1.0 - d).view(1, -1, 1) * lstm_state[1],
+                ),
+            )
+            new_hidden += [h]
+        new_hidden = torch.flatten(torch.cat(new_hidden), 0, 1)
+        return new_hidden, lstm_state
+
+    def forward(self, x, lstm_state, done):
+        hidden, lstm_state = self.get_states(x, lstm_state, done)
+        return self.q_func(hidden), lstm_state
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = envpool.make(
+        args.env_id,
+        env_type="gym",
+        num_envs=args.num_envs,
+        episodic_life=True,
+        reward_clip=True,
+        seed=args.seed,
+        stack_num=1,
+    )
+    envs.num_envs = args.num_envs
+    envs.single_action_space = envs.action_space
+    envs.single_observation_space = envs.observation_space
+    envs = RecordEpisodeStatistics(envs)
+    assert isinstance(envs.action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.RAdam(q_network.parameters(), lr=args.learning_rate)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    avg_returns = deque(maxlen=20)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs = torch.Tensor(envs.reset()).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    next_lstm_state = (
+        torch.zeros(q_network.lstm.num_layers, args.num_envs, q_network.lstm.hidden_size).to(device),
+        torch.zeros(q_network.lstm.num_layers, args.num_envs, q_network.lstm.hidden_size).to(device),
+    )  # hidden and cell states (see https://youtu.be/8HyCNIVRbSU)
+
+    for iteration in range(1, args.num_iterations + 1):
+        initial_lstm_state = (next_lstm_state[0].clone(), next_lstm_state[1].clone())
+
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+
+            random_actions = torch.randint(0, envs.single_action_space.n, (args.num_envs,)).to(device)
+            with torch.no_grad():
+                q_values, next_lstm_state = q_network(next_obs, next_lstm_state, next_done)
+                max_actions = torch.argmax(q_values, dim=1)
+                values[step] = q_values[torch.arange(args.num_envs), max_actions].flatten()
+
+            explore = torch.rand((args.num_envs,)).to(device) < epsilon
+            action = torch.where(explore, random_actions, max_actions)
+            actions[step] = action
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, next_done, info = envs.step(action.cpu().numpy())
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            for idx, d in enumerate(next_done):
+                if d and info["lives"][idx] == 0:
+                    print(f"global_step={global_step}, episodic_return={info['r'][idx]}")
+                    avg_returns.append(info["r"][idx])
+                    writer.add_scalar("charts/avg_episodic_return", np.average(avg_returns), global_step)
+                    writer.add_scalar("charts/episodic_return", info["r"][idx], global_step)
+                    writer.add_scalar("charts/episodic_length", info["l"][idx], global_step)
+
+        # Compute Q(lambda) targets
+        with torch.no_grad():
+            returns = torch.zeros_like(rewards).to(device)
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    next_value, _ = torch.max(q_network(next_obs, next_lstm_state, next_done)[0], dim=-1)
+                    nextnonterminal = 1.0 - next_done
+                    returns[t] = rewards[t] + args.gamma * next_value * nextnonterminal
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    next_value = values[t + 1]
+                    returns[t] = (
+                        rewards[t]
+                        + args.gamma * (args.q_lambda * returns[t + 1] + (1 - args.q_lambda) * next_value) * nextnonterminal
+                    )
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_returns = returns.reshape(-1)
+        b_dones = dones.reshape(-1)
+
+        assert args.num_envs % args.num_minibatches == 0
+        envsperbatch = args.num_envs // args.num_minibatches
+        envinds = np.arange(args.num_envs)
+        flatinds = np.arange(args.batch_size).reshape(args.num_steps, args.num_envs)
+
+        # Optimizing the Q-network
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(envinds)
+            for start in range(0, args.num_envs, envsperbatch):
+                end = start + envsperbatch
+                mbenvinds = envinds[start:end]
+                mb_inds = flatinds[:, mbenvinds].ravel()  # be really careful about the index
+
+                old_val, _ = q_network(
+                    b_obs[mb_inds],
+                    (initial_lstm_state[0][:, mbenvinds], initial_lstm_state[1][:, mbenvinds]),
+                    b_dones[mb_inds],
+                )
+                old_val = old_val.gather(1, b_actions[mb_inds].unsqueeze(-1).long()).squeeze()
+
+                loss = F.mse_loss(b_returns[mb_inds], old_val)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(q_network.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+        writer.add_scalar("losses/td_loss", loss, global_step)
+        writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+        print("SPS:", int(global_step / (time.time() - start_time)))
+        writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/qdagger_dqn_atari_impalacnn.py b/cleanrl/cleanrl/qdagger_dqn_atari_impalacnn.py
new file mode 100644
index 0000000000000000000000000000000000000000..fe9e5abf619f3915cf68d870e44b8d95764cf738
--- /dev/null
+++ b/cleanrl/cleanrl/qdagger_dqn_atari_impalacnn.py
@@ -0,0 +1,466 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/qdagger/#qdagger_dqn_atari_jax_impalacnnpy
+import os
+import random
+import time
+from collections import deque
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from huggingface_hub import hf_hub_download
+from rich.progress import track
+from torch.utils.tensorboard import SummaryWriter
+
+from cleanrl.dqn_atari import QNetwork as TeacherModel
+from cleanrl_utils.atari_wrappers import (
+    ClipRewardEnv,
+    EpisodicLifeEnv,
+    FireResetEnv,
+    MaxAndSkipEnv,
+    NoopResetEnv,
+)
+from cleanrl_utils.buffers import ReplayBuffer
+from cleanrl_utils.evals.dqn_eval import evaluate
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+    upload_model: bool = False
+    """whether to upload the saved model to huggingface"""
+    hf_entity: str = ""
+    """the user or org name of the model repository from the Hugging Face Hub"""
+
+    # Algorithm specific arguments
+    env_id: str = "BreakoutNoFrameskip-v4"
+    """the id of the environment"""
+    total_timesteps: int = 10000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 1e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 1
+    """the number of parallel game environments"""
+    buffer_size: int = 1000000
+    """the replay memory buffer size"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    tau: float = 1.0
+    """the target network update rate"""
+    target_network_frequency: int = 1000
+    """the timesteps it takes to update the target network"""
+    batch_size: int = 32
+    """the batch size of sample from the reply memory"""
+    start_e: float = 1.0
+    """the starting epsilon for exploration"""
+    end_e: float = 0.01
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.10
+    """the fraction of `total-timesteps` it takes from start-e to go end-e"""
+    learning_starts: int = 80000
+    """timestep to start learning"""
+    train_frequency: int = 4
+    """the frequency of training"""
+
+    # QDagger specific arguments
+    teacher_policy_hf_repo: str = None
+    """the huggingface repo of the teacher policy"""
+    teacher_model_exp_name: str = "dqn_atari"
+    """the experiment name of the teacher model"""
+    teacher_eval_episodes: int = 10
+    """the number of episodes to run the teacher policy evaluate"""
+    teacher_steps: int = 500000
+    """the number of steps to run the teacher policy to generate the replay buffer"""
+    offline_steps: int = 500000
+    """the number of steps to run the student policy with the teacher's replay buffer"""
+    temperature: float = 1.0
+    """the temperature parameter for qdagger"""
+
+
+def make_env(env_id, seed, idx, capture_video, run_name):
+    def thunk():
+        if capture_video and idx == 0:
+            env = gym.make(env_id, render_mode="rgb_array")
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        else:
+            env = gym.make(env_id)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        env = NoopResetEnv(env, noop_max=30)
+        env = MaxAndSkipEnv(env, skip=4)
+        env = EpisodicLifeEnv(env)
+        if "FIRE" in env.unwrapped.get_action_meanings():
+            env = FireResetEnv(env)
+        env = ClipRewardEnv(env)
+        env = gym.wrappers.ResizeObservation(env, (84, 84))
+        env = gym.wrappers.GrayScaleObservation(env)
+        env = gym.wrappers.FrameStack(env, 4)
+        env.action_space.seed(seed)
+
+        return env
+
+    return thunk
+
+
+# taken from https://github.com/AIcrowd/neurips2020-procgen-starter-kit/blob/142d09586d2272a17f44481a115c4bd817cf6a94/models/impala_cnn_torch.py
+class ResidualBlock(nn.Module):
+    def __init__(self, channels):
+        super().__init__()
+        self.conv0 = nn.Conv2d(in_channels=channels, out_channels=channels, kernel_size=3, padding=1)
+        self.conv1 = nn.Conv2d(in_channels=channels, out_channels=channels, kernel_size=3, padding=1)
+
+    def forward(self, x):
+        inputs = x
+        x = nn.functional.relu(x)
+        x = self.conv0(x)
+        x = nn.functional.relu(x)
+        x = self.conv1(x)
+        return x + inputs
+
+
+class ConvSequence(nn.Module):
+    def __init__(self, input_shape, out_channels):
+        super().__init__()
+        self._input_shape = input_shape
+        self._out_channels = out_channels
+        self.conv = nn.Conv2d(in_channels=self._input_shape[0], out_channels=self._out_channels, kernel_size=3, padding=1)
+        self.res_block0 = ResidualBlock(self._out_channels)
+        self.res_block1 = ResidualBlock(self._out_channels)
+
+    def forward(self, x):
+        x = self.conv(x)
+        x = nn.functional.max_pool2d(x, kernel_size=3, stride=2, padding=1)
+        x = self.res_block0(x)
+        x = self.res_block1(x)
+        assert x.shape[1:] == self.get_output_shape()
+        return x
+
+    def get_output_shape(self):
+        _c, h, w = self._input_shape
+        return (self._out_channels, (h + 1) // 2, (w + 1) // 2)
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    def __init__(self, env):
+        super().__init__()
+        c, h, w = envs.single_observation_space.shape
+        shape = (c, h, w)
+        conv_seqs = []
+        for out_channels in [16, 32, 32]:
+            conv_seq = ConvSequence(shape, out_channels)
+            shape = conv_seq.get_output_shape()
+            conv_seqs.append(conv_seq)
+        conv_seqs += [
+            nn.Flatten(),
+            nn.ReLU(),
+            nn.Linear(in_features=shape[0] * shape[1] * shape[2], out_features=256),
+            nn.ReLU(),
+            nn.Linear(in_features=256, out_features=env.single_action_space.n),
+        ]
+        self.network = nn.Sequential(*conv_seqs)
+
+    def forward(self, x):
+        return self.network(x / 255.0)
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+def kl_divergence_with_logits(target_logits, prediction_logits):
+    """Implementation of on-policy distillation loss."""
+    out = -F.softmax(target_logits, dim=-1) * (F.log_softmax(prediction_logits, dim=-1) - F.log_softmax(target_logits, dim=-1))
+    return torch.sum(out)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    assert args.num_envs == 1, "vectorized envs are not supported at the moment"
+    if args.teacher_policy_hf_repo is None:
+        args.teacher_policy_hf_repo = f"cleanrl/{args.env_id}-{args.teacher_model_exp_name}-seed1"
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.Adam(q_network.parameters(), lr=args.learning_rate)
+    target_network = QNetwork(envs).to(device)
+    target_network.load_state_dict(q_network.state_dict())
+
+    # QDAGGER LOGIC:
+    teacher_model_path = hf_hub_download(
+        repo_id=args.teacher_policy_hf_repo, filename=f"{args.teacher_model_exp_name}.cleanrl_model"
+    )
+    teacher_model = TeacherModel(envs).to(device)
+    teacher_model.load_state_dict(torch.load(teacher_model_path, map_location=device))
+    teacher_model.eval()
+
+    # evaluate the teacher model
+    teacher_episodic_returns = evaluate(
+        teacher_model_path,
+        make_env,
+        args.env_id,
+        eval_episodes=args.teacher_eval_episodes,
+        run_name=f"{run_name}-teacher-eval",
+        Model=TeacherModel,
+        epsilon=args.end_e,
+        capture_video=False,
+    )
+    writer.add_scalar("charts/teacher/avg_episodic_return", np.mean(teacher_episodic_returns), 0)
+
+    # collect teacher data for args.teacher_steps
+    # we assume we don't have access to the teacher's replay buffer
+    # see Fig. A.19 in Agarwal et al. 2022 for more detail
+    teacher_rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        optimize_memory_usage=True,
+        handle_timeout_termination=False,
+    )
+
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in track(range(args.teacher_steps), description="filling teacher's replay buffer"):
+        epsilon = linear_schedule(args.start_e, args.end_e, args.teacher_steps, global_step)
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = teacher_model(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        teacher_rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+        obs = next_obs
+
+    # offline training phase: train the student model using the qdagger loss
+    for global_step in track(range(args.offline_steps), description="offline student training"):
+        data = teacher_rb.sample(args.batch_size)
+        # perform a gradient-descent step
+        with torch.no_grad():
+            target_max, _ = target_network(data.next_observations).max(dim=1)
+            td_target = data.rewards.flatten() + args.gamma * target_max * (1 - data.dones.flatten())
+            teacher_q_values = teacher_model(data.observations) / args.temperature
+
+        student_q_values = q_network(data.observations)
+        old_val = student_q_values.gather(1, data.actions).squeeze()
+        q_loss = F.mse_loss(td_target, old_val)
+
+        student_q_values = student_q_values / args.temperature
+        distill_loss = torch.mean(kl_divergence_with_logits(teacher_q_values, student_q_values))
+
+        loss = q_loss + 1.0 * distill_loss
+
+        optimizer.zero_grad()
+        loss.backward()
+        optimizer.step()
+
+        # update the target network
+        if global_step % args.target_network_frequency == 0:
+            for target_network_param, q_network_param in zip(target_network.parameters(), q_network.parameters()):
+                target_network_param.data.copy_(args.tau * q_network_param.data + (1.0 - args.tau) * target_network_param.data)
+
+        if global_step % 100 == 0:
+            writer.add_scalar("charts/offline/loss", loss, global_step)
+            writer.add_scalar("charts/offline/q_loss", q_loss, global_step)
+            writer.add_scalar("charts/offline/distill_loss", distill_loss, global_step)
+
+        if global_step % 100000 == 0:
+            # evaluate the student model
+            model_path = f"runs/{run_name}/{args.exp_name}-offline-{global_step}.cleanrl_model"
+            torch.save(q_network.state_dict(), model_path)
+            print(f"model saved to {model_path}")
+
+            episodic_returns = evaluate(
+                model_path,
+                make_env,
+                args.env_id,
+                eval_episodes=10,
+                run_name=f"{run_name}-eval",
+                Model=QNetwork,
+                device=device,
+                epsilon=args.end_e,
+            )
+            print(episodic_returns)
+            writer.add_scalar("charts/offline/avg_episodic_return", np.mean(episodic_returns), global_step)
+
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        optimize_memory_usage=True,
+        handle_timeout_termination=False,
+    )
+    start_time = time.time()
+
+    # TRY NOT TO MODIFY: start the game
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
+    )
+    obs, _ = envs.reset(seed=args.seed)
+    episodic_returns = deque(maxlen=10)
+    # online training phase
+    for global_step in track(range(args.total_timesteps), description="online student training"):
+        global_step += args.offline_steps
+        # ALGO LOGIC: put action logic here
+        epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = q_network(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                # Skip the envs that are not done
+                if "episode" not in info:
+                    continue
+                print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+                writer.add_scalar("charts/epsilon", epsilon, global_step)
+                episodic_returns.append(info["episode"]["r"])
+                break
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            if global_step % args.train_frequency == 0:
+                data = rb.sample(args.batch_size)
+                # perform a gradient-descent step
+                if len(episodic_returns) < 10:
+                    distill_coeff = 1.0
+                else:
+                    distill_coeff = max(1 - np.mean(episodic_returns) / np.mean(teacher_episodic_returns), 0)
+                with torch.no_grad():
+                    target_max, _ = target_network(data.next_observations).max(dim=1)
+                    td_target = data.rewards.flatten() + args.gamma * target_max * (1 - data.dones.flatten())
+                    teacher_q_values = teacher_model(data.observations) / args.temperature
+
+                student_q_values = q_network(data.observations)
+                old_val = student_q_values.gather(1, data.actions).squeeze()
+                q_loss = F.mse_loss(td_target, old_val)
+
+                student_q_values = student_q_values / args.temperature
+                distill_loss = torch.mean(kl_divergence_with_logits(teacher_q_values, student_q_values))
+
+                loss = q_loss + distill_coeff * distill_loss
+
+                if global_step % 100 == 0:
+                    writer.add_scalar("losses/loss", loss, global_step)
+                    writer.add_scalar("losses/td_loss", q_loss, global_step)
+                    writer.add_scalar("losses/distill_loss", distill_loss, global_step)
+                    writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+                    writer.add_scalar("charts/distill_coeff", distill_coeff, global_step)
+                    print("SPS:", int(global_step / (time.time() - start_time)))
+                    print(distill_coeff)
+                    writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                optimizer.step()
+
+            # update the target network
+            if global_step % args.target_network_frequency == 0:
+                for target_network_param, q_network_param in zip(target_network.parameters(), q_network.parameters()):
+                    target_network_param.data.copy_(
+                        args.tau * q_network_param.data + (1.0 - args.tau) * target_network_param.data
+                    )
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        torch.save(q_network.state_dict(), model_path)
+        print(f"model saved to {model_path}")
+        from cleanrl_utils.evals.dqn_eval import evaluate
+
+        episodic_returns = evaluate(
+            model_path,
+            make_env,
+            args.env_id,
+            eval_episodes=10,
+            run_name=f"{run_name}-eval",
+            Model=QNetwork,
+            device=device,
+            epsilon=args.end_e,
+        )
+        for idx, episodic_return in enumerate(episodic_returns):
+            writer.add_scalar("eval/episodic_return", episodic_return, idx)
+
+        if args.upload_model:
+            from cleanrl_utils.huggingface import push_to_hub
+
+            repo_name = f"{args.env_id}-{args.exp_name}-seed{args.seed}"
+            repo_id = f"{args.hf_entity}/{repo_name}" if args.hf_entity else repo_name
+            push_to_hub(args, episodic_returns, repo_id, "Qdagger", f"runs/{run_name}", f"videos/{run_name}-eval")
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/ragen_wrappers.py b/cleanrl/cleanrl/ragen_wrappers.py
new file mode 100644
index 0000000000000000000000000000000000000000..af63107e99b08764b8410c77be7c66ddbb52899d
--- /dev/null
+++ b/cleanrl/cleanrl/ragen_wrappers.py
@@ -0,0 +1,235 @@
+"""
+Gymnasium-compatible wrappers for RAGEN environments to enable traditional RL training.
+These wrappers convert text-based observations to numerical representations suitable for MLP networks.
+"""
+import gymnasium as gym
+import numpy as np
+from typing import Any, Dict, Tuple
+
+
+class BanditWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Bandit that uses only observable text.
+    Converts text observations to a fixed-size one-hot hash vector.
+    Does not alter episode semantics and does not inspect env internals.
+    """
+    def __init__(self, env, feature_dim_per_name: int = 16):
+        super().__init__(env)
+        # Two name slots (first/second), each hashed to one-hot of size K
+        self.k = feature_dim_per_name
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(2 * self.k,), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(2)
+
+    def _parse_names(self, text_obs: str):
+        """Extract the two arm names from the prompt text purely via regex/string ops."""
+        # Heuristic: look for the segment after "named " and split by " and "
+        try:
+            anchor = "named "
+            if anchor in text_obs:
+                segment = text_obs.split(anchor, 1)[1]
+                # Cut at newline if present
+                segment = segment.split("\n", 1)[0]
+                # Now split by " and " to get two names; also strip punctuation
+                parts = segment.split(" and ")
+                if len(parts) >= 2:
+                    name_a = parts[0].strip().strip(' .!?,')
+                    name_b = parts[1].strip().strip(' .!?,')
+                    return name_a, name_b
+        except Exception:
+            pass
+        # Fallback: no names found
+        return "", ""
+
+    def _names_to_vector(self, name_a: str, name_b: str) -> np.ndarray:
+        vec = np.zeros(2 * self.k, dtype=np.float32)
+        idx_a = (hash(name_a) % self.k)
+        idx_b = (hash(name_b) % self.k)
+        vec[idx_a] = 1.0
+        vec[self.k + idx_b] = 1.0
+        return vec
+
+    def reset(self, **kwargs):
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        name_a, name_b = self._parse_names(text_obs)
+        return self._names_to_vector(name_a, name_b), {}
+
+    def step(self, action):
+        ragen_action = int(action) + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        name_a, name_b = self._parse_names(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return self._names_to_vector(name_a, name_b), reward, terminated, truncated, info
+
+
+class FrozenLakeWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN FrozenLake environment.
+    Converts grid-based text observations to numerical state representation.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine grid size from text only
+        bootstrap_text = self.env.reset()
+        flat, _ = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        # Serve the bootstrapped obs on first reset without calling env.reset again
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, Tuple[int, int]]:
+        """Parse text observation into numerical state (one-hot grid + player pos)."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        player_pos = None
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Parse grid
+        for i, line in enumerate(lines):
+            row = []
+            for j, char in enumerate(line):
+                if char == 'P':  # Player
+                    row.append(0)
+                    player_pos = (i, j)
+                elif char == '_':  # Frozen
+                    row.append(1)
+                elif char == 'O':  # Hole
+                    row.append(2)
+                elif char == 'G':  # Goal
+                    row.append(3)
+                elif char == 'X':  # Player in hole
+                    row.append(2)
+                    player_pos = (i, j)
+                elif char == '√':  # Player on goal
+                    row.append(3)
+                    player_pos = (i, j)
+                else:
+                    row.append(1)  # Default to frozen
+            grid.append(row)
+        # Pad ragged rows if needed
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        grid_size = (rows, cols)
+        # One-hot encode grid over 4 cell types
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 4), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        # Flatten grid
+        flat_grid = one_hot_grid.flatten()
+        # Add normalized player position
+        if player_pos is None:
+            player_pos = (0, 0)
+        player_pos_norm = np.array([
+            0.0 if rows <= 1 else player_pos[0] / max(1, rows - 1),
+            0.0 if cols <= 1 else player_pos[1] / max(1, cols - 1),
+        ], dtype=np.float32)
+        flat = np.concatenate([flat_grid, player_pos_norm])
+        return flat, grid_size
+    
+    def reset(self, **kwargs):
+        # Filter out 'options' parameter that gymnasium passes but RAGEN doesn't support
+        if self._bootstrap_ready:
+            # First call returns the bootstrapped observation to avoid double reset
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+
+
+class SokobanWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Sokoban environment.
+    Converts grid-based text observations to numerical state representation.
+    Note: Does not inherit from gym.Wrapper due to old gym vs gymnasium compatibility.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine room size from text only
+        bootstrap_text = self.env.reset()
+        flat, rows, cols = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        self.metadata = getattr(env, 'metadata', {})
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, int, int]:
+        """Parse text observation into numerical state and return dims."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Mapping from characters to cell types
+        char_to_type = {
+            '#': 0,  # wall
+            '_': 1,  # empty
+            'O': 2,  # target
+            '√': 3,  # box on target
+            'X': 4,  # box
+            'P': 5,  # player
+            'S': 6,  # player on target
+        }
+        
+        for line in lines:
+            row = []
+            for char in line:
+                row.append(char_to_type.get(char, 1))  # Default to empty
+            grid.append(row)
+        # Pad ragged rows
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 7), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        return one_hot_grid.flatten(), rows, cols
+    
+    def reset(self, **kwargs):
+        if self._bootstrap_ready:
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+    
+    def close(self):
+        if hasattr(self.env, 'close'):
+            self.env.close()
+    
+    def render(self):
+        if hasattr(self.env, 'render'):
+            return self.env.render()
+        return None
diff --git a/cleanrl/cleanrl/sac_atari.py b/cleanrl/cleanrl/sac_atari.py
new file mode 100644
index 0000000000000000000000000000000000000000..96d95a7fd85f00b31e21bbb3e06ddf87874f4c8f
--- /dev/null
+++ b/cleanrl/cleanrl/sac_atari.py
@@ -0,0 +1,343 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/sac/#sac_ataripy
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+from cleanrl_utils.atari_wrappers import (
+    ClipRewardEnv,
+    EpisodicLifeEnv,
+    FireResetEnv,
+    MaxAndSkipEnv,
+    NoopResetEnv,
+)
+from cleanrl_utils.buffers import ReplayBuffer
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "BeamRiderNoFrameskip-v4"
+    """the id of the environment"""
+    total_timesteps: int = 5000000
+    """total timesteps of the experiments"""
+    buffer_size: int = int(1e6)
+    """the replay memory buffer size"""  # smaller than in original paper but evaluation is done only for 100k steps anyway
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    tau: float = 1.0
+    """target smoothing coefficient (default: 1)"""
+    batch_size: int = 64
+    """the batch size of sample from the reply memory"""
+    learning_starts: int = 2e4
+    """timestep to start learning"""
+    policy_lr: float = 3e-4
+    """the learning rate of the policy network optimizer"""
+    q_lr: float = 3e-4
+    """the learning rate of the Q network network optimizer"""
+    update_frequency: int = 4
+    """the frequency of training updates"""
+    target_network_frequency: int = 8000
+    """the frequency of updates for the target networks"""
+    alpha: float = 0.2
+    """Entropy regularization coefficient."""
+    autotune: bool = True
+    """automatic tuning of the entropy coefficient"""
+    target_entropy_scale: float = 0.89
+    """coefficient for scaling the autotune entropy target"""
+
+
+def make_env(env_id, seed, idx, capture_video, run_name):
+    def thunk():
+        if capture_video and idx == 0:
+            env = gym.make(env_id, render_mode="rgb_array")
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        else:
+            env = gym.make(env_id)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+
+        env = NoopResetEnv(env, noop_max=30)
+        env = MaxAndSkipEnv(env, skip=4)
+        env = EpisodicLifeEnv(env)
+        if "FIRE" in env.unwrapped.get_action_meanings():
+            env = FireResetEnv(env)
+        env = ClipRewardEnv(env)
+        env = gym.wrappers.ResizeObservation(env, (84, 84))
+        env = gym.wrappers.GrayScaleObservation(env)
+        env = gym.wrappers.FrameStack(env, 4)
+
+        env.action_space.seed(seed)
+        return env
+
+    return thunk
+
+
+def layer_init(layer, bias_const=0.0):
+    nn.init.kaiming_normal_(layer.weight)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+# ALGO LOGIC: initialize agent here:
+# NOTE: Sharing a CNN encoder between Actor and Critics is not recommended for SAC without stopping actor gradients
+# See the SAC+AE paper https://arxiv.org/abs/1910.01741 for more info
+# TL;DR The actor's gradients mess up the representation when using a joint encoder
+class SoftQNetwork(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = envs.single_observation_space.shape
+        self.conv = nn.Sequential(
+            layer_init(nn.Conv2d(obs_shape[0], 32, kernel_size=8, stride=4)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(32, 64, kernel_size=4, stride=2)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 64, kernel_size=3, stride=1)),
+            nn.Flatten(),
+        )
+
+        with torch.inference_mode():
+            output_dim = self.conv(torch.zeros(1, *obs_shape)).shape[1]
+
+        self.fc1 = layer_init(nn.Linear(output_dim, 512))
+        self.fc_q = layer_init(nn.Linear(512, envs.single_action_space.n))
+
+    def forward(self, x):
+        x = F.relu(self.conv(x / 255.0))
+        x = F.relu(self.fc1(x))
+        q_vals = self.fc_q(x)
+        return q_vals
+
+
+class Actor(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = envs.single_observation_space.shape
+        self.conv = nn.Sequential(
+            layer_init(nn.Conv2d(obs_shape[0], 32, kernel_size=8, stride=4)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(32, 64, kernel_size=4, stride=2)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 64, kernel_size=3, stride=1)),
+            nn.Flatten(),
+        )
+
+        with torch.inference_mode():
+            output_dim = self.conv(torch.zeros(1, *obs_shape)).shape[1]
+
+        self.fc1 = layer_init(nn.Linear(output_dim, 512))
+        self.fc_logits = layer_init(nn.Linear(512, envs.single_action_space.n))
+
+    def forward(self, x):
+        x = F.relu(self.conv(x))
+        x = F.relu(self.fc1(x))
+        logits = self.fc_logits(x)
+
+        return logits
+
+    def get_action(self, x):
+        logits = self(x / 255.0)
+        policy_dist = Categorical(logits=logits)
+        action = policy_dist.sample()
+        # Action probabilities for calculating the adapted soft-Q loss
+        action_probs = policy_dist.probs
+        log_prob = F.log_softmax(logits, dim=1)
+        return action, log_prob, action_probs
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv([make_env(args.env_id, args.seed, 0, args.capture_video, run_name)])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    actor = Actor(envs).to(device)
+    qf1 = SoftQNetwork(envs).to(device)
+    qf2 = SoftQNetwork(envs).to(device)
+    qf1_target = SoftQNetwork(envs).to(device)
+    qf2_target = SoftQNetwork(envs).to(device)
+    qf1_target.load_state_dict(qf1.state_dict())
+    qf2_target.load_state_dict(qf2.state_dict())
+    # TRY NOT TO MODIFY: eps=1e-4 increases numerical stability
+    q_optimizer = optim.Adam(list(qf1.parameters()) + list(qf2.parameters()), lr=args.q_lr, eps=1e-4)
+    actor_optimizer = optim.Adam(list(actor.parameters()), lr=args.policy_lr, eps=1e-4)
+
+    # Automatic entropy tuning
+    if args.autotune:
+        target_entropy = -args.target_entropy_scale * torch.log(1 / torch.tensor(envs.single_action_space.n))
+        log_alpha = torch.zeros(1, requires_grad=True, device=device)
+        alpha = log_alpha.exp().item()
+        a_optimizer = optim.Adam([log_alpha], lr=args.q_lr, eps=1e-4)
+    else:
+        alpha = args.alpha
+
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        handle_timeout_termination=False,
+    )
+    start_time = time.time()
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        if global_step < args.learning_starts:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            actions, _, _ = actor.get_action(torch.Tensor(obs).to(device))
+            actions = actions.detach().cpu().numpy()
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                # Skip the envs that are not done
+                if "episode" not in info:
+                    continue
+                print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+                break
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            if global_step % args.update_frequency == 0:
+                data = rb.sample(args.batch_size)
+                # CRITIC training
+                with torch.no_grad():
+                    _, next_state_log_pi, next_state_action_probs = actor.get_action(data.next_observations)
+                    qf1_next_target = qf1_target(data.next_observations)
+                    qf2_next_target = qf2_target(data.next_observations)
+                    # we can use the action probabilities instead of MC sampling to estimate the expectation
+                    min_qf_next_target = next_state_action_probs * (
+                        torch.min(qf1_next_target, qf2_next_target) - alpha * next_state_log_pi
+                    )
+                    # adapt Q-target for discrete Q-function
+                    min_qf_next_target = min_qf_next_target.sum(dim=1)
+                    next_q_value = data.rewards.flatten() + (1 - data.dones.flatten()) * args.gamma * (min_qf_next_target)
+
+                # use Q-values only for the taken actions
+                qf1_values = qf1(data.observations)
+                qf2_values = qf2(data.observations)
+                qf1_a_values = qf1_values.gather(1, data.actions.long()).view(-1)
+                qf2_a_values = qf2_values.gather(1, data.actions.long()).view(-1)
+                qf1_loss = F.mse_loss(qf1_a_values, next_q_value)
+                qf2_loss = F.mse_loss(qf2_a_values, next_q_value)
+                qf_loss = qf1_loss + qf2_loss
+
+                q_optimizer.zero_grad()
+                qf_loss.backward()
+                q_optimizer.step()
+
+                # ACTOR training
+                _, log_pi, action_probs = actor.get_action(data.observations)
+                with torch.no_grad():
+                    qf1_values = qf1(data.observations)
+                    qf2_values = qf2(data.observations)
+                    min_qf_values = torch.min(qf1_values, qf2_values)
+                # no need for reparameterization, the expectation can be calculated for discrete actions
+                actor_loss = (action_probs * ((alpha * log_pi) - min_qf_values)).mean()
+
+                actor_optimizer.zero_grad()
+                actor_loss.backward()
+                actor_optimizer.step()
+
+                if args.autotune:
+                    # reuse action probabilities for temperature loss
+                    alpha_loss = (action_probs.detach() * (-log_alpha.exp() * (log_pi + target_entropy).detach())).mean()
+
+                    a_optimizer.zero_grad()
+                    alpha_loss.backward()
+                    a_optimizer.step()
+                    alpha = log_alpha.exp().item()
+
+            # update the target networks
+            if global_step % args.target_network_frequency == 0:
+                for param, target_param in zip(qf1.parameters(), qf1_target.parameters()):
+                    target_param.data.copy_(args.tau * param.data + (1 - args.tau) * target_param.data)
+                for param, target_param in zip(qf2.parameters(), qf2_target.parameters()):
+                    target_param.data.copy_(args.tau * param.data + (1 - args.tau) * target_param.data)
+
+            if global_step % 100 == 0:
+                writer.add_scalar("losses/qf1_values", qf1_a_values.mean().item(), global_step)
+                writer.add_scalar("losses/qf2_values", qf2_a_values.mean().item(), global_step)
+                writer.add_scalar("losses/qf1_loss", qf1_loss.item(), global_step)
+                writer.add_scalar("losses/qf2_loss", qf2_loss.item(), global_step)
+                writer.add_scalar("losses/qf_loss", qf_loss.item() / 2.0, global_step)
+                writer.add_scalar("losses/actor_loss", actor_loss.item(), global_step)
+                writer.add_scalar("losses/alpha", alpha, global_step)
+                print("SPS:", int(global_step / (time.time() - start_time)))
+                writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+                if args.autotune:
+                    writer.add_scalar("losses/alpha_loss", alpha_loss.item(), global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/scout_dqn/dqn_bandit_nochangeenv.py b/cleanrl/cleanrl/scout_dqn/dqn_bandit_nochangeenv.py
new file mode 100644
index 0000000000000000000000000000000000000000..5dc0c01246e574b188f8473ae82c7cfd74b7a351
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/dqn_bandit_nochangeenv.py
@@ -0,0 +1,422 @@
+# DQN with small MLP for RAGEN Bandit using the existing env (no env edits)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Deque, Tuple
+from collections import deque
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+import json
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.bandit.env import BanditEnv
+from ragen.env.bandit.config import BanditEnvConfig
+from ragen_wrappers import BanditWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "Subagent"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "BanditDQN"
+    total_timesteps: int = 50_000
+    learning_rate: float = 1e-3
+    gamma: float = 0.0  # bandit is single-step
+    batch_size: int = 64
+    buffer_size: int = 10_000
+    target_network_frequency: int = 500
+    train_frequency: int = 1
+
+    # Epsilon-greedy
+    start_e: float = 1.0
+    end_e: float = 0.05
+    exploration_fraction: float = 0.2  # fraction of total timesteps over which to anneal epsilon
+
+    # Model size
+    hidden_size: int = 32
+    feature_dim_per_name: int = 16
+
+
+def make_env(run_name: str, seed: int, feature_dim_per_name: int, capture_video: bool = False):
+    cfg = BanditEnvConfig()
+    env = BanditEnv(cfg)
+    env = BanditWrapper(env, feature_dim_per_name=feature_dim_per_name)
+    env = gym.wrappers.RecordEpisodeStatistics(env)
+    if capture_video:
+        env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+    return env
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QNetwork(nn.Module):
+    def __init__(self, obs_dim: int, act_dim: int, hidden: int):
+        super().__init__()
+        self.net = nn.Sequential(
+            layer_init(nn.Linear(obs_dim, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, act_dim), std=0.01),
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.net(x)
+
+
+class ReplayBuffer:
+    def __init__(self, capacity: int):
+        self.capacity = capacity
+        self.ptr = 0
+        self.full = False
+        self.obs_buf = None
+        self.next_obs_buf = None
+        self.act_buf = None
+        self.rew_buf = None
+        self.done_buf = None
+
+    def add(self, obs: np.ndarray, act: int, rew: float, done: bool, next_obs: np.ndarray):
+        if self.obs_buf is None:
+            obs_shape = obs.shape
+            self.obs_buf = np.zeros((self.capacity,) + obs_shape, dtype=np.float32)
+            self.next_obs_buf = np.zeros((self.capacity,) + obs_shape, dtype=np.float32)
+            self.act_buf = np.zeros((self.capacity,), dtype=np.int64)
+            self.rew_buf = np.zeros((self.capacity,), dtype=np.float32)
+            self.done_buf = np.zeros((self.capacity,), dtype=np.float32)
+        self.obs_buf[self.ptr] = obs
+        self.next_obs_buf[self.ptr] = next_obs
+        self.act_buf[self.ptr] = act
+        self.rew_buf[self.ptr] = rew
+        self.done_buf[self.ptr] = 1.0 if done else 0.0
+        self.ptr = (self.ptr + 1) % self.capacity
+        if self.ptr == 0:
+            self.full = True
+
+    def can_sample(self, batch_size: int) -> bool:
+        return (self.capacity if self.full else self.ptr) >= batch_size
+
+    def sample(self, batch_size: int) -> Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray, np.ndarray]:
+        size = self.capacity if self.full else self.ptr
+        idxs = np.random.randint(0, size, size=batch_size)
+        return (
+            self.obs_buf[idxs],
+            self.act_buf[idxs],
+            self.rew_buf[idxs],
+            self.done_buf[idxs],
+            self.next_obs_buf[idxs],
+        )
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env
+    env = make_env(run_name, args.seed, args.feature_dim_per_name, args.capture_video)
+    obs_shape = env.observation_space.shape
+    act_dim = env.action_space.n
+
+    # networks
+    policy_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net.load_state_dict(policy_net.state_dict())
+    target_net.eval()
+
+    optimizer = optim.Adam(policy_net.parameters(), lr=args.learning_rate)
+    criterion = nn.SmoothL1Loss()
+
+    rb = ReplayBuffer(args.buffer_size)
+
+    # epsilon schedule
+    exploration_steps = max(1, int(args.exploration_fraction * args.total_timesteps))
+    epsilon_by_step = lambda t: args.end_e + (args.start_e - args.end_e) * max(0.0, (exploration_steps - t) / exploration_steps)
+
+    # periodic eval setup (mirrors noisy_dqn_sokoban_small.py)
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes: int, step_tag: int):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env_eval.reset(seed=args.seed + 100000 + collected)
+                # try to fetch the original prompt text from underlying env
+                prompt_text = None
+                try:
+                    if hasattr(env_eval, 'env') and hasattr(env_eval.env, 'render'):
+                        p = env_eval.env.render()
+                        if isinstance(p, str):
+                            prompt_text = p
+                except Exception:
+                    pass
+                # lightweight name parser (aligned with converter/wrapper)
+                def _parse_names(text: str):
+                    try:
+                        anchor = "named "
+                        if isinstance(text, str) and anchor in text:
+                            segment = text.split(anchor, 1)[1]
+                            segment = segment.split("\n", 1)[0]
+                            parts = segment.split(" and ")
+                            if len(parts) >= 2:
+                                a = parts[0].strip().strip(' .!?,')
+                                b = parts[1].strip().strip(' .!?,')
+                                if a and b:
+                                    return a, b
+                    except Exception:
+                        pass
+                    return None
+                names_tuple = _parse_names(prompt_text) if prompt_text else None
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                # bandit is single-step, but keep loop for generality
+                while not done:
+                    with torch.no_grad():
+                        q = agent_model(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(q, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env_eval.step(action)
+                    d = bool(terminated) or bool(truncated)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                    "prompt": prompt_text,
+                    "names": list(names_tuple) if names_tuple is not None else None,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    # choose periodicity similar to reference
+    eval_splits = 2
+    eval_episodes = 4000
+    eval_every_steps = max(1, args.total_timesteps // eval_splits)
+
+    # training loop
+    global_step = 0
+    start_time = time.time()
+
+    obs, _ = env.reset(seed=args.seed)
+    # success tracking
+    ep_success_window = deque(maxlen=100)
+    total_episodes = 0
+    total_successes = 0
+
+    while global_step < args.total_timesteps:
+        epsilon = epsilon_by_step(global_step)
+        if np.random.rand() < epsilon:
+            action = env.action_space.sample()
+        else:
+            with torch.no_grad():
+                q_values = policy_net(torch.tensor(obs, dtype=torch.float32, device=device).unsqueeze(0))
+                action = int(torch.argmax(q_values, dim=1).item())
+        next_obs, reward, terminated, truncated, info = env.step(action)
+        done = bool(terminated) or bool(truncated)
+
+        rb.add(obs.astype(np.float32), action, float(reward), done, next_obs.astype(np.float32))
+
+        obs = next_obs
+        global_step += 1
+
+        # Bandit episodes end in one step; reset immediately
+        if done:
+            # record success
+            succ = bool(info.get('success', False))
+            ep_success_window.append(1.0 if succ else 0.0)
+            total_episodes += 1
+            total_successes += (1 if succ else 0)
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "rollout/success": float(1.0 if succ else 0.0),
+                        "rollout/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                        "rollout/episodes": int(total_episodes),
+                    }, step=global_step)
+                except Exception:
+                    pass
+            obs, _ = env.reset()
+
+        # optimize
+        if rb.can_sample(args.batch_size) and (global_step % args.train_frequency == 0):
+            batch_obs, batch_act, batch_rew, batch_done, batch_next_obs = rb.sample(args.batch_size)
+            b_obs = torch.tensor(batch_obs, dtype=torch.float32, device=device)
+            b_act = torch.tensor(batch_act, dtype=torch.int64, device=device)
+            b_rew = torch.tensor(batch_rew, dtype=torch.float32, device=device)
+            b_done = torch.tensor(batch_done, dtype=torch.float32, device=device)
+            b_next_obs = torch.tensor(batch_next_obs, dtype=torch.float32, device=device)
+
+            # Q-learning target: r + gamma * max_a' Q_target(s', a') * (1-done)
+            with torch.no_grad():
+                next_q = target_net(b_next_obs).max(dim=1)[0]
+                target_q = b_rew + args.gamma * (1.0 - b_done) * next_q
+
+            current_q = policy_net(b_obs).gather(1, b_act.view(-1, 1)).squeeze(1)
+            loss = criterion(current_q, target_q)
+
+            optimizer.zero_grad()
+            loss.backward()
+            nn.utils.clip_grad_norm_(policy_net.parameters(), max_norm=10.0)
+            optimizer.step()
+
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "charts/epsilon": float(epsilon),
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+        # target network update
+        if global_step % args.target_network_frequency == 0:
+            target_net.load_state_dict(policy_net.state_dict())
+
+        # occasional print
+        if global_step % 1000 == 0:
+            sps = int(global_step / (time.time() - start_time))
+            sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+            print(f"Step {global_step} | SPS: {sps} | Epsilon: {epsilon:.3f} | SR@100: {sr100:.3f}")
+
+        # periodic evaluation and trajectory dump (mirrors reference)
+        if global_step == 1 or (global_step % eval_every_steps == 0):
+            try:
+                def eval_thunk():
+                    return make_env(run_name, args.seed + 9999, args.feature_dim_per_name, False)
+                collect_eval_trajectories(policy_net, eval_thunk, n_episodes=eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import wandb
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {eval_episodes} eval trajectories at step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    # simple evaluation after training
+    def evaluate(n_episodes=200):
+        returns = []
+        successes = []
+        for i in range(n_episodes):
+            s, _ = env.reset(seed=args.seed + 100000 + i)
+            done = False
+            G = 0.0
+            while not done:
+                with torch.no_grad():
+                    q = policy_net(torch.tensor(s, dtype=torch.float32, device=device).unsqueeze(0))
+                    a = int(torch.argmax(q, dim=1).item())
+                s, r, term, trunc, info = env.step(a)
+                G += float(r)
+                done = bool(term) or bool(trunc)
+            successes.append(1.0 if bool(info.get('success', False)) else 0.0)
+            returns.append(G)
+        return float(np.mean(returns)), float(np.std(returns)), float(np.mean(successes))
+
+    avg_ret, std_ret, succ_rate = evaluate(4000)
+    if args.track:
+        try:
+            import wandb
+            wandb.log({
+                "global_step": int(global_step),
+                "eval/avg_return": float(avg_ret),
+                "eval/std_return": float(std_ret),
+                "eval/episodes": int(4000),
+                "eval/success_rate": float(succ_rate),
+            }, step=global_step)
+        except Exception:
+            pass
+
+    env.close()
diff --git a/cleanrl/cleanrl/scout_dqn/dqn_frozenlake.py b/cleanrl/cleanrl/scout_dqn/dqn_frozenlake.py
new file mode 100644
index 0000000000000000000000000000000000000000..db72b76cee463563aad13172b5723c1b1a9f5163
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/dqn_frozenlake.py
@@ -0,0 +1,428 @@
+# DQN with small MLP for RAGEN FrozenLake using the existing env (no env edits)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any
+import json
+import re  # Added for ANSI strip
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.frozen_lake.env import FrozenLakeEnv
+from ragen.env.frozen_lake.config import FrozenLakeEnvConfig
+
+
+class FrozenLakeWrapper(gym.Env):
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: FrozenLakeEnv):
+        super().__init__()
+        self._env = env
+        self._size = int(self._env.nrow)
+        self._tokens = ['P', '_', 'O', 'G', 'X', '√']
+        self._token_to_idx = {t: i for i, t in enumerate(self._tokens)}
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(self._size * self._size * len(self._tokens),), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+
+    def _encode_obs(self, text_obs: str) -> np.ndarray:
+        # 1. 去除 ANSI 颜色代码
+        ansi_escape = re.compile(r'\x1B(?:[@-Z\\-_]|\[[0-?]*[ -/]*[@-~])')
+        text_obs = ansi_escape.sub('', text_obs)
+
+        # 2. 清理边框和分割行
+        raw_rows = text_obs.split('\n')
+        rows = []
+        for r in raw_rows:
+            clean_r = r.strip()
+            # 跳过边框行 (如 +---+) 或空行
+            if not clean_r or set(clean_r).issubset({'+', '-', ' '}):
+                continue
+            # 去除行首行尾的竖线 (如 | P | ->  P )
+            clean_r = clean_r.strip('|')
+            rows.append(list(clean_r))
+
+        # 3. 构建 Grid
+        h = self._size
+        w = self._size
+        grid = np.zeros((h, w, len(self._tokens)), dtype=np.float32)
+        
+        # 安全填充，防止索引越界
+        for i in range(min(h, len(rows))):
+            for j in range(min(w, len(rows[i]))):
+                ch = rows[i][j]
+                # 修复: 只有在 token 列表中才置1，遇到未知字符(如墙壁)不默认为 Player(0)
+                if ch in self._token_to_idx:
+                    idx = self._token_to_idx[ch]
+                    grid[i, j, idx] = 1.0
+                    
+        return grid.reshape(-1).astype(np.float32)
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        mapped = int(action) + 1
+        text_obs, reward, done, info = self._env.step(mapped)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "FrozenLakeDQN"
+    total_timesteps: int = 4000_000
+    learning_rate: float = 2.5e-4
+    gamma: float = 0.99
+    batch_size: int = 128
+    buffer_size: int = 200_000
+    target_network_frequency: int = 2000
+    train_frequency: int = 4
+    learning_starts: int = 5000
+
+    # Epsilon-greedy
+    start_e: float = 1.0
+    end_e: float = 0.05
+    exploration_fraction: float = 0.2
+
+    # Model size
+    hidden_size: int = 128
+
+    # FrozenLake specific
+    grid_size: int = 4
+    is_slippery: bool = False
+
+    # Eval
+    eval_splits: int = 1
+    eval_episodes: int = 4000
+
+
+def make_env(idx, run_name, seed, grid_size, is_slippery, capture_video=False):
+    def thunk():
+        config = FrozenLakeEnvConfig(size=grid_size, p=0.9, success_rate=0.8, is_slippery=is_slippery, map_seed=seed + idx, render_mode='text')
+        env = FrozenLakeEnv(config)
+        env = FrozenLakeWrapper(env)
+        max_steps = int(grid_size * grid_size * 4)
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QNetwork(nn.Module):
+    def __init__(self, obs_dim: int, act_dim: int, hidden: int):
+        super().__init__()
+        self.net = nn.Sequential(
+            layer_init(nn.Linear(obs_dim, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, act_dim), std=0.01),
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.net(x)
+
+
+class ReplayBuffer:
+    def __init__(self, capacity: int, obs_shape: Tuple[int, ...]):
+        self.capacity = capacity
+        self.ptr = 0
+        self.full = False
+        self.obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.next_obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.act_buf = np.zeros((capacity,), dtype=np.int64)
+        self.rew_buf = np.zeros((capacity,), dtype=np.float32)
+        self.done_buf = np.zeros((capacity,), dtype=np.float32)
+
+    def add(self, obs: np.ndarray, act: int, rew: float, done: bool, next_obs: np.ndarray):
+        i = self.ptr
+        self.obs_buf[i] = obs
+        self.next_obs_buf[i] = next_obs
+        self.act_buf[i] = act
+        self.rew_buf[i] = rew
+        self.done_buf[i] = 1.0 if done else 0.0
+        self.ptr = (self.ptr + 1) % self.capacity
+        if self.ptr == 0:
+            self.full = True
+
+    def can_sample(self, batch_size: int) -> bool:
+        return (self.capacity if self.full else self.ptr) >= batch_size
+
+    def sample(self, batch_size: int):
+        size = self.capacity if self.full else self.ptr
+        idxs = np.random.randint(0, size, size=batch_size)
+        return (
+            self.obs_buf[idxs],
+            self.act_buf[idxs],
+            self.rew_buf[idxs],
+            self.done_buf[idxs],
+            self.next_obs_buf[idxs],
+        )
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    env = make_env(0, run_name, args.seed, args.grid_size, args.is_slippery, args.capture_video)()
+    obs_shape = env.observation_space.shape
+    act_dim = env.action_space.n
+
+    policy_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net.load_state_dict(policy_net.state_dict())
+    target_net.eval()
+
+    optimizer = optim.Adam(policy_net.parameters(), lr=args.learning_rate)
+    criterion = nn.SmoothL1Loss()
+
+    rb = ReplayBuffer(args.buffer_size, obs_shape)
+
+    exploration_steps = max(1, int(args.exploration_fraction * args.total_timesteps))
+    epsilon_by_step = lambda t: args.end_e + (args.start_e - args.end_e) * max(0.0, (exploration_steps - t) / exploration_steps)
+
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env_eval.reset(seed=args.seed + 100000 + collected)
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env_eval, '_max_episode_steps', None) or int(args.grid_size * args.grid_size * 4)
+                while not done:
+                    with torch.no_grad():
+                        q = agent_model(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(q, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env_eval.step(action)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    global_step = 0
+    start_time = time.time()
+
+    obs, _ = env.reset(seed=args.seed)
+    ep_success_window = []
+    ep_return = 0.0
+    ep_len = 0
+
+    eval_every_steps = max(1, args.total_timesteps // args.eval_splits)
+
+    while global_step < args.total_timesteps:
+        epsilon = epsilon_by_step(global_step)
+        if np.random.rand() < epsilon or global_step < args.learning_starts:
+            action = env.action_space.sample()
+        else:
+            with torch.no_grad():
+                q_values = policy_net(torch.tensor(obs, dtype=torch.float32, device=device).unsqueeze(0))
+                action = int(torch.argmax(q_values, dim=1).item())
+
+        next_obs, reward, terminated, truncated, info = env.step(action)
+        # 注意：这里 done 仅用于控制循环和 logging，不用于 ReplayBuffer 的逻辑判断
+        done = bool(terminated) or bool(truncated)
+
+        # 修复：Buffer 中只存储真正的 termination (死亡或到达)，不存 truncation (超时)
+        # 这样 Q-learning 在超时时不会错误地认为价值归零
+        rb.add(obs.astype(np.float32), int(action), float(reward), bool(terminated), next_obs.astype(np.float32))
+
+        obs = next_obs
+        ep_return += float(reward)
+        ep_len += 1
+        global_step += 1
+
+        if done:
+            succ = bool((info or {}).get('success', False))
+            ep_success_window.append(1.0 if succ else 0.0)
+            if len(ep_success_window) > 100:
+                ep_success_window.pop(0)
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "rollout/success": float(1.0 if succ else 0.0),
+                        "rollout/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                        # PPO-compatible episodic keys
+                        "train/episodic_return": float(ep_return),
+                        "train/episodic_length": int(ep_len),
+                        "train/success": float(1.0 if succ else 0.0),
+                        "train/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                    }, step=global_step)
+                except Exception:
+                    pass
+            obs, _ = env.reset()
+            ep_return, ep_len = 0.0, 0
+
+        if rb.can_sample(args.batch_size) and (global_step % args.train_frequency == 0) and (global_step > args.learning_starts):
+            batch_obs, batch_act, batch_rew, batch_done, batch_next_obs = rb.sample(args.batch_size)
+            b_obs = torch.tensor(batch_obs, dtype=torch.float32, device=device)
+            b_act = torch.tensor(batch_act, dtype=torch.int64, device=device)
+            b_rew = torch.tensor(batch_rew, dtype=torch.float32, device=device)
+            b_done = torch.tensor(batch_done, dtype=torch.float32, device=device)
+            b_next_obs = torch.tensor(batch_next_obs, dtype=torch.float32, device=device)
+
+            with torch.no_grad():
+                next_q = target_net(b_next_obs).max(dim=1)[0]
+                target_q = b_rew + args.gamma * (1.0 - b_done) * next_q
+
+            current_q = policy_net(b_obs).gather(1, b_act.view(-1, 1)).squeeze(1)
+            loss = criterion(current_q, target_q)
+
+            optimizer.zero_grad()
+            loss.backward()
+            nn.utils.clip_grad_norm_(policy_net.parameters(), max_norm=10.0)
+            optimizer.step()
+
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "charts/epsilon": float(epsilon),
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                        "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+        if global_step % args.target_network_frequency == 0:
+            target_net.load_state_dict(policy_net.state_dict())
+
+        if global_step % 1000 == 0:
+            sps = int(global_step / (time.time() - start_time))
+            sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+            print(f"Step {global_step} | SPS: {sps} | Epsilon: {epsilon:.3f} | SR@100: {sr100:.3f}")
+
+        if global_step == 1 or (global_step % eval_every_steps == 0):
+            try:
+                def eval_thunk():
+                    return make_env(0, run_name, args.seed + 9999, args.grid_size, args.is_slippery, False)()
+                collect_eval_trajectories(policy_net, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import wandb
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    env.close()
\ No newline at end of file
diff --git a/cleanrl/cleanrl/scout_dqn/dqn_rubikscube.py b/cleanrl/cleanrl/scout_dqn/dqn_rubikscube.py
new file mode 100644
index 0000000000000000000000000000000000000000..d1dca9015d4b401d1a1f71d6a50157f10680269e
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/dqn_rubikscube.py
@@ -0,0 +1,454 @@
+# DQN with Step Penalty & Time Limit Bootstrap for RAGEN Rubik's Cube 2x2
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+import re
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+
+import sys
+# 假设 ragen 库在当前目录的上两级，请根据实际情况调整
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.rubikscube.env import RubiksCube2x2Env
+from ragen.env.rubikscube.config import RubiksCube2x2Config
+
+
+class RubiksCubeWrapper(gym.Env):
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: RubiksCube2x2Env):
+        super().__init__()
+        self._env = env
+        self._colors = ['W', 'O', 'G', 'R', 'B', 'Y']
+        self._color_to_idx = {c: i for i, c in enumerate(self._colors)}
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(24 * len(self._colors),), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(12)
+        self._face_pat = re.compile(r"\[(?:\s*([A-Z])\s*,\s*([A-Z])\s*\])\n\s*\[(?:\s*([A-Z])\s*,\s*([A-Z])\s*\])")
+        
+        # [关键修改] 每步惩罚
+        # 如果20步都没解出来，累计惩罚是 20 * -0.05 = -1.0，抵消掉最后可能获得的 +1.0
+        # 这迫使 agent 寻找更短路径
+        self.step_penalty = -0.044
+
+    def _encode_obs(self, text_obs: str) -> np.ndarray:
+        faces_order = ["Up (U):", "Left (L):", "Front (F):", "Right (R):", "Back (B):", "Down (D):"]
+        lines = text_obs.splitlines()
+        blocks: List[str] = []
+        i = 0
+        while i < len(lines):
+            line = lines[i]
+            for header in faces_order:
+                if line.startswith(header):
+                    content = line[len(header):].strip()
+                    next_line = lines[i + 1] if i + 1 < len(lines) else ""
+                    block = f"{content}\n{next_line}"
+                    blocks.append(block)
+                    break
+            i += 1
+        if len(blocks) != 6:
+            return np.zeros(24 * len(self._colors), dtype=np.float32)
+        stickers: List[int] = []
+        for blk in blocks:
+            m = self._face_pat.search(blk)
+            if not m:
+                return np.zeros(24 * len(self._colors), dtype=np.float32)
+            c0, c1, c2, c3 = m.group(1), m.group(2), m.group(3), m.group(4)
+            stickers.extend([c0, c1, c2, c3])
+        grid = np.zeros((24, len(self._colors)), dtype=np.float32)
+        for idx, ch in enumerate(stickers):
+            cidx = self._color_to_idx.get(ch, None)
+            if cidx is not None:
+                grid[idx, cidx] = 1.0
+        return grid.reshape(-1)
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        mapped = int(action) + 1
+        text_obs, reward, done, info = self._env.step(mapped)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        
+        # [关键修改] 应用惩罚
+        # 此时 reward 包含原本的稀疏奖励 (0 或 1) 加上每步惩罚
+        reward = float(reward) + self.step_penalty
+        
+        return obs, reward, terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "RubiksCube2x2DQN"
+    total_timesteps: int = 400_000
+    learning_rate: float = 2.5e-4
+    gamma: float = 0.99
+    batch_size: int = 128
+    buffer_size: int = 200_000
+    target_network_frequency: int = 4000
+    train_frequency: int = 4
+    learning_starts: int = 5000
+
+    # Epsilon-greedy
+    start_e: float = 1.0
+    end_e: float = 0.05
+    exploration_fraction: float = 0.2
+
+    # Model size
+    hidden_size: int = 256
+
+    # Rubik specific
+    scramble_depth: int = 3 # 初始难度
+    max_steps_env: int = 20 # 硬性时间限制
+
+    # Eval
+    eval_splits: int = 2
+    eval_episodes: int = 4000
+
+
+def make_env(idx, run_name, seed, scramble_depth, max_steps_env, capture_video=False):
+    def thunk():
+        # [关键修改] 内部 max_steps 设为 1000，防止 env 内部发出 done=True
+        # 我们完全依赖外部 TimeLimit wrapper 来处理超时
+        config = RubiksCube2x2Config(scramble_depth=scramble_depth, max_steps=20, render_mode='text')
+        env = RubiksCube2x2Env(config)
+        env = RubiksCubeWrapper(env)
+        # 外部限制设为 20
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps_env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QNetwork(nn.Module):
+    def __init__(self, obs_dim: int, act_dim: int, hidden: int):
+        super().__init__()
+        self.net = nn.Sequential(
+            layer_init(nn.Linear(obs_dim, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, act_dim), std=0.01),
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.net(x)
+
+
+class ReplayBuffer:
+    def __init__(self, capacity: int, obs_shape: Tuple[int, ...]):
+        self.capacity = capacity
+        self.ptr = 0
+        self.full = False
+        self.obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.next_obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.act_buf = np.zeros((capacity,), dtype=np.int64)
+        self.rew_buf = np.zeros((capacity,), dtype=np.float32)
+        self.done_buf = np.zeros((capacity,), dtype=np.float32)
+
+    def add(self, obs: np.ndarray, act: int, rew: float, done: bool, next_obs: np.ndarray):
+        i = self.ptr
+        self.obs_buf[i] = obs
+        self.next_obs_buf[i] = next_obs
+        self.act_buf[i] = act
+        self.rew_buf[i] = rew
+        self.done_buf[i] = 1.0 if done else 0.0
+        self.ptr = (self.ptr + 1) % self.capacity
+        if self.ptr == 0:
+            self.full = True
+
+    def can_sample(self, batch_size: int) -> bool:
+        return (self.capacity if self.full else self.ptr) >= batch_size
+
+    def sample(self, batch_size: int):
+        size = self.capacity if self.full else self.ptr
+        idxs = np.random.randint(0, size, size=batch_size)
+        return (
+            self.obs_buf[idxs],
+            self.act_buf[idxs],
+            self.rew_buf[idxs],
+            self.done_buf[idxs],
+            self.next_obs_buf[idxs],
+        )
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    env = make_env(0, run_name, args.seed, args.scramble_depth, args.max_steps_env, args.capture_video)()
+    obs_shape = env.observation_space.shape
+    act_dim = env.action_space.n
+
+    policy_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net = QNetwork(int(np.prod(obs_shape)), act_dim, args.hidden_size).to(device)
+    target_net.load_state_dict(policy_net.state_dict())
+    target_net.eval()
+
+    optimizer = optim.Adam(policy_net.parameters(), lr=args.learning_rate)
+    criterion = nn.SmoothL1Loss()
+
+    rb = ReplayBuffer(args.buffer_size, obs_shape)
+
+    exploration_steps = max(1, int(args.exploration_fraction * args.total_timesteps))
+    epsilon_by_step = lambda t: args.end_e + (args.start_e - args.end_e) * max(0.0, (exploration_steps - t) / exploration_steps)
+
+    # 评估函数
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env_eval.reset(seed=args.seed + 100000 + collected)
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env_eval, '_max_episode_steps', None) or int(args.max_steps_env)
+                while not done:
+                    with torch.no_grad():
+                        q = agent_model(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(q, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env_eval.step(action)
+                    
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    # 主训练循环
+    global_step = 0
+    start_time = time.time()
+
+    obs, _ = env.reset(seed=args.seed)
+    ep_return = 0.0
+    ep_len = 0
+    ep_success_window: List[float] = []
+
+    eval_every_steps = max(1, args.total_timesteps // args.eval_splits)
+
+    while global_step < args.total_timesteps:
+        epsilon = epsilon_by_step(global_step)
+        if np.random.rand() < epsilon or global_step < args.learning_starts:
+            action = env.action_space.sample()
+        else:
+            with torch.no_grad():
+                q_values = policy_net(torch.tensor(obs, dtype=torch.float32, device=device).unsqueeze(0))
+                action = int(torch.argmax(q_values, dim=1).item())
+
+        next_obs, reward, terminated, truncated, info = env.step(action)
+        
+        # 判断是否需要 Reset (任何结束都 Reset)
+        real_done = bool(terminated) or bool(truncated)
+
+        # [核心逻辑]
+        # 1. 即使是 truncated (超时)，done 也记为 False，以便进行 bootstrap (计算未来价值)
+        # 2. 只有 terminated (真正解开了)，done 才记为 True
+        # 3. Step Penalty 已经包含在 reward 中，DQN 会学到"为了避免扣分，必须在未来几步内解决"
+        rb.add(
+            obs.astype(np.float32), 
+            int(action), 
+            float(reward), 
+            bool(terminated), # 注意：只用 terminated
+            next_obs.astype(np.float32)
+        )
+
+        obs = next_obs
+        ep_return += float(reward)
+        ep_len += 1
+        global_step += 1
+
+        # 训练过程
+        if rb.can_sample(args.batch_size) and (global_step % args.train_frequency == 0) and (global_step > args.learning_starts):
+            batch_obs, batch_act, batch_rew, batch_done, batch_next_obs = rb.sample(args.batch_size)
+            b_obs = torch.tensor(batch_obs, dtype=torch.float32, device=device)
+            b_act = torch.tensor(batch_act, dtype=torch.int64, device=device)
+            b_rew = torch.tensor(batch_rew, dtype=torch.float32, device=device)
+            b_done = torch.tensor(batch_done, dtype=torch.float32, device=device)
+            b_next_obs = torch.tensor(batch_next_obs, dtype=torch.float32, device=device)
+
+            with torch.no_grad():
+                next_q = target_net(b_next_obs).max(dim=1)[0]
+                target_q = b_rew + args.gamma * (1.0 - b_done) * next_q
+
+            current_q = policy_net(b_obs).gather(1, b_act.view(-1, 1)).squeeze(1)
+            loss = criterion(current_q, target_q)
+
+            optimizer.zero_grad()
+            loss.backward()
+            nn.utils.clip_grad_norm_(policy_net.parameters(), max_norm=10.0)
+            optimizer.step()
+
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "charts/epsilon": float(epsilon),
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                        "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+        # 结束时 Reset
+        if real_done:
+            succ = bool((info or {}).get('success', False))
+            ep_success_window.append(1.0 if succ else 0.0)
+            if len(ep_success_window) > 100:
+                ep_success_window.pop(0)
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/episodic_return": float(ep_return),
+                        "train/episodic_length": int(ep_len),
+                        "train/success": float(1.0 if succ else 0.0),
+                        "train/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) >= 1 else None,
+                    }, step=global_step)
+                except Exception:
+                    pass
+            obs, _ = env.reset()
+            ep_return, ep_len = 0.0, 0
+
+        # 更新目标网络
+        if global_step % args.target_network_frequency == 0:
+            target_net.load_state_dict(policy_net.state_dict())
+
+        # 打印日志
+        if global_step % 1000 == 0:
+            sps = int(global_step / (time.time() - start_time))
+            sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+            print(f"Step {global_step} | SPS: {sps} | Epsilon: {epsilon:.3f} | SR@100: {sr100:.3f}")
+
+        # 评估
+        if global_step == 1 or (global_step % eval_every_steps == 0):
+            try:
+                def eval_thunk():
+                    # 评估环境也同样设置：内部1000，外部20
+                    return make_env(0, run_name, args.seed + 9999, args.scramble_depth, args.max_steps_env, False)()
+                collect_eval_trajectories(policy_net, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import wandb
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    env.close()
\ No newline at end of file
diff --git a/cleanrl/cleanrl/scout_dqn/dqn_sudoku.py b/cleanrl/cleanrl/scout_dqn/dqn_sudoku.py
new file mode 100644
index 0000000000000000000000000000000000000000..c53b17cfab9258630c3faeab398780f0dde279a0
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/dqn_sudoku.py
@@ -0,0 +1,485 @@
+# DQN with Action Masking for RAGEN Sudoku (4x4, max_step=20)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sudoku.env import SudokuEnv
+from ragen.env.sudoku.config import SudokuEnvConfig
+
+
+class SudokuWrapper(gym.Env):
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: SudokuEnv, grid_size: int):
+        super().__init__()
+        self._env = env
+        self._size = grid_size
+        self._val_dim = self._size + 1
+        self._act_n = self._size * self._size * self._size
+        self.action_space = gym.spaces.Discrete(self._act_n)
+        self.observation_space = gym.spaces.Dict({
+            "observation": gym.spaces.Box(low=0.0, high=1.0, shape=(self._size * self._size * self._val_dim,), dtype=np.float32),
+            "action_mask": gym.spaces.Box(low=0.0, high=1.0, shape=(self._act_n,), dtype=np.float32)
+        })
+
+    def _encode_obs(self, text_obs: str) -> Dict[str, np.ndarray]:
+        vals: List[int] = []
+        for line in text_obs.splitlines():
+            ls = line.strip()
+            if len(ls) == 0:
+                continue
+            if set(ls) <= {'-'}:
+                continue
+            tokens = [t for t in ls.split() if t != '|']
+            if len(tokens) == 0:
+                continue
+            for t in tokens:
+                if t == '.':
+                    vals.append(0)
+                else:
+                    try:
+                        v = int(t)
+                    except ValueError:
+                        v = 0
+                    vals.append(v)
+        target = self._size * self._size
+        if len(vals) < target:
+            vals.extend([0] * (target - len(vals)))
+        if len(vals) > target:
+            vals = vals[:target]
+        grid = np.zeros((target, self._val_dim), dtype=np.float32)
+        mask = np.zeros(self._act_n, dtype=np.float32)
+        for i, v in enumerate(vals):
+            v_clamped = int(v)
+            if v_clamped < 0 or v_clamped > self._size:
+                v_clamped = 0
+            grid[i, v_clamped] = 1.0
+            if v_clamped != 0:
+                start_idx = i * self._size
+                end_idx = start_idx + self._size
+                mask[start_idx:end_idx] = 0.0
+        cur = np.array(vals, dtype=np.int64).reshape(self._size, self._size)
+        box = int(np.sqrt(self._size))
+        for i, v in enumerate(vals):
+            if int(v) == 0:
+                r = i // self._size
+                c = i % self._size
+                row_vals = set(cur[r, :].tolist())
+                col_vals = set(cur[:, c].tolist())
+                br = (r // box) * box
+                bc = (c // box) * box
+                box_vals = set(cur[br:br + box, bc:bc + box].reshape(-1).tolist())
+                start_idx = i * self._size
+                for num in range(1, self._size + 1):
+                    if (num not in row_vals) and (num not in col_vals) and (num not in box_vals):
+                        mask[start_idx + (num - 1)] = 1.0
+        if mask.sum() == 0:
+            for i, v in enumerate(vals):
+                if int(v) == 0:
+                    start_idx = i * self._size
+                    end_idx = start_idx + self._size
+                    mask[start_idx:end_idx] = 1.0
+            if mask.sum() == 0:
+                mask[:] = 1.0
+        return {
+            "observation": grid.reshape(-1),
+            "action_mask": mask,
+        }
+
+    @staticmethod
+    def _decode_action(action_id: int, grid_size: int) -> Tuple[int, int, int]:
+        g = grid_size
+        row = action_id // (g * g)
+        rem = action_id % (g * g)
+        col = rem // g
+        num = (rem % g) + 1
+        return row, col, num
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        row, col, num = self._decode_action(int(action), self._size)
+        act_str = f"{row + 1},{col + 1},{num}"
+        text_obs, reward, done, info = self._env.step(act_str)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "SudokuDQN"
+    total_timesteps: int = 200000
+    learning_rate: float = 3e-4
+    gamma: float = 0.99
+    batch_size: int = 128
+    buffer_size: int = 400_000
+    target_network_frequency: int = 4000
+    train_frequency: int = 1
+    learning_starts: int = 10000
+
+    # Epsilon-greedy
+    start_e: float = 1.0
+    end_e: float = 0.05
+    exploration_fraction: float = 0.2
+
+    # Model size
+    hidden_size: int = 256
+
+    # Sudoku specific
+    grid_size: int = 4
+    difficulty: str = "easy"
+
+    # Eval
+    eval_splits: int = 2
+    eval_episodes: int = 4000
+
+
+def make_env(idx, run_name, seed, grid_size, difficulty, capture_video=False):
+    def thunk():
+        config = SudokuEnvConfig(
+            grid_size=grid_size,
+            difficulty=difficulty,
+            render_mode='text',
+            render_format='simple',
+        )
+        env = SudokuEnv(config)
+        env = SudokuWrapper(env, grid_size)
+        max_steps = 20
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QNetwork(nn.Module):
+    def __init__(self, obs_dim: int, act_dim: int, hidden: int):
+        super().__init__()
+        self.net = nn.Sequential(
+            layer_init(nn.Linear(obs_dim, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.ReLU(),
+            layer_init(nn.Linear(hidden, act_dim), std=0.01),
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.net(x)
+
+
+class ReplayBuffer:
+    def __init__(self, capacity: int, obs_dim: int):
+        self.capacity = capacity
+        self.ptr = 0
+        self.full = False
+        self.obs_buf = np.zeros((capacity, obs_dim), dtype=np.float32)
+        self.next_obs_buf = np.zeros((capacity, obs_dim), dtype=np.float32)
+        self.act_buf = np.zeros((capacity,), dtype=np.int64)
+        self.rew_buf = np.zeros((capacity,), dtype=np.float32)
+        self.done_buf = np.zeros((capacity,), dtype=np.float32)
+
+    def add(self, obs: np.ndarray, act: int, rew: float, done: bool, next_obs: np.ndarray):
+        i = self.ptr
+        self.obs_buf[i] = obs
+        self.next_obs_buf[i] = next_obs
+        self.act_buf[i] = act
+        self.rew_buf[i] = rew
+        self.done_buf[i] = 1.0 if done else 0.0
+        self.ptr = (self.ptr + 1) % self.capacity
+        if self.ptr == 0:
+            self.full = True
+
+    def can_sample(self, batch_size: int) -> bool:
+        return (self.capacity if self.full else self.ptr) >= batch_size
+
+    def sample(self, batch_size: int):
+        size = self.capacity if self.full else self.ptr
+        idxs = np.random.randint(0, size, size=batch_size)
+        return (
+            self.obs_buf[idxs],
+            self.act_buf[idxs],
+            self.rew_buf[idxs],
+            self.done_buf[idxs],
+            self.next_obs_buf[idxs],
+        )
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    env = make_env(0, run_name, args.seed, args.grid_size, args.difficulty, args.capture_video)()
+    # Flattened grid only from Dict observation
+    sample_obs, _ = env.reset(seed=args.seed)
+    obs_dim = int(np.prod(sample_obs["observation"].shape))
+    act_dim = env.action_space.n
+
+    policy_net = QNetwork(obs_dim, act_dim, args.hidden_size).to(device)
+    target_net = QNetwork(obs_dim, act_dim, args.hidden_size).to(device)
+    target_net.load_state_dict(policy_net.state_dict())
+    target_net.eval()
+
+    optimizer = optim.Adam(policy_net.parameters(), lr=args.learning_rate)
+    criterion = nn.SmoothL1Loss()
+
+    rb = ReplayBuffer(args.buffer_size, obs_dim)
+
+    exploration_steps = max(1, int(args.exploration_fraction * args.total_timesteps))
+    epsilon_by_step = lambda t: args.end_e + (args.start_e - args.end_e) * max(0.0, (exploration_steps - t) / exploration_steps)
+
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                obs_dict, _ = env_eval.reset(seed=args.seed + 100000 + collected)
+                state = obs_dict['observation']
+                mask = obs_dict['action_mask']
+                traj_states = [state.tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env_eval, '_max_episode_steps', None) or int(args.grid_size * args.grid_size * 6)
+                while not done:
+                    with torch.no_grad():
+                        q = agent_model(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        mask_t = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+                        masked_q = q + (mask_t - 1.0) * 1e8
+                        action = int(torch.argmax(masked_q, dim=1).item())
+                    next_obs_dict, reward, terminated, truncated, info = env_eval.step(action)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_obs_dict['observation']
+                    mask = next_obs_dict['action_mask']
+                    traj_states.append(state.tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    global_step = 0
+    start_time = time.time()
+
+    obs_dict, _ = env.reset(seed=args.seed)
+    obs = obs_dict['observation']
+    mask = obs_dict['action_mask']
+    ep_success_window: List[float] = []
+
+    eval_every_steps = max(1, args.total_timesteps // args.eval_splits)
+
+    while global_step < args.total_timesteps:
+        epsilon = epsilon_by_step(global_step)
+        with torch.no_grad():
+            q_values = policy_net(torch.tensor(obs, dtype=torch.float32, device=device).unsqueeze(0))
+            mask_t = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+            masked_q = q_values + (mask_t - 1.0) * 1e8
+            greedy_action = int(torch.argmax(masked_q, dim=1).item())
+        if (np.random.rand() < epsilon) or (global_step < args.learning_starts):
+            valid = np.where(mask > 0.5)[0]
+            if len(valid) > 0:
+                action = int(np.random.choice(valid))
+            else:
+                action = int(np.random.randint(0, act_dim))
+        else:
+            action = greedy_action
+
+        next_obs_dict, reward, terminated, truncated, info = env.step(action)
+        done = bool(terminated) or bool(truncated)
+
+        next_obs = next_obs_dict['observation']
+        next_mask = next_obs_dict['action_mask']
+
+        rb.add(obs.astype(np.float32), int(action), float(reward), bool(done), next_obs.astype(np.float32))
+
+        obs = next_obs
+        mask = next_mask
+        global_step += 1
+
+        if done:
+            succ = bool((info or {}).get('success', False))
+            ep_success_window.append(1.0 if succ else 0.0)
+            if len(ep_success_window) > 100:
+                ep_success_window.pop(0)
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "rollout/success": float(1.0 if succ else 0.0),
+                        "rollout/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                    }, step=global_step)
+                except Exception:
+                    pass
+            obs_dict, _ = env.reset()
+            obs = obs_dict['observation']
+            mask = obs_dict['action_mask']
+
+        if rb.can_sample(args.batch_size) and (global_step % args.train_frequency == 0) and (global_step > args.learning_starts):
+            batch_obs, batch_act, batch_rew, batch_done, batch_next_obs = rb.sample(args.batch_size)
+            b_obs = torch.tensor(batch_obs, dtype=torch.float32, device=device)
+            b_act = torch.tensor(batch_act, dtype=torch.int64, device=device)
+            b_rew = torch.tensor(batch_rew, dtype=torch.float32, device=device)
+            b_done = torch.tensor(batch_done, dtype=torch.float32, device=device)
+            b_next_obs = torch.tensor(batch_next_obs, dtype=torch.float32, device=device)
+
+            with torch.no_grad():
+                next_q = target_net(b_next_obs).max(dim=1)[0]
+                target_q = b_rew + args.gamma * (1.0 - b_done) * next_q
+
+            current_q = policy_net(b_obs).gather(1, b_act.view(-1, 1)).squeeze(1)
+            loss = criterion(current_q, target_q)
+
+            optimizer.zero_grad()
+            loss.backward()
+            nn.utils.clip_grad_norm_(policy_net.parameters(), max_norm=10.0)
+            optimizer.step()
+
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "charts/epsilon": float(epsilon),
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                        "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+        if global_step % args.target_network_frequency == 0:
+            target_net.load_state_dict(policy_net.state_dict())
+
+        if global_step % 1000 == 0:
+            sps = int(global_step / (time.time() - start_time))
+            sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+            print(f"Step {global_step} | SPS: {sps} | Epsilon: {epsilon:.3f} | SR@100: {sr100:.3f}")
+
+        if global_step == 1 or (global_step % eval_every_steps == 0):
+            try:
+                def eval_thunk():
+                    return make_env(0, run_name, args.seed + 9999, args.grid_size, args.difficulty, False)()
+                collect_eval_trajectories(policy_net, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import wandb
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    env.close()
diff --git a/cleanrl/cleanrl/scout_dqn/noisy_dqn_sokoban.py b/cleanrl/cleanrl/scout_dqn/noisy_dqn_sokoban.py
new file mode 100644
index 0000000000000000000000000000000000000000..182b6247d9ba826bc85c52ae0b038ae6df8d32bc
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/noisy_dqn_sokoban.py
@@ -0,0 +1,541 @@
+# NoisyNet DQN (dueling CNN) for RAGEN Sokoban, tuned for box=2
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Dict, Any, Tuple
+from collections import deque
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+import json
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sokoban.env import SokobanEnv
+from ragen.env.sokoban.config import SokobanEnvConfig
+
+
+class SokobanWrapper(gym.Env):
+    metadata = {"render_modes": ["rgb_array", "human", "ansi", "text"]}
+
+    def __init__(self, env: SokobanEnv):
+        super().__init__()
+        self._env = env
+        self._h = int(self._env.dim_room[0])
+        self._w = int(self._env.dim_room[1])
+        self._tokens = ['#', '_', 'O', '√', 'X', 'P', 'S']
+        self._token_to_idx = {t: i for i, t in enumerate(self._tokens)}
+        self._c = len(self._tokens)
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(self._c, self._h, self._w), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+
+    def _encode_obs(self, text_obs: str) -> np.ndarray:
+        rows = text_obs.split('\n')
+        rows = [list(r) for r in rows if len(r) > 0]
+        h = len(rows)
+        w = len(rows[0]) if h > 0 else self._w
+        grid = np.zeros((self._c, self._h, self._w), dtype=np.float32)
+        for i in range(min(h, self._h)):
+            for j in range(min(w, self._w)):
+                ch = rows[i][j]
+                idx = self._token_to_idx.get(ch, 0)
+                grid[idx, i, j] = 1.0
+        return grid
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        mapped = int(action) + 1  # env expects 1..4
+        text_obs, reward, done, info = self._env.step(mapped)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "SokobanNoisyDQN"
+    total_timesteps: int = 1_000_000
+    learning_rate: float = 2.5e-4
+    gamma: float = 0.99
+    batch_size: int = 128
+    buffer_size: int = 200_000
+    target_network_frequency: int = 8000
+    train_frequency: int = 4
+    learning_starts: int = 20_000
+
+    # Epsilon-greedy (used lightly for warmup)
+    start_e: float = 1.0
+    end_e: float = 0.1
+    exploration_fraction: float = 0.8
+
+    # Model
+    dueling: bool = True
+    reward_clip_abs: float | None = 1.0
+
+    # Eval config
+    eval_splits: int = 2
+    eval_episodes: int = 8000
+
+    # Sokoban env config (default for harder task)
+    grid_h: int = 6
+    grid_w: int = 6
+    num_boxes: int = 2
+    max_steps_env: int = 100
+    search_depth: int = 300
+
+
+def make_env(run_name: str, seed: int, args: Args, capture_video: bool = False):
+    cfg = SokobanEnvConfig(
+        dim_room=(args.grid_h, args.grid_w),
+        max_steps=args.max_steps_env,
+        num_boxes=args.num_boxes,
+        search_depth=args.search_depth,
+        render_mode='text',
+        observation_format='grid',
+    )
+    env = SokobanEnv(cfg)
+    env = SokobanWrapper(env)
+    env = gym.wrappers.RecordEpisodeStatistics(env)
+    if capture_video:
+        env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+    return env
+
+
+class NoisyLinear(nn.Module):
+    def __init__(self, in_features: int, out_features: int, std_init: float = 0.5):
+        super().__init__()
+        self.in_features = in_features
+        self.out_features = out_features
+        self.weight_mu = nn.Parameter(torch.empty(out_features, in_features))
+        self.weight_sigma = nn.Parameter(torch.empty(out_features, in_features))
+        self.register_buffer('weight_epsilon', torch.empty(out_features, in_features))
+        self.bias_mu = nn.Parameter(torch.empty(out_features))
+        self.bias_sigma = nn.Parameter(torch.empty(out_features))
+        self.register_buffer('bias_epsilon', torch.empty(out_features))
+        self.std_init = std_init / np.sqrt(in_features)
+        self.reset_parameters()
+        self.reset_noise()
+
+    def reset_parameters(self):
+        mu_range = 1 / np.sqrt(self.in_features)
+        self.weight_mu.data.uniform_(-mu_range, mu_range)
+        self.weight_sigma.data.fill_(self.std_init)
+        self.bias_mu.data.uniform_(-mu_range, mu_range)
+        self.bias_sigma.data.fill_(self.std_init)
+
+    def reset_noise(self):
+        epsilon_in = torch.randn(self.in_features, device=self.weight_mu.device)
+        epsilon_out = torch.randn(self.out_features, device=self.weight_mu.device)
+        self.weight_epsilon.copy_(epsilon_out.ger(epsilon_in))
+        self.bias_epsilon.copy_(epsilon_out)
+
+    def forward(self, x):
+        if self.training:
+            w = self.weight_mu + self.weight_sigma * self.weight_epsilon
+            b = self.bias_mu + self.bias_sigma * self.bias_epsilon
+        else:
+            w = self.weight_mu
+            b = self.bias_mu
+        return torch.nn.functional.linear(x, w, b)
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    if isinstance(layer, NoisyLinear):
+        nn.init.orthogonal_(layer.weight_mu, std)
+        nn.init.constant_(layer.bias_mu, bias_const)
+        layer.weight_sigma.data.fill_(layer.std_init)
+        layer.bias_sigma.data.fill_(layer.std_init)
+    else:
+        nn.init.orthogonal_(layer.weight, std)
+        nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class QConvNoisy(nn.Module):
+    def __init__(self, obs_shape: Tuple[int, int, int], act_dim: int, dueling: bool = True):
+        super().__init__()
+        c, h, w = obs_shape
+        self.dueling = dueling
+        self._act_dim = act_dim
+        self.features = nn.Sequential(
+            layer_init(nn.Conv2d(c, 32, 3, 1, 1)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(32, 64, 3, 1, 1)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 64, 3, 1, 1)),
+            nn.ReLU(),
+            nn.Flatten(),
+        )
+        fc_in = 64 * h * w
+        if self.dueling:
+            self.adv_head = nn.Sequential(
+                layer_init(NoisyLinear(fc_in, 512)),
+                nn.ReLU(),
+                layer_init(NoisyLinear(512, act_dim), std=0.01),
+            )
+            self.val_head = nn.Sequential(
+                layer_init(NoisyLinear(fc_in, 512)),
+                nn.ReLU(),
+                layer_init(NoisyLinear(512, 1), std=0.01),
+            )
+        else:
+            self.head = nn.Sequential(
+                layer_init(NoisyLinear(fc_in, 512)),
+                nn.ReLU(),
+                layer_init(NoisyLinear(512, act_dim), std=0.01),
+            )
+
+    def reset_noise(self):
+        for m in self.modules():
+            if isinstance(m, NoisyLinear):
+                m.reset_noise()
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.features(x)
+        if self.dueling:
+            adv = self.adv_head(x)
+            val = self.val_head(x)
+            q = val + adv - adv.mean(dim=1, keepdim=True)
+            return q
+        else:
+            q = self.head(x)
+            return q
+
+
+class ReplayBuffer:
+    def __init__(self, capacity: int, obs_shape: Tuple[int, int, int]):
+        self.capacity = capacity
+        self.ptr = 0
+        self.full = False
+        self.obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.next_obs_buf = np.zeros((capacity,) + obs_shape, dtype=np.float32)
+        self.act_buf = np.zeros((capacity,), dtype=np.int64)
+        self.rew_buf = np.zeros((capacity,), dtype=np.float32)
+        self.done_buf = np.zeros((capacity,), dtype=np.float32)
+
+    def add(self, obs: np.ndarray, act: int, rew: float, done: bool, next_obs: np.ndarray):
+        self.obs_buf[self.ptr] = obs
+        self.next_obs_buf[self.ptr] = next_obs
+        self.act_buf[self.ptr] = act
+        self.rew_buf[self.ptr] = rew
+        self.done_buf[self.ptr] = 1.0 if done else 0.0
+        self.ptr = (self.ptr + 1) % self.capacity
+        if self.ptr == 0:
+            self.full = True
+
+    def can_sample(self, batch_size: int) -> bool:
+        return (self.capacity if self.full else self.ptr) >= batch_size
+
+    def sample(self, batch_size: int):
+        size = self.capacity if self.full else self.ptr
+        idxs = np.random.randint(0, size, size=batch_size)
+        return (
+            self.obs_buf[idxs],
+            self.act_buf[idxs],
+            self.rew_buf[idxs],
+            self.done_buf[idxs],
+            self.next_obs_buf[idxs],
+        )
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env
+    env = make_env(run_name, args.seed, args, args.capture_video)
+    obs_shape = env.observation_space.shape  # (C,H,W)
+    act_dim = env.action_space.n
+
+    # networks
+    policy_net = QConvNoisy(obs_shape, act_dim, dueling=args.dueling).to(device)
+    target_net = QConvNoisy(obs_shape, act_dim, dueling=args.dueling).to(device)
+    target_net.load_state_dict(policy_net.state_dict())
+    target_net.eval()
+
+    optimizer = optim.Adam(policy_net.parameters(), lr=args.learning_rate)
+    criterion = nn.SmoothL1Loss()
+
+    rb = ReplayBuffer(args.buffer_size, obs_shape)
+
+    # periodic eval setup
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes: int, step_tag: int):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env_eval.reset(seed=args.seed + 100000 + collected)
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env_eval, '_max_episode_steps', None) or (args.grid_h * args.grid_w * 6)
+                while not done:
+                    with torch.no_grad():
+                        q = agent_model(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(q, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env_eval.step(action)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    # epsilon schedule (log only; noisy nets handle exploration)
+    exploration_steps = max(1, int(args.exploration_fraction * args.total_timesteps))
+    def epsilon_by_step(t: int):
+        return args.end_e + (args.start_e - args.end_e) * max(0.0, (exploration_steps - t) / exploration_steps)
+
+    # training loop
+    global_step = 0
+    start_time = time.time()
+
+    obs, _ = env.reset(seed=args.seed)
+    ep_return = 0.0
+    ep_len = 0
+    ep_success_window = deque(maxlen=100)
+
+    eval_every_steps = max(1, args.total_timesteps // args.eval_splits)
+
+    while global_step < args.total_timesteps:
+        epsilon = epsilon_by_step(global_step)
+        with torch.no_grad():
+            q_values = policy_net(torch.tensor(obs, dtype=torch.float32, device=device).unsqueeze(0))
+            action_greedy = int(torch.argmax(q_values, dim=1).item())
+        if (global_step < args.learning_starts) and (np.random.rand() < 0.5):
+            action = env.action_space.sample()
+        else:
+            action = action_greedy
+        next_obs, reward, terminated, truncated, info = env.step(action)
+        done = bool(terminated) or bool(truncated)
+
+        r = float(reward)
+        if args.reward_clip_abs is not None:
+            cap = float(args.reward_clip_abs)
+            r = max(-cap, min(cap, r))
+
+        rb.add(obs.astype(np.float32), action, r, done, next_obs.astype(np.float32))
+
+        obs = next_obs
+        ep_return += float(reward)
+        ep_len += 1
+        global_step += 1
+
+        # optimize
+        if (global_step > args.learning_starts) and rb.can_sample(args.batch_size) and (global_step % args.train_frequency == 0):
+            batch_obs, batch_act, batch_rew, batch_done, batch_next_obs = rb.sample(args.batch_size)
+            b_obs = torch.tensor(batch_obs, dtype=torch.float32, device=device)
+            b_act = torch.tensor(batch_act, dtype=torch.int64, device=device)
+            b_rew = torch.tensor(batch_rew, dtype=torch.float32, device=device)
+            b_done = torch.tensor(batch_done, dtype=torch.float32, device=device)
+            b_next_obs = torch.tensor(batch_next_obs, dtype=torch.float32, device=device)
+
+            with torch.no_grad():
+                next_actions = policy_net(b_next_obs).argmax(dim=1)
+                next_q = target_net(b_next_obs).gather(1, next_actions.view(-1, 1)).squeeze(1)
+                target_q = b_rew + args.gamma * (1.0 - b_done) * next_q
+
+            current_q = policy_net(b_obs).gather(1, b_act.view(-1, 1)).squeeze(1)
+            loss = criterion(current_q, target_q)
+
+            optimizer.zero_grad()
+            loss.backward()
+            nn.utils.clip_grad_norm_(policy_net.parameters(), max_norm=10.0)
+            optimizer.step()
+
+            # reset noisy parameters
+            policy_net.reset_noise()
+            target_net.reset_noise()
+
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "charts/epsilon": float(epsilon),
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+        # target network update
+        if global_step % args.target_network_frequency == 0:
+            target_net.load_state_dict(policy_net.state_dict())
+
+        if done:
+            succ = bool((info or {}).get('success', False))
+            ep_success_window.append(1.0 if succ else 0.0)
+            if args.track:
+                try:
+                    import wandb
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "rollout/episodic_return": float(ep_return),
+                        "rollout/episodic_length": int(ep_len),
+                        "rollout/success": float(1.0 if succ else 0.0),
+                        "rollout/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                    }, step=global_step)
+                except Exception:
+                    pass
+            obs, _ = env.reset()
+            ep_return, ep_len = 0.0, 0
+
+        # occasional print
+        if global_step % 1000 == 0:
+            sps = int(global_step / (time.time() - start_time))
+            sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+            print(f"Step {global_step} | SPS: {sps} | Epsilon: {epsilon:.3f} | SR@100: {sr100:.3f}")
+
+        # periodic evaluation and trajectory dump
+        if global_step==0 or (global_step % eval_every_steps == 0):
+            try:
+                def eval_thunk():
+                    return make_env(run_name, args.seed + 9999, args, False)
+                collect_eval_trajectories(policy_net, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import wandb
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    # simple evaluation after training
+    def evaluate(n_episodes=200):
+        returns = []
+        successes = []
+        for i in range(n_episodes):
+            s, _ = env.reset(seed=args.seed + 100000 + i)
+            done = False
+            G = 0.0
+            while not done:
+                with torch.no_grad():
+                    q = policy_net(torch.tensor(s, dtype=torch.float32, device=device).unsqueeze(0))
+                    a = int(torch.argmax(q, dim=1).item())
+                s, r, term, trunc, info = env.step(a)
+                G += float(r)
+                done = bool(term) or bool(trunc)
+            successes.append(1.0 if bool((info or {}).get('success', False)) else 0.0)
+            returns.append(G)
+        return float(np.mean(returns)), float(np.std(returns)), float(np.mean(successes))
+
+    avg_ret, std_ret, succ_rate = evaluate(400)
+    if args.track:
+        try:
+            import wandb
+            wandb.log({
+                "global_step": int(global_step),
+                "eval/avg_return": float(avg_ret),
+                "eval/std_return": float(std_ret),
+                "eval/episodes": int(400),
+                "eval/success_rate": float(succ_rate),
+            }, step=global_step)
+        except Exception:
+            pass
+
+    env.close()
diff --git a/cleanrl/cleanrl/scout_dqn/ragen_wrappers.py b/cleanrl/cleanrl/scout_dqn/ragen_wrappers.py
new file mode 100644
index 0000000000000000000000000000000000000000..af63107e99b08764b8410c77be7c66ddbb52899d
--- /dev/null
+++ b/cleanrl/cleanrl/scout_dqn/ragen_wrappers.py
@@ -0,0 +1,235 @@
+"""
+Gymnasium-compatible wrappers for RAGEN environments to enable traditional RL training.
+These wrappers convert text-based observations to numerical representations suitable for MLP networks.
+"""
+import gymnasium as gym
+import numpy as np
+from typing import Any, Dict, Tuple
+
+
+class BanditWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Bandit that uses only observable text.
+    Converts text observations to a fixed-size one-hot hash vector.
+    Does not alter episode semantics and does not inspect env internals.
+    """
+    def __init__(self, env, feature_dim_per_name: int = 16):
+        super().__init__(env)
+        # Two name slots (first/second), each hashed to one-hot of size K
+        self.k = feature_dim_per_name
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(2 * self.k,), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(2)
+
+    def _parse_names(self, text_obs: str):
+        """Extract the two arm names from the prompt text purely via regex/string ops."""
+        # Heuristic: look for the segment after "named " and split by " and "
+        try:
+            anchor = "named "
+            if anchor in text_obs:
+                segment = text_obs.split(anchor, 1)[1]
+                # Cut at newline if present
+                segment = segment.split("\n", 1)[0]
+                # Now split by " and " to get two names; also strip punctuation
+                parts = segment.split(" and ")
+                if len(parts) >= 2:
+                    name_a = parts[0].strip().strip(' .!?,')
+                    name_b = parts[1].strip().strip(' .!?,')
+                    return name_a, name_b
+        except Exception:
+            pass
+        # Fallback: no names found
+        return "", ""
+
+    def _names_to_vector(self, name_a: str, name_b: str) -> np.ndarray:
+        vec = np.zeros(2 * self.k, dtype=np.float32)
+        idx_a = (hash(name_a) % self.k)
+        idx_b = (hash(name_b) % self.k)
+        vec[idx_a] = 1.0
+        vec[self.k + idx_b] = 1.0
+        return vec
+
+    def reset(self, **kwargs):
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        name_a, name_b = self._parse_names(text_obs)
+        return self._names_to_vector(name_a, name_b), {}
+
+    def step(self, action):
+        ragen_action = int(action) + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        name_a, name_b = self._parse_names(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return self._names_to_vector(name_a, name_b), reward, terminated, truncated, info
+
+
+class FrozenLakeWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN FrozenLake environment.
+    Converts grid-based text observations to numerical state representation.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine grid size from text only
+        bootstrap_text = self.env.reset()
+        flat, _ = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        # Serve the bootstrapped obs on first reset without calling env.reset again
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, Tuple[int, int]]:
+        """Parse text observation into numerical state (one-hot grid + player pos)."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        player_pos = None
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Parse grid
+        for i, line in enumerate(lines):
+            row = []
+            for j, char in enumerate(line):
+                if char == 'P':  # Player
+                    row.append(0)
+                    player_pos = (i, j)
+                elif char == '_':  # Frozen
+                    row.append(1)
+                elif char == 'O':  # Hole
+                    row.append(2)
+                elif char == 'G':  # Goal
+                    row.append(3)
+                elif char == 'X':  # Player in hole
+                    row.append(2)
+                    player_pos = (i, j)
+                elif char == '√':  # Player on goal
+                    row.append(3)
+                    player_pos = (i, j)
+                else:
+                    row.append(1)  # Default to frozen
+            grid.append(row)
+        # Pad ragged rows if needed
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        grid_size = (rows, cols)
+        # One-hot encode grid over 4 cell types
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 4), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        # Flatten grid
+        flat_grid = one_hot_grid.flatten()
+        # Add normalized player position
+        if player_pos is None:
+            player_pos = (0, 0)
+        player_pos_norm = np.array([
+            0.0 if rows <= 1 else player_pos[0] / max(1, rows - 1),
+            0.0 if cols <= 1 else player_pos[1] / max(1, cols - 1),
+        ], dtype=np.float32)
+        flat = np.concatenate([flat_grid, player_pos_norm])
+        return flat, grid_size
+    
+    def reset(self, **kwargs):
+        # Filter out 'options' parameter that gymnasium passes but RAGEN doesn't support
+        if self._bootstrap_ready:
+            # First call returns the bootstrapped observation to avoid double reset
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+
+
+class SokobanWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Sokoban environment.
+    Converts grid-based text observations to numerical state representation.
+    Note: Does not inherit from gym.Wrapper due to old gym vs gymnasium compatibility.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine room size from text only
+        bootstrap_text = self.env.reset()
+        flat, rows, cols = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        self.metadata = getattr(env, 'metadata', {})
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, int, int]:
+        """Parse text observation into numerical state and return dims."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Mapping from characters to cell types
+        char_to_type = {
+            '#': 0,  # wall
+            '_': 1,  # empty
+            'O': 2,  # target
+            '√': 3,  # box on target
+            'X': 4,  # box
+            'P': 5,  # player
+            'S': 6,  # player on target
+        }
+        
+        for line in lines:
+            row = []
+            for char in line:
+                row.append(char_to_type.get(char, 1))  # Default to empty
+            grid.append(row)
+        # Pad ragged rows
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 7), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        return one_hot_grid.flatten(), rows, cols
+    
+    def reset(self, **kwargs):
+        if self._bootstrap_ready:
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+    
+    def close(self):
+        if hasattr(self.env, 'close'):
+            self.env.close()
+    
+    def render(self):
+        if hasattr(self.env, 'render'):
+            return self.env.render()
+        return None
diff --git a/cleanrl/cleanrl/scout_ppo/ppo_2048.py b/cleanrl/cleanrl/scout_ppo/ppo_2048.py
new file mode 100644
index 0000000000000000000000000000000000000000..3e91fec32f8c5b419446642e1ca6539330ae3181
--- /dev/null
+++ b/cleanrl/cleanrl/scout_ppo/ppo_2048.py
@@ -0,0 +1,514 @@
+# PPO (CNN actor-critic) for RAGEN 2048, matching NoisyNet DQN args and wandb logging
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Dict, Any, Tuple
+from collections import deque
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import torch.nn.functional as F
+import tyro
+import json
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+# env and config
+from ragen.env.game_2048.env import Game2048Env
+from ragen.env.game_2048.config import Game2048EnvConfig
+
+
+class Game2048Wrapper(gym.Env):
+    metadata = {"render_modes": ["text"]}
+
+    def __init__(self, env: Game2048Env, n_channels: int = 16):
+        super().__init__()
+        self._env = env
+        self._n_channels = int(n_channels)
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(self._n_channels, 4, 4), dtype=np.float32)
+        self.action_space = self._env.action_space
+        self._last_info: Dict[str, Any] | None = None
+
+    def _encode_grid(self, grid: np.ndarray) -> np.ndarray:
+        grid_flat = grid.flatten()
+        with np.errstate(divide='ignore'):
+            power_grid = np.log2(grid_flat, where=(grid_flat > 0)).astype(int)
+            power_grid[grid_flat == 0] = 0
+            power_grid = np.clip(power_grid, 0, self._n_channels - 1)
+            one_hot = np.eye(self._n_channels)[power_grid]
+            obs = one_hot.reshape(4, 4, self._n_channels).transpose(2, 0, 1)
+        return obs.astype(np.float32)
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs, info = self._env.reset(seed=seed, options=options)
+        self._last_info = info
+        grid = info.get('grid', np.zeros((4, 4), dtype=np.int64))
+        obs = self._encode_grid(grid)
+        ret_info = {k: v for k, v in (info or {}).items() if k != 'grid'}
+        try:
+            ret_info['max_tile'] = int(np.max(grid))
+        except Exception:
+            ret_info['max_tile'] = int(ret_info.get('max_tile', 0))
+        return obs, ret_info
+
+    def step(self, action: int):
+        text_obs, reward, done, info = self._env.step(int(action))
+        self._last_info = info
+        grid = info.get('grid', np.zeros((4, 4), dtype=np.int64))
+        obs = self._encode_grid(grid)
+        ret_info = {k: v for k, v in (info or {}).items() if k != 'grid'}
+        try:
+            ret_info['max_tile'] = int(np.max(grid))
+        except Exception:
+            ret_info['max_tile'] = int(ret_info.get('max_tile', 0))
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, ret_info
+
+    def get_action_mask(self) -> np.ndarray:
+        if self._last_info is None:
+            return np.ones((4,), dtype=bool)
+        mask = self._last_info.get('action_mask', None)
+        if mask is None:
+            return np.ones((4,), dtype=bool)
+        return np.asarray(mask, dtype=bool)
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    # mirror DQN args for wandb compatibility
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "2048-RL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm identifiers
+    env_id: str = "Game2048PPO"
+    total_timesteps: int = 3_000_000
+    learning_rate: float = 2.5e-4
+    gamma: float = 0.997
+
+    # DQN-only args kept for wandb/backward-compat (unused here)
+    batch_size: int = 512
+    buffer_size: int = 2400_000
+    target_network_frequency: int = 15000
+    train_frequency: int = 4
+    learning_starts: int = 20_000
+    start_e: float = 1.0
+    end_e: float = 0.05
+    exploration_fraction: float = 0.8
+    dueling: bool = True
+    n_step: int = 10
+    per_alpha: float = 0.5
+    per_beta_start: float = 0.4
+    per_beta_frames: int = 1_000_000
+    per_eps: float = 1e-6
+
+    # Env config
+    two_prob: float = 0.9
+    max_steps_env: int = 1000
+
+    # Eval config
+    eval_splits: int = 1
+    eval_episodes: int = 400
+
+    # PPO specific
+    num_steps: int = 256
+    num_minibatches: int = 8
+    update_epochs: int = 4
+    gae_lambda: float = 0.95
+    clip_coef: float = 0.2
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    anneal_lr: bool = True
+
+
+def make_env(run_name: str, seed: int, args: Args, capture_video: bool = False):
+    cfg = Game2048EnvConfig(size=4, two_prob=args.two_prob, use_log_reward=True)
+    base = Game2048Env(cfg)
+    env = Game2048Wrapper(base)
+    env = gym.wrappers.TimeLimit(env, max_episode_steps=args.max_steps_env)
+    env = gym.wrappers.RecordEpisodeStatistics(env)
+    if capture_video:
+        env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+    return env
+
+
+class ActorCriticCNN(nn.Module):
+    def __init__(self, obs_shape: Tuple[int, int, int], act_dim: int):
+        super().__init__()
+        c, h, w = obs_shape
+        self.features = nn.Sequential(
+            nn.Conv2d(c, 64, 2, 1, 0),
+            nn.ReLU(),
+            nn.Conv2d(64, 128, 2, 1, 1),
+            nn.ReLU(),
+            nn.Conv2d(128, 128, 2, 1, 0),
+            nn.ReLU(),
+            nn.Flatten(),
+        )
+        with torch.no_grad():
+            fc_in = int(self.features(torch.zeros(1, *obs_shape)).shape[1])
+        self.pi = nn.Sequential(
+            nn.Linear(fc_in, 512), nn.ReLU(), nn.Linear(512, act_dim)
+        )
+        self.v = nn.Sequential(
+            nn.Linear(fc_in, 512), nn.ReLU(), nn.Linear(512, 1)
+        )
+
+    def get_value(self, x):
+        x = self.features(x)
+        return self.v(x).squeeze(-1)
+
+    def get_action_and_value(self, x, action=None, action_mask=None):
+        x = self.features(x)
+        logits = self.pi(x)
+        if action_mask is not None:
+            mask = action_mask.bool()
+            logits = torch.where(mask, logits, torch.full_like(logits, -1e9))
+        probs = torch.distributions.Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.v(x).squeeze(-1)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    try:
+        wandb.define_metric("global_step")
+        for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+            wandb.define_metric(prefix, step_metric="global_step")
+    except Exception:
+        pass
+
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    env = make_env(run_name, args.seed, args, args.capture_video)
+    obs_shape = env.observation_space.shape
+    act_dim = env.action_space.n
+
+    agent = ActorCriticCNN(obs_shape, act_dim).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # Storage
+    num_steps = args.num_steps
+    obs = np.zeros((num_steps,) + obs_shape, dtype=np.float32)
+    actions = np.zeros((num_steps,), dtype=np.int64)
+    logprobs = np.zeros((num_steps,), dtype=np.float32)
+    rewards = np.zeros((num_steps,), dtype=np.float32)
+    dones = np.zeros((num_steps,), dtype=np.float32)
+    values = np.zeros((num_steps,), dtype=np.float32)
+    masks_buf = np.zeros((num_steps, act_dim), dtype=bool)
+
+    global_step = 0
+    start_time = time.time()
+
+    next_obs, info = env.reset(seed=args.seed)
+    current_info = info or {}
+    next_done = False
+
+    ep_return = 0.0
+    ep_len = 0
+    ep_success_window = deque(maxlen=100)
+    ep_return_window = deque(maxlen=100)
+    step_reward_window = deque(maxlen=2048)
+
+    eval_every_steps = max(1, args.total_timesteps // args.eval_splits)
+
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes: int, step_tag: int):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env_eval = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, info = env_eval.reset(seed=args.seed + 100000 + collected)
+                current_info = info or {}
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env_eval, '_max_episode_steps', None) or args.max_steps_env
+                while not done:
+                    with torch.no_grad():
+                        s = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                        mask_np = current_info.get('action_mask', np.ones(act_dim, dtype=bool))
+                        mask = torch.tensor(mask_np, dtype=torch.bool, device=device).unsqueeze(0)
+                        logits = agent_model.pi(agent_model.features(s))
+                        logits = torch.where(mask, logits, torch.full_like(logits, -1e9))
+                        action = int(torch.argmax(logits, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env_eval.step(action)
+                    traj_actions.append(int(action))
+                    raw_r = info.get('raw_reward', reward) if info else reward
+                    traj_rewards.append(float(raw_r))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    current_info = info or {}
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env_eval.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    num_updates = args.total_timesteps // num_steps
+
+    for update in range(1, num_updates + 1):
+        if args.anneal_lr:
+            frac = 1.0 - (update - 1.0) / float(max(1, num_updates))
+            lrnow = args.learning_rate * frac
+            for pg in optimizer.param_groups:
+                pg['lr'] = lrnow
+
+        for step in range(num_steps):
+            obs[step] = next_obs
+            dones[step] = float(next_done)
+            with torch.no_grad():
+                s = torch.tensor(next_obs, dtype=torch.float32, device=device).unsqueeze(0)
+                mask_np = current_info.get('action_mask', np.ones(act_dim, dtype=bool))
+                masks_buf[step] = mask_np
+                mask = torch.tensor(mask_np, dtype=torch.bool, device=device).unsqueeze(0)
+                a, lp, ent, val = agent.get_action_and_value(s, action_mask=mask)
+            action = int(a.item())
+            next_obs, reward, terminated, truncated, info = env.step(action)
+            done = bool(terminated) or bool(truncated)
+
+            rewards[step] = float(reward)  # training reward (log-scale per env)
+            actions[step] = action
+            logprobs[step] = float(lp.item())
+            values[step] = float(val.item())
+
+            # logging raw reward for charts
+            raw_r = float((info or {}).get('raw_reward', reward))
+            ep_return += raw_r
+            try:
+                step_reward_window.append(float(reward))
+            except Exception:
+                pass
+            ep_len += 1
+            global_step += 1
+
+            if done:
+                succ = bool((info or {}).get('success', False))
+                max_tile = int((info or {}).get('max_tile', 0))
+                ep_success_window.append(1.0 if succ else 0.0)
+                ep_return_window.append(float(ep_return))
+                try:
+                    if max_tile is not None:
+                        print(f"global_step={global_step}, episodic_return={ep_return:.1f}, length={ep_len}, max_tile={int(max_tile)}, success={succ}")
+                    else:
+                        print(f"global_step={global_step}, episodic_return={ep_return:.1f}, length={ep_len}, success={succ}")
+                except Exception:
+                    pass
+                if args.track:
+                    try:
+                        import wandb
+                        avg_ep_ret = float(np.mean(ep_return_window)) if len(ep_return_window) > 0 else 0.0
+                        wandb.log({
+                            "global_step": int(global_step),
+                            "rollout/episodic_return": float(ep_return),
+                            "rollout/episodic_length": int(ep_len),
+                            "rollout/success": float(1.0 if succ else 0.0),
+                            "rollout/max_tile": int(max_tile),
+                            "rollout/success_rate_100": float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else None,
+                            "charts/episodic_return": float(ep_return),
+                            "charts/episodic_length": int(ep_len),
+                            "charts/success": float(1.0 if succ else 0.0),
+                            "charts/max_tile": int(max_tile),
+                            "charts/avg_episode_return": avg_ep_ret,
+                        }, step=global_step)
+                    except Exception:
+                        pass
+                next_obs, info = env.reset()
+                current_info = info or {}
+                ep_return, ep_len = 0.0, 0
+            else:
+                current_info = info or {}
+                next_done = False
+
+            if global_step % 1000 == 0:
+                sps = int(global_step / (time.time() - start_time))
+                sr100 = float(np.mean(ep_success_window)) if len(ep_success_window) > 0 else 0.0
+                print(f"Step {global_step} | SPS: {sps} | SR@100: {sr100:.3f}")
+
+            if (global_step % eval_every_steps == 0):
+                try:
+                    def eval_thunk():
+                        return make_env(run_name, args.seed + 9999, args, False)
+                    collect_eval_trajectories(agent, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                    if args.track:
+                        try:
+                            import wandb
+                            mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                            if mpath.exists():
+                                with mpath.open("r") as mf:
+                                    metrics = json.load(mf)
+                                wandb.log({
+                                    "eval/success_rate": metrics.get("success_rate"),
+                                    "eval/avg_return": metrics.get("avg_return"),
+                                    "eval/std_return": metrics.get("std_return"),
+                                    "eval/episodes": metrics.get("episodes"),
+                                }, step=global_step)
+                        except Exception:
+                            pass
+                    print(f"Collected {args.eval_episodes} eval trajectories at step {global_step}")
+                except Exception as e:
+                    print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+        # Compute GAE
+        with torch.no_grad():
+            s = torch.tensor(next_obs, dtype=torch.float32, device=device).unsqueeze(0)
+            next_value = agent.get_value(s).item()
+        advantages = np.zeros_like(rewards)
+        lastgaelam = 0.0
+        for t in reversed(range(num_steps)):
+            if t == num_steps - 1:
+                nextnonterminal = 1.0 - next_done
+                nextvalues = next_value
+            else:
+                nextnonterminal = 1.0 - dones[t + 1]
+                nextvalues = values[t + 1]
+            delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+            advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+        returns = advantages + values
+
+        # Flatten
+        b_obs = torch.tensor(obs, dtype=torch.float32, device=device)
+        b_actions = torch.tensor(actions, dtype=torch.int64, device=device)
+        b_logprobs = torch.tensor(logprobs, dtype=torch.float32, device=device)
+        b_returns = torch.tensor(returns, dtype=torch.float32, device=device)
+        b_values = torch.tensor(values, dtype=torch.float32, device=device)
+        b_advantages = torch.tensor(advantages, dtype=torch.float32, device=device)
+        b_masks = torch.tensor(masks_buf, dtype=torch.bool, device=device)
+
+        b_advantages = (b_advantages - b_advantages.mean()) / (b_advantages.std() + 1e-8)
+
+        # PPO epochs
+        batch_size = num_steps
+        minibatch_size = batch_size // args.num_minibatches
+        inds = np.arange(batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(inds)
+            for start in range(0, batch_size, minibatch_size):
+                end = start + minibatch_size
+                mb_inds = inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(
+                    b_obs[mb_inds], action=b_actions[mb_inds], action_mask=b_masks[mb_inds]
+                )
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+                with torch.no_grad():
+                    approx_kl = ((ratio - 1) - logratio).mean().item()
+                mb_adv = b_advantages[mb_inds]
+                pg_loss1 = -mb_adv * ratio
+                pg_loss2 = -mb_adv * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                v_clipped = b_values[mb_inds] + torch.clamp(
+                    newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef
+                )
+                v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                v_loss = 0.5 * v_loss_max
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + args.vf_coef * v_loss
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.track:
+                try:
+                    import wandb
+                    try:
+                        avg_reward_val = float(np.mean(step_reward_window)) if len(step_reward_window) > 0 else 0.0
+                    except Exception:
+                        avg_reward_val = 0.0
+                    wandb.log({
+                        "global_step": int(global_step),
+                        "train/loss": float(loss.item()),
+                        "train/value_loss": float(v_loss.item()),
+                        "train/policy_loss": float(pg_loss.item()),
+                        "train/entropy": float(entropy_loss.item()),
+                        "losses/explained_variance": None,
+                        "charts/avg_reward": avg_reward_val,
+                        "charts/avg_value": float(b_values.mean().item()) if b_values.numel() > 0 else None,
+                        "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                        "charts/epsilon": None,
+                        "perf/SPS": int(global_step / (time.time() - start_time)),
+                    }, step=global_step)
+                except Exception:
+                    pass
+
+    env.close()
diff --git a/cleanrl/cleanrl/scout_ppo/ppo_bandit_small.py b/cleanrl/cleanrl/scout_ppo/ppo_bandit_small.py
new file mode 100644
index 0000000000000000000000000000000000000000..7da74b77a8e14a5bf000c86ba3390dfebcab1913
--- /dev/null
+++ b/cleanrl/cleanrl/scout_ppo/ppo_bandit_small.py
@@ -0,0 +1,410 @@
+# PPO with small MLP for RAGEN Bandit using the existing env (no env edits)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any, Dict
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.bandit.env import BanditEnv
+from ragen.env.bandit.config import BanditEnvConfig
+from ragen.env.base import BaseDiscreteActionEnv
+from ragen_wrappers import BanditWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "Bandit"
+    total_timesteps: int = 50_000
+    learning_rate: float = 3e-4
+    num_envs: int = 16
+    num_steps: int = 16
+    anneal_lr: bool = True
+    gamma: float = 0.0  # single-step bandit; no bootstrapping
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    # Model size
+    hidden_size: int = 32
+    feature_dim_per_name: int = 16  # BanditWrapper setting
+
+    # runtime filled
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+
+
+def make_env(idx, run_name, seed, feature_dim_per_name: int, capture_video=False):
+    def thunk():
+        cfg = BanditEnvConfig()
+        env = BanditEnv(cfg)
+        env = BanditWrapper(env, feature_dim_per_name=feature_dim_per_name)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs, hidden: int):
+        super().__init__()
+        obs_shape = int(np.array(envs.single_observation_space.shape).prod())
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None):
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # envs
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args.feature_dim_per_name, args.capture_video)
+        for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs, hidden=args.hidden_size).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # storage
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # start
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    eval_out_dir = Path(f"runs/{run_name}/trajectories")
+    eval_out_dir.mkdir(parents=True, exist_ok=True)
+    episode_successes = []
+
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env.reset(seed=args.seed + 200000 + collected)
+                traj_rewards = []
+                done = False
+                while not done:
+                    with torch.no_grad():
+                        logits = agent_model.actor(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0))
+                        action = int(torch.argmax(logits, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env.step(action)
+                    traj_rewards.append(float(reward))
+                    done = bool(terminated) or bool(truncated)
+                    state = next_state
+                ep_ret = float(sum(traj_rewards))
+                succ = False
+                try:
+                    succ = bool((info or {}).get('success', False))
+                except Exception:
+                    pass
+                f.write(json.dumps({"episode_return": ep_ret, "success": succ}) + "\n")
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if succ else 0.0)
+                collected += 1
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    eval_splits = 2
+    eval_episodes = 4000
+    eval_every_iters = max(1, (args.total_timesteps // args.batch_size) // eval_splits)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Anneal LR
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+            # Episode stats logging: compute success and success_rate_100 if available
+            try:
+                mask = None
+                if isinstance(infos, dict):
+                    if "_episode" in infos:
+                        mask = np.asarray(infos["_episode"]).astype(bool)
+                    elif "episode" in infos and isinstance(infos["episode"], dict) and "_l" in infos["episode"]:
+                        mask = np.asarray(infos["episode"]["_l"]).astype(bool)
+                if mask is not None and np.any(mask):
+                    r_arr = np.asarray(infos.get("episode", {}).get("r", np.zeros_like(mask, dtype=float)))
+                    l_arr = np.asarray(infos.get("episode", {}).get("l", np.zeros_like(mask, dtype=int)))
+                    if "success" in infos:
+                        succ_arr = np.asarray(infos.get("success", np.zeros_like(mask, dtype=bool))).astype(float)
+                    else:
+                        try:
+                            succ_arr = (np.asarray(r_arr) > 0).astype(float)
+                        except Exception:
+                            succ_arr = np.zeros_like(mask, dtype=float)
+                    for s in np.asarray(succ_arr)[mask]:
+                        episode_successes.append(float(s))
+                    if args.track:
+                        try:
+                            import wandb
+                            log_dict = {
+                                "global_step": int(global_step),
+                                "rollout/ep_rew_mean": float(np.mean(r_arr[mask])) if np.any(mask) else None,
+                                "rollout/ep_len_mean": float(np.mean(l_arr[mask])) if np.any(mask) else None,
+                                "rollout/success_rate": float(np.mean(succ_arr[mask])) if np.any(mask) else None,
+                            }
+                            if np.any(mask):
+                                last_idx = np.where(mask)[0][-1]
+                                log_dict.update({
+                                    "train/episodic_return": float(r_arr[last_idx]),
+                                    "train/episodic_length": int(l_arr[last_idx]),
+                                    "train/success": float(succ_arr[last_idx]),
+                                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 1 else None,
+                                })
+                            wandb.log(log_dict, step=global_step)
+                        except Exception:
+                            pass
+            except Exception:
+                pass
+
+        # Since gamma=0 for bandit, GAE simplifies but we keep general code
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # update
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss = 0.5 * torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Value: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(values.mean().item()),
+                    "perf/SPS": int(sps),
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                }, step=global_step)
+            except Exception:
+                pass
+
+        # periodic evaluation collection
+        if iteration % eval_every_iters == 0:
+            try:
+                eval_thunk = make_env(0, run_name, args.seed + 9999, args.feature_dim_per_name, False)
+                collect_eval_trajectories(agent, eval_thunk, n_episodes=eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import json as _json
+                        from pathlib import Path as _Path
+                        mpath = _Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = _json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
diff --git a/cleanrl/cleanrl/scout_ppo/ppo_sokoban.py b/cleanrl/cleanrl/scout_ppo/ppo_sokoban.py
new file mode 100644
index 0000000000000000000000000000000000000000..2a210ef7e57218808049d71e293861bcfef65e9c
--- /dev/null
+++ b/cleanrl/cleanrl/scout_ppo/ppo_sokoban.py
@@ -0,0 +1,501 @@
+import os
+import random
+import time
+from dataclasses import dataclass
+from typing import Dict, Any, Tuple
+from pathlib import Path
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from torch.distributions.categorical import Categorical
+import tyro
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sokoban.env import SokobanEnv
+from ragen.env.sokoban.config import SokobanEnvConfig
+
+
+class SokobanWrapper(gym.Env):
+    metadata = {"render_modes": ["rgb_array", "human", "ansi", "text"]}
+
+    def __init__(self, env: SokobanEnv):
+        super().__init__()
+        self._env = env
+        self._h = int(self._env.dim_room[0])
+        self._w = int(self._env.dim_room[1])
+        self._tokens = ['#', '_', 'O', '√', 'X', 'P', 'S']
+        self._token_to_idx = {t: i for i, t in enumerate(self._tokens)}
+        self._c = len(self._tokens)
+        self.observation_space = gym.spaces.Box(low=0.0, high=1.0, shape=(self._c, self._h, self._w), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+
+    def _encode_obs(self, text_obs: str) -> np.ndarray:
+        rows = text_obs.split('\n')
+        rows = [list(r) for r in rows if len(r) > 0]
+        h = len(rows)
+        w = len(rows[0]) if h > 0 else self._w
+        grid = np.zeros((self._c, self._h, self._w), dtype=np.float32)
+        for i in range(min(h, self._h)):
+            for j in range(min(w, self._w)):
+                ch = rows[i][j]
+                idx = self._token_to_idx.get(ch, 0)
+                grid[idx, i, j] = 1.0
+        return grid
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        mapped = int(action) + 1
+        text_obs, reward, done, info = self._env.step(mapped)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    env_id: str = "Sokoban"
+    total_timesteps: int = 10_000_000
+    learning_rate: float = 2.5e-4
+    num_envs: int = 8
+    num_steps: int = 128
+    anneal_lr: bool = True
+    gamma: float = 0.99
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.05
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    grid_h: int = 6
+    grid_w: int = 6
+    num_boxes: int = 2
+    max_steps_env: int = 150
+    search_depth: int = 500
+
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+    # eval config to mirror reference script
+    eval_splits: int = 20
+    eval_episodes: int = 400
+
+
+def make_env(idx: int, run_name: str, seed: int, args: Args, capture_video=False):
+    def thunk():
+        cfg = SokobanEnvConfig(
+            dim_room=(args.grid_h, args.grid_w),
+            max_steps=args.max_steps_env,
+            num_boxes=args.num_boxes,
+            search_depth=args.search_depth,
+            render_mode='text',
+            observation_format='grid',
+        )
+        env = SokobanEnv(cfg)
+        env = SokobanWrapper(env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    nn.init.orthogonal_(layer.weight, std)
+    nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        c, h, w = envs.single_observation_space.shape
+        hidden = 1024
+        self.net = nn.Sequential(
+            layer_init(nn.Conv2d(c, 64, kernel_size=3, stride=1, padding=1)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1)),
+            nn.ReLU(),
+            layer_init(nn.Conv2d(128, 128, kernel_size=3, stride=1, padding=1)),
+            nn.ReLU(),
+            nn.Flatten(),
+            layer_init(nn.Linear(128 * h * w, hidden)),
+            nn.ReLU(),
+        )
+        # twin critics (double value heads)
+        self.critic1 = layer_init(nn.Linear(hidden, 1), std=1.0)
+        self.critic2 = layer_init(nn.Linear(hidden, 1), std=1.0)
+        self.actor = layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01)
+
+    def get_values(self, x):
+        x = self.net(x)
+        return self.critic1(x), self.critic2(x)
+
+    def get_action_and_value(self, x, action=None):
+        x = self.net(x)
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        v1 = self.critic1(x)
+        v2 = self.critic2(x)
+        return action, probs.log_prob(action), probs.entropy(), v1, v2
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args, args.capture_video) for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values1 = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values2 = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.tensor(next_obs, dtype=torch.float32).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    
+    episode_returns = []
+    episode_steps = []
+    episode_successes = []
+
+    # Eval helper mirroring reference implementation
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                state, _ = env.reset(seed=args.seed + 100000 + collected)
+                traj_states = [np.asarray(state).tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                # rely on env internal max steps; add a safety cap
+                max_eval_steps = getattr(env, '_max_episode_steps', None) or (args.grid_h * args.grid_w * 6)
+                while not done:
+                    with torch.no_grad():
+                        logits = agent_model.actor(agent_model.net(torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)))
+                        action = int(torch.argmax(logits, dim=1).item())
+                    next_state, reward, terminated, truncated, info = env.step(action)
+                    traj_actions.append(int(action))
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool((info or {}).get('success', False)))
+                    state = next_state
+                    traj_states.append(np.asarray(state).tolist())
+                    done = d
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    # Eval cadence identical to reference style
+    eval_every_iters = max(1, args.num_iterations // args.eval_splits)
+
+    for iteration in range(1, args.num_iterations + 1):
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            with torch.no_grad():
+                action, logprob, _, v1, v2 = agent.get_action_and_value(next_obs)
+                values1[step] = v1.flatten()
+                values2[step] = v2.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs_np, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs = torch.tensor(next_obs_np, dtype=torch.float32).to(device)
+            next_done = torch.tensor(next_done, dtype=torch.float32).to(device)
+
+            # Episode stats logging using Gymnasium vector env final_info
+            try:
+                if isinstance(infos, dict) and "final_info" in infos and infos["final_info"] is not None:
+                    finals = infos["final_info"]
+                    for fi in finals:
+                        if fi is None:
+                            continue
+                        ep_r = float(fi.get("episode", {}).get("r", 0.0)) if isinstance(fi.get("episode"), dict) else float(fi.get("reward", 0.0))
+                        ep_l = int(fi.get("episode", {}).get("l", 0)) if isinstance(fi.get("episode"), dict) else int(fi.get("length", 0))
+                        # infer success from positive episodic return if the env doesn't set it
+                        ep_succ = float(fi.get("success", 1.0 if ep_r > 0.0 else 0.0))
+                        episode_returns.append(ep_r)
+                        episode_steps.append(global_step)
+                        episode_successes.append(ep_succ)
+                        if args.track:
+                            try:
+                                import wandb
+                                wandb.log({
+                                    "global_step": int(global_step),
+                                    # DQN-aligned episodic keys
+                                    "rollout/episodic_return": float(ep_r),
+                                    "rollout/episodic_length": int(ep_l),
+                                    "rollout/success": float(ep_succ),
+                                    "rollout/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 100 else None,
+                                    # duplicate train-prefixed keys as in DQN
+                                    "train/episodic_return": float(ep_r),
+                                    "train/episodic_length": int(ep_l),
+                                    "train/success": float(ep_succ),
+                                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 100 else None,
+                                }, step=global_step)
+                            except Exception:
+                                pass
+            except Exception:
+                pass
+
+        with torch.no_grad():
+            nv1, nv2 = agent.get_values(next_obs)
+            next_value_min = torch.minimum(nv1, nv2).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues_min = next_value_min
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues_min = torch.minimum(values1[t + 1], values2[t + 1])
+                values_min_t = torch.minimum(values1[t], values2[t])
+                delta = rewards[t] + args.gamma * nextvalues_min * nextnonterminal - values_min_t
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            values_min = torch.minimum(values1, values2)
+            returns = advantages + values_min
+
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values1 = values1.reshape(-1)
+        b_values2 = values2.reshape(-1)
+        b_values_min = torch.minimum(b_values1, b_values2)
+
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue1, newvalue2 = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue1 = newvalue1.view(-1)
+                newvalue2 = newvalue2.view(-1)
+                if args.clip_vloss:
+                    # head 1
+                    v1_unclipped = (newvalue1 - b_returns[mb_inds]) ** 2
+                    v1_clipped_old = b_values1[mb_inds]
+                    v1_clipped = v1_clipped_old + torch.clamp(newvalue1 - v1_clipped_old, -args.clip_coef, args.clip_coef)
+                    v1_loss = torch.max(v1_unclipped, (v1_clipped - b_returns[mb_inds]) ** 2).mean()
+                    # head 2
+                    v2_unclipped = (newvalue2 - b_returns[mb_inds]) ** 2
+                    v2_clipped_old = b_values2[mb_inds]
+                    v2_clipped = v2_clipped_old + torch.clamp(newvalue2 - v2_clipped_old, -args.clip_coef, args.clip_coef)
+                    v2_loss = torch.max(v2_unclipped, (v2_clipped - b_returns[mb_inds]) ** 2).mean()
+                    v_loss = 0.5 * (v1_loss + v2_loss)
+                else:
+                    v1_loss = ((newvalue1 - b_returns[mb_inds]) ** 2).mean()
+                    v2_loss = ((newvalue2 - b_returns[mb_inds]) ** 2).mean()
+                    v_loss = 0.5 * (v1_loss + v2_loss)
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values_min.detach().cpu().numpy(), b_returns.detach().cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Value(min): {b_values_min.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "train/loss": float(loss.item()),
+                    "charts/epsilon": None,
+                    "perf/SPS": int(sps),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(b_values_min.mean().item()),
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                    # training success rate over all finished episodes so far
+                    "train/success_rate": float(np.mean(episode_successes)) if len(episode_successes) > 0 else None,
+                    # keep a short-horizon success rate to monitor recent progress
+                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) > 0 else None,
+                }, step=global_step)
+            except Exception:
+                pass
+
+        # periodic evaluation collection and logging
+        if iteration % eval_every_iters == 0:
+            try:
+                def eval_thunk():
+                    # reuse same config and wrapper as training, but single env
+                    return make_env(0, run_name, args.seed + 9999, args, False)()
+                collect_eval_trajectories(agent, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        mpath = Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at global_step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
diff --git a/cleanrl/cleanrl/scout_ppo/ppo_sudoku_actionmask.py b/cleanrl/cleanrl/scout_ppo/ppo_sudoku_actionmask.py
new file mode 100644
index 0000000000000000000000000000000000000000..1c8015e746431024a871eb7ced372e89aa501bb6
--- /dev/null
+++ b/cleanrl/cleanrl/scout_ppo/ppo_sudoku_actionmask.py
@@ -0,0 +1,588 @@
+# PPO with Action Masking for RAGEN Sudoku (4x4, max_step=20)
+import os
+import random
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Tuple, Dict, Any, List
+import json
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+
+import sys
+# 假设 ragen 库在两级目录之上，请根据实际情况调整
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sudoku.env import SudokuEnv
+from ragen.env.sudoku.config import SudokuEnvConfig
+
+
+class SudokuWrapper(gym.Env):
+    """
+    Adapter to use ragen SudokuEnv with Gymnasium vector API.
+    Improvements: Returns a Dict observation with 'action_mask' to prevent
+    the agent from modifying cells that are already filled.
+    """
+    metadata = {"render_modes": ["rgb_array", "human", "ansi"]}
+
+    def __init__(self, env: SudokuEnv, grid_size: int):
+        super().__init__()
+        self._env = env
+        self._size = grid_size
+        # 0 denotes empty, 1..grid_size denote values
+        self._val_dim = self._size + 1
+        
+        # Actions: (row, col, num) -> Flattened
+        self._act_n = self._size * self._size * self._size
+        self.action_space = gym.spaces.Discrete(self._act_n)
+
+        # Observation: Dict with mask
+        self.observation_space = gym.spaces.Dict({
+            "observation": gym.spaces.Box(low=0.0, high=1.0, shape=(self._size * self._size * self._val_dim,), dtype=np.float32),
+            "action_mask": gym.spaces.Box(low=0.0, high=1.0, shape=(self._act_n,), dtype=np.float32)
+        })
+
+    def _encode_obs(self, text_obs: str) -> Dict[str, np.ndarray]:
+        # Parse the 'simple' grid format
+        vals: List[int] = []
+        for line in text_obs.splitlines():
+            ls = line.strip()
+            if len(ls) == 0: continue
+            if set(ls) <= {'-'}: continue
+            tokens = [t for t in ls.split() if t != '|']
+            if len(tokens) == 0: continue
+            for t in tokens:
+                if t == '.': vals.append(0)
+                else:
+                    try: v = int(t)
+                    except ValueError: v = 0
+                    vals.append(v)
+        
+        target = self._size * self._size
+        if len(vals) < target: vals.extend([0] * (target - len(vals)))
+        if len(vals) > target: vals = vals[:target]
+        
+        # One-hot encode grid
+        grid = np.zeros((target, self._val_dim), dtype=np.float32)
+        # Initialize mask (1.0 = valid, 0.0 = invalid)
+        mask = np.ones(self._act_n, dtype=np.float32)
+
+        for i, v in enumerate(vals):
+            v_clamped = int(v)
+            if v_clamped < 0 or v_clamped > self._size:
+                v_clamped = 0
+            grid[i, v_clamped] = 1.0
+
+            # If a cell is NOT empty (v_clamped != 0), mask all actions for this cell.
+            # Agent should not overwrite existing numbers.
+            if v_clamped != 0:
+                start_idx = i * self._size
+                end_idx = start_idx + self._size
+                mask[start_idx:end_idx] = 0.0
+
+        return {
+            "observation": grid.reshape(-1),
+            "action_mask": mask
+        }
+
+    @staticmethod
+    def _decode_action(action_id: int, grid_size: int) -> Tuple[int, int, int]:
+        g = grid_size
+        row = action_id // (g * g)
+        rem = action_id % (g * g)
+        col = rem // g
+        num = (rem % g) + 1
+        return row, col, num
+
+    def reset(self, *, seed: int | None = None, options: Dict[str, Any] | None = None):
+        text_obs = self._env.reset(seed=seed)
+        obs = self._encode_obs(text_obs)
+        return obs, {}
+
+    def step(self, action: int):
+        row, col, num = self._decode_action(int(action), self._size)
+        act_str = f"{row+1},{col+1},{num}"
+        text_obs, reward, done, info = self._env.step(act_str)
+        obs = self._encode_obs(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return obs, float(reward), terminated, truncated, info or {}
+
+    def render(self):
+        return self._env.render()
+
+    def close(self):
+        self._env.close()
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    seed: int = 1
+    torch_deterministic: bool = True
+    cuda: bool = True
+    track: bool = True
+    wandb_project_name: str = "cleanRL"
+    wandb_entity: str | None = None
+    capture_video: bool = False
+
+    # Algorithm
+    env_id: str = "Sudoku"
+    total_timesteps: int = 10000_000
+    learning_rate: float = 3e-4
+    num_envs: int = 8
+    num_steps: int = 128
+    anneal_lr: bool = True
+    gamma: float = 0.99
+    gae_lambda: float = 0.95
+    num_minibatches: int = 4
+    update_epochs: int = 4
+    norm_adv: bool = True
+    clip_coef: float = 0.2
+    clip_vloss: bool = True
+    ent_coef: float = 0.01
+    vf_coef: float = 0.5
+    max_grad_norm: float = 0.5
+    target_kl: float | None = None
+
+    # Sudoku specific
+    grid_size: int = 4
+    difficulty: str = "easy"
+
+    # runtime filled
+    batch_size: int = 0
+    minibatch_size: int = 0
+    num_iterations: int = 0
+
+    # eval
+    eval_splits: int = 2
+    eval_episodes: int = 4000
+
+
+def make_env(idx, run_name, seed, grid_size, difficulty, capture_video=False):
+    def thunk():
+        config = SudokuEnvConfig(
+            grid_size=grid_size,
+            difficulty=difficulty,
+            render_mode='text',
+            render_format='simple',
+        )
+        env = SudokuEnv(config)
+        env = SudokuWrapper(env, grid_size)
+        # Use env's own max_steps default if available, otherwise a sane cap
+        # Keeping your request for strict step limit logic, although wrapper enforces logic
+        max_steps = 81
+        # max_steps = int(grid_size * grid_size * 6)
+        env = gym.wrappers.TimeLimit(env, max_episode_steps=max_steps)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        # Accessing the shape from the Dict space
+        obs_shape = int(np.array(envs.single_observation_space['observation'].shape).prod())
+        hidden = 256 # Increased hidden size slightly for better capacity
+
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, hidden)),
+            nn.Tanh(),
+            layer_init(nn.Linear(hidden, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None, action_mask=None):
+        logits = self.actor(x)
+        
+        # Apply Action Masking
+        if action_mask is not None:
+            # Set logits of invalid actions to a very large negative number
+            logits = logits + (action_mask - 1.0) * 1e8
+
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+
+    if args.track:
+        import wandb
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+        try:
+            wandb.define_metric("global_step")
+            for prefix in ["train/*", "rollout/*", "eval/*", "losses/*", "charts/*", "perf/*"]:
+                wandb.define_metric(prefix, step_metric="global_step")
+        except Exception:
+            pass
+
+    # seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # envs
+    envs = gym.vector.SyncVectorEnv([
+        make_env(i, run_name, args.seed, args.grid_size, args.difficulty, args.capture_video)
+        for i in range(args.num_envs)
+    ])
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete)
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # storage
+    # Note: obs storage now only stores the flattened grid part
+    obs_shape = envs.single_observation_space['observation'].shape
+    mask_shape = envs.single_observation_space['action_mask'].shape
+    
+    obs = torch.zeros((args.num_steps, args.num_envs) + obs_shape).to(device)
+    masks = torch.zeros((args.num_steps, args.num_envs) + mask_shape).to(device) # Storage for masks
+    
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # start
+    global_step = 0
+    start_time = time.time()
+    
+    # envs.reset() returns a Dict of stacked arrays
+    next_obs_dict, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+    next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+    
+    episode_returns = []
+    episode_steps = []
+    episode_successes = []
+
+    # Eval helper
+    def collect_eval_trajectories(agent_model, make_env_fn, n_episodes, step_tag):
+        out_dir = Path(f"runs/{run_name}/trajectories/step_{step_tag}")
+        out_dir.mkdir(parents=True, exist_ok=True)
+        out_path = out_dir / "trajectories.jsonl"
+        env = make_env_fn()
+        collected = 0
+        summary_returns = []
+        summary_success = []
+        with out_path.open("w") as f:
+            while collected < n_episodes:
+                obs_dict, _ = env.reset(seed=args.seed + collected)
+                # Handle single env dict unpacking
+                state = obs_dict['observation']
+                mask = obs_dict['action_mask']
+                
+                traj_states = [state.tolist()]
+                traj_actions = []
+                traj_rewards = []
+                traj_dones = []
+                traj_success = []
+                done = False
+                step_count = 0
+                max_eval_steps = getattr(env, '_max_episode_steps', None) or int(args.grid_size * args.grid_size * 6)
+                
+                # Eval loop
+                current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+
+                while not done:
+                    with torch.no_grad():
+                        # Pass mask to actor during eval
+                        action, _, _, _ = agent_model.get_action_and_value(current_obs, action_mask=current_mask)
+                        action_item = int(action.item())
+                        
+                    next_obs_dict, reward, terminated, truncated, info = env.step(action_item)
+                    
+                    traj_actions.append(action_item)
+                    traj_rewards.append(float(reward))
+                    step_count += 1
+                    d = bool(terminated) or bool(truncated) or (step_count >= max_eval_steps)
+                    traj_dones.append(d)
+                    traj_success.append(bool(info.get('success', False)))
+                    
+                    state = next_obs_dict['observation']
+                    mask = next_obs_dict['action_mask']
+                    current_obs = torch.tensor(state, dtype=torch.float32, device=device).unsqueeze(0)
+                    current_mask = torch.tensor(mask, dtype=torch.float32, device=device).unsqueeze(0)
+                    
+                    traj_states.append(state.tolist())
+                    done = d
+                    
+                ep_ret = float(sum(traj_rewards))
+                ep_succ = bool(any(traj_success))
+                record = {
+                    "states": traj_states,
+                    "actions": traj_actions,
+                    "rewards": traj_rewards,
+                    "dones": traj_dones,
+                    "success": traj_success,
+                    "episode_return": ep_ret,
+                    "episode_success": ep_succ,
+                }
+                f.write(json.dumps(record) + "\n")
+                collected += 1
+                summary_returns.append(ep_ret)
+                summary_success.append(1.0 if ep_succ else 0.0)
+        env.close()
+        try:
+            metrics = {
+                "global_step": int(step_tag),
+                "episodes": int(n_episodes),
+                "success_rate": float(np.mean(summary_success)) if len(summary_success) else 0.0,
+                "avg_return": float(np.mean(summary_returns)) if len(summary_returns) else 0.0,
+                "std_return": float(np.std(summary_returns)) if len(summary_returns) else 0.0,
+            }
+            with (out_dir / "metrics.json").open("w") as mf:
+                json.dump(metrics, mf)
+        except Exception as e:
+            print(f"Warning: failed to write eval metrics: {e}")
+
+    eval_every_iters = max(1, args.num_iterations // args.eval_splits)
+
+    # training loop
+    for iteration in range(1, args.num_iterations + 1):
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            optimizer.param_groups[0]["lr"] = frac * args.learning_rate
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            masks[step] = next_mask # Store mask
+            dones[step] = next_done
+
+            with torch.no_grad():
+                # PASS MASK HERE
+                action, logprob, _, value = agent.get_action_and_value(next_obs, action_mask=next_mask)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            next_obs_dict, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            
+            # Unpack dict again
+            next_obs = torch.Tensor(next_obs_dict['observation']).to(device)
+            next_mask = torch.Tensor(next_obs_dict['action_mask']).to(device)
+            next_done = torch.Tensor(next_done).to(device)
+
+            try:
+                mask = None
+                if isinstance(infos, dict):
+                    if "_episode" in infos:
+                        mask = np.asarray(infos["_episode"]).astype(bool)
+                    elif "episode" in infos and isinstance(infos["episode"], dict) and "_l" in infos["episode"]:
+                        mask = np.asarray(infos["episode"]["_l"]).astype(bool)
+                if mask is not None and np.any(mask):
+                    r_arr = np.asarray(infos.get("episode", {}).get("r", np.zeros_like(mask, dtype=float)))
+                    l_arr = np.asarray(infos.get("episode", {}).get("l", np.zeros_like(mask, dtype=int)))
+                    succ_arr = np.asarray(infos.get("success", np.zeros_like(mask, dtype=bool))).astype(float)
+                    for i in np.where(mask)[0]:
+                        episode_returns.append(float(r_arr[i]))
+                        episode_steps.append(global_step)
+                        episode_successes.append(float(succ_arr[i]))
+                    if args.track:
+                        try:
+                            import wandb
+                            log_dict = {
+                                "global_step": int(global_step),
+                                "rollout/ep_rew_mean": float(np.mean(r_arr[mask])) if np.any(mask) else None,
+                                "rollout/ep_len_mean": float(np.mean(l_arr[mask])) if np.any(mask) else None,
+                                "rollout/success_rate": float(np.mean(succ_arr[mask])) if np.any(mask) else None,
+                            }
+                            if np.any(mask):
+                                last_idx = np.where(mask)[0][-1]
+                                log_dict.update({
+                                    "train/episodic_return": float(r_arr[last_idx]),
+                                    "train/episodic_length": int(l_arr[last_idx]),
+                                    "train/success": float(succ_arr[last_idx]),
+                                    "train/success_rate_100": float(np.mean(episode_successes[-100:])) if len(episode_successes) >= 100 else None,
+                                })
+                            wandb.log(log_dict, step=global_step)
+                        except Exception:
+                            pass
+            except Exception:
+                pass
+
+        # GAE
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten batch
+        b_obs = obs.reshape((-1,) + obs_shape)
+        b_masks = masks.reshape((-1,) + mask_shape) # Flatten masks
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # update
+        b_inds = np.arange(args.batch_size)
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                # PASS MASK HERE
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(
+                    b_obs[mb_inds], 
+                    action=b_actions.long()[mb_inds],
+                    action_mask=b_masks[mb_inds]
+                )
+                
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds], -args.clip_coef, args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss = 0.5 * torch.max(v_loss_unclipped, v_loss_clipped).mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        # logging
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Val: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        if args.track:
+            try:
+                import wandb
+                wandb.log({
+                    "global_step": int(global_step),
+                    "train/value_loss": float(v_loss.item()),
+                    "train/policy_loss": float(pg_loss.item()),
+                    "train/entropy": float(entropy_loss.item()),
+                    "train/old_approx_kl": float(old_approx_kl.item()),
+                    "train/approx_kl": float(approx_kl.item()),
+                    "losses/explained_variance": float(explained_var),
+                    "charts/avg_reward": float(rewards.mean().item()),
+                    "charts/avg_value": float(values.mean().item()),
+                    "perf/SPS": int(sps),
+                    "train/learning_rate": float(optimizer.param_groups[0]["lr"]),
+                }, step=global_step)
+            except Exception:
+                pass
+
+        if iteration % eval_every_iters == 0:
+            try:
+                eval_thunk = make_env(0, run_name, args.seed + 9999, args.grid_size, args.difficulty, False)
+                collect_eval_trajectories(agent, eval_thunk, n_episodes=args.eval_episodes, step_tag=global_step)
+                if args.track:
+                    try:
+                        import json as _json
+                        from pathlib import Path as _Path
+                        mpath = _Path(f"runs/{run_name}/trajectories/step_{global_step}/metrics.json")
+                        if mpath.exists():
+                            with mpath.open("r") as mf:
+                                metrics = _json.load(mf)
+                            wandb.log({
+                                "eval/success_rate": metrics.get("success_rate"),
+                                "eval/avg_return": metrics.get("avg_return"),
+                                "eval/std_return": metrics.get("std_return"),
+                                "eval/episodes": metrics.get("episodes"),
+                            }, step=global_step)
+                    except Exception:
+                        pass
+                print(f"Collected {args.eval_episodes} eval trajectories at global_step {global_step}")
+            except Exception as e:
+                print(f"Warning: eval trajectory collection failed at step {global_step}: {e}")
+
+    envs.close()
\ No newline at end of file
diff --git a/cleanrl/cleanrl/scout_ppo/ragen_wrappers.py b/cleanrl/cleanrl/scout_ppo/ragen_wrappers.py
new file mode 100644
index 0000000000000000000000000000000000000000..af63107e99b08764b8410c77be7c66ddbb52899d
--- /dev/null
+++ b/cleanrl/cleanrl/scout_ppo/ragen_wrappers.py
@@ -0,0 +1,235 @@
+"""
+Gymnasium-compatible wrappers for RAGEN environments to enable traditional RL training.
+These wrappers convert text-based observations to numerical representations suitable for MLP networks.
+"""
+import gymnasium as gym
+import numpy as np
+from typing import Any, Dict, Tuple
+
+
+class BanditWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Bandit that uses only observable text.
+    Converts text observations to a fixed-size one-hot hash vector.
+    Does not alter episode semantics and does not inspect env internals.
+    """
+    def __init__(self, env, feature_dim_per_name: int = 16):
+        super().__init__(env)
+        # Two name slots (first/second), each hashed to one-hot of size K
+        self.k = feature_dim_per_name
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(2 * self.k,), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(2)
+
+    def _parse_names(self, text_obs: str):
+        """Extract the two arm names from the prompt text purely via regex/string ops."""
+        # Heuristic: look for the segment after "named " and split by " and "
+        try:
+            anchor = "named "
+            if anchor in text_obs:
+                segment = text_obs.split(anchor, 1)[1]
+                # Cut at newline if present
+                segment = segment.split("\n", 1)[0]
+                # Now split by " and " to get two names; also strip punctuation
+                parts = segment.split(" and ")
+                if len(parts) >= 2:
+                    name_a = parts[0].strip().strip(' .!?,')
+                    name_b = parts[1].strip().strip(' .!?,')
+                    return name_a, name_b
+        except Exception:
+            pass
+        # Fallback: no names found
+        return "", ""
+
+    def _names_to_vector(self, name_a: str, name_b: str) -> np.ndarray:
+        vec = np.zeros(2 * self.k, dtype=np.float32)
+        idx_a = (hash(name_a) % self.k)
+        idx_b = (hash(name_b) % self.k)
+        vec[idx_a] = 1.0
+        vec[self.k + idx_b] = 1.0
+        return vec
+
+    def reset(self, **kwargs):
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        name_a, name_b = self._parse_names(text_obs)
+        return self._names_to_vector(name_a, name_b), {}
+
+    def step(self, action):
+        ragen_action = int(action) + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        name_a, name_b = self._parse_names(text_obs)
+        terminated = bool(done)
+        truncated = False
+        return self._names_to_vector(name_a, name_b), reward, terminated, truncated, info
+
+
+class FrozenLakeWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN FrozenLake environment.
+    Converts grid-based text observations to numerical state representation.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine grid size from text only
+        bootstrap_text = self.env.reset()
+        flat, _ = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        # Serve the bootstrapped obs on first reset without calling env.reset again
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, Tuple[int, int]]:
+        """Parse text observation into numerical state (one-hot grid + player pos)."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        player_pos = None
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Parse grid
+        for i, line in enumerate(lines):
+            row = []
+            for j, char in enumerate(line):
+                if char == 'P':  # Player
+                    row.append(0)
+                    player_pos = (i, j)
+                elif char == '_':  # Frozen
+                    row.append(1)
+                elif char == 'O':  # Hole
+                    row.append(2)
+                elif char == 'G':  # Goal
+                    row.append(3)
+                elif char == 'X':  # Player in hole
+                    row.append(2)
+                    player_pos = (i, j)
+                elif char == '√':  # Player on goal
+                    row.append(3)
+                    player_pos = (i, j)
+                else:
+                    row.append(1)  # Default to frozen
+            grid.append(row)
+        # Pad ragged rows if needed
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        grid_size = (rows, cols)
+        # One-hot encode grid over 4 cell types
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 4), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        # Flatten grid
+        flat_grid = one_hot_grid.flatten()
+        # Add normalized player position
+        if player_pos is None:
+            player_pos = (0, 0)
+        player_pos_norm = np.array([
+            0.0 if rows <= 1 else player_pos[0] / max(1, rows - 1),
+            0.0 if cols <= 1 else player_pos[1] / max(1, cols - 1),
+        ], dtype=np.float32)
+        flat = np.concatenate([flat_grid, player_pos_norm])
+        return flat, grid_size
+    
+    def reset(self, **kwargs):
+        # Filter out 'options' parameter that gymnasium passes but RAGEN doesn't support
+        if self._bootstrap_ready:
+            # First call returns the bootstrapped observation to avoid double reset
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+
+
+class SokobanWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Sokoban environment.
+    Converts grid-based text observations to numerical state representation.
+    Note: Does not inherit from gym.Wrapper due to old gym vs gymnasium compatibility.
+    """
+    def __init__(self, env):
+        super().__init__(env)
+        # Bootstrap an observation to determine room size from text only
+        bootstrap_text = self.env.reset()
+        flat, rows, cols = self._parse_observation_and_meta(bootstrap_text)
+        self.observation_space = gym.spaces.Box(low=0, high=1, shape=(flat.shape[0],), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(4)
+        self.metadata = getattr(env, 'metadata', {})
+        self._bootstrap_obs = flat
+        self._bootstrap_ready = True
+        
+    def _parse_observation_and_meta(self, text_obs: str) -> Tuple[np.ndarray, int, int]:
+        """Parse text observation into numerical state and return dims."""
+        lines = text_obs.strip().split('\n')
+        grid = []
+        rows = len(lines)
+        cols = max(len(line) for line in lines) if rows > 0 else 0
+        # Mapping from characters to cell types
+        char_to_type = {
+            '#': 0,  # wall
+            '_': 1,  # empty
+            'O': 2,  # target
+            '√': 3,  # box on target
+            'X': 4,  # box
+            'P': 5,  # player
+            'S': 6,  # player on target
+        }
+        
+        for line in lines:
+            row = []
+            for char in line:
+                row.append(char_to_type.get(char, 1))  # Default to empty
+            grid.append(row)
+        # Pad ragged rows
+        grid = np.array([r + [1] * (cols - len(r)) for r in grid], dtype=np.int32)
+        # One-hot encode grid
+        one_hot_grid = np.zeros((rows, cols, 7), dtype=np.float32)
+        for i in range(rows):
+            for j in range(cols):
+                cell_type = grid[i, j]
+                one_hot_grid[i, j, cell_type] = 1.0
+        return one_hot_grid.flatten(), rows, cols
+    
+    def reset(self, **kwargs):
+        if self._bootstrap_ready:
+            self._bootstrap_ready = False
+            return self._bootstrap_obs.copy(), {}
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        text_obs = self.env.reset(seed=seed, mode=mode)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        return state, {}
+    
+    def step(self, action):
+        # Map action from 0,1,2,3 to 1,2,3,4 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        text_obs, reward, done, info = self.env.step(ragen_action)
+        state, _, _ = self._parse_observation_and_meta(text_obs)
+        
+        terminated = done
+        truncated = False
+        
+        return state, reward, terminated, truncated, info
+    
+    def close(self):
+        if hasattr(self.env, 'close'):
+            self.env.close()
+    
+    def render(self):
+        if hasattr(self.env, 'render'):
+            return self.env.render()
+        return None
diff --git a/cleanrl/cleanrl/td3_continuous_action.py b/cleanrl/cleanrl/td3_continuous_action.py
new file mode 100644
index 0000000000000000000000000000000000000000..65d7e690466594ae58c99d07eca3ad1ce28f34e3
--- /dev/null
+++ b/cleanrl/cleanrl/td3_continuous_action.py
@@ -0,0 +1,317 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/td3/#td3_continuous_actionpy
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+from cleanrl_utils.buffers import ReplayBuffer
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+    upload_model: bool = False
+    """whether to upload the saved model to huggingface"""
+    hf_entity: str = ""
+    """the user or org name of the model repository from the Hugging Face Hub"""
+
+    # Algorithm specific arguments
+    env_id: str = "Hopper-v4"
+    """the id of the environment"""
+    total_timesteps: int = 1000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 3e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 1
+    """the number of parallel game environments"""
+    buffer_size: int = int(1e6)
+    """the replay memory buffer size"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    tau: float = 0.005
+    """target smoothing coefficient (default: 0.005)"""
+    batch_size: int = 256
+    """the batch size of sample from the reply memory"""
+    policy_noise: float = 0.2
+    """the scale of policy noise"""
+    exploration_noise: float = 0.1
+    """the scale of exploration noise"""
+    learning_starts: int = 25e3
+    """timestep to start learning"""
+    policy_frequency: int = 2
+    """the frequency of training policy (delayed)"""
+    noise_clip: float = 0.5
+    """noise clip parameter of the Target Policy Smoothing Regularization"""
+
+
+def make_env(env_id, seed, idx, capture_video, run_name):
+    def thunk():
+        if capture_video and idx == 0:
+            env = gym.make(env_id, render_mode="rgb_array")
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        else:
+            env = gym.make(env_id)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        env.action_space.seed(seed)
+        return env
+
+    return thunk
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    def __init__(self, env):
+        super().__init__()
+        self.fc1 = nn.Linear(
+            np.array(env.single_observation_space.shape).prod() + np.prod(env.single_action_space.shape),
+            256,
+        )
+        self.fc2 = nn.Linear(256, 256)
+        self.fc3 = nn.Linear(256, 1)
+
+    def forward(self, x, a):
+        x = torch.cat([x, a], 1)
+        x = F.relu(self.fc1(x))
+        x = F.relu(self.fc2(x))
+        x = self.fc3(x)
+        return x
+
+
+class Actor(nn.Module):
+    def __init__(self, env):
+        super().__init__()
+        self.fc1 = nn.Linear(np.array(env.single_observation_space.shape).prod(), 256)
+        self.fc2 = nn.Linear(256, 256)
+        self.fc_mu = nn.Linear(256, np.prod(env.single_action_space.shape))
+        # action rescaling
+        self.register_buffer(
+            "action_scale",
+            torch.tensor(
+                (env.single_action_space.high - env.single_action_space.low) / 2.0,
+                dtype=torch.float32,
+            ),
+        )
+        self.register_buffer(
+            "action_bias",
+            torch.tensor(
+                (env.single_action_space.high + env.single_action_space.low) / 2.0,
+                dtype=torch.float32,
+            ),
+        )
+
+    def forward(self, x):
+        x = F.relu(self.fc1(x))
+        x = F.relu(self.fc2(x))
+        x = torch.tanh(self.fc_mu(x))
+        return x * self.action_scale + self.action_bias
+
+
+if __name__ == "__main__":
+
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Box), "only continuous action space is supported"
+
+    actor = Actor(envs).to(device)
+    qf1 = QNetwork(envs).to(device)
+    qf2 = QNetwork(envs).to(device)
+    qf1_target = QNetwork(envs).to(device)
+    qf2_target = QNetwork(envs).to(device)
+    target_actor = Actor(envs).to(device)
+    target_actor.load_state_dict(actor.state_dict())
+    qf1_target.load_state_dict(qf1.state_dict())
+    qf2_target.load_state_dict(qf2.state_dict())
+    q_optimizer = optim.Adam(list(qf1.parameters()) + list(qf2.parameters()), lr=args.learning_rate)
+    actor_optimizer = optim.Adam(list(actor.parameters()), lr=args.learning_rate)
+
+    envs.single_observation_space.dtype = np.float32
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        n_envs=args.num_envs,
+        handle_timeout_termination=False,
+    )
+    start_time = time.time()
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        if global_step < args.learning_starts:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            with torch.no_grad():
+                actions = actor(torch.Tensor(obs).to(device))
+                actions += torch.normal(0, actor.action_scale * args.exploration_noise)
+                actions = actions.cpu().numpy().clip(envs.single_action_space.low, envs.single_action_space.high)
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if info is not None:
+                    print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                    writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                    writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+                    break
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            data = rb.sample(args.batch_size)
+            with torch.no_grad():
+                clipped_noise = (torch.randn_like(data.actions, device=device) * args.policy_noise).clamp(
+                    -args.noise_clip, args.noise_clip
+                ) * target_actor.action_scale
+
+                next_state_actions = (target_actor(data.next_observations) + clipped_noise).clamp(
+                    envs.single_action_space.low[0], envs.single_action_space.high[0]
+                )
+                qf1_next_target = qf1_target(data.next_observations, next_state_actions)
+                qf2_next_target = qf2_target(data.next_observations, next_state_actions)
+                min_qf_next_target = torch.min(qf1_next_target, qf2_next_target)
+                next_q_value = data.rewards.flatten() + (1 - data.dones.flatten()) * args.gamma * (min_qf_next_target).view(-1)
+
+            qf1_a_values = qf1(data.observations, data.actions).view(-1)
+            qf2_a_values = qf2(data.observations, data.actions).view(-1)
+            qf1_loss = F.mse_loss(qf1_a_values, next_q_value)
+            qf2_loss = F.mse_loss(qf2_a_values, next_q_value)
+            qf_loss = qf1_loss + qf2_loss
+
+            # optimize the model
+            q_optimizer.zero_grad()
+            qf_loss.backward()
+            q_optimizer.step()
+
+            if global_step % args.policy_frequency == 0:
+                actor_loss = -qf1(data.observations, actor(data.observations)).mean()
+                actor_optimizer.zero_grad()
+                actor_loss.backward()
+                actor_optimizer.step()
+
+                # update the target network
+                for param, target_param in zip(actor.parameters(), target_actor.parameters()):
+                    target_param.data.copy_(args.tau * param.data + (1 - args.tau) * target_param.data)
+                for param, target_param in zip(qf1.parameters(), qf1_target.parameters()):
+                    target_param.data.copy_(args.tau * param.data + (1 - args.tau) * target_param.data)
+                for param, target_param in zip(qf2.parameters(), qf2_target.parameters()):
+                    target_param.data.copy_(args.tau * param.data + (1 - args.tau) * target_param.data)
+
+            if global_step % 100 == 0:
+                writer.add_scalar("losses/qf1_values", qf1_a_values.mean().item(), global_step)
+                writer.add_scalar("losses/qf2_values", qf2_a_values.mean().item(), global_step)
+                writer.add_scalar("losses/qf1_loss", qf1_loss.item(), global_step)
+                writer.add_scalar("losses/qf2_loss", qf2_loss.item(), global_step)
+                writer.add_scalar("losses/qf_loss", qf_loss.item() / 2.0, global_step)
+                writer.add_scalar("losses/actor_loss", actor_loss.item(), global_step)
+                print("SPS:", int(global_step / (time.time() - start_time)))
+                writer.add_scalar(
+                    "charts/SPS",
+                    int(global_step / (time.time() - start_time)),
+                    global_step,
+                )
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        torch.save((actor.state_dict(), qf1.state_dict(), qf2.state_dict()), model_path)
+        print(f"model saved to {model_path}")
+        from cleanrl_utils.evals.td3_eval import evaluate
+
+        episodic_returns = evaluate(
+            model_path,
+            make_env,
+            args.env_id,
+            eval_episodes=10,
+            run_name=f"{run_name}-eval",
+            Model=(Actor, QNetwork),
+            device=device,
+            exploration_noise=args.exploration_noise,
+        )
+        for idx, episodic_return in enumerate(episodic_returns):
+            writer.add_scalar("eval/episodic_return", episodic_return, idx)
+
+        if args.upload_model:
+            from cleanrl_utils.huggingface import push_to_hub
+
+            repo_name = f"{args.env_id}-{args.exp_name}-seed{args.seed}"
+            repo_id = f"{args.hf_entity}/{repo_name}" if args.hf_entity else repo_name
+            push_to_hub(
+                args,
+                episodic_returns,
+                repo_id,
+                "TD3",
+                f"runs/{run_name}",
+                f"videos/{run_name}-eval",
+            )
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/td3_continuous_action_jax.py b/cleanrl/cleanrl/td3_continuous_action_jax.py
new file mode 100644
index 0000000000000000000000000000000000000000..9daf9f461ce1478ae1157e6d1e9d876bb1bcce25
--- /dev/null
+++ b/cleanrl/cleanrl/td3_continuous_action_jax.py
@@ -0,0 +1,361 @@
+# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/td3/#td3_continuous_action_jaxpy
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import flax
+import flax.linen as nn
+import gymnasium as gym
+import jax
+import jax.numpy as jnp
+import numpy as np
+import optax
+import tyro
+from flax.training.train_state import TrainState
+from torch.utils.tensorboard import SummaryWriter
+
+from cleanrl_utils.buffers import ReplayBuffer
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+    upload_model: bool = False
+    """whether to upload the saved model to huggingface"""
+    hf_entity: str = ""
+    """the user or org name of the model repository from the Hugging Face Hub"""
+
+    # Algorithm specific arguments
+    env_id: str = "Hopper-v4"
+    """the id of the environment"""
+    total_timesteps: int = 1000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 3e-4
+    """the learning rate of the optimizer"""
+    buffer_size: int = int(1e6)
+    """the replay memory buffer size"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    tau: float = 0.005
+    """target smoothing coefficient (default: 0.005)"""
+    batch_size: int = 256
+    """the batch size of sample from the reply memory"""
+    policy_noise: float = 0.2
+    """the scale of policy noise"""
+    exploration_noise: float = 0.1
+    """the scale of exploration noise"""
+    learning_starts: int = 25e3
+    """timestep to start learning"""
+    policy_frequency: int = 2
+    """the frequency of training policy (delayed)"""
+    noise_clip: float = 0.5
+    """noise clip parameter of the Target Policy Smoothing Regularization"""
+
+
+def make_env(env_id, seed, idx, capture_video, run_name):
+    def thunk():
+        if capture_video and idx == 0:
+            env = gym.make(env_id, render_mode="rgb_array")
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        else:
+            env = gym.make(env_id)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        env.action_space.seed(seed)
+        return env
+
+    return thunk
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    @nn.compact
+    def __call__(self, x: jnp.ndarray, a: jnp.ndarray):
+        x = jnp.concatenate([x, a], -1)
+        x = nn.Dense(256)(x)
+        x = nn.relu(x)
+        x = nn.Dense(256)(x)
+        x = nn.relu(x)
+        x = nn.Dense(1)(x)
+        return x
+
+
+class Actor(nn.Module):
+    action_dim: int
+    action_scale: jnp.ndarray
+    action_bias: jnp.ndarray
+
+    @nn.compact
+    def __call__(self, x):
+        x = nn.Dense(256)(x)
+        x = nn.relu(x)
+        x = nn.Dense(256)(x)
+        x = nn.relu(x)
+        x = nn.Dense(self.action_dim)(x)
+        x = nn.tanh(x)
+        x = x * self.action_scale + self.action_bias
+        return x
+
+
+class TrainState(TrainState):
+    target_params: flax.core.FrozenDict
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    key = jax.random.PRNGKey(args.seed)
+    key, actor_key, qf1_key, qf2_key = jax.random.split(key, 4)
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv([make_env(args.env_id, args.seed, 0, args.capture_video, run_name)])
+    assert isinstance(envs.single_action_space, gym.spaces.Box), "only continuous action space is supported"
+
+    max_action = float(envs.single_action_space.high[0])
+    envs.single_observation_space.dtype = np.float32
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device="cpu",
+        handle_timeout_termination=False,
+    )
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+
+    actor = Actor(
+        action_dim=np.prod(envs.single_action_space.shape),
+        action_scale=jnp.array((envs.action_space.high - envs.action_space.low) / 2.0),
+        action_bias=jnp.array((envs.action_space.high + envs.action_space.low) / 2.0),
+    )
+    actor_state = TrainState.create(
+        apply_fn=actor.apply,
+        params=actor.init(actor_key, obs),
+        target_params=actor.init(actor_key, obs),
+        tx=optax.adam(learning_rate=args.learning_rate),
+    )
+    qf = QNetwork()
+    qf1_state = TrainState.create(
+        apply_fn=qf.apply,
+        params=qf.init(qf1_key, obs, envs.action_space.sample()),
+        target_params=qf.init(qf1_key, obs, envs.action_space.sample()),
+        tx=optax.adam(learning_rate=args.learning_rate),
+    )
+    qf2_state = TrainState.create(
+        apply_fn=qf.apply,
+        params=qf.init(qf2_key, obs, envs.action_space.sample()),
+        target_params=qf.init(qf2_key, obs, envs.action_space.sample()),
+        tx=optax.adam(learning_rate=args.learning_rate),
+    )
+    actor.apply = jax.jit(actor.apply)
+    qf.apply = jax.jit(qf.apply)
+
+    @jax.jit
+    def update_critic(
+        actor_state: TrainState,
+        qf1_state: TrainState,
+        qf2_state: TrainState,
+        observations: np.ndarray,
+        actions: np.ndarray,
+        next_observations: np.ndarray,
+        rewards: np.ndarray,
+        terminations: np.ndarray,
+        key: jnp.ndarray,
+    ):
+        # TODO Maybe pre-generate a lot of random keys
+        # also check https://jax.readthedocs.io/en/latest/jax.random.html
+        key, noise_key = jax.random.split(key, 2)
+        clipped_noise = (
+            jnp.clip(
+                (jax.random.normal(noise_key, actions.shape) * args.policy_noise),
+                -args.noise_clip,
+                args.noise_clip,
+            )
+            * actor.action_scale
+        )
+        next_state_actions = jnp.clip(
+            actor.apply(actor_state.target_params, next_observations) + clipped_noise,
+            envs.single_action_space.low,
+            envs.single_action_space.high,
+        )
+        qf1_next_target = qf.apply(qf1_state.target_params, next_observations, next_state_actions).reshape(-1)
+        qf2_next_target = qf.apply(qf2_state.target_params, next_observations, next_state_actions).reshape(-1)
+        min_qf_next_target = jnp.minimum(qf1_next_target, qf2_next_target)
+        next_q_value = (rewards + (1 - terminations) * args.gamma * (min_qf_next_target)).reshape(-1)
+
+        def mse_loss(params):
+            qf_a_values = qf.apply(params, observations, actions).squeeze()
+            return ((qf_a_values - next_q_value) ** 2).mean(), qf_a_values.mean()
+
+        (qf1_loss_value, qf1_a_values), grads1 = jax.value_and_grad(mse_loss, has_aux=True)(qf1_state.params)
+        (qf2_loss_value, qf2_a_values), grads2 = jax.value_and_grad(mse_loss, has_aux=True)(qf2_state.params)
+        qf1_state = qf1_state.apply_gradients(grads=grads1)
+        qf2_state = qf2_state.apply_gradients(grads=grads2)
+
+        return (qf1_state, qf2_state), (qf1_loss_value, qf2_loss_value), (qf1_a_values, qf2_a_values), key
+
+    @jax.jit
+    def update_actor(
+        actor_state: TrainState,
+        qf1_state: TrainState,
+        qf2_state: TrainState,
+        observations: np.ndarray,
+    ):
+        def actor_loss(params):
+            return -qf.apply(qf1_state.params, observations, actor.apply(params, observations)).mean()
+
+        actor_loss_value, grads = jax.value_and_grad(actor_loss)(actor_state.params)
+        actor_state = actor_state.apply_gradients(grads=grads)
+        actor_state = actor_state.replace(
+            target_params=optax.incremental_update(actor_state.params, actor_state.target_params, args.tau)
+        )
+
+        qf1_state = qf1_state.replace(
+            target_params=optax.incremental_update(qf1_state.params, qf1_state.target_params, args.tau)
+        )
+        qf2_state = qf2_state.replace(
+            target_params=optax.incremental_update(qf2_state.params, qf2_state.target_params, args.tau)
+        )
+        return actor_state, (qf1_state, qf2_state), actor_loss_value
+
+    start_time = time.time()
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        if global_step < args.learning_starts:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            actions = actor.apply(actor_state.params, obs)
+            actions = np.array(
+                [
+                    (
+                        jax.device_get(actions)[0]
+                        + np.random.normal(0, max_action * args.exploration_noise, size=envs.single_action_space.shape)
+                    ).clip(envs.single_action_space.low, envs.single_action_space.high)
+                ]
+            )
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+                break
+
+        # TRY NOT TO MODIFY: save data to replay buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            data = rb.sample(args.batch_size)
+
+            (qf1_state, qf2_state), (qf1_loss_value, qf2_loss_value), (qf1_a_values, qf2_a_values), key = update_critic(
+                actor_state,
+                qf1_state,
+                qf2_state,
+                data.observations.numpy(),
+                data.actions.numpy(),
+                data.next_observations.numpy(),
+                data.rewards.flatten().numpy(),
+                data.dones.flatten().numpy(),
+                key,
+            )
+
+            if global_step % args.policy_frequency == 0:
+                actor_state, (qf1_state, qf2_state), actor_loss_value = update_actor(
+                    actor_state,
+                    qf1_state,
+                    qf2_state,
+                    data.observations.numpy(),
+                )
+
+            if global_step % 100 == 0:
+                writer.add_scalar("losses/qf1_loss", qf1_loss_value.item(), global_step)
+                writer.add_scalar("losses/qf2_loss", qf2_loss_value.item(), global_step)
+                writer.add_scalar("losses/qf1_values", qf1_a_values.item(), global_step)
+                writer.add_scalar("losses/qf2_values", qf2_a_values.item(), global_step)
+                writer.add_scalar("losses/actor_loss", actor_loss_value.item(), global_step)
+                print("SPS:", int(global_step / (time.time() - start_time)))
+                writer.add_scalar("charts/SPS", int(global_step / (time.time() - start_time)), global_step)
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        with open(model_path, "wb") as f:
+            f.write(
+                flax.serialization.to_bytes(
+                    [
+                        actor_state.params,
+                        qf1_state.params,
+                        qf2_state.params,
+                    ]
+                )
+            )
+        print(f"model saved to {model_path}")
+        from cleanrl_utils.evals.td3_jax_eval import evaluate
+
+        episodic_returns = evaluate(
+            model_path,
+            make_env,
+            args.env_id,
+            eval_episodes=10,
+            run_name=f"{run_name}-eval",
+            Model=(Actor, QNetwork),
+            exploration_noise=args.exploration_noise,
+        )
+        for idx, episodic_return in enumerate(episodic_returns):
+            writer.add_scalar("eval/episodic_return", episodic_return, idx)
+
+        if args.upload_model:
+            from cleanrl_utils.huggingface import push_to_hub
+
+            repo_name = f"{args.env_id}-{args.exp_name}-seed{args.seed}"
+            repo_id = f"{args.hf_entity}/{repo_name}" if args.hf_entity else repo_name
+            push_to_hub(args, episodic_returns, repo_id, "TD3", f"runs/{run_name}", f"videos/{run_name}-eval")
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/wandb/debug-internal.log b/cleanrl/cleanrl/wandb/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..a61ba666dd77c3c9e5ab5f378fed38b993c06108
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/debug-internal.log
@@ -0,0 +1,15 @@
+{"time":"2025-11-07T12:56:47.957309906+08:00","level":"INFO","msg":"stream: starting","core version":"0.22.3"}
+{"time":"2025-11-07T12:56:48.33142385+08:00","level":"INFO","msg":"stream: created new stream","id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.331674194+08:00","level":"INFO","msg":"handler: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335368381+08:00","level":"INFO","msg":"stream: started","id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335386015+08:00","level":"INFO","msg":"writer: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335389832+08:00","level":"INFO","msg":"sender: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.726770104+08:00","level":"ERROR","msg":"git repo not found","error":"repository does not exist"}
+{"time":"2025-11-07T12:56:58.861202459+08:00","level":"WARN","msg":"tensorboard: no root directory","error":"timed out after 10s"}
+{"time":"2025-11-07T12:56:58.862839009+08:00","level":"INFO","msg":"tensorboard: tracking new log directory","rootDir":{},"logDir":"LocalOrCloudPath(LocalPath=\"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491400\")","namespace":""}
+{"time":"2025-11-07T12:56:58.864923146+08:00","level":"INFO","msg":"tensorboard: saving file","fileLocation":"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491400/events.out.tfevents.1762491408.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.234009.0","runPath":"runs/Bandit__dqn_bandit__1__1762491400/events.out.tfevents.1762491408.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.234009.0"}
+{"time":"2025-11-07T12:56:58.868040818+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":122}
+{"time":"2025-11-07T12:56:58.871369044+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":311}
+{"time":"2025-11-07T12:57:03.576733669+08:00","level":"INFO","msg":"stream: closing","id":"g1edw7ov"}
+{"time":"2025-11-07T12:57:03.579454439+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":506}
+{"time":"2025-11-07T12:57:03.581505913+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":20}
diff --git a/cleanrl/cleanrl/wandb/debug.log b/cleanrl/cleanrl/wandb/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..cf9eb1cc4ac7daf727c39f2eb082010f08927c7e
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/debug.log
@@ -0,0 +1,388 @@
+2025-11-07 12:56:47,735 INFO    MainThread:234009 [wandb_setup.py:_flush():81] Current SDK version is 0.22.3
+2025-11-07 12:56:47,736 INFO    MainThread:234009 [wandb_setup.py:_flush():81] Configure stats pid to 234009
+2025-11-07 12:56:47,736 INFO    MainThread:234009 [wandb_setup.py:_flush():81] Loading settings from /root/.config/wandb/settings
+2025-11-07 12:56:47,736 INFO    MainThread:234009 [wandb_setup.py:_flush():81] Loading settings from /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/settings
+2025-11-07 12:56:47,736 INFO    MainThread:234009 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2025-11-07 12:56:47,736 INFO    MainThread:234009 [wandb_init.py:setup_run_log_directory():706] Logging user logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/logs/debug.log
+2025-11-07 12:56:47,737 INFO    MainThread:234009 [wandb_init.py:setup_run_log_directory():707] Logging internal logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/logs/debug-internal.log
+2025-11-07 12:56:47,738 INFO    MainThread:234009 [wandb_init.py:init():833] calling init triggers
+2025-11-07 12:56:47,738 INFO    MainThread:234009 [wandb_init.py:init():838] wandb.init called with sweep_config: {}
+config: {'exp_name': 'dqn_bandit', 'seed': 1, 'torch_deterministic': True, 'cuda': True, 'track': True, 'wandb_project_name': 'ragen-bandit', 'wandb_entity': None, 'capture_video': False, 'save_model': False, 'env_id': 'Bandit', 'total_timesteps': 100000, 'learning_rate': 0.001, 'num_envs': 1, 'buffer_size': 10000, 'gamma': 0.99, 'tau': 1.0, 'target_network_frequency': 500, 'batch_size': 32, 'start_e': 1.0, 'end_e': 0.05, 'exploration_fraction': 0.5, 'learning_starts': 1000, 'train_frequency': 1, 'steps_per_episode': 10, '_wandb': {'code_path': 'code/cleanrl/dqn_bandit.py'}}
+2025-11-07 12:56:47,738 INFO    MainThread:234009 [wandb_init.py:init():881] starting backend
+2025-11-07 12:56:47,944 INFO    MainThread:234009 [wandb_init.py:init():884] sending inform_init request
+2025-11-07 12:56:47,954 INFO    MainThread:234009 [wandb_init.py:init():892] backend started and connected
+2025-11-07 12:56:47,956 INFO    MainThread:234009 [wandb_init.py:init():962] updated telemetry
+2025-11-07 12:56:47,990 INFO    MainThread:234009 [wandb_init.py:init():986] communicating run to backend with 90.0 second timeout
+2025-11-07 12:56:48,714 INFO    MainThread:234009 [wandb_init.py:init():1033] starting run threads in backend
+2025-11-07 12:56:48,856 INFO    MainThread:234009 [wandb_run.py:_console_start():2506] atexit reg
+2025-11-07 12:56:48,857 INFO    MainThread:234009 [wandb_run.py:_redirect():2354] redirect: wrap_raw
+2025-11-07 12:56:48,857 INFO    MainThread:234009 [wandb_run.py:_redirect():2423] Wrapping output streams.
+2025-11-07 12:56:48,857 INFO    MainThread:234009 [wandb_run.py:_redirect():2446] Redirects installed.
+2025-11-07 12:56:48,859 INFO    MainThread:234009 [wandb_init.py:init():1073] run started, returning control to user process
+2025-11-07 12:56:48,860 INFO    MainThread:234009 [wandb_run.py:_tensorboard_callback():1598] tensorboard callback: runs/Bandit__dqn_bandit__1__1762491400, True
+2025-11-07 12:57:03,576 INFO    wandb-AsyncioManager-main:234009 [service_client.py:_forward_responses():80] Reached EOF.
+2025-11-07 12:57:03,577 INFO    wandb-AsyncioManager-main:234009 [mailbox.py:close():137] Closing mailbox, abandoning 1 handles.
+2025-11-07 12:57:03,925 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,933 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,933 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,934 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,934 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,935 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,935 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,935 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,936 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,936 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,937 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,937 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,937 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,938 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,938 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,938 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,939 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,939 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,948 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,948 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,950 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,951 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,952 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,953 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,954 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,955 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,955 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:57:03,957 ERROR   wandb-AsyncioManager-main:234009 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
diff --git a/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d331f007a448c4fffb97beaf95c1fa15ae977963
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/config.yaml
@@ -0,0 +1,153 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/ppo_bandit.py
+        e:
+            6ymqnqubvr99vvj2hajaxngee30bwzdk:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-bandit
+                codePath: cleanrl/ppo_bandit.py
+                codePathLocal: ppo_bandit.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "30509887488"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_bandit.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T02:26:33.725426Z"
+                writerId: 6ymqnqubvr99vvj2hajaxngee30bwzdk
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+anneal_lr:
+    value: true
+batch_size:
+    value: 16384
+capture_video:
+    value: false
+clip_coef:
+    value: 0.2
+clip_vloss:
+    value: true
+cuda:
+    value: true
+ent_coef:
+    value: 0.01
+env_id:
+    value: Bandit
+exp_name:
+    value: ppo_bandit
+gae_lambda:
+    value: 0.95
+gamma:
+    value: 0.99
+learning_rate:
+    value: 0.00025
+max_grad_norm:
+    value: 0.5
+minibatch_size:
+    value: 4096
+norm_adv:
+    value: true
+num_envs:
+    value: 32
+num_iterations:
+    value: 610
+num_minibatches:
+    value: 4
+num_steps:
+    value: 512
+seed:
+    value: 1
+target_kl:
+    value: null
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 10000000
+track:
+    value: true
+update_epochs:
+    value: 4
+vf_coef:
+    value: 0.5
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-bandit
diff --git a/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/wandb-summary.json b/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..a4d46dca73af2b07cc8110c0475616f9104561fb
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_102633-lu2iu68t/files/wandb-summary.json
@@ -0,0 +1 @@
+{"losses/value_loss":0.044632442,"charts/min_reward":0,"losses/old_approx_kl":-1.379085e-07,"charts/avg_reward":0.089166254,"_wandb":{"runtime":353},"losses/clipfrac":0,"global_step":9994240,"losses/policy_loss":-2.789311e-07,"_step":7929,"_runtime":353,"losses/approx_kl":3.2465323e-08,"losses/explained_variance":5.1021576e-05,"charts/max_reward":1,"losses/entropy":0.69236815,"_timestamp":1.7624827486915712e+09,"charts/SPS":28390,"charts/learning_rate":4.0983608e-07,"charts/avg_value":0.17397192}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/code/cleanrl/ppo_frozenlake.py b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/code/cleanrl/ppo_frozenlake.py
new file mode 100644
index 0000000000000000000000000000000000000000..c4aa75a01b6e37727e409ec55033613494df184a
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/code/cleanrl/ppo_frozenlake.py
@@ -0,0 +1,347 @@
+# PPO implementation for RAGEN FrozenLake environment
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.frozen_lake.env import FrozenLakeEnv
+from ragen.env.frozen_lake.config import FrozenLakeEnvConfig
+from ragen_wrappers import FrozenLakeWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "FrozenLake"
+    """the id of the environment"""
+    total_timesteps: int = 1000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 8
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.2
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.01
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+    
+    # FrozenLake specific
+    grid_size: int = 4
+    """size of the frozen lake grid"""
+    is_slippery: bool = True
+    """whether the ice is slippery"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, grid_size, is_slippery):
+    def thunk():
+        config = FrozenLakeEnvConfig(
+            size=grid_size,
+            p=0.8,
+            is_slippery=is_slippery,
+            map_seed=seed + idx
+        )
+        env = FrozenLakeEnv(config)
+        env = FrozenLakeWrapper(env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = np.array(envs.single_observation_space.shape).prod()
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None):
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed, args.grid_size, args.is_slippery) 
+         for i in range(args.num_envs)],
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                        writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                        writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped)
+                    v_loss = 0.5 * v_loss_max.mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        
+        # Additional useful metrics
+        writer.add_scalar("charts/avg_reward", rewards.mean().item(), global_step)
+        writer.add_scalar("charts/avg_value", values.mean().item(), global_step)
+        writer.add_scalar("charts/max_reward", rewards.max().item(), global_step)
+        writer.add_scalar("charts/min_reward", rewards.min().item(), global_step)
+        
+        # Console output with key metrics
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Value: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        writer.add_scalar("charts/SPS", sps, global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b7f01c169dfce0c33c36b1fcf2d7ccd659aacdc5
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/config.yaml
@@ -0,0 +1,157 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/ppo_frozenlake.py
+        e:
+            e26crksekquh3exjmxygxro2irogd3e1:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-bandit
+                codePath: cleanrl/ppo_frozenlake.py
+                codePathLocal: ppo_frozenlake.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "30509891584"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_frozenlake.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T02:33:31.320641Z"
+                writerId: e26crksekquh3exjmxygxro2irogd3e1
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+anneal_lr:
+    value: true
+batch_size:
+    value: 1024
+capture_video:
+    value: false
+clip_coef:
+    value: 0.2
+clip_vloss:
+    value: true
+cuda:
+    value: true
+ent_coef:
+    value: 0.01
+env_id:
+    value: FrozenLake
+exp_name:
+    value: ppo_frozenlake
+gae_lambda:
+    value: 0.95
+gamma:
+    value: 0.99
+grid_size:
+    value: 4
+is_slippery:
+    value: true
+learning_rate:
+    value: 0.00025
+max_grad_norm:
+    value: 0.5
+minibatch_size:
+    value: 256
+norm_adv:
+    value: true
+num_envs:
+    value: 8
+num_iterations:
+    value: 976
+num_minibatches:
+    value: 4
+num_steps:
+    value: 128
+seed:
+    value: 1
+target_kl:
+    value: null
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 1000000
+track:
+    value: true
+update_epochs:
+    value: 4
+vf_coef:
+    value: 0.5
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-bandit
diff --git a/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/output.log b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..fd1b7d58ba0e394281677caaa85dbe563bc7abc5
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/output.log
@@ -0,0 +1,976 @@
+[  0.1%] Iter    1/976 | SPS:  1656 | Reward:  0.040 | Value:  0.488 | VLoss: 0.0679 | PLoss: -0.0096 | Ent: 1.3836
+[  0.2%] Iter    2/976 | SPS:  2307 | Reward:  0.028 | Value:  0.440 | VLoss: 0.0411 | PLoss: -0.0020 | Ent: 1.3806
+[  0.3%] Iter    3/976 | SPS:  2667 | Reward:  0.024 | Value:  0.492 | VLoss: 0.0341 | PLoss: -0.0039 | Ent: 1.3784
+[  0.4%] Iter    4/976 | SPS:  2848 | Reward:  0.025 | Value:  0.381 | VLoss: 0.0418 | PLoss: -0.0020 | Ent: 1.3730
+[  0.5%] Iter    5/976 | SPS:  2984 | Reward:  0.022 | Value:  0.318 | VLoss: 0.0353 | PLoss: -0.0024 | Ent: 1.3742
+[  0.6%] Iter    6/976 | SPS:  3068 | Reward:  0.030 | Value:  0.242 | VLoss: 0.0481 | PLoss: -0.0041 | Ent: 1.3771
+[  0.7%] Iter    7/976 | SPS:  3141 | Reward:  0.025 | Value:  0.356 | VLoss: 0.0281 | PLoss: -0.0027 | Ent: 1.3781
+[  0.8%] Iter    8/976 | SPS:  3190 | Reward:  0.024 | Value:  0.288 | VLoss: 0.0383 | PLoss: -0.0024 | Ent: 1.3780
+[  0.9%] Iter    9/976 | SPS:  3240 | Reward:  0.028 | Value:  0.264 | VLoss: 0.0424 | PLoss: -0.0039 | Ent: 1.3737
+[  1.0%] Iter   10/976 | SPS:  3288 | Reward:  0.026 | Value:  0.365 | VLoss: 0.0351 | PLoss: -0.0173 | Ent: 1.3478
+[  1.1%] Iter   11/976 | SPS:  3321 | Reward:  0.030 | Value:  0.400 | VLoss: 0.0352 | PLoss: -0.0070 | Ent: 1.3563
+[  1.2%] Iter   12/976 | SPS:  3343 | Reward:  0.038 | Value:  0.360 | VLoss: 0.0500 | PLoss: -0.0133 | Ent: 1.3592
+[  1.3%] Iter   13/976 | SPS:  3359 | Reward:  0.025 | Value:  0.333 | VLoss: 0.0387 | PLoss: -0.0039 | Ent: 1.3414
+[  1.4%] Iter   14/976 | SPS:  3376 | Reward:  0.028 | Value:  0.347 | VLoss: 0.0375 | PLoss: -0.0045 | Ent: 1.3484
+[  1.5%] Iter   15/976 | SPS:  3393 | Reward:  0.031 | Value:  0.347 | VLoss: 0.0358 | PLoss: -0.0101 | Ent: 1.3481
+[  1.6%] Iter   16/976 | SPS:  3407 | Reward:  0.037 | Value:  0.356 | VLoss: 0.0431 | PLoss: 0.0009 | Ent: 1.3317
+[  1.7%] Iter   17/976 | SPS:  3428 | Reward:  0.032 | Value:  0.400 | VLoss: 0.0312 | PLoss: -0.0064 | Ent: 1.3081
+[  1.8%] Iter   18/976 | SPS:  3442 | Reward:  0.030 | Value:  0.454 | VLoss: 0.0348 | PLoss: -0.0078 | Ent: 1.3382
+[  1.9%] Iter   19/976 | SPS:  3460 | Reward:  0.030 | Value:  0.440 | VLoss: 0.0397 | PLoss: -0.0053 | Ent: 1.3214
+[  2.0%] Iter   20/976 | SPS:  3472 | Reward:  0.025 | Value:  0.388 | VLoss: 0.0402 | PLoss: -0.0013 | Ent: 1.2874
+[  2.2%] Iter   21/976 | SPS:  3487 | Reward:  0.024 | Value:  0.365 | VLoss: 0.0386 | PLoss: -0.0105 | Ent: 1.3403
+[  2.3%] Iter   22/976 | SPS:  3489 | Reward:  0.035 | Value:  0.352 | VLoss: 0.0473 | PLoss: -0.0029 | Ent: 1.3389
+[  2.4%] Iter   23/976 | SPS:  3489 | Reward:  0.038 | Value:  0.359 | VLoss: 0.0539 | PLoss: -0.0060 | Ent: 1.3497
+[  2.5%] Iter   24/976 | SPS:  3497 | Reward:  0.025 | Value:  0.391 | VLoss: 0.0305 | PLoss: -0.0026 | Ent: 1.3670
+[  2.6%] Iter   25/976 | SPS:  3501 | Reward:  0.031 | Value:  0.305 | VLoss: 0.0445 | PLoss: -0.0047 | Ent: 1.3656
+[  2.7%] Iter   26/976 | SPS:  3509 | Reward:  0.033 | Value:  0.329 | VLoss: 0.0402 | PLoss: -0.0053 | Ent: 1.3707
+[  2.8%] Iter   27/976 | SPS:  3513 | Reward:  0.025 | Value:  0.423 | VLoss: 0.0448 | PLoss: -0.0060 | Ent: 1.3593
+[  2.9%] Iter   28/976 | SPS:  3523 | Reward:  0.029 | Value:  0.401 | VLoss: 0.0402 | PLoss: -0.0104 | Ent: 1.3408
+[  3.0%] Iter   29/976 | SPS:  3531 | Reward:  0.027 | Value:  0.394 | VLoss: 0.0377 | PLoss: -0.0056 | Ent: 1.3466
+[  3.1%] Iter   30/976 | SPS:  3533 | Reward:  0.033 | Value:  0.382 | VLoss: 0.0424 | PLoss: -0.0001 | Ent: 1.3631
+[  3.2%] Iter   31/976 | SPS:  3541 | Reward:  0.023 | Value:  0.377 | VLoss: 0.0298 | PLoss: -0.0071 | Ent: 1.3573
+[  3.3%] Iter   32/976 | SPS:  3545 | Reward:  0.027 | Value:  0.323 | VLoss: 0.0410 | PLoss: -0.0096 | Ent: 1.3517
+[  3.4%] Iter   33/976 | SPS:  3547 | Reward:  0.040 | Value:  0.328 | VLoss: 0.0427 | PLoss: -0.0019 | Ent: 1.3466
+[  3.5%] Iter   34/976 | SPS:  3546 | Reward:  0.038 | Value:  0.390 | VLoss: 0.0459 | PLoss: -0.0127 | Ent: 1.3443
+[  3.6%] Iter   35/976 | SPS:  3549 | Reward:  0.027 | Value:  0.460 | VLoss: 0.0365 | PLoss: -0.0195 | Ent: 1.3184
+[  3.7%] Iter   36/976 | SPS:  3553 | Reward:  0.026 | Value:  0.411 | VLoss: 0.0343 | PLoss: -0.0050 | Ent: 1.2980
+[  3.8%] Iter   37/976 | SPS:  3555 | Reward:  0.036 | Value:  0.360 | VLoss: 0.0467 | PLoss: 0.0010 | Ent: 1.3258
+[  3.9%] Iter   38/976 | SPS:  3555 | Reward:  0.030 | Value:  0.313 | VLoss: 0.0343 | PLoss: -0.0010 | Ent: 1.3251
+[  4.0%] Iter   39/976 | SPS:  3554 | Reward:  0.028 | Value:  0.313 | VLoss: 0.0411 | PLoss: -0.0069 | Ent: 1.3120
+[  4.1%] Iter   40/976 | SPS:  3558 | Reward:  0.022 | Value:  0.370 | VLoss: 0.0308 | PLoss: -0.0060 | Ent: 1.3009
+[  4.2%] Iter   41/976 | SPS:  3565 | Reward:  0.021 | Value:  0.367 | VLoss: 0.0303 | PLoss: -0.0049 | Ent: 1.3415
+[  4.3%] Iter   42/976 | SPS:  3562 | Reward:  0.040 | Value:  0.318 | VLoss: 0.0476 | PLoss: -0.0078 | Ent: 1.3573
+[  4.4%] Iter   43/976 | SPS:  3562 | Reward:  0.027 | Value:  0.341 | VLoss: 0.0310 | PLoss: -0.0052 | Ent: 1.3618
+[  4.5%] Iter   44/976 | SPS:  3568 | Reward:  0.033 | Value:  0.335 | VLoss: 0.0451 | PLoss: -0.0033 | Ent: 1.3627
+[  4.6%] Iter   45/976 | SPS:  3568 | Reward:  0.028 | Value:  0.383 | VLoss: 0.0350 | PLoss: -0.0077 | Ent: 1.3537
+[  4.7%] Iter   46/976 | SPS:  3570 | Reward:  0.027 | Value:  0.328 | VLoss: 0.0373 | PLoss: -0.0147 | Ent: 1.3342
+[  4.8%] Iter   47/976 | SPS:  3569 | Reward:  0.038 | Value:  0.308 | VLoss: 0.0510 | PLoss: -0.0060 | Ent: 1.3445
+[  4.9%] Iter   48/976 | SPS:  3574 | Reward:  0.021 | Value:  0.330 | VLoss: 0.0336 | PLoss: -0.0090 | Ent: 1.3592
+[  5.0%] Iter   49/976 | SPS:  3578 | Reward:  0.029 | Value:  0.346 | VLoss: 0.0389 | PLoss: -0.0083 | Ent: 1.3483
+[  5.1%] Iter   50/976 | SPS:  3582 | Reward:  0.029 | Value:  0.434 | VLoss: 0.0338 | PLoss: -0.0120 | Ent: 1.3311
+[  5.2%] Iter   51/976 | SPS:  3584 | Reward:  0.034 | Value:  0.437 | VLoss: 0.0340 | PLoss: -0.0154 | Ent: 1.3385
+[  5.3%] Iter   52/976 | SPS:  3585 | Reward:  0.035 | Value:  0.378 | VLoss: 0.0336 | PLoss: -0.0093 | Ent: 1.3306
+[  5.4%] Iter   53/976 | SPS:  3588 | Reward:  0.024 | Value:  0.359 | VLoss: 0.0341 | PLoss: -0.0046 | Ent: 1.3348
+[  5.5%] Iter   54/976 | SPS:  3589 | Reward:  0.030 | Value:  0.379 | VLoss: 0.0423 | PLoss: -0.0119 | Ent: 1.3232
+[  5.6%] Iter   55/976 | SPS:  3589 | Reward:  0.030 | Value:  0.384 | VLoss: 0.0429 | PLoss: -0.0005 | Ent: 1.3409
+[  5.7%] Iter   56/976 | SPS:  3591 | Reward:  0.036 | Value:  0.315 | VLoss: 0.0472 | PLoss: -0.0107 | Ent: 1.3358
+[  5.8%] Iter   57/976 | SPS:  3594 | Reward:  0.032 | Value:  0.330 | VLoss: 0.0381 | PLoss: -0.0050 | Ent: 1.3259
+[  5.9%] Iter   58/976 | SPS:  3595 | Reward:  0.023 | Value:  0.427 | VLoss: 0.0361 | PLoss: -0.0080 | Ent: 1.3089
+[  6.0%] Iter   59/976 | SPS:  3596 | Reward:  0.033 | Value:  0.342 | VLoss: 0.0443 | PLoss: -0.0189 | Ent: 1.3243
+[  6.1%] Iter   60/976 | SPS:  3594 | Reward:  0.042 | Value:  0.363 | VLoss: 0.0544 | PLoss: -0.0065 | Ent: 1.3360
+[  6.2%] Iter   61/976 | SPS:  3595 | Reward:  0.035 | Value:  0.379 | VLoss: 0.0421 | PLoss: -0.0109 | Ent: 1.2967
+[  6.4%] Iter   62/976 | SPS:  3593 | Reward:  0.041 | Value:  0.427 | VLoss: 0.0469 | PLoss: -0.0140 | Ent: 1.3060
+[  6.5%] Iter   63/976 | SPS:  3596 | Reward:  0.031 | Value:  0.450 | VLoss: 0.0323 | PLoss: -0.0189 | Ent: 1.2699
+[  6.6%] Iter   64/976 | SPS:  3596 | Reward:  0.025 | Value:  0.408 | VLoss: 0.0379 | PLoss: -0.0047 | Ent: 1.2427
+[  6.7%] Iter   65/976 | SPS:  3598 | Reward:  0.026 | Value:  0.336 | VLoss: 0.0340 | PLoss: -0.0089 | Ent: 1.2498
+[  6.8%] Iter   66/976 | SPS:  3598 | Reward:  0.036 | Value:  0.349 | VLoss: 0.0437 | PLoss: -0.0018 | Ent: 1.2662
+[  6.9%] Iter   67/976 | SPS:  3598 | Reward:  0.034 | Value:  0.339 | VLoss: 0.0422 | PLoss: -0.0076 | Ent: 1.2473
+[  7.0%] Iter   68/976 | SPS:  3596 | Reward:  0.030 | Value:  0.360 | VLoss: 0.0419 | PLoss: -0.0133 | Ent: 1.2446
+[  7.1%] Iter   69/976 | SPS:  3595 | Reward:  0.037 | Value:  0.337 | VLoss: 0.0490 | PLoss: 0.0054 | Ent: 1.2588
+[  7.2%] Iter   70/976 | SPS:  3597 | Reward:  0.032 | Value:  0.345 | VLoss: 0.0428 | PLoss: -0.0157 | Ent: 1.2487
+[  7.3%] Iter   71/976 | SPS:  3597 | Reward:  0.037 | Value:  0.371 | VLoss: 0.0307 | PLoss: -0.0075 | Ent: 1.2554
+[  7.4%] Iter   72/976 | SPS:  3599 | Reward:  0.032 | Value:  0.425 | VLoss: 0.0345 | PLoss: -0.0044 | Ent: 1.2872
+[  7.5%] Iter   73/976 | SPS:  3601 | Reward:  0.031 | Value:  0.390 | VLoss: 0.0412 | PLoss: -0.0106 | Ent: 1.2348
+[  7.6%] Iter   74/976 | SPS:  3600 | Reward:  0.033 | Value:  0.404 | VLoss: 0.0436 | PLoss: -0.0158 | Ent: 1.2401
+[  7.7%] Iter   75/976 | SPS:  3601 | Reward:  0.039 | Value:  0.359 | VLoss: 0.0438 | PLoss: -0.0029 | Ent: 1.2246
+[  7.8%] Iter   76/976 | SPS:  3602 | Reward:  0.031 | Value:  0.400 | VLoss: 0.0358 | PLoss: -0.0146 | Ent: 1.2252
+[  7.9%] Iter   77/976 | SPS:  3603 | Reward:  0.047 | Value:  0.423 | VLoss: 0.0402 | PLoss: -0.0072 | Ent: 1.2330
+[  8.0%] Iter   78/976 | SPS:  3603 | Reward:  0.037 | Value:  0.474 | VLoss: 0.0388 | PLoss: -0.0015 | Ent: 1.2262
+[  8.1%] Iter   79/976 | SPS:  3603 | Reward:  0.037 | Value:  0.446 | VLoss: 0.0421 | PLoss: -0.0109 | Ent: 1.1916
+[  8.2%] Iter   80/976 | SPS:  3602 | Reward:  0.042 | Value:  0.371 | VLoss: 0.0482 | PLoss: -0.0150 | Ent: 1.1728
+[  8.3%] Iter   81/976 | SPS:  3603 | Reward:  0.038 | Value:  0.455 | VLoss: 0.0435 | PLoss: -0.0052 | Ent: 1.1940
+[  8.4%] Iter   82/976 | SPS:  3604 | Reward:  0.031 | Value:  0.408 | VLoss: 0.0408 | PLoss: -0.0093 | Ent: 1.1874
+[  8.5%] Iter   83/976 | SPS:  3606 | Reward:  0.037 | Value:  0.416 | VLoss: 0.0398 | PLoss: -0.0028 | Ent: 1.2047
+[  8.6%] Iter   84/976 | SPS:  3606 | Reward:  0.040 | Value:  0.467 | VLoss: 0.0406 | PLoss: -0.0006 | Ent: 1.1277
+[  8.7%] Iter   85/976 | SPS:  3608 | Reward:  0.038 | Value:  0.486 | VLoss: 0.0329 | PLoss: 0.0004 | Ent: 1.1875
+[  8.8%] Iter   86/976 | SPS:  3608 | Reward:  0.037 | Value:  0.466 | VLoss: 0.0422 | PLoss: -0.0104 | Ent: 1.1898
+[  8.9%] Iter   87/976 | SPS:  3607 | Reward:  0.044 | Value:  0.465 | VLoss: 0.0494 | PLoss: -0.0114 | Ent: 1.1254
+[  9.0%] Iter   88/976 | SPS:  3607 | Reward:  0.033 | Value:  0.432 | VLoss: 0.0413 | PLoss: -0.0220 | Ent: 1.0914
+[  9.1%] Iter   89/976 | SPS:  3608 | Reward:  0.033 | Value:  0.399 | VLoss: 0.0428 | PLoss: -0.0068 | Ent: 1.0847
+[  9.2%] Iter   90/976 | SPS:  3607 | Reward:  0.042 | Value:  0.340 | VLoss: 0.0484 | PLoss: -0.0089 | Ent: 1.1182
+[  9.3%] Iter   91/976 | SPS:  3607 | Reward:  0.041 | Value:  0.418 | VLoss: 0.0450 | PLoss: -0.0098 | Ent: 1.0837
+[  9.4%] Iter   92/976 | SPS:  3607 | Reward:  0.035 | Value:  0.417 | VLoss: 0.0476 | PLoss: -0.0123 | Ent: 1.0562
+[  9.5%] Iter   93/976 | SPS:  3609 | Reward:  0.029 | Value:  0.448 | VLoss: 0.0408 | PLoss: -0.0069 | Ent: 1.1138
+[  9.6%] Iter   94/976 | SPS:  3608 | Reward:  0.035 | Value:  0.371 | VLoss: 0.0404 | PLoss: -0.0128 | Ent: 1.1118
+[  9.7%] Iter   95/976 | SPS:  3608 | Reward:  0.040 | Value:  0.400 | VLoss: 0.0378 | PLoss: -0.0021 | Ent: 1.0924
+[  9.8%] Iter   96/976 | SPS:  3608 | Reward:  0.039 | Value:  0.444 | VLoss: 0.0410 | PLoss: -0.0125 | Ent: 1.1025
+[  9.9%] Iter   97/976 | SPS:  3609 | Reward:  0.038 | Value:  0.447 | VLoss: 0.0415 | PLoss: -0.0119 | Ent: 1.1242
+[ 10.0%] Iter   98/976 | SPS:  3611 | Reward:  0.042 | Value:  0.456 | VLoss: 0.0420 | PLoss: -0.0070 | Ent: 1.0487
+[ 10.1%] Iter   99/976 | SPS:  3613 | Reward:  0.033 | Value:  0.517 | VLoss: 0.0332 | PLoss: -0.0122 | Ent: 1.1079
+[ 10.2%] Iter  100/976 | SPS:  3614 | Reward:  0.034 | Value:  0.564 | VLoss: 0.0513 | PLoss: -0.0108 | Ent: 1.1087
+[ 10.3%] Iter  101/976 | SPS:  3616 | Reward:  0.035 | Value:  0.480 | VLoss: 0.0327 | PLoss: -0.0098 | Ent: 1.1525
+[ 10.5%] Iter  102/976 | SPS:  3616 | Reward:  0.037 | Value:  0.396 | VLoss: 0.0420 | PLoss: -0.0179 | Ent: 1.0964
+[ 10.6%] Iter  103/976 | SPS:  3615 | Reward:  0.041 | Value:  0.436 | VLoss: 0.0478 | PLoss: -0.0112 | Ent: 1.0933
+[ 10.7%] Iter  104/976 | SPS:  3614 | Reward:  0.045 | Value:  0.442 | VLoss: 0.0472 | PLoss: -0.0164 | Ent: 1.1369
+[ 10.8%] Iter  105/976 | SPS:  3616 | Reward:  0.040 | Value:  0.452 | VLoss: 0.0379 | PLoss: -0.0136 | Ent: 1.1244
+[ 10.9%] Iter  106/976 | SPS:  3617 | Reward:  0.037 | Value:  0.481 | VLoss: 0.0400 | PLoss: -0.0141 | Ent: 1.0892
+[ 11.0%] Iter  107/976 | SPS:  3615 | Reward:  0.056 | Value:  0.482 | VLoss: 0.0476 | PLoss: 0.0136 | Ent: 1.0745
+[ 11.1%] Iter  108/976 | SPS:  3615 | Reward:  0.040 | Value:  0.528 | VLoss: 0.0415 | PLoss: -0.0122 | Ent: 1.0930
+[ 11.2%] Iter  109/976 | SPS:  3615 | Reward:  0.036 | Value:  0.500 | VLoss: 0.0348 | PLoss: -0.0153 | Ent: 1.0995
+[ 11.3%] Iter  110/976 | SPS:  3614 | Reward:  0.043 | Value:  0.448 | VLoss: 0.0371 | PLoss: -0.0155 | Ent: 1.0459
+[ 11.4%] Iter  111/976 | SPS:  3615 | Reward:  0.033 | Value:  0.484 | VLoss: 0.0362 | PLoss: -0.0149 | Ent: 0.9808
+[ 11.5%] Iter  112/976 | SPS:  3616 | Reward:  0.032 | Value:  0.470 | VLoss: 0.0340 | PLoss: -0.0209 | Ent: 0.9281
+[ 11.6%] Iter  113/976 | SPS:  3617 | Reward:  0.035 | Value:  0.429 | VLoss: 0.0459 | PLoss: -0.0075 | Ent: 0.9856
+[ 11.7%] Iter  114/976 | SPS:  3617 | Reward:  0.045 | Value:  0.502 | VLoss: 0.0409 | PLoss: -0.0121 | Ent: 0.9637
+[ 11.8%] Iter  115/976 | SPS:  3617 | Reward:  0.042 | Value:  0.536 | VLoss: 0.0406 | PLoss: -0.0192 | Ent: 0.9085
+[ 11.9%] Iter  116/976 | SPS:  3617 | Reward:  0.042 | Value:  0.479 | VLoss: 0.0428 | PLoss: -0.0141 | Ent: 0.9797
+[ 12.0%] Iter  117/976 | SPS:  3616 | Reward:  0.048 | Value:  0.422 | VLoss: 0.0481 | PLoss: -0.0078 | Ent: 0.9716
+[ 12.1%] Iter  118/976 | SPS:  3616 | Reward:  0.048 | Value:  0.477 | VLoss: 0.0437 | PLoss: -0.0035 | Ent: 1.0162
+[ 12.2%] Iter  119/976 | SPS:  3616 | Reward:  0.041 | Value:  0.522 | VLoss: 0.0356 | PLoss: -0.0127 | Ent: 1.0430
+[ 12.3%] Iter  120/976 | SPS:  3617 | Reward:  0.042 | Value:  0.492 | VLoss: 0.0412 | PLoss: -0.0150 | Ent: 0.9025
+[ 12.4%] Iter  121/976 | SPS:  3619 | Reward:  0.039 | Value:  0.486 | VLoss: 0.0320 | PLoss: -0.0061 | Ent: 0.9372
+[ 12.5%] Iter  122/976 | SPS:  3620 | Reward:  0.034 | Value:  0.492 | VLoss: 0.0362 | PLoss: -0.0109 | Ent: 0.9401
+[ 12.6%] Iter  123/976 | SPS:  3621 | Reward:  0.033 | Value:  0.440 | VLoss: 0.0306 | PLoss: -0.0173 | Ent: 0.8596
+[ 12.7%] Iter  124/976 | SPS:  3622 | Reward:  0.034 | Value:  0.459 | VLoss: 0.0348 | PLoss: -0.0109 | Ent: 0.9720
+[ 12.8%] Iter  125/976 | SPS:  3624 | Reward:  0.043 | Value:  0.479 | VLoss: 0.0338 | PLoss: -0.0075 | Ent: 0.8698
+[ 12.9%] Iter  126/976 | SPS:  3624 | Reward:  0.030 | Value:  0.471 | VLoss: 0.0380 | PLoss: -0.0137 | Ent: 0.9521
+[ 13.0%] Iter  127/976 | SPS:  3624 | Reward:  0.044 | Value:  0.412 | VLoss: 0.0407 | PLoss: -0.0111 | Ent: 0.9711
+[ 13.1%] Iter  128/976 | SPS:  3624 | Reward:  0.047 | Value:  0.516 | VLoss: 0.0471 | PLoss: 0.0000 | Ent: 0.9945
+[ 13.2%] Iter  129/976 | SPS:  3626 | Reward:  0.035 | Value:  0.527 | VLoss: 0.0272 | PLoss: -0.0168 | Ent: 0.9633
+[ 13.3%] Iter  130/976 | SPS:  3626 | Reward:  0.044 | Value:  0.544 | VLoss: 0.0476 | PLoss: -0.0216 | Ent: 0.9435
+[ 13.4%] Iter  131/976 | SPS:  3627 | Reward:  0.044 | Value:  0.522 | VLoss: 0.0367 | PLoss: -0.0155 | Ent: 0.9554
+[ 13.5%] Iter  132/976 | SPS:  3627 | Reward:  0.044 | Value:  0.529 | VLoss: 0.0429 | PLoss: -0.0098 | Ent: 0.9239
+[ 13.6%] Iter  133/976 | SPS:  3628 | Reward:  0.036 | Value:  0.492 | VLoss: 0.0356 | PLoss: -0.0086 | Ent: 0.8508
+[ 13.7%] Iter  134/976 | SPS:  3629 | Reward:  0.033 | Value:  0.552 | VLoss: 0.0391 | PLoss: -0.0209 | Ent: 0.9161
+[ 13.8%] Iter  135/976 | SPS:  3629 | Reward:  0.047 | Value:  0.461 | VLoss: 0.0411 | PLoss: -0.0003 | Ent: 0.9033
+[ 13.9%] Iter  136/976 | SPS:  3630 | Reward:  0.036 | Value:  0.531 | VLoss: 0.0397 | PLoss: -0.0093 | Ent: 0.9047
+[ 14.0%] Iter  137/976 | SPS:  3630 | Reward:  0.036 | Value:  0.505 | VLoss: 0.0354 | PLoss: -0.0056 | Ent: 0.9284
+[ 14.1%] Iter  138/976 | SPS:  3629 | Reward:  0.044 | Value:  0.526 | VLoss: 0.0508 | PLoss: -0.0202 | Ent: 0.8781
+[ 14.2%] Iter  139/976 | SPS:  3630 | Reward:  0.042 | Value:  0.463 | VLoss: 0.0447 | PLoss: -0.0213 | Ent: 0.8591
+[ 14.3%] Iter  140/976 | SPS:  3631 | Reward:  0.034 | Value:  0.519 | VLoss: 0.0283 | PLoss: -0.0131 | Ent: 0.9776
+[ 14.4%] Iter  141/976 | SPS:  3630 | Reward:  0.048 | Value:  0.458 | VLoss: 0.0470 | PLoss: -0.0078 | Ent: 0.9055
+[ 14.5%] Iter  142/976 | SPS:  3632 | Reward:  0.027 | Value:  0.487 | VLoss: 0.0341 | PLoss: -0.0109 | Ent: 0.7795
+[ 14.7%] Iter  143/976 | SPS:  3631 | Reward:  0.060 | Value:  0.454 | VLoss: 0.0488 | PLoss: -0.0084 | Ent: 0.8291
+[ 14.8%] Iter  144/976 | SPS:  3631 | Reward:  0.047 | Value:  0.536 | VLoss: 0.0393 | PLoss: -0.0167 | Ent: 0.8766
+[ 14.9%] Iter  145/976 | SPS:  3630 | Reward:  0.045 | Value:  0.536 | VLoss: 0.0362 | PLoss: -0.0123 | Ent: 0.9398
+[ 15.0%] Iter  146/976 | SPS:  3631 | Reward:  0.036 | Value:  0.531 | VLoss: 0.0326 | PLoss: -0.0138 | Ent: 0.8371
+[ 15.1%] Iter  147/976 | SPS:  3630 | Reward:  0.057 | Value:  0.495 | VLoss: 0.0406 | PLoss: -0.0075 | Ent: 0.8094
+[ 15.2%] Iter  148/976 | SPS:  3630 | Reward:  0.047 | Value:  0.550 | VLoss: 0.0366 | PLoss: 0.0044 | Ent: 0.8222
+[ 15.3%] Iter  149/976 | SPS:  3630 | Reward:  0.047 | Value:  0.504 | VLoss: 0.0393 | PLoss: -0.0046 | Ent: 0.8255
+[ 15.4%] Iter  150/976 | SPS:  3631 | Reward:  0.042 | Value:  0.516 | VLoss: 0.0406 | PLoss: -0.0171 | Ent: 0.8024
+[ 15.5%] Iter  151/976 | SPS:  3631 | Reward:  0.047 | Value:  0.484 | VLoss: 0.0337 | PLoss: -0.0204 | Ent: 0.9086
+[ 15.6%] Iter  152/976 | SPS:  3632 | Reward:  0.037 | Value:  0.537 | VLoss: 0.0389 | PLoss: -0.0183 | Ent: 0.9135
+[ 15.7%] Iter  153/976 | SPS:  3634 | Reward:  0.039 | Value:  0.510 | VLoss: 0.0263 | PLoss: -0.0052 | Ent: 0.7872
+[ 15.8%] Iter  154/976 | SPS:  3635 | Reward:  0.033 | Value:  0.467 | VLoss: 0.0249 | PLoss: -0.0019 | Ent: 0.8366
+[ 15.9%] Iter  155/976 | SPS:  3634 | Reward:  0.048 | Value:  0.524 | VLoss: 0.0469 | PLoss: -0.0039 | Ent: 0.7985
+[ 16.0%] Iter  156/976 | SPS:  3633 | Reward:  0.051 | Value:  0.527 | VLoss: 0.0524 | PLoss: -0.0062 | Ent: 0.8473
+[ 16.1%] Iter  157/976 | SPS:  3633 | Reward:  0.033 | Value:  0.530 | VLoss: 0.0379 | PLoss: -0.0194 | Ent: 0.7754
+[ 16.2%] Iter  158/976 | SPS:  3633 | Reward:  0.051 | Value:  0.489 | VLoss: 0.0491 | PLoss: -0.0150 | Ent: 0.8189
+[ 16.3%] Iter  159/976 | SPS:  3633 | Reward:  0.048 | Value:  0.516 | VLoss: 0.0428 | PLoss: -0.0129 | Ent: 0.8456
+[ 16.4%] Iter  160/976 | SPS:  3632 | Reward:  0.055 | Value:  0.466 | VLoss: 0.0404 | PLoss: 0.0013 | Ent: 0.9461
+[ 16.5%] Iter  161/976 | SPS:  3632 | Reward:  0.048 | Value:  0.519 | VLoss: 0.0323 | PLoss: -0.0107 | Ent: 0.8905
+[ 16.6%] Iter  162/976 | SPS:  3633 | Reward:  0.043 | Value:  0.530 | VLoss: 0.0343 | PLoss: -0.0061 | Ent: 0.7917
+[ 16.7%] Iter  163/976 | SPS:  3634 | Reward:  0.038 | Value:  0.530 | VLoss: 0.0384 | PLoss: -0.0138 | Ent: 0.7724
+[ 16.8%] Iter  164/976 | SPS:  3634 | Reward:  0.048 | Value:  0.540 | VLoss: 0.0376 | PLoss: -0.0210 | Ent: 0.7525
+[ 16.9%] Iter  165/976 | SPS:  3634 | Reward:  0.040 | Value:  0.538 | VLoss: 0.0346 | PLoss: -0.0086 | Ent: 0.7551
+[ 17.0%] Iter  166/976 | SPS:  3634 | Reward:  0.044 | Value:  0.509 | VLoss: 0.0409 | PLoss: -0.0149 | Ent: 0.8293
+[ 17.1%] Iter  167/976 | SPS:  3635 | Reward:  0.038 | Value:  0.527 | VLoss: 0.0287 | PLoss: -0.0113 | Ent: 0.8124
+[ 17.2%] Iter  168/976 | SPS:  3635 | Reward:  0.045 | Value:  0.525 | VLoss: 0.0324 | PLoss: -0.0131 | Ent: 0.7832
+[ 17.3%] Iter  169/976 | SPS:  3637 | Reward:  0.041 | Value:  0.560 | VLoss: 0.0387 | PLoss: -0.0114 | Ent: 0.8682
+[ 17.4%] Iter  170/976 | SPS:  3636 | Reward:  0.062 | Value:  0.548 | VLoss: 0.0397 | PLoss: -0.0011 | Ent: 0.7829
+[ 17.5%] Iter  171/976 | SPS:  3636 | Reward:  0.048 | Value:  0.532 | VLoss: 0.0388 | PLoss: -0.0074 | Ent: 0.7564
+[ 17.6%] Iter  172/976 | SPS:  3637 | Reward:  0.034 | Value:  0.554 | VLoss: 0.0366 | PLoss: -0.0034 | Ent: 0.8234
+[ 17.7%] Iter  173/976 | SPS:  3637 | Reward:  0.047 | Value:  0.554 | VLoss: 0.0340 | PLoss: -0.0130 | Ent: 0.7936
+[ 17.8%] Iter  174/976 | SPS:  3637 | Reward:  0.045 | Value:  0.560 | VLoss: 0.0362 | PLoss: -0.0141 | Ent: 0.8130
+[ 17.9%] Iter  175/976 | SPS:  3637 | Reward:  0.053 | Value:  0.553 | VLoss: 0.0408 | PLoss: -0.0096 | Ent: 0.8031
+[ 18.0%] Iter  176/976 | SPS:  3637 | Reward:  0.053 | Value:  0.528 | VLoss: 0.0450 | PLoss: -0.0124 | Ent: 0.7591
+[ 18.1%] Iter  177/976 | SPS:  3637 | Reward:  0.041 | Value:  0.529 | VLoss: 0.0333 | PLoss: -0.0145 | Ent: 0.7824
+[ 18.2%] Iter  178/976 | SPS:  3638 | Reward:  0.032 | Value:  0.494 | VLoss: 0.0262 | PLoss: 0.0024 | Ent: 0.7295
+[ 18.3%] Iter  179/976 | SPS:  3640 | Reward:  0.038 | Value:  0.452 | VLoss: 0.0301 | PLoss: -0.0031 | Ent: 0.5687
+[ 18.4%] Iter  180/976 | SPS:  3640 | Reward:  0.038 | Value:  0.470 | VLoss: 0.0357 | PLoss: -0.0137 | Ent: 0.6757
+[ 18.5%] Iter  181/976 | SPS:  3641 | Reward:  0.046 | Value:  0.444 | VLoss: 0.0343 | PLoss: -0.0163 | Ent: 0.7710
+[ 18.6%] Iter  182/976 | SPS:  3641 | Reward:  0.045 | Value:  0.515 | VLoss: 0.0323 | PLoss: -0.0126 | Ent: 0.5793
+[ 18.8%] Iter  183/976 | SPS:  3642 | Reward:  0.029 | Value:  0.505 | VLoss: 0.0309 | PLoss: -0.0086 | Ent: 0.7169
+[ 18.9%] Iter  184/976 | SPS:  3645 | Reward:  0.021 | Value:  0.534 | VLoss: 0.0195 | PLoss: -0.0066 | Ent: 0.7260
+[ 19.0%] Iter  185/976 | SPS:  3645 | Reward:  0.046 | Value:  0.512 | VLoss: 0.0432 | PLoss: -0.0054 | Ent: 0.6558
+[ 19.1%] Iter  186/976 | SPS:  3645 | Reward:  0.047 | Value:  0.506 | VLoss: 0.0380 | PLoss: -0.0074 | Ent: 0.7145
+[ 19.2%] Iter  187/976 | SPS:  3646 | Reward:  0.032 | Value:  0.490 | VLoss: 0.0354 | PLoss: -0.0160 | Ent: 0.6568
+[ 19.3%] Iter  188/976 | SPS:  3647 | Reward:  0.035 | Value:  0.481 | VLoss: 0.0386 | PLoss: -0.0122 | Ent: 0.6760
+[ 19.4%] Iter  189/976 | SPS:  3647 | Reward:  0.047 | Value:  0.509 | VLoss: 0.0415 | PLoss: -0.0233 | Ent: 0.7844
+[ 19.5%] Iter  190/976 | SPS:  3648 | Reward:  0.043 | Value:  0.502 | VLoss: 0.0407 | PLoss: -0.0107 | Ent: 0.6578
+[ 19.6%] Iter  191/976 | SPS:  3649 | Reward:  0.041 | Value:  0.546 | VLoss: 0.0385 | PLoss: -0.0188 | Ent: 0.6588
+[ 19.7%] Iter  192/976 | SPS:  3650 | Reward:  0.041 | Value:  0.503 | VLoss: 0.0294 | PLoss: -0.0100 | Ent: 0.7584
+[ 19.8%] Iter  193/976 | SPS:  3650 | Reward:  0.042 | Value:  0.493 | VLoss: 0.0312 | PLoss: -0.0132 | Ent: 0.5823
+[ 19.9%] Iter  194/976 | SPS:  3652 | Reward:  0.033 | Value:  0.540 | VLoss: 0.0271 | PLoss: -0.0096 | Ent: 0.6171
+[ 20.0%] Iter  195/976 | SPS:  3653 | Reward:  0.053 | Value:  0.485 | VLoss: 0.0412 | PLoss: -0.0025 | Ent: 0.6541
+[ 20.1%] Iter  196/976 | SPS:  3654 | Reward:  0.040 | Value:  0.578 | VLoss: 0.0318 | PLoss: -0.0106 | Ent: 0.6925
+[ 20.2%] Iter  197/976 | SPS:  3655 | Reward:  0.036 | Value:  0.588 | VLoss: 0.0233 | PLoss: -0.0104 | Ent: 0.7476
+[ 20.3%] Iter  198/976 | SPS:  3656 | Reward:  0.047 | Value:  0.580 | VLoss: 0.0394 | PLoss: -0.0215 | Ent: 0.7343
+[ 20.4%] Iter  199/976 | SPS:  3656 | Reward:  0.048 | Value:  0.627 | VLoss: 0.0363 | PLoss: -0.0111 | Ent: 0.6291
+[ 20.5%] Iter  200/976 | SPS:  3656 | Reward:  0.054 | Value:  0.563 | VLoss: 0.0429 | PLoss: -0.0127 | Ent: 0.6578
+[ 20.6%] Iter  201/976 | SPS:  3655 | Reward:  0.051 | Value:  0.601 | VLoss: 0.0423 | PLoss: -0.0082 | Ent: 0.6651
+[ 20.7%] Iter  202/976 | SPS:  3654 | Reward:  0.053 | Value:  0.555 | VLoss: 0.0457 | PLoss: -0.0023 | Ent: 0.6651
+[ 20.8%] Iter  203/976 | SPS:  3654 | Reward:  0.047 | Value:  0.457 | VLoss: 0.0311 | PLoss: -0.0123 | Ent: 0.7208
+[ 20.9%] Iter  204/976 | SPS:  3655 | Reward:  0.045 | Value:  0.616 | VLoss: 0.0372 | PLoss: -0.0136 | Ent: 0.6116
+[ 21.0%] Iter  205/976 | SPS:  3656 | Reward:  0.041 | Value:  0.606 | VLoss: 0.0335 | PLoss: -0.0097 | Ent: 0.6217
+[ 21.1%] Iter  206/976 | SPS:  3657 | Reward:  0.037 | Value:  0.574 | VLoss: 0.0282 | PLoss: -0.0083 | Ent: 0.6151
+[ 21.2%] Iter  207/976 | SPS:  3657 | Reward:  0.052 | Value:  0.575 | VLoss: 0.0374 | PLoss: -0.0148 | Ent: 0.6196
+[ 21.3%] Iter  208/976 | SPS:  3658 | Reward:  0.042 | Value:  0.574 | VLoss: 0.0260 | PLoss: -0.0082 | Ent: 0.6152
+[ 21.4%] Iter  209/976 | SPS:  3658 | Reward:  0.049 | Value:  0.560 | VLoss: 0.0293 | PLoss: -0.0172 | Ent: 0.6234
+[ 21.5%] Iter  210/976 | SPS:  3658 | Reward:  0.051 | Value:  0.585 | VLoss: 0.0473 | PLoss: -0.0084 | Ent: 0.7527
+[ 21.6%] Iter  211/976 | SPS:  3657 | Reward:  0.055 | Value:  0.565 | VLoss: 0.0427 | PLoss: -0.0152 | Ent: 0.7510
+[ 21.7%] Iter  212/976 | SPS:  3658 | Reward:  0.044 | Value:  0.568 | VLoss: 0.0329 | PLoss: -0.0081 | Ent: 0.5964
+[ 21.8%] Iter  213/976 | SPS:  3658 | Reward:  0.046 | Value:  0.549 | VLoss: 0.0352 | PLoss: -0.0128 | Ent: 0.6696
+[ 21.9%] Iter  214/976 | SPS:  3658 | Reward:  0.055 | Value:  0.551 | VLoss: 0.0396 | PLoss: -0.0061 | Ent: 0.6999
+[ 22.0%] Iter  215/976 | SPS:  3658 | Reward:  0.043 | Value:  0.614 | VLoss: 0.0345 | PLoss: -0.0054 | Ent: 0.6459
+[ 22.1%] Iter  216/976 | SPS:  3659 | Reward:  0.040 | Value:  0.618 | VLoss: 0.0383 | PLoss: -0.0121 | Ent: 0.7271
+[ 22.2%] Iter  217/976 | SPS:  3660 | Reward:  0.038 | Value:  0.580 | VLoss: 0.0316 | PLoss: -0.0105 | Ent: 0.6758
+[ 22.3%] Iter  218/976 | SPS:  3659 | Reward:  0.046 | Value:  0.574 | VLoss: 0.0483 | PLoss: -0.0051 | Ent: 0.6748
+[ 22.4%] Iter  219/976 | SPS:  3659 | Reward:  0.037 | Value:  0.547 | VLoss: 0.0294 | PLoss: -0.0050 | Ent: 0.6588
+[ 22.5%] Iter  220/976 | SPS:  3659 | Reward:  0.056 | Value:  0.556 | VLoss: 0.0351 | PLoss: -0.0149 | Ent: 0.5462
+[ 22.6%] Iter  221/976 | SPS:  3658 | Reward:  0.052 | Value:  0.608 | VLoss: 0.0407 | PLoss: -0.0070 | Ent: 0.5694
+[ 22.7%] Iter  222/976 | SPS:  3659 | Reward:  0.042 | Value:  0.579 | VLoss: 0.0318 | PLoss: -0.0039 | Ent: 0.5506
+[ 22.8%] Iter  223/976 | SPS:  3658 | Reward:  0.055 | Value:  0.541 | VLoss: 0.0443 | PLoss: -0.0132 | Ent: 0.6625
+[ 23.0%] Iter  224/976 | SPS:  3658 | Reward:  0.037 | Value:  0.565 | VLoss: 0.0361 | PLoss: -0.0188 | Ent: 0.7321
+[ 23.1%] Iter  225/976 | SPS:  3658 | Reward:  0.059 | Value:  0.541 | VLoss: 0.0401 | PLoss: -0.0054 | Ent: 0.6677
+[ 23.2%] Iter  226/976 | SPS:  3658 | Reward:  0.051 | Value:  0.616 | VLoss: 0.0300 | PLoss: -0.0122 | Ent: 0.6468
+[ 23.3%] Iter  227/976 | SPS:  3657 | Reward:  0.061 | Value:  0.541 | VLoss: 0.0406 | PLoss: -0.0174 | Ent: 0.6495
+[ 23.4%] Iter  228/976 | SPS:  3657 | Reward:  0.053 | Value:  0.571 | VLoss: 0.0343 | PLoss: -0.0105 | Ent: 0.6878
+[ 23.5%] Iter  229/976 | SPS:  3657 | Reward:  0.049 | Value:  0.564 | VLoss: 0.0341 | PLoss: 0.0019 | Ent: 0.6797
+[ 23.6%] Iter  230/976 | SPS:  3657 | Reward:  0.054 | Value:  0.587 | VLoss: 0.0420 | PLoss: -0.0100 | Ent: 0.6783
+[ 23.7%] Iter  231/976 | SPS:  3657 | Reward:  0.054 | Value:  0.584 | VLoss: 0.0338 | PLoss: -0.0083 | Ent: 0.6787
+[ 23.8%] Iter  232/976 | SPS:  3657 | Reward:  0.055 | Value:  0.640 | VLoss: 0.0366 | PLoss: -0.0164 | Ent: 0.6869
+[ 23.9%] Iter  233/976 | SPS:  3657 | Reward:  0.052 | Value:  0.635 | VLoss: 0.0306 | PLoss: -0.0096 | Ent: 0.7125
+[ 24.0%] Iter  234/976 | SPS:  3657 | Reward:  0.054 | Value:  0.611 | VLoss: 0.0356 | PLoss: -0.0129 | Ent: 0.7304
+[ 24.1%] Iter  235/976 | SPS:  3657 | Reward:  0.050 | Value:  0.642 | VLoss: 0.0337 | PLoss: -0.0124 | Ent: 0.6776
+[ 24.2%] Iter  236/976 | SPS:  3658 | Reward:  0.044 | Value:  0.599 | VLoss: 0.0362 | PLoss: -0.0110 | Ent: 0.6994
+[ 24.3%] Iter  237/976 | SPS:  3657 | Reward:  0.055 | Value:  0.609 | VLoss: 0.0393 | PLoss: -0.0072 | Ent: 0.6793
+[ 24.4%] Iter  238/976 | SPS:  3656 | Reward:  0.064 | Value:  0.621 | VLoss: 0.0419 | PLoss: -0.0112 | Ent: 0.6055
+[ 24.5%] Iter  239/976 | SPS:  3657 | Reward:  0.038 | Value:  0.591 | VLoss: 0.0377 | PLoss: -0.0186 | Ent: 0.7526
+[ 24.6%] Iter  240/976 | SPS:  3657 | Reward:  0.051 | Value:  0.624 | VLoss: 0.0416 | PLoss: -0.0095 | Ent: 0.6318
+[ 24.7%] Iter  241/976 | SPS:  3657 | Reward:  0.045 | Value:  0.559 | VLoss: 0.0379 | PLoss: -0.0147 | Ent: 0.6463
+[ 24.8%] Iter  242/976 | SPS:  3658 | Reward:  0.050 | Value:  0.568 | VLoss: 0.0373 | PLoss: -0.0074 | Ent: 0.6148
+[ 24.9%] Iter  243/976 | SPS:  3657 | Reward:  0.056 | Value:  0.621 | VLoss: 0.0369 | PLoss: -0.0094 | Ent: 0.5655
+[ 25.0%] Iter  244/976 | SPS:  3658 | Reward:  0.045 | Value:  0.595 | VLoss: 0.0287 | PLoss: -0.0140 | Ent: 0.6638
+[ 25.1%] Iter  245/976 | SPS:  3658 | Reward:  0.039 | Value:  0.559 | VLoss: 0.0367 | PLoss: -0.0080 | Ent: 0.6306
+[ 25.2%] Iter  246/976 | SPS:  3659 | Reward:  0.033 | Value:  0.531 | VLoss: 0.0280 | PLoss: -0.0077 | Ent: 0.5451
+[ 25.3%] Iter  247/976 | SPS:  3660 | Reward:  0.033 | Value:  0.497 | VLoss: 0.0290 | PLoss: -0.0140 | Ent: 0.6206
+[ 25.4%] Iter  248/976 | SPS:  3661 | Reward:  0.032 | Value:  0.442 | VLoss: 0.0297 | PLoss: -0.0082 | Ent: 0.5462
+[ 25.5%] Iter  249/976 | SPS:  3661 | Reward:  0.056 | Value:  0.581 | VLoss: 0.0404 | PLoss: -0.0036 | Ent: 0.5711
+[ 25.6%] Iter  250/976 | SPS:  3661 | Reward:  0.044 | Value:  0.549 | VLoss: 0.0257 | PLoss: -0.0168 | Ent: 0.6730
+[ 25.7%] Iter  251/976 | SPS:  3662 | Reward:  0.045 | Value:  0.570 | VLoss: 0.0297 | PLoss: -0.0102 | Ent: 0.6010
+[ 25.8%] Iter  252/976 | SPS:  3662 | Reward:  0.041 | Value:  0.576 | VLoss: 0.0379 | PLoss: -0.0084 | Ent: 0.5861
+[ 25.9%] Iter  253/976 | SPS:  3661 | Reward:  0.063 | Value:  0.554 | VLoss: 0.0447 | PLoss: -0.0002 | Ent: 0.6199
+[ 26.0%] Iter  254/976 | SPS:  3662 | Reward:  0.039 | Value:  0.593 | VLoss: 0.0286 | PLoss: -0.0051 | Ent: 0.5656
+[ 26.1%] Iter  255/976 | SPS:  3662 | Reward:  0.051 | Value:  0.524 | VLoss: 0.0303 | PLoss: -0.0101 | Ent: 0.6363
+[ 26.2%] Iter  256/976 | SPS:  3661 | Reward:  0.054 | Value:  0.602 | VLoss: 0.0326 | PLoss: -0.0226 | Ent: 0.5862
+[ 26.3%] Iter  257/976 | SPS:  3662 | Reward:  0.038 | Value:  0.591 | VLoss: 0.0371 | PLoss: -0.0099 | Ent: 0.5595
+[ 26.4%] Iter  258/976 | SPS:  3663 | Reward:  0.036 | Value:  0.567 | VLoss: 0.0348 | PLoss: -0.0015 | Ent: 0.5789
+[ 26.5%] Iter  259/976 | SPS:  3662 | Reward:  0.054 | Value:  0.499 | VLoss: 0.0386 | PLoss: -0.0109 | Ent: 0.5577
+[ 26.6%] Iter  260/976 | SPS:  3662 | Reward:  0.046 | Value:  0.595 | VLoss: 0.0339 | PLoss: -0.0079 | Ent: 0.6055
+[ 26.7%] Iter  261/976 | SPS:  3662 | Reward:  0.047 | Value:  0.613 | VLoss: 0.0364 | PLoss: -0.0055 | Ent: 0.6489
+[ 26.8%] Iter  262/976 | SPS:  3662 | Reward:  0.059 | Value:  0.615 | VLoss: 0.0325 | PLoss: -0.0044 | Ent: 0.5745
+[ 26.9%] Iter  263/976 | SPS:  3662 | Reward:  0.051 | Value:  0.615 | VLoss: 0.0394 | PLoss: -0.0039 | Ent: 0.5755
+[ 27.0%] Iter  264/976 | SPS:  3662 | Reward:  0.055 | Value:  0.625 | VLoss: 0.0418 | PLoss: -0.0082 | Ent: 0.5276
+[ 27.2%] Iter  265/976 | SPS:  3662 | Reward:  0.056 | Value:  0.601 | VLoss: 0.0314 | PLoss: -0.0120 | Ent: 0.6444
+[ 27.3%] Iter  266/976 | SPS:  3661 | Reward:  0.056 | Value:  0.576 | VLoss: 0.0331 | PLoss: -0.0107 | Ent: 0.6782
+[ 27.4%] Iter  267/976 | SPS:  3662 | Reward:  0.048 | Value:  0.574 | VLoss: 0.0365 | PLoss: -0.0105 | Ent: 0.6472
+[ 27.5%] Iter  268/976 | SPS:  3662 | Reward:  0.047 | Value:  0.565 | VLoss: 0.0326 | PLoss: -0.0103 | Ent: 0.5767
+[ 27.6%] Iter  269/976 | SPS:  3662 | Reward:  0.045 | Value:  0.576 | VLoss: 0.0323 | PLoss: -0.0092 | Ent: 0.6181
+[ 27.7%] Iter  270/976 | SPS:  3662 | Reward:  0.053 | Value:  0.574 | VLoss: 0.0501 | PLoss: -0.0027 | Ent: 0.6139
+[ 27.8%] Iter  271/976 | SPS:  3662 | Reward:  0.054 | Value:  0.569 | VLoss: 0.0392 | PLoss: -0.0077 | Ent: 0.5858
+[ 27.9%] Iter  272/976 | SPS:  3662 | Reward:  0.051 | Value:  0.587 | VLoss: 0.0379 | PLoss: -0.0072 | Ent: 0.6366
+[ 28.0%] Iter  273/976 | SPS:  3663 | Reward:  0.046 | Value:  0.643 | VLoss: 0.0312 | PLoss: -0.0064 | Ent: 0.5688
+[ 28.1%] Iter  274/976 | SPS:  3662 | Reward:  0.062 | Value:  0.627 | VLoss: 0.0393 | PLoss: -0.0166 | Ent: 0.6256
+[ 28.2%] Iter  275/976 | SPS:  3662 | Reward:  0.048 | Value:  0.579 | VLoss: 0.0342 | PLoss: -0.0069 | Ent: 0.5792
+[ 28.3%] Iter  276/976 | SPS:  3662 | Reward:  0.046 | Value:  0.551 | VLoss: 0.0329 | PLoss: -0.0055 | Ent: 0.6150
+[ 28.4%] Iter  277/976 | SPS:  3664 | Reward:  0.028 | Value:  0.546 | VLoss: 0.0232 | PLoss: -0.0065 | Ent: 0.5858
+[ 28.5%] Iter  278/976 | SPS:  3663 | Reward:  0.048 | Value:  0.531 | VLoss: 0.0416 | PLoss: -0.0092 | Ent: 0.5416
+[ 28.6%] Iter  279/976 | SPS:  3663 | Reward:  0.055 | Value:  0.517 | VLoss: 0.0459 | PLoss: -0.0130 | Ent: 0.5927
+[ 28.7%] Iter  280/976 | SPS:  3664 | Reward:  0.049 | Value:  0.558 | VLoss: 0.0341 | PLoss: -0.0158 | Ent: 0.5910
+[ 28.8%] Iter  281/976 | SPS:  3664 | Reward:  0.046 | Value:  0.566 | VLoss: 0.0285 | PLoss: -0.0137 | Ent: 0.6069
+[ 28.9%] Iter  282/976 | SPS:  3664 | Reward:  0.057 | Value:  0.511 | VLoss: 0.0310 | PLoss: -0.0161 | Ent: 0.5846
+[ 29.0%] Iter  283/976 | SPS:  3665 | Reward:  0.029 | Value:  0.535 | VLoss: 0.0205 | PLoss: -0.0134 | Ent: 0.4890
+[ 29.1%] Iter  284/976 | SPS:  3665 | Reward:  0.039 | Value:  0.475 | VLoss: 0.0342 | PLoss: -0.0065 | Ent: 0.5845
+[ 29.2%] Iter  285/976 | SPS:  3666 | Reward:  0.040 | Value:  0.464 | VLoss: 0.0234 | PLoss: 0.0026 | Ent: 0.5820
+[ 29.3%] Iter  286/976 | SPS:  3667 | Reward:  0.042 | Value:  0.538 | VLoss: 0.0356 | PLoss: -0.0140 | Ent: 0.6157
+[ 29.4%] Iter  287/976 | SPS:  3668 | Reward:  0.035 | Value:  0.526 | VLoss: 0.0262 | PLoss: -0.0035 | Ent: 0.4815
+[ 29.5%] Iter  288/976 | SPS:  3669 | Reward:  0.038 | Value:  0.590 | VLoss: 0.0256 | PLoss: -0.0106 | Ent: 0.5681
+[ 29.6%] Iter  289/976 | SPS:  3670 | Reward:  0.035 | Value:  0.492 | VLoss: 0.0221 | PLoss: -0.0048 | Ent: 0.5709
+[ 29.7%] Iter  290/976 | SPS:  3670 | Reward:  0.040 | Value:  0.511 | VLoss: 0.0321 | PLoss: -0.0170 | Ent: 0.6573
+[ 29.8%] Iter  291/976 | SPS:  3670 | Reward:  0.048 | Value:  0.540 | VLoss: 0.0374 | PLoss: -0.0132 | Ent: 0.6492
+[ 29.9%] Iter  292/976 | SPS:  3670 | Reward:  0.056 | Value:  0.587 | VLoss: 0.0389 | PLoss: -0.0039 | Ent: 0.5456
+[ 30.0%] Iter  293/976 | SPS:  3670 | Reward:  0.052 | Value:  0.611 | VLoss: 0.0372 | PLoss: -0.0065 | Ent: 0.5866
+[ 30.1%] Iter  294/976 | SPS:  3669 | Reward:  0.063 | Value:  0.604 | VLoss: 0.0346 | PLoss: -0.0124 | Ent: 0.5365
+[ 30.2%] Iter  295/976 | SPS:  3669 | Reward:  0.053 | Value:  0.646 | VLoss: 0.0322 | PLoss: -0.0007 | Ent: 0.5748
+[ 30.3%] Iter  296/976 | SPS:  3669 | Reward:  0.053 | Value:  0.585 | VLoss: 0.0350 | PLoss: -0.0148 | Ent: 0.5178
+[ 30.4%] Iter  297/976 | SPS:  3669 | Reward:  0.062 | Value:  0.611 | VLoss: 0.0328 | PLoss: -0.0192 | Ent: 0.5521
+[ 30.5%] Iter  298/976 | SPS:  3669 | Reward:  0.044 | Value:  0.650 | VLoss: 0.0306 | PLoss: -0.0129 | Ent: 0.5608
+[ 30.6%] Iter  299/976 | SPS:  3668 | Reward:  0.058 | Value:  0.658 | VLoss: 0.0333 | PLoss: -0.0122 | Ent: 0.4741
+[ 30.7%] Iter  300/976 | SPS:  3668 | Reward:  0.056 | Value:  0.608 | VLoss: 0.0388 | PLoss: -0.0127 | Ent: 0.6490
+[ 30.8%] Iter  301/976 | SPS:  3668 | Reward:  0.057 | Value:  0.634 | VLoss: 0.0366 | PLoss: -0.0124 | Ent: 0.5648
+[ 30.9%] Iter  302/976 | SPS:  3668 | Reward:  0.047 | Value:  0.620 | VLoss: 0.0352 | PLoss: -0.0124 | Ent: 0.6013
+[ 31.0%] Iter  303/976 | SPS:  3668 | Reward:  0.051 | Value:  0.616 | VLoss: 0.0328 | PLoss: -0.0056 | Ent: 0.5464
+[ 31.1%] Iter  304/976 | SPS:  3667 | Reward:  0.058 | Value:  0.609 | VLoss: 0.0380 | PLoss: -0.0079 | Ent: 0.6301
+[ 31.2%] Iter  305/976 | SPS:  3668 | Reward:  0.048 | Value:  0.582 | VLoss: 0.0402 | PLoss: -0.0130 | Ent: 0.5626
+[ 31.4%] Iter  306/976 | SPS:  3668 | Reward:  0.053 | Value:  0.557 | VLoss: 0.0336 | PLoss: -0.0074 | Ent: 0.5804
+[ 31.5%] Iter  307/976 | SPS:  3667 | Reward:  0.069 | Value:  0.605 | VLoss: 0.0424 | PLoss: -0.0125 | Ent: 0.5509
+[ 31.6%] Iter  308/976 | SPS:  3666 | Reward:  0.068 | Value:  0.638 | VLoss: 0.0372 | PLoss: -0.0004 | Ent: 0.5681
+[ 31.7%] Iter  309/976 | SPS:  3666 | Reward:  0.049 | Value:  0.655 | VLoss: 0.0333 | PLoss: -0.0032 | Ent: 0.5130
+[ 31.8%] Iter  310/976 | SPS:  3665 | Reward:  0.067 | Value:  0.605 | VLoss: 0.0434 | PLoss: -0.0077 | Ent: 0.5484
+[ 31.9%] Iter  311/976 | SPS:  3665 | Reward:  0.048 | Value:  0.640 | VLoss: 0.0328 | PLoss: -0.0069 | Ent: 0.6024
+[ 32.0%] Iter  312/976 | SPS:  3665 | Reward:  0.049 | Value:  0.599 | VLoss: 0.0359 | PLoss: -0.0020 | Ent: 0.5864
+[ 32.1%] Iter  313/976 | SPS:  3665 | Reward:  0.050 | Value:  0.600 | VLoss: 0.0311 | PLoss: -0.0069 | Ent: 0.6415
+[ 32.2%] Iter  314/976 | SPS:  3665 | Reward:  0.054 | Value:  0.572 | VLoss: 0.0375 | PLoss: 0.0033 | Ent: 0.6012
+[ 32.3%] Iter  315/976 | SPS:  3664 | Reward:  0.060 | Value:  0.634 | VLoss: 0.0404 | PLoss: -0.0075 | Ent: 0.5498
+[ 32.4%] Iter  316/976 | SPS:  3663 | Reward:  0.061 | Value:  0.609 | VLoss: 0.0489 | PLoss: -0.0135 | Ent: 0.5564
+[ 32.5%] Iter  317/976 | SPS:  3663 | Reward:  0.061 | Value:  0.612 | VLoss: 0.0534 | PLoss: -0.0062 | Ent: 0.5444
+[ 32.6%] Iter  318/976 | SPS:  3662 | Reward:  0.080 | Value:  0.629 | VLoss: 0.0401 | PLoss: -0.0141 | Ent: 0.5931
+[ 32.7%] Iter  319/976 | SPS:  3661 | Reward:  0.052 | Value:  0.595 | VLoss: 0.0524 | PLoss: -0.0068 | Ent: 0.5712
+[ 32.8%] Iter  320/976 | SPS:  3661 | Reward:  0.058 | Value:  0.612 | VLoss: 0.0277 | PLoss: -0.0117 | Ent: 0.5631
+[ 32.9%] Iter  321/976 | SPS:  3661 | Reward:  0.060 | Value:  0.609 | VLoss: 0.0336 | PLoss: -0.0070 | Ent: 0.5670
+[ 33.0%] Iter  322/976 | SPS:  3661 | Reward:  0.050 | Value:  0.674 | VLoss: 0.0360 | PLoss: -0.0108 | Ent: 0.4978
+[ 33.1%] Iter  323/976 | SPS:  3661 | Reward:  0.061 | Value:  0.601 | VLoss: 0.0348 | PLoss: -0.0026 | Ent: 0.5366
+[ 33.2%] Iter  324/976 | SPS:  3660 | Reward:  0.058 | Value:  0.613 | VLoss: 0.0384 | PLoss: -0.0074 | Ent: 0.4787
+[ 33.3%] Iter  325/976 | SPS:  3659 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0386 | PLoss: -0.0061 | Ent: 0.5239
+[ 33.4%] Iter  326/976 | SPS:  3660 | Reward:  0.044 | Value:  0.611 | VLoss: 0.0290 | PLoss: -0.0119 | Ent: 0.6012
+[ 33.5%] Iter  327/976 | SPS:  3659 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0396 | PLoss: -0.0131 | Ent: 0.5094
+[ 33.6%] Iter  328/976 | SPS:  3659 | Reward:  0.045 | Value:  0.631 | VLoss: 0.0275 | PLoss: -0.0070 | Ent: 0.5801
+[ 33.7%] Iter  329/976 | SPS:  3659 | Reward:  0.057 | Value:  0.590 | VLoss: 0.0281 | PLoss: -0.0040 | Ent: 0.5073
+[ 33.8%] Iter  330/976 | SPS:  3659 | Reward:  0.050 | Value:  0.657 | VLoss: 0.0306 | PLoss: -0.0124 | Ent: 0.5377
+[ 33.9%] Iter  331/976 | SPS:  3658 | Reward:  0.068 | Value:  0.626 | VLoss: 0.0450 | PLoss: -0.0032 | Ent: 0.4975
+[ 34.0%] Iter  332/976 | SPS:  3658 | Reward:  0.070 | Value:  0.659 | VLoss: 0.0305 | PLoss: -0.0150 | Ent: 0.5508
+[ 34.1%] Iter  333/976 | SPS:  3657 | Reward:  0.060 | Value:  0.639 | VLoss: 0.0418 | PLoss: -0.0102 | Ent: 0.5157
+[ 34.2%] Iter  334/976 | SPS:  3656 | Reward:  0.076 | Value:  0.661 | VLoss: 0.0420 | PLoss: -0.0118 | Ent: 0.5616
+[ 34.3%] Iter  335/976 | SPS:  3656 | Reward:  0.048 | Value:  0.672 | VLoss: 0.0422 | PLoss: -0.0090 | Ent: 0.4736
+[ 34.4%] Iter  336/976 | SPS:  3656 | Reward:  0.052 | Value:  0.626 | VLoss: 0.0378 | PLoss: -0.0124 | Ent: 0.5451
+[ 34.5%] Iter  337/976 | SPS:  3656 | Reward:  0.048 | Value:  0.606 | VLoss: 0.0339 | PLoss: -0.0078 | Ent: 0.5216
+[ 34.6%] Iter  338/976 | SPS:  3656 | Reward:  0.054 | Value:  0.598 | VLoss: 0.0373 | PLoss: -0.0058 | Ent: 0.5466
+[ 34.7%] Iter  339/976 | SPS:  3656 | Reward:  0.057 | Value:  0.670 | VLoss: 0.0404 | PLoss: -0.0068 | Ent: 0.5506
+[ 34.8%] Iter  340/976 | SPS:  3656 | Reward:  0.052 | Value:  0.631 | VLoss: 0.0321 | PLoss: -0.0010 | Ent: 0.5506
+[ 34.9%] Iter  341/976 | SPS:  3655 | Reward:  0.068 | Value:  0.542 | VLoss: 0.0378 | PLoss: -0.0124 | Ent: 0.5823
+[ 35.0%] Iter  342/976 | SPS:  3655 | Reward:  0.057 | Value:  0.632 | VLoss: 0.0414 | PLoss: -0.0123 | Ent: 0.5200
+[ 35.1%] Iter  343/976 | SPS:  3654 | Reward:  0.057 | Value:  0.625 | VLoss: 0.0337 | PLoss: -0.0032 | Ent: 0.5499
+[ 35.2%] Iter  344/976 | SPS:  3654 | Reward:  0.062 | Value:  0.642 | VLoss: 0.0405 | PLoss: -0.0055 | Ent: 0.4607
+[ 35.3%] Iter  345/976 | SPS:  3654 | Reward:  0.054 | Value:  0.637 | VLoss: 0.0315 | PLoss: -0.0136 | Ent: 0.5144
+[ 35.5%] Iter  346/976 | SPS:  3653 | Reward:  0.066 | Value:  0.604 | VLoss: 0.0464 | PLoss: -0.0054 | Ent: 0.5435
+[ 35.6%] Iter  347/976 | SPS:  3653 | Reward:  0.064 | Value:  0.606 | VLoss: 0.0483 | PLoss: -0.0061 | Ent: 0.4839
+[ 35.7%] Iter  348/976 | SPS:  3652 | Reward:  0.063 | Value:  0.616 | VLoss: 0.0460 | PLoss: -0.0166 | Ent: 0.5187
+[ 35.8%] Iter  349/976 | SPS:  3653 | Reward:  0.055 | Value:  0.644 | VLoss: 0.0366 | PLoss: -0.0029 | Ent: 0.5156
+[ 35.9%] Iter  350/976 | SPS:  3653 | Reward:  0.059 | Value:  0.624 | VLoss: 0.0367 | PLoss: -0.0043 | Ent: 0.4917
+[ 36.0%] Iter  351/976 | SPS:  3652 | Reward:  0.066 | Value:  0.627 | VLoss: 0.0410 | PLoss: -0.0091 | Ent: 0.5443
+[ 36.1%] Iter  352/976 | SPS:  3652 | Reward:  0.058 | Value:  0.575 | VLoss: 0.0473 | PLoss: -0.0080 | Ent: 0.4450
+[ 36.2%] Iter  353/976 | SPS:  3652 | Reward:  0.052 | Value:  0.590 | VLoss: 0.0313 | PLoss: -0.0033 | Ent: 0.4603
+[ 36.3%] Iter  354/976 | SPS:  3651 | Reward:  0.058 | Value:  0.605 | VLoss: 0.0407 | PLoss: -0.0100 | Ent: 0.4959
+[ 36.4%] Iter  355/976 | SPS:  3651 | Reward:  0.068 | Value:  0.592 | VLoss: 0.0450 | PLoss: -0.0110 | Ent: 0.4850
+[ 36.5%] Iter  356/976 | SPS:  3650 | Reward:  0.059 | Value:  0.582 | VLoss: 0.0408 | PLoss: -0.0033 | Ent: 0.4773
+[ 36.6%] Iter  357/976 | SPS:  3650 | Reward:  0.064 | Value:  0.588 | VLoss: 0.0345 | PLoss: -0.0183 | Ent: 0.4775
+[ 36.7%] Iter  358/976 | SPS:  3650 | Reward:  0.053 | Value:  0.577 | VLoss: 0.0395 | PLoss: -0.0069 | Ent: 0.4927
+[ 36.8%] Iter  359/976 | SPS:  3650 | Reward:  0.051 | Value:  0.551 | VLoss: 0.0334 | PLoss: -0.0024 | Ent: 0.4898
+[ 36.9%] Iter  360/976 | SPS:  3649 | Reward:  0.065 | Value:  0.593 | VLoss: 0.0470 | PLoss: -0.0070 | Ent: 0.4966
+[ 37.0%] Iter  361/976 | SPS:  3649 | Reward:  0.054 | Value:  0.597 | VLoss: 0.0302 | PLoss: -0.0031 | Ent: 0.3950
+[ 37.1%] Iter  362/976 | SPS:  3649 | Reward:  0.060 | Value:  0.614 | VLoss: 0.0397 | PLoss: -0.0164 | Ent: 0.4878
+[ 37.2%] Iter  363/976 | SPS:  3649 | Reward:  0.068 | Value:  0.654 | VLoss: 0.0381 | PLoss: 0.0036 | Ent: 0.4553
+[ 37.3%] Iter  364/976 | SPS:  3648 | Reward:  0.059 | Value:  0.655 | VLoss: 0.0455 | PLoss: -0.0116 | Ent: 0.4551
+[ 37.4%] Iter  365/976 | SPS:  3648 | Reward:  0.065 | Value:  0.604 | VLoss: 0.0440 | PLoss: -0.0118 | Ent: 0.4548
+[ 37.5%] Iter  366/976 | SPS:  3648 | Reward:  0.056 | Value:  0.651 | VLoss: 0.0411 | PLoss: -0.0082 | Ent: 0.4476
+[ 37.6%] Iter  367/976 | SPS:  3647 | Reward:  0.064 | Value:  0.575 | VLoss: 0.0405 | PLoss: -0.0026 | Ent: 0.4775
+[ 37.7%] Iter  368/976 | SPS:  3646 | Reward:  0.070 | Value:  0.599 | VLoss: 0.0383 | PLoss: -0.0078 | Ent: 0.4831
+[ 37.8%] Iter  369/976 | SPS:  3646 | Reward:  0.059 | Value:  0.639 | VLoss: 0.0497 | PLoss: -0.0033 | Ent: 0.4793
+[ 37.9%] Iter  370/976 | SPS:  3646 | Reward:  0.062 | Value:  0.591 | VLoss: 0.0427 | PLoss: 0.0010 | Ent: 0.4455
+[ 38.0%] Iter  371/976 | SPS:  3645 | Reward:  0.058 | Value:  0.626 | VLoss: 0.0407 | PLoss: -0.0029 | Ent: 0.4574
+[ 38.1%] Iter  372/976 | SPS:  3645 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0326 | PLoss: -0.0084 | Ent: 0.4124
+[ 38.2%] Iter  373/976 | SPS:  3645 | Reward:  0.064 | Value:  0.627 | VLoss: 0.0350 | PLoss: -0.0084 | Ent: 0.3835
+[ 38.3%] Iter  374/976 | SPS:  3644 | Reward:  0.066 | Value:  0.677 | VLoss: 0.0423 | PLoss: -0.0104 | Ent: 0.4460
+[ 38.4%] Iter  375/976 | SPS:  3644 | Reward:  0.058 | Value:  0.644 | VLoss: 0.0402 | PLoss: -0.0006 | Ent: 0.4337
+[ 38.5%] Iter  376/976 | SPS:  3644 | Reward:  0.066 | Value:  0.612 | VLoss: 0.0340 | PLoss: -0.0090 | Ent: 0.4327
+[ 38.6%] Iter  377/976 | SPS:  3644 | Reward:  0.051 | Value:  0.635 | VLoss: 0.0302 | PLoss: -0.0055 | Ent: 0.4553
+[ 38.7%] Iter  378/976 | SPS:  3643 | Reward:  0.064 | Value:  0.643 | VLoss: 0.0484 | PLoss: -0.0054 | Ent: 0.5199
+[ 38.8%] Iter  379/976 | SPS:  3643 | Reward:  0.068 | Value:  0.611 | VLoss: 0.0343 | PLoss: -0.0171 | Ent: 0.5248
+[ 38.9%] Iter  380/976 | SPS:  3642 | Reward:  0.066 | Value:  0.660 | VLoss: 0.0345 | PLoss: 0.0035 | Ent: 0.4320
+[ 39.0%] Iter  381/976 | SPS:  3642 | Reward:  0.057 | Value:  0.661 | VLoss: 0.0427 | PLoss: -0.0122 | Ent: 0.4031
+[ 39.1%] Iter  382/976 | SPS:  3641 | Reward:  0.072 | Value:  0.659 | VLoss: 0.0373 | PLoss: -0.0091 | Ent: 0.4387
+[ 39.2%] Iter  383/976 | SPS:  3641 | Reward:  0.067 | Value:  0.678 | VLoss: 0.0374 | PLoss: -0.0104 | Ent: 0.4367
+[ 39.3%] Iter  384/976 | SPS:  3640 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0373 | PLoss: -0.0133 | Ent: 0.4313
+[ 39.4%] Iter  385/976 | SPS:  3640 | Reward:  0.059 | Value:  0.731 | VLoss: 0.0325 | PLoss: -0.0107 | Ent: 0.3993
+[ 39.5%] Iter  386/976 | SPS:  3640 | Reward:  0.051 | Value:  0.632 | VLoss: 0.0454 | PLoss: -0.0115 | Ent: 0.4676
+[ 39.7%] Iter  387/976 | SPS:  3640 | Reward:  0.064 | Value:  0.659 | VLoss: 0.0455 | PLoss: -0.0057 | Ent: 0.4579
+[ 39.8%] Iter  388/976 | SPS:  3639 | Reward:  0.067 | Value:  0.618 | VLoss: 0.0472 | PLoss: -0.0134 | Ent: 0.4457
+[ 39.9%] Iter  389/976 | SPS:  3639 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0417 | PLoss: -0.0082 | Ent: 0.4398
+[ 40.0%] Iter  390/976 | SPS:  3638 | Reward:  0.067 | Value:  0.655 | VLoss: 0.0432 | PLoss: -0.0110 | Ent: 0.3995
+[ 40.1%] Iter  391/976 | SPS:  3638 | Reward:  0.053 | Value:  0.649 | VLoss: 0.0272 | PLoss: -0.0115 | Ent: 0.4760
+[ 40.2%] Iter  392/976 | SPS:  3638 | Reward:  0.055 | Value:  0.636 | VLoss: 0.0434 | PLoss: -0.0073 | Ent: 0.4365
+[ 40.3%] Iter  393/976 | SPS:  3637 | Reward:  0.067 | Value:  0.570 | VLoss: 0.0507 | PLoss: -0.0051 | Ent: 0.4448
+[ 40.4%] Iter  394/976 | SPS:  3637 | Reward:  0.069 | Value:  0.626 | VLoss: 0.0414 | PLoss: 0.0001 | Ent: 0.4363
+[ 40.5%] Iter  395/976 | SPS:  3636 | Reward:  0.062 | Value:  0.646 | VLoss: 0.0410 | PLoss: -0.0076 | Ent: 0.4581
+[ 40.6%] Iter  396/976 | SPS:  3635 | Reward:  0.066 | Value:  0.642 | VLoss: 0.0367 | PLoss: -0.0071 | Ent: 0.4679
+[ 40.7%] Iter  397/976 | SPS:  3635 | Reward:  0.061 | Value:  0.647 | VLoss: 0.0389 | PLoss: -0.0076 | Ent: 0.4161
+[ 40.8%] Iter  398/976 | SPS:  3635 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0366 | PLoss: -0.0179 | Ent: 0.4274
+[ 40.9%] Iter  399/976 | SPS:  3634 | Reward:  0.057 | Value:  0.650 | VLoss: 0.0389 | PLoss: -0.0107 | Ent: 0.4317
+[ 41.0%] Iter  400/976 | SPS:  3634 | Reward:  0.060 | Value:  0.665 | VLoss: 0.0357 | PLoss: -0.0113 | Ent: 0.4234
+[ 41.1%] Iter  401/976 | SPS:  3634 | Reward:  0.053 | Value:  0.633 | VLoss: 0.0277 | PLoss: 0.0024 | Ent: 0.4276
+[ 41.2%] Iter  402/976 | SPS:  3634 | Reward:  0.061 | Value:  0.653 | VLoss: 0.0385 | PLoss: -0.0044 | Ent: 0.3950
+[ 41.3%] Iter  403/976 | SPS:  3634 | Reward:  0.056 | Value:  0.658 | VLoss: 0.0406 | PLoss: -0.0079 | Ent: 0.4027
+[ 41.4%] Iter  404/976 | SPS:  3633 | Reward:  0.070 | Value:  0.639 | VLoss: 0.0542 | PLoss: -0.0076 | Ent: 0.3685
+[ 41.5%] Iter  405/976 | SPS:  3632 | Reward:  0.069 | Value:  0.566 | VLoss: 0.0497 | PLoss: -0.0072 | Ent: 0.4513
+[ 41.6%] Iter  406/976 | SPS:  3632 | Reward:  0.074 | Value:  0.631 | VLoss: 0.0347 | PLoss: -0.0085 | Ent: 0.3356
+[ 41.7%] Iter  407/976 | SPS:  3631 | Reward:  0.069 | Value:  0.636 | VLoss: 0.0403 | PLoss: -0.0025 | Ent: 0.3446
+[ 41.8%] Iter  408/976 | SPS:  3630 | Reward:  0.067 | Value:  0.634 | VLoss: 0.0413 | PLoss: -0.0102 | Ent: 0.3953
+[ 41.9%] Iter  409/976 | SPS:  3630 | Reward:  0.054 | Value:  0.739 | VLoss: 0.0311 | PLoss: -0.0126 | Ent: 0.3799
+[ 42.0%] Iter  410/976 | SPS:  3630 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0354 | PLoss: -0.0068 | Ent: 0.3957
+[ 42.1%] Iter  411/976 | SPS:  3630 | Reward:  0.059 | Value:  0.602 | VLoss: 0.0322 | PLoss: -0.0144 | Ent: 0.4174
+[ 42.2%] Iter  412/976 | SPS:  3629 | Reward:  0.070 | Value:  0.617 | VLoss: 0.0443 | PLoss: -0.0100 | Ent: 0.3705
+[ 42.3%] Iter  413/976 | SPS:  3629 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0451 | PLoss: -0.0120 | Ent: 0.4407
+[ 42.4%] Iter  414/976 | SPS:  3629 | Reward:  0.063 | Value:  0.650 | VLoss: 0.0376 | PLoss: -0.0082 | Ent: 0.3615
+[ 42.5%] Iter  415/976 | SPS:  3628 | Reward:  0.065 | Value:  0.644 | VLoss: 0.0430 | PLoss: -0.0099 | Ent: 0.4522
+[ 42.6%] Iter  416/976 | SPS:  3628 | Reward:  0.066 | Value:  0.638 | VLoss: 0.0356 | PLoss: -0.0082 | Ent: 0.3833
+[ 42.7%] Iter  417/976 | SPS:  3628 | Reward:  0.056 | Value:  0.608 | VLoss: 0.0383 | PLoss: -0.0035 | Ent: 0.4086
+[ 42.8%] Iter  418/976 | SPS:  3627 | Reward:  0.065 | Value:  0.609 | VLoss: 0.0381 | PLoss: -0.0134 | Ent: 0.3867
+[ 42.9%] Iter  419/976 | SPS:  3627 | Reward:  0.058 | Value:  0.543 | VLoss: 0.0471 | PLoss: -0.0085 | Ent: 0.4060
+[ 43.0%] Iter  420/976 | SPS:  3627 | Reward:  0.069 | Value:  0.615 | VLoss: 0.0422 | PLoss: -0.0047 | Ent: 0.4426
+[ 43.1%] Iter  421/976 | SPS:  3627 | Reward:  0.061 | Value:  0.637 | VLoss: 0.0355 | PLoss: -0.0008 | Ent: 0.3837
+[ 43.2%] Iter  422/976 | SPS:  3627 | Reward:  0.053 | Value:  0.654 | VLoss: 0.0353 | PLoss: -0.0109 | Ent: 0.3996
+[ 43.3%] Iter  423/976 | SPS:  3627 | Reward:  0.062 | Value:  0.631 | VLoss: 0.0379 | PLoss: -0.0067 | Ent: 0.3675
+[ 43.4%] Iter  424/976 | SPS:  3627 | Reward:  0.049 | Value:  0.654 | VLoss: 0.0305 | PLoss: -0.0089 | Ent: 0.3798
+[ 43.5%] Iter  425/976 | SPS:  3627 | Reward:  0.054 | Value:  0.531 | VLoss: 0.0403 | PLoss: -0.0172 | Ent: 0.4565
+[ 43.6%] Iter  426/976 | SPS:  3627 | Reward:  0.064 | Value:  0.603 | VLoss: 0.0497 | PLoss: -0.0066 | Ent: 0.4406
+[ 43.8%] Iter  427/976 | SPS:  3626 | Reward:  0.068 | Value:  0.662 | VLoss: 0.0356 | PLoss: -0.0075 | Ent: 0.3616
+[ 43.9%] Iter  428/976 | SPS:  3627 | Reward:  0.043 | Value:  0.662 | VLoss: 0.0328 | PLoss: -0.0145 | Ent: 0.4183
+[ 44.0%] Iter  429/976 | SPS:  3626 | Reward:  0.069 | Value:  0.641 | VLoss: 0.0342 | PLoss: 0.0013 | Ent: 0.4146
+[ 44.1%] Iter  430/976 | SPS:  3626 | Reward:  0.063 | Value:  0.684 | VLoss: 0.0382 | PLoss: -0.0115 | Ent: 0.4029
+[ 44.2%] Iter  431/976 | SPS:  3626 | Reward:  0.066 | Value:  0.668 | VLoss: 0.0341 | PLoss: -0.0071 | Ent: 0.4199
+[ 44.3%] Iter  432/976 | SPS:  3626 | Reward:  0.058 | Value:  0.696 | VLoss: 0.0326 | PLoss: -0.0041 | Ent: 0.3957
+[ 44.4%] Iter  433/976 | SPS:  3625 | Reward:  0.066 | Value:  0.609 | VLoss: 0.0469 | PLoss: 0.0013 | Ent: 0.4290
+[ 44.5%] Iter  434/976 | SPS:  3625 | Reward:  0.065 | Value:  0.610 | VLoss: 0.0392 | PLoss: -0.0076 | Ent: 0.4288
+[ 44.6%] Iter  435/976 | SPS:  3625 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0384 | PLoss: -0.0069 | Ent: 0.3977
+[ 44.7%] Iter  436/976 | SPS:  3625 | Reward:  0.066 | Value:  0.645 | VLoss: 0.0346 | PLoss: -0.0104 | Ent: 0.4094
+[ 44.8%] Iter  437/976 | SPS:  3625 | Reward:  0.063 | Value:  0.652 | VLoss: 0.0301 | PLoss: -0.0061 | Ent: 0.3676
+[ 44.9%] Iter  438/976 | SPS:  3624 | Reward:  0.066 | Value:  0.690 | VLoss: 0.0472 | PLoss: -0.0112 | Ent: 0.3495
+[ 45.0%] Iter  439/976 | SPS:  3624 | Reward:  0.056 | Value:  0.646 | VLoss: 0.0345 | PLoss: -0.0069 | Ent: 0.3537
+[ 45.1%] Iter  440/976 | SPS:  3624 | Reward:  0.060 | Value:  0.602 | VLoss: 0.0424 | PLoss: -0.0102 | Ent: 0.4224
+[ 45.2%] Iter  441/976 | SPS:  3624 | Reward:  0.054 | Value:  0.628 | VLoss: 0.0413 | PLoss: -0.0073 | Ent: 0.4057
+[ 45.3%] Iter  442/976 | SPS:  3624 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0393 | PLoss: -0.0062 | Ent: 0.3545
+[ 45.4%] Iter  443/976 | SPS:  3624 | Reward:  0.065 | Value:  0.671 | VLoss: 0.0277 | PLoss: -0.0027 | Ent: 0.3310
+[ 45.5%] Iter  444/976 | SPS:  3624 | Reward:  0.046 | Value:  0.649 | VLoss: 0.0376 | PLoss: -0.0133 | Ent: 0.3355
+[ 45.6%] Iter  445/976 | SPS:  3624 | Reward:  0.071 | Value:  0.592 | VLoss: 0.0414 | PLoss: -0.0062 | Ent: 0.3445
+[ 45.7%] Iter  446/976 | SPS:  3623 | Reward:  0.070 | Value:  0.669 | VLoss: 0.0424 | PLoss: -0.0205 | Ent: 0.4009
+[ 45.8%] Iter  447/976 | SPS:  3623 | Reward:  0.053 | Value:  0.629 | VLoss: 0.0357 | PLoss: -0.0065 | Ent: 0.3877
+[ 45.9%] Iter  448/976 | SPS:  3623 | Reward:  0.058 | Value:  0.638 | VLoss: 0.0330 | PLoss: -0.0084 | Ent: 0.3865
+[ 46.0%] Iter  449/976 | SPS:  3623 | Reward:  0.054 | Value:  0.648 | VLoss: 0.0307 | PLoss: -0.0057 | Ent: 0.3855
+[ 46.1%] Iter  450/976 | SPS:  3623 | Reward:  0.060 | Value:  0.629 | VLoss: 0.0410 | PLoss: -0.0134 | Ent: 0.3471
+[ 46.2%] Iter  451/976 | SPS:  3623 | Reward:  0.059 | Value:  0.610 | VLoss: 0.0344 | PLoss: -0.0041 | Ent: 0.3950
+[ 46.3%] Iter  452/976 | SPS:  3623 | Reward:  0.070 | Value:  0.607 | VLoss: 0.0355 | PLoss: -0.0119 | Ent: 0.3436
+[ 46.4%] Iter  453/976 | SPS:  3623 | Reward:  0.067 | Value:  0.684 | VLoss: 0.0371 | PLoss: -0.0047 | Ent: 0.3393
+[ 46.5%] Iter  454/976 | SPS:  3623 | Reward:  0.067 | Value:  0.729 | VLoss: 0.0349 | PLoss: -0.0114 | Ent: 0.3700
+[ 46.6%] Iter  455/976 | SPS:  3622 | Reward:  0.064 | Value:  0.673 | VLoss: 0.0406 | PLoss: -0.0085 | Ent: 0.4199
+[ 46.7%] Iter  456/976 | SPS:  3622 | Reward:  0.065 | Value:  0.633 | VLoss: 0.0368 | PLoss: -0.0066 | Ent: 0.2935
+[ 46.8%] Iter  457/976 | SPS:  3622 | Reward:  0.063 | Value:  0.647 | VLoss: 0.0499 | PLoss: -0.0117 | Ent: 0.3013
+[ 46.9%] Iter  458/976 | SPS:  3622 | Reward:  0.060 | Value:  0.660 | VLoss: 0.0248 | PLoss: -0.0005 | Ent: 0.3411
+[ 47.0%] Iter  459/976 | SPS:  3622 | Reward:  0.067 | Value:  0.639 | VLoss: 0.0407 | PLoss: -0.0103 | Ent: 0.3917
+[ 47.1%] Iter  460/976 | SPS:  3622 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0407 | PLoss: -0.0096 | Ent: 0.3791
+[ 47.2%] Iter  461/976 | SPS:  3622 | Reward:  0.059 | Value:  0.650 | VLoss: 0.0319 | PLoss: -0.0101 | Ent: 0.3568
+[ 47.3%] Iter  462/976 | SPS:  3622 | Reward:  0.072 | Value:  0.693 | VLoss: 0.0340 | PLoss: -0.0114 | Ent: 0.3468
+[ 47.4%] Iter  463/976 | SPS:  3621 | Reward:  0.073 | Value:  0.712 | VLoss: 0.0431 | PLoss: -0.0056 | Ent: 0.3889
+[ 47.5%] Iter  464/976 | SPS:  3621 | Reward:  0.060 | Value:  0.702 | VLoss: 0.0355 | PLoss: -0.0193 | Ent: 0.3503
+[ 47.6%] Iter  465/976 | SPS:  3621 | Reward:  0.071 | Value:  0.669 | VLoss: 0.0375 | PLoss: -0.0121 | Ent: 0.3758
+[ 47.7%] Iter  466/976 | SPS:  3620 | Reward:  0.060 | Value:  0.609 | VLoss: 0.0372 | PLoss: -0.0061 | Ent: 0.4005
+[ 47.8%] Iter  467/976 | SPS:  3620 | Reward:  0.062 | Value:  0.652 | VLoss: 0.0384 | PLoss: -0.0111 | Ent: 0.4008
+[ 48.0%] Iter  468/976 | SPS:  3620 | Reward:  0.052 | Value:  0.656 | VLoss: 0.0399 | PLoss: -0.0088 | Ent: 0.3110
+[ 48.1%] Iter  469/976 | SPS:  3620 | Reward:  0.073 | Value:  0.613 | VLoss: 0.0378 | PLoss: -0.0007 | Ent: 0.2954
+[ 48.2%] Iter  470/976 | SPS:  3620 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0484 | PLoss: -0.0015 | Ent: 0.3658
+[ 48.3%] Iter  471/976 | SPS:  3620 | Reward:  0.061 | Value:  0.721 | VLoss: 0.0257 | PLoss: -0.0046 | Ent: 0.3027
+[ 48.4%] Iter  472/976 | SPS:  3620 | Reward:  0.064 | Value:  0.685 | VLoss: 0.0262 | PLoss: -0.0103 | Ent: 0.3979
+[ 48.5%] Iter  473/976 | SPS:  3619 | Reward:  0.068 | Value:  0.703 | VLoss: 0.0434 | PLoss: -0.0088 | Ent: 0.3253
+[ 48.6%] Iter  474/976 | SPS:  3619 | Reward:  0.055 | Value:  0.657 | VLoss: 0.0283 | PLoss: -0.0092 | Ent: 0.3111
+[ 48.7%] Iter  475/976 | SPS:  3620 | Reward:  0.053 | Value:  0.604 | VLoss: 0.0292 | PLoss: -0.0109 | Ent: 0.3804
+[ 48.8%] Iter  476/976 | SPS:  3619 | Reward:  0.068 | Value:  0.655 | VLoss: 0.0451 | PLoss: -0.0097 | Ent: 0.3561
+[ 48.9%] Iter  477/976 | SPS:  3618 | Reward:  0.069 | Value:  0.605 | VLoss: 0.0446 | PLoss: -0.0044 | Ent: 0.3503
+[ 49.0%] Iter  478/976 | SPS:  3619 | Reward:  0.051 | Value:  0.619 | VLoss: 0.0287 | PLoss: -0.0030 | Ent: 0.3362
+[ 49.1%] Iter  479/976 | SPS:  3619 | Reward:  0.051 | Value:  0.594 | VLoss: 0.0306 | PLoss: -0.0111 | Ent: 0.3316
+[ 49.2%] Iter  480/976 | SPS:  3619 | Reward:  0.048 | Value:  0.583 | VLoss: 0.0253 | PLoss: -0.0012 | Ent: 0.2884
+[ 49.3%] Iter  481/976 | SPS:  3619 | Reward:  0.047 | Value:  0.551 | VLoss: 0.0302 | PLoss: -0.0123 | Ent: 0.3175
+[ 49.4%] Iter  482/976 | SPS:  3619 | Reward:  0.064 | Value:  0.576 | VLoss: 0.0455 | PLoss: -0.0054 | Ent: 0.3449
+[ 49.5%] Iter  483/976 | SPS:  3619 | Reward:  0.072 | Value:  0.612 | VLoss: 0.0333 | PLoss: -0.0091 | Ent: 0.3620
+[ 49.6%] Iter  484/976 | SPS:  3619 | Reward:  0.045 | Value:  0.664 | VLoss: 0.0347 | PLoss: 0.0002 | Ent: 0.3616
+[ 49.7%] Iter  485/976 | SPS:  3619 | Reward:  0.067 | Value:  0.643 | VLoss: 0.0250 | PLoss: -0.0079 | Ent: 0.3447
+[ 49.8%] Iter  486/976 | SPS:  3619 | Reward:  0.055 | Value:  0.670 | VLoss: 0.0311 | PLoss: -0.0155 | Ent: 0.3552
+[ 49.9%] Iter  487/976 | SPS:  3619 | Reward:  0.073 | Value:  0.686 | VLoss: 0.0344 | PLoss: -0.0058 | Ent: 0.3514
+[ 50.0%] Iter  488/976 | SPS:  3619 | Reward:  0.075 | Value:  0.666 | VLoss: 0.0465 | PLoss: -0.0084 | Ent: 0.3535
+[ 50.1%] Iter  489/976 | SPS:  3618 | Reward:  0.059 | Value:  0.645 | VLoss: 0.0381 | PLoss: -0.0101 | Ent: 0.3877
+[ 50.2%] Iter  490/976 | SPS:  3619 | Reward:  0.056 | Value:  0.679 | VLoss: 0.0304 | PLoss: -0.0069 | Ent: 0.3773
+[ 50.3%] Iter  491/976 | SPS:  3619 | Reward:  0.047 | Value:  0.604 | VLoss: 0.0307 | PLoss: -0.0052 | Ent: 0.3836
+[ 50.4%] Iter  492/976 | SPS:  3618 | Reward:  0.064 | Value:  0.536 | VLoss: 0.0451 | PLoss: -0.0059 | Ent: 0.4045
+[ 50.5%] Iter  493/976 | SPS:  3618 | Reward:  0.053 | Value:  0.585 | VLoss: 0.0336 | PLoss: -0.0012 | Ent: 0.3346
+[ 50.6%] Iter  494/976 | SPS:  3618 | Reward:  0.050 | Value:  0.569 | VLoss: 0.0422 | PLoss: -0.0020 | Ent: 0.3945
+[ 50.7%] Iter  495/976 | SPS:  3618 | Reward:  0.066 | Value:  0.588 | VLoss: 0.0379 | PLoss: -0.0044 | Ent: 0.3420
+[ 50.8%] Iter  496/976 | SPS:  3618 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0312 | PLoss: 0.0008 | Ent: 0.3587
+[ 50.9%] Iter  497/976 | SPS:  3618 | Reward:  0.051 | Value:  0.687 | VLoss: 0.0370 | PLoss: -0.0133 | Ent: 0.3565
+[ 51.0%] Iter  498/976 | SPS:  3618 | Reward:  0.054 | Value:  0.638 | VLoss: 0.0233 | PLoss: 0.0078 | Ent: 0.3792
+[ 51.1%] Iter  499/976 | SPS:  3618 | Reward:  0.049 | Value:  0.623 | VLoss: 0.0358 | PLoss: -0.0096 | Ent: 0.4190
+[ 51.2%] Iter  500/976 | SPS:  3618 | Reward:  0.062 | Value:  0.663 | VLoss: 0.0356 | PLoss: -0.0017 | Ent: 0.3876
+[ 51.3%] Iter  501/976 | SPS:  3618 | Reward:  0.056 | Value:  0.652 | VLoss: 0.0311 | PLoss: -0.0137 | Ent: 0.3496
+[ 51.4%] Iter  502/976 | SPS:  3618 | Reward:  0.048 | Value:  0.633 | VLoss: 0.0299 | PLoss: -0.0093 | Ent: 0.3780
+[ 51.5%] Iter  503/976 | SPS:  3618 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0376 | PLoss: -0.0134 | Ent: 0.3672
+[ 51.6%] Iter  504/976 | SPS:  3619 | Reward:  0.044 | Value:  0.654 | VLoss: 0.0265 | PLoss: -0.0116 | Ent: 0.3524
+[ 51.7%] Iter  505/976 | SPS:  3619 | Reward:  0.055 | Value:  0.685 | VLoss: 0.0351 | PLoss: -0.0045 | Ent: 0.2966
+[ 51.8%] Iter  506/976 | SPS:  3619 | Reward:  0.041 | Value:  0.620 | VLoss: 0.0322 | PLoss: -0.0053 | Ent: 0.4129
+[ 51.9%] Iter  507/976 | SPS:  3619 | Reward:  0.062 | Value:  0.553 | VLoss: 0.0377 | PLoss: -0.0096 | Ent: 0.3874
+[ 52.0%] Iter  508/976 | SPS:  3619 | Reward:  0.049 | Value:  0.584 | VLoss: 0.0433 | PLoss: -0.0076 | Ent: 0.3596
+[ 52.2%] Iter  509/976 | SPS:  3619 | Reward:  0.053 | Value:  0.600 | VLoss: 0.0374 | PLoss: -0.0032 | Ent: 0.3456
+[ 52.3%] Iter  510/976 | SPS:  3619 | Reward:  0.061 | Value:  0.573 | VLoss: 0.0426 | PLoss: -0.0084 | Ent: 0.3389
+[ 52.4%] Iter  511/976 | SPS:  3619 | Reward:  0.059 | Value:  0.646 | VLoss: 0.0319 | PLoss: -0.0069 | Ent: 0.3011
+[ 52.5%] Iter  512/976 | SPS:  3618 | Reward:  0.077 | Value:  0.631 | VLoss: 0.0459 | PLoss: -0.0104 | Ent: 0.3520
+[ 52.6%] Iter  513/976 | SPS:  3618 | Reward:  0.053 | Value:  0.655 | VLoss: 0.0328 | PLoss: -0.0075 | Ent: 0.3742
+[ 52.7%] Iter  514/976 | SPS:  3618 | Reward:  0.061 | Value:  0.578 | VLoss: 0.0417 | PLoss: -0.0114 | Ent: 0.3475
+[ 52.8%] Iter  515/976 | SPS:  3618 | Reward:  0.066 | Value:  0.687 | VLoss: 0.0357 | PLoss: -0.0156 | Ent: 0.3288
+[ 52.9%] Iter  516/976 | SPS:  3618 | Reward:  0.051 | Value:  0.633 | VLoss: 0.0214 | PLoss: -0.0041 | Ent: 0.3753
+[ 53.0%] Iter  517/976 | SPS:  3618 | Reward:  0.057 | Value:  0.635 | VLoss: 0.0260 | PLoss: -0.0017 | Ent: 0.3200
+[ 53.1%] Iter  518/976 | SPS:  3618 | Reward:  0.054 | Value:  0.552 | VLoss: 0.0457 | PLoss: -0.0059 | Ent: 0.4430
+[ 53.2%] Iter  519/976 | SPS:  3618 | Reward:  0.059 | Value:  0.585 | VLoss: 0.0331 | PLoss: -0.0060 | Ent: 0.3969
+[ 53.3%] Iter  520/976 | SPS:  3618 | Reward:  0.063 | Value:  0.618 | VLoss: 0.0339 | PLoss: -0.0105 | Ent: 0.3697
+[ 53.4%] Iter  521/976 | SPS:  3618 | Reward:  0.047 | Value:  0.674 | VLoss: 0.0244 | PLoss: 0.0002 | Ent: 0.3407
+[ 53.5%] Iter  522/976 | SPS:  3618 | Reward:  0.055 | Value:  0.553 | VLoss: 0.0423 | PLoss: -0.0052 | Ent: 0.3278
+[ 53.6%] Iter  523/976 | SPS:  3619 | Reward:  0.051 | Value:  0.624 | VLoss: 0.0287 | PLoss: -0.0055 | Ent: 0.3341
+[ 53.7%] Iter  524/976 | SPS:  3619 | Reward:  0.056 | Value:  0.645 | VLoss: 0.0344 | PLoss: -0.0086 | Ent: 0.3692
+[ 53.8%] Iter  525/976 | SPS:  3618 | Reward:  0.072 | Value:  0.629 | VLoss: 0.0349 | PLoss: -0.0100 | Ent: 0.3234
+[ 53.9%] Iter  526/976 | SPS:  3618 | Reward:  0.065 | Value:  0.586 | VLoss: 0.0396 | PLoss: -0.0037 | Ent: 0.3960
+[ 54.0%] Iter  527/976 | SPS:  3618 | Reward:  0.057 | Value:  0.705 | VLoss: 0.0333 | PLoss: -0.0055 | Ent: 0.3165
+[ 54.1%] Iter  528/976 | SPS:  3618 | Reward:  0.055 | Value:  0.638 | VLoss: 0.0329 | PLoss: -0.0056 | Ent: 0.3709
+[ 54.2%] Iter  529/976 | SPS:  3618 | Reward:  0.065 | Value:  0.633 | VLoss: 0.0345 | PLoss: -0.0095 | Ent: 0.3610
+[ 54.3%] Iter  530/976 | SPS:  3619 | Reward:  0.042 | Value:  0.656 | VLoss: 0.0203 | PLoss: -0.0102 | Ent: 0.4454
+[ 54.4%] Iter  531/976 | SPS:  3619 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0331 | PLoss: -0.0052 | Ent: 0.3343
+[ 54.5%] Iter  532/976 | SPS:  3619 | Reward:  0.041 | Value:  0.679 | VLoss: 0.0134 | PLoss: -0.0078 | Ent: 0.3261
+[ 54.6%] Iter  533/976 | SPS:  3620 | Reward:  0.047 | Value:  0.587 | VLoss: 0.0432 | PLoss: -0.0159 | Ent: 0.3392
+[ 54.7%] Iter  534/976 | SPS:  3619 | Reward:  0.058 | Value:  0.565 | VLoss: 0.0434 | PLoss: -0.0066 | Ent: 0.3889
+[ 54.8%] Iter  535/976 | SPS:  3619 | Reward:  0.071 | Value:  0.666 | VLoss: 0.0365 | PLoss: -0.0066 | Ent: 0.3465
+[ 54.9%] Iter  536/976 | SPS:  3618 | Reward:  0.062 | Value:  0.595 | VLoss: 0.0375 | PLoss: -0.0103 | Ent: 0.4072
+[ 55.0%] Iter  537/976 | SPS:  3619 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0329 | PLoss: -0.0122 | Ent: 0.3558
+[ 55.1%] Iter  538/976 | SPS:  3618 | Reward:  0.063 | Value:  0.670 | VLoss: 0.0288 | PLoss: -0.0078 | Ent: 0.3406
+[ 55.2%] Iter  539/976 | SPS:  3618 | Reward:  0.051 | Value:  0.614 | VLoss: 0.0410 | PLoss: -0.0043 | Ent: 0.3828
+[ 55.3%] Iter  540/976 | SPS:  3618 | Reward:  0.062 | Value:  0.557 | VLoss: 0.0416 | PLoss: -0.0015 | Ent: 0.3439
+[ 55.4%] Iter  541/976 | SPS:  3618 | Reward:  0.059 | Value:  0.639 | VLoss: 0.0284 | PLoss: -0.0077 | Ent: 0.2802
+[ 55.5%] Iter  542/976 | SPS:  3619 | Reward:  0.060 | Value:  0.699 | VLoss: 0.0432 | PLoss: -0.0073 | Ent: 0.2883
+[ 55.6%] Iter  543/976 | SPS:  3619 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0282 | PLoss: -0.0101 | Ent: 0.3017
+[ 55.7%] Iter  544/976 | SPS:  3619 | Reward:  0.055 | Value:  0.684 | VLoss: 0.0355 | PLoss: -0.0049 | Ent: 0.2899
+[ 55.8%] Iter  545/976 | SPS:  3615 | Reward:  0.057 | Value:  0.643 | VLoss: 0.0348 | PLoss: -0.0096 | Ent: 0.3113
+[ 55.9%] Iter  546/976 | SPS:  3615 | Reward:  0.066 | Value:  0.666 | VLoss: 0.0252 | PLoss: -0.0186 | Ent: 0.4211
+[ 56.0%] Iter  547/976 | SPS:  3615 | Reward:  0.067 | Value:  0.673 | VLoss: 0.0473 | PLoss: -0.0093 | Ent: 0.3938
+[ 56.1%] Iter  548/976 | SPS:  3615 | Reward:  0.051 | Value:  0.641 | VLoss: 0.0310 | PLoss: -0.0076 | Ent: 0.3868
+[ 56.2%] Iter  549/976 | SPS:  3615 | Reward:  0.048 | Value:  0.681 | VLoss: 0.0283 | PLoss: -0.0083 | Ent: 0.2875
+[ 56.4%] Iter  550/976 | SPS:  3615 | Reward:  0.058 | Value:  0.654 | VLoss: 0.0377 | PLoss: -0.0132 | Ent: 0.3339
+[ 56.5%] Iter  551/976 | SPS:  3615 | Reward:  0.067 | Value:  0.617 | VLoss: 0.0381 | PLoss: -0.0093 | Ent: 0.3271
+[ 56.6%] Iter  552/976 | SPS:  3615 | Reward:  0.052 | Value:  0.650 | VLoss: 0.0276 | PLoss: -0.0106 | Ent: 0.3210
+[ 56.7%] Iter  553/976 | SPS:  3615 | Reward:  0.073 | Value:  0.664 | VLoss: 0.0369 | PLoss: -0.0088 | Ent: 0.3243
+[ 56.8%] Iter  554/976 | SPS:  3614 | Reward:  0.063 | Value:  0.731 | VLoss: 0.0304 | PLoss: -0.0091 | Ent: 0.3413
+[ 56.9%] Iter  555/976 | SPS:  3614 | Reward:  0.074 | Value:  0.667 | VLoss: 0.0327 | PLoss: -0.0025 | Ent: 0.3196
+[ 57.0%] Iter  556/976 | SPS:  3614 | Reward:  0.062 | Value:  0.667 | VLoss: 0.0403 | PLoss: -0.0114 | Ent: 0.3478
+[ 57.1%] Iter  557/976 | SPS:  3614 | Reward:  0.051 | Value:  0.676 | VLoss: 0.0372 | PLoss: -0.0088 | Ent: 0.3990
+[ 57.2%] Iter  558/976 | SPS:  3615 | Reward:  0.045 | Value:  0.606 | VLoss: 0.0383 | PLoss: -0.0042 | Ent: 0.3827
+[ 57.3%] Iter  559/976 | SPS:  3614 | Reward:  0.067 | Value:  0.578 | VLoss: 0.0422 | PLoss: -0.0045 | Ent: 0.3483
+[ 57.4%] Iter  560/976 | SPS:  3614 | Reward:  0.073 | Value:  0.621 | VLoss: 0.0449 | PLoss: -0.0105 | Ent: 0.3003
+[ 57.5%] Iter  561/976 | SPS:  3614 | Reward:  0.073 | Value:  0.727 | VLoss: 0.0293 | PLoss: 0.0016 | Ent: 0.3314
+[ 57.6%] Iter  562/976 | SPS:  3614 | Reward:  0.053 | Value:  0.734 | VLoss: 0.0332 | PLoss: -0.0094 | Ent: 0.3253
+[ 57.7%] Iter  563/976 | SPS:  3614 | Reward:  0.053 | Value:  0.602 | VLoss: 0.0431 | PLoss: -0.0083 | Ent: 0.3648
+[ 57.8%] Iter  564/976 | SPS:  3614 | Reward:  0.069 | Value:  0.568 | VLoss: 0.0369 | PLoss: -0.0085 | Ent: 0.3229
+[ 57.9%] Iter  565/976 | SPS:  3614 | Reward:  0.073 | Value:  0.607 | VLoss: 0.0358 | PLoss: -0.0020 | Ent: 0.3024
+[ 58.0%] Iter  566/976 | SPS:  3613 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0393 | PLoss: -0.0071 | Ent: 0.3268
+[ 58.1%] Iter  567/976 | SPS:  3613 | Reward:  0.059 | Value:  0.606 | VLoss: 0.0388 | PLoss: -0.0101 | Ent: 0.4424
+[ 58.2%] Iter  568/976 | SPS:  3613 | Reward:  0.060 | Value:  0.606 | VLoss: 0.0343 | PLoss: -0.0068 | Ent: 0.3574
+[ 58.3%] Iter  569/976 | SPS:  3613 | Reward:  0.068 | Value:  0.600 | VLoss: 0.0363 | PLoss: -0.0080 | Ent: 0.3446
+[ 58.4%] Iter  570/976 | SPS:  3612 | Reward:  0.075 | Value:  0.643 | VLoss: 0.0447 | PLoss: -0.0098 | Ent: 0.3417
+[ 58.5%] Iter  571/976 | SPS:  3612 | Reward:  0.054 | Value:  0.621 | VLoss: 0.0305 | PLoss: -0.0049 | Ent: 0.3462
+[ 58.6%] Iter  572/976 | SPS:  3612 | Reward:  0.056 | Value:  0.604 | VLoss: 0.0323 | PLoss: -0.0068 | Ent: 0.3362
+[ 58.7%] Iter  573/976 | SPS:  3612 | Reward:  0.065 | Value:  0.613 | VLoss: 0.0459 | PLoss: -0.0046 | Ent: 0.2648
+[ 58.8%] Iter  574/976 | SPS:  3612 | Reward:  0.063 | Value:  0.623 | VLoss: 0.0370 | PLoss: -0.0021 | Ent: 0.2884
+[ 58.9%] Iter  575/976 | SPS:  3611 | Reward:  0.073 | Value:  0.647 | VLoss: 0.0339 | PLoss: -0.0082 | Ent: 0.3135
+[ 59.0%] Iter  576/976 | SPS:  3612 | Reward:  0.047 | Value:  0.681 | VLoss: 0.0343 | PLoss: -0.0114 | Ent: 0.3087
+[ 59.1%] Iter  577/976 | SPS:  3611 | Reward:  0.077 | Value:  0.665 | VLoss: 0.0311 | PLoss: -0.0055 | Ent: 0.3609
+[ 59.2%] Iter  578/976 | SPS:  3610 | Reward:  0.068 | Value:  0.614 | VLoss: 0.0443 | PLoss: -0.0055 | Ent: 0.3571
+[ 59.3%] Iter  579/976 | SPS:  3610 | Reward:  0.075 | Value:  0.596 | VLoss: 0.0452 | PLoss: -0.0048 | Ent: 0.3144
+[ 59.4%] Iter  580/976 | SPS:  3610 | Reward:  0.066 | Value:  0.641 | VLoss: 0.0366 | PLoss: -0.0107 | Ent: 0.2698
+[ 59.5%] Iter  581/976 | SPS:  3610 | Reward:  0.053 | Value:  0.670 | VLoss: 0.0370 | PLoss: -0.0027 | Ent: 0.3141
+[ 59.6%] Iter  582/976 | SPS:  3610 | Reward:  0.057 | Value:  0.566 | VLoss: 0.0404 | PLoss: 0.0023 | Ent: 0.2613
+[ 59.7%] Iter  583/976 | SPS:  3610 | Reward:  0.064 | Value:  0.634 | VLoss: 0.0272 | PLoss: -0.0060 | Ent: 0.2824
+[ 59.8%] Iter  584/976 | SPS:  3610 | Reward:  0.066 | Value:  0.724 | VLoss: 0.0365 | PLoss: -0.0058 | Ent: 0.2989
+[ 59.9%] Iter  585/976 | SPS:  3610 | Reward:  0.056 | Value:  0.738 | VLoss: 0.0324 | PLoss: -0.0071 | Ent: 0.3069
+[ 60.0%] Iter  586/976 | SPS:  3610 | Reward:  0.046 | Value:  0.695 | VLoss: 0.0307 | PLoss: -0.0036 | Ent: 0.2630
+[ 60.1%] Iter  587/976 | SPS:  3610 | Reward:  0.055 | Value:  0.603 | VLoss: 0.0485 | PLoss: -0.0064 | Ent: 0.3268
+[ 60.2%] Iter  588/976 | SPS:  3610 | Reward:  0.061 | Value:  0.567 | VLoss: 0.0366 | PLoss: -0.0020 | Ent: 0.3023
+[ 60.3%] Iter  589/976 | SPS:  3610 | Reward:  0.067 | Value:  0.632 | VLoss: 0.0253 | PLoss: -0.0052 | Ent: 0.2420
+[ 60.5%] Iter  590/976 | SPS:  3610 | Reward:  0.058 | Value:  0.702 | VLoss: 0.0385 | PLoss: -0.0078 | Ent: 0.3150
+[ 60.6%] Iter  591/976 | SPS:  3610 | Reward:  0.064 | Value:  0.663 | VLoss: 0.0296 | PLoss: -0.0102 | Ent: 0.3347
+[ 60.7%] Iter  592/976 | SPS:  3610 | Reward:  0.053 | Value:  0.644 | VLoss: 0.0309 | PLoss: -0.0109 | Ent: 0.3389
+[ 60.8%] Iter  593/976 | SPS:  3610 | Reward:  0.069 | Value:  0.647 | VLoss: 0.0468 | PLoss: -0.0084 | Ent: 0.3121
+[ 60.9%] Iter  594/976 | SPS:  3610 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0318 | PLoss: -0.0100 | Ent: 0.3011
+[ 61.0%] Iter  595/976 | SPS:  3610 | Reward:  0.062 | Value:  0.638 | VLoss: 0.0459 | PLoss: -0.0113 | Ent: 0.3752
+[ 61.1%] Iter  596/976 | SPS:  3609 | Reward:  0.057 | Value:  0.558 | VLoss: 0.0442 | PLoss: -0.0025 | Ent: 0.3291
+[ 61.2%] Iter  597/976 | SPS:  3609 | Reward:  0.048 | Value:  0.580 | VLoss: 0.0228 | PLoss: -0.0039 | Ent: 0.2742
+[ 61.3%] Iter  598/976 | SPS:  3609 | Reward:  0.051 | Value:  0.547 | VLoss: 0.0335 | PLoss: -0.0083 | Ent: 0.3789
+[ 61.4%] Iter  599/976 | SPS:  3609 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0369 | PLoss: -0.0090 | Ent: 0.3125
+[ 61.5%] Iter  600/976 | SPS:  3609 | Reward:  0.066 | Value:  0.657 | VLoss: 0.0281 | PLoss: -0.0052 | Ent: 0.3268
+[ 61.6%] Iter  601/976 | SPS:  3609 | Reward:  0.060 | Value:  0.660 | VLoss: 0.0269 | PLoss: -0.0043 | Ent: 0.2838
+[ 61.7%] Iter  602/976 | SPS:  3609 | Reward:  0.060 | Value:  0.684 | VLoss: 0.0462 | PLoss: -0.0037 | Ent: 0.2993
+[ 61.8%] Iter  603/976 | SPS:  3609 | Reward:  0.057 | Value:  0.630 | VLoss: 0.0343 | PLoss: -0.0060 | Ent: 0.2934
+[ 61.9%] Iter  604/976 | SPS:  3609 | Reward:  0.067 | Value:  0.641 | VLoss: 0.0384 | PLoss: -0.0073 | Ent: 0.2988
+[ 62.0%] Iter  605/976 | SPS:  3608 | Reward:  0.066 | Value:  0.651 | VLoss: 0.0349 | PLoss: -0.0064 | Ent: 0.2626
+[ 62.1%] Iter  606/976 | SPS:  3608 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0329 | PLoss: -0.0039 | Ent: 0.3037
+[ 62.2%] Iter  607/976 | SPS:  3608 | Reward:  0.064 | Value:  0.656 | VLoss: 0.0336 | PLoss: -0.0064 | Ent: 0.3463
+[ 62.3%] Iter  608/976 | SPS:  3608 | Reward:  0.063 | Value:  0.640 | VLoss: 0.0394 | PLoss: -0.0045 | Ent: 0.3108
+[ 62.4%] Iter  609/976 | SPS:  3608 | Reward:  0.061 | Value:  0.661 | VLoss: 0.0370 | PLoss: -0.0070 | Ent: 0.3442
+[ 62.5%] Iter  610/976 | SPS:  3608 | Reward:  0.064 | Value:  0.641 | VLoss: 0.0385 | PLoss: -0.0029 | Ent: 0.2811
+[ 62.6%] Iter  611/976 | SPS:  3608 | Reward:  0.075 | Value:  0.631 | VLoss: 0.0412 | PLoss: -0.0054 | Ent: 0.2639
+[ 62.7%] Iter  612/976 | SPS:  3607 | Reward:  0.065 | Value:  0.664 | VLoss: 0.0368 | PLoss: -0.0031 | Ent: 0.3470
+[ 62.8%] Iter  613/976 | SPS:  3607 | Reward:  0.057 | Value:  0.653 | VLoss: 0.0327 | PLoss: -0.0058 | Ent: 0.3190
+[ 62.9%] Iter  614/976 | SPS:  3607 | Reward:  0.067 | Value:  0.589 | VLoss: 0.0487 | PLoss: -0.0112 | Ent: 0.2868
+[ 63.0%] Iter  615/976 | SPS:  3607 | Reward:  0.080 | Value:  0.626 | VLoss: 0.0362 | PLoss: -0.0104 | Ent: 0.3187
+[ 63.1%] Iter  616/976 | SPS:  3607 | Reward:  0.057 | Value:  0.692 | VLoss: 0.0422 | PLoss: -0.0031 | Ent: 0.3337
+[ 63.2%] Iter  617/976 | SPS:  3607 | Reward:  0.063 | Value:  0.592 | VLoss: 0.0417 | PLoss: -0.0100 | Ent: 0.3392
+[ 63.3%] Iter  618/976 | SPS:  3607 | Reward:  0.069 | Value:  0.666 | VLoss: 0.0301 | PLoss: -0.0096 | Ent: 0.3002
+[ 63.4%] Iter  619/976 | SPS:  3606 | Reward:  0.067 | Value:  0.704 | VLoss: 0.0371 | PLoss: -0.0055 | Ent: 0.3060
+[ 63.5%] Iter  620/976 | SPS:  3606 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0287 | PLoss: -0.0037 | Ent: 0.3141
+[ 63.6%] Iter  621/976 | SPS:  3605 | Reward:  0.081 | Value:  0.616 | VLoss: 0.0393 | PLoss: -0.0034 | Ent: 0.3422
+[ 63.7%] Iter  622/976 | SPS:  3605 | Reward:  0.060 | Value:  0.696 | VLoss: 0.0424 | PLoss: -0.0094 | Ent: 0.3335
+[ 63.8%] Iter  623/976 | SPS:  3605 | Reward:  0.062 | Value:  0.660 | VLoss: 0.0327 | PLoss: -0.0027 | Ent: 0.3926
+[ 63.9%] Iter  624/976 | SPS:  3605 | Reward:  0.062 | Value:  0.636 | VLoss: 0.0392 | PLoss: -0.0016 | Ent: 0.3228
+[ 64.0%] Iter  625/976 | SPS:  3605 | Reward:  0.063 | Value:  0.589 | VLoss: 0.0481 | PLoss: -0.0099 | Ent: 0.3303
+[ 64.1%] Iter  626/976 | SPS:  3604 | Reward:  0.070 | Value:  0.619 | VLoss: 0.0388 | PLoss: -0.0067 | Ent: 0.3263
+[ 64.2%] Iter  627/976 | SPS:  3604 | Reward:  0.054 | Value:  0.613 | VLoss: 0.0315 | PLoss: -0.0053 | Ent: 0.2825
+[ 64.3%] Iter  628/976 | SPS:  3604 | Reward:  0.068 | Value:  0.680 | VLoss: 0.0474 | PLoss: -0.0061 | Ent: 0.3281
+[ 64.4%] Iter  629/976 | SPS:  3604 | Reward:  0.057 | Value:  0.636 | VLoss: 0.0375 | PLoss: -0.0019 | Ent: 0.3380
+[ 64.5%] Iter  630/976 | SPS:  3604 | Reward:  0.059 | Value:  0.621 | VLoss: 0.0282 | PLoss: 0.0002 | Ent: 0.3676
+[ 64.7%] Iter  631/976 | SPS:  3604 | Reward:  0.063 | Value:  0.660 | VLoss: 0.0401 | PLoss: -0.0083 | Ent: 0.3644
+[ 64.8%] Iter  632/976 | SPS:  3604 | Reward:  0.063 | Value:  0.626 | VLoss: 0.0391 | PLoss: -0.0087 | Ent: 0.3444
+[ 64.9%] Iter  633/976 | SPS:  3603 | Reward:  0.066 | Value:  0.609 | VLoss: 0.0449 | PLoss: -0.0016 | Ent: 0.3802
+[ 65.0%] Iter  634/976 | SPS:  3603 | Reward:  0.065 | Value:  0.648 | VLoss: 0.0412 | PLoss: -0.0052 | Ent: 0.3051
+[ 65.1%] Iter  635/976 | SPS:  3603 | Reward:  0.071 | Value:  0.617 | VLoss: 0.0460 | PLoss: -0.0038 | Ent: 0.3165
+[ 65.2%] Iter  636/976 | SPS:  3603 | Reward:  0.056 | Value:  0.641 | VLoss: 0.0390 | PLoss: -0.0019 | Ent: 0.3372
+[ 65.3%] Iter  637/976 | SPS:  3603 | Reward:  0.047 | Value:  0.622 | VLoss: 0.0368 | PLoss: -0.0044 | Ent: 0.3489
+[ 65.4%] Iter  638/976 | SPS:  3603 | Reward:  0.067 | Value:  0.613 | VLoss: 0.0431 | PLoss: -0.0034 | Ent: 0.3261
+[ 65.5%] Iter  639/976 | SPS:  3603 | Reward:  0.067 | Value:  0.593 | VLoss: 0.0381 | PLoss: -0.0061 | Ent: 0.3109
+[ 65.6%] Iter  640/976 | SPS:  3603 | Reward:  0.053 | Value:  0.658 | VLoss: 0.0322 | PLoss: -0.0086 | Ent: 0.3814
+[ 65.7%] Iter  641/976 | SPS:  3603 | Reward:  0.066 | Value:  0.701 | VLoss: 0.0272 | PLoss: -0.0071 | Ent: 0.3195
+[ 65.8%] Iter  642/976 | SPS:  3603 | Reward:  0.056 | Value:  0.716 | VLoss: 0.0259 | PLoss: -0.0129 | Ent: 0.4201
+[ 65.9%] Iter  643/976 | SPS:  3603 | Reward:  0.069 | Value:  0.643 | VLoss: 0.0402 | PLoss: -0.0023 | Ent: 0.2941
+[ 66.0%] Iter  644/976 | SPS:  3602 | Reward:  0.064 | Value:  0.638 | VLoss: 0.0385 | PLoss: -0.0037 | Ent: 0.2966
+[ 66.1%] Iter  645/976 | SPS:  3602 | Reward:  0.072 | Value:  0.631 | VLoss: 0.0336 | PLoss: -0.0066 | Ent: 0.3272
+[ 66.2%] Iter  646/976 | SPS:  3602 | Reward:  0.059 | Value:  0.694 | VLoss: 0.0358 | PLoss: -0.0077 | Ent: 0.3122
+[ 66.3%] Iter  647/976 | SPS:  3602 | Reward:  0.073 | Value:  0.594 | VLoss: 0.0346 | PLoss: 0.0007 | Ent: 0.3559
+[ 66.4%] Iter  648/976 | SPS:  3602 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0305 | PLoss: -0.0139 | Ent: 0.3536
+[ 66.5%] Iter  649/976 | SPS:  3602 | Reward:  0.061 | Value:  0.624 | VLoss: 0.0436 | PLoss: -0.0086 | Ent: 0.3257
+[ 66.6%] Iter  650/976 | SPS:  3602 | Reward:  0.071 | Value:  0.656 | VLoss: 0.0408 | PLoss: -0.0103 | Ent: 0.3029
+[ 66.7%] Iter  651/976 | SPS:  3602 | Reward:  0.064 | Value:  0.637 | VLoss: 0.0351 | PLoss: -0.0057 | Ent: 0.3957
+[ 66.8%] Iter  652/976 | SPS:  3602 | Reward:  0.059 | Value:  0.621 | VLoss: 0.0426 | PLoss: -0.0163 | Ent: 0.3621
+[ 66.9%] Iter  653/976 | SPS:  3601 | Reward:  0.078 | Value:  0.589 | VLoss: 0.0430 | PLoss: -0.0040 | Ent: 0.3498
+[ 67.0%] Iter  654/976 | SPS:  3601 | Reward:  0.066 | Value:  0.630 | VLoss: 0.0375 | PLoss: 0.0004 | Ent: 0.3359
+[ 67.1%] Iter  655/976 | SPS:  3601 | Reward:  0.065 | Value:  0.607 | VLoss: 0.0397 | PLoss: -0.0074 | Ent: 0.3802
+[ 67.2%] Iter  656/976 | SPS:  3601 | Reward:  0.067 | Value:  0.671 | VLoss: 0.0473 | PLoss: -0.0002 | Ent: 0.3064
+[ 67.3%] Iter  657/976 | SPS:  3600 | Reward:  0.058 | Value:  0.605 | VLoss: 0.0448 | PLoss: -0.0043 | Ent: 0.3031
+[ 67.4%] Iter  658/976 | SPS:  3600 | Reward:  0.056 | Value:  0.584 | VLoss: 0.0335 | PLoss: -0.0081 | Ent: 0.3350
+[ 67.5%] Iter  659/976 | SPS:  3600 | Reward:  0.058 | Value:  0.590 | VLoss: 0.0478 | PLoss: -0.0117 | Ent: 0.3243
+[ 67.6%] Iter  660/976 | SPS:  3600 | Reward:  0.069 | Value:  0.562 | VLoss: 0.0443 | PLoss: -0.0068 | Ent: 0.3011
+[ 67.7%] Iter  661/976 | SPS:  3600 | Reward:  0.088 | Value:  0.611 | VLoss: 0.0400 | PLoss: -0.0102 | Ent: 0.3318
+[ 67.8%] Iter  662/976 | SPS:  3600 | Reward:  0.063 | Value:  0.701 | VLoss: 0.0454 | PLoss: -0.0032 | Ent: 0.3476
+[ 67.9%] Iter  663/976 | SPS:  3599 | Reward:  0.064 | Value:  0.628 | VLoss: 0.0489 | PLoss: -0.0092 | Ent: 0.3404
+[ 68.0%] Iter  664/976 | SPS:  3599 | Reward:  0.062 | Value:  0.624 | VLoss: 0.0378 | PLoss: -0.0050 | Ent: 0.2799
+[ 68.1%] Iter  665/976 | SPS:  3599 | Reward:  0.053 | Value:  0.681 | VLoss: 0.0350 | PLoss: -0.0093 | Ent: 0.3152
+[ 68.2%] Iter  666/976 | SPS:  3599 | Reward:  0.063 | Value:  0.600 | VLoss: 0.0340 | PLoss: -0.0117 | Ent: 0.3086
+[ 68.3%] Iter  667/976 | SPS:  3599 | Reward:  0.068 | Value:  0.642 | VLoss: 0.0407 | PLoss: -0.0057 | Ent: 0.3140
+[ 68.4%] Iter  668/976 | SPS:  3599 | Reward:  0.060 | Value:  0.742 | VLoss: 0.0306 | PLoss: -0.0000 | Ent: 0.3816
+[ 68.5%] Iter  669/976 | SPS:  3599 | Reward:  0.049 | Value:  0.651 | VLoss: 0.0390 | PLoss: 0.0012 | Ent: 0.3357
+[ 68.6%] Iter  670/976 | SPS:  3599 | Reward:  0.067 | Value:  0.649 | VLoss: 0.0308 | PLoss: -0.0024 | Ent: 0.3139
+[ 68.8%] Iter  671/976 | SPS:  3599 | Reward:  0.071 | Value:  0.618 | VLoss: 0.0500 | PLoss: -0.0026 | Ent: 0.2839
+[ 68.9%] Iter  672/976 | SPS:  3598 | Reward:  0.062 | Value:  0.620 | VLoss: 0.0376 | PLoss: -0.0037 | Ent: 0.3239
+[ 69.0%] Iter  673/976 | SPS:  3598 | Reward:  0.066 | Value:  0.658 | VLoss: 0.0383 | PLoss: -0.0053 | Ent: 0.2977
+[ 69.1%] Iter  674/976 | SPS:  3598 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0430 | PLoss: -0.0063 | Ent: 0.2983
+[ 69.2%] Iter  675/976 | SPS:  3598 | Reward:  0.074 | Value:  0.696 | VLoss: 0.0405 | PLoss: -0.0017 | Ent: 0.3180
+[ 69.3%] Iter  676/976 | SPS:  3598 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0418 | PLoss: -0.0083 | Ent: 0.2912
+[ 69.4%] Iter  677/976 | SPS:  3598 | Reward:  0.062 | Value:  0.626 | VLoss: 0.0312 | PLoss: -0.0059 | Ent: 0.3058
+[ 69.5%] Iter  678/976 | SPS:  3598 | Reward:  0.065 | Value:  0.653 | VLoss: 0.0404 | PLoss: -0.0053 | Ent: 0.3735
+[ 69.6%] Iter  679/976 | SPS:  3598 | Reward:  0.056 | Value:  0.603 | VLoss: 0.0359 | PLoss: -0.0101 | Ent: 0.3711
+[ 69.7%] Iter  680/976 | SPS:  3598 | Reward:  0.076 | Value:  0.683 | VLoss: 0.0404 | PLoss: -0.0021 | Ent: 0.3075
+[ 69.8%] Iter  681/976 | SPS:  3597 | Reward:  0.064 | Value:  0.609 | VLoss: 0.0403 | PLoss: 0.0026 | Ent: 0.3010
+[ 69.9%] Iter  682/976 | SPS:  3597 | Reward:  0.074 | Value:  0.651 | VLoss: 0.0418 | PLoss: -0.0041 | Ent: 0.3092
+[ 70.0%] Iter  683/976 | SPS:  3597 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0329 | PLoss: -0.0064 | Ent: 0.3457
+[ 70.1%] Iter  684/976 | SPS:  3597 | Reward:  0.060 | Value:  0.684 | VLoss: 0.0402 | PLoss: -0.0068 | Ent: 0.2778
+[ 70.2%] Iter  685/976 | SPS:  3597 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0312 | PLoss: -0.0016 | Ent: 0.3124
+[ 70.3%] Iter  686/976 | SPS:  3597 | Reward:  0.066 | Value:  0.657 | VLoss: 0.0306 | PLoss: -0.0118 | Ent: 0.2848
+[ 70.4%] Iter  687/976 | SPS:  3597 | Reward:  0.058 | Value:  0.625 | VLoss: 0.0387 | PLoss: -0.0042 | Ent: 0.3187
+[ 70.5%] Iter  688/976 | SPS:  3597 | Reward:  0.064 | Value:  0.656 | VLoss: 0.0335 | PLoss: -0.0048 | Ent: 0.3423
+[ 70.6%] Iter  689/976 | SPS:  3597 | Reward:  0.054 | Value:  0.645 | VLoss: 0.0377 | PLoss: -0.0075 | Ent: 0.3128
+[ 70.7%] Iter  690/976 | SPS:  3597 | Reward:  0.055 | Value:  0.629 | VLoss: 0.0319 | PLoss: -0.0012 | Ent: 0.3071
+[ 70.8%] Iter  691/976 | SPS:  3596 | Reward:  0.076 | Value:  0.634 | VLoss: 0.0436 | PLoss: 0.0033 | Ent: 0.3092
+[ 70.9%] Iter  692/976 | SPS:  3596 | Reward:  0.069 | Value:  0.727 | VLoss: 0.0359 | PLoss: -0.0032 | Ent: 0.2704
+[ 71.0%] Iter  693/976 | SPS:  3596 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0344 | PLoss: -0.0107 | Ent: 0.3050
+[ 71.1%] Iter  694/976 | SPS:  3596 | Reward:  0.061 | Value:  0.634 | VLoss: 0.0316 | PLoss: -0.0034 | Ent: 0.2510
+[ 71.2%] Iter  695/976 | SPS:  3596 | Reward:  0.067 | Value:  0.651 | VLoss: 0.0371 | PLoss: -0.0023 | Ent: 0.2819
+[ 71.3%] Iter  696/976 | SPS:  3596 | Reward:  0.062 | Value:  0.633 | VLoss: 0.0471 | PLoss: -0.0035 | Ent: 0.3626
+[ 71.4%] Iter  697/976 | SPS:  3596 | Reward:  0.072 | Value:  0.605 | VLoss: 0.0401 | PLoss: -0.0057 | Ent: 0.3019
+[ 71.5%] Iter  698/976 | SPS:  3596 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0439 | PLoss: -0.0080 | Ent: 0.3553
+[ 71.6%] Iter  699/976 | SPS:  3596 | Reward:  0.061 | Value:  0.554 | VLoss: 0.0405 | PLoss: -0.0042 | Ent: 0.3096
+[ 71.7%] Iter  700/976 | SPS:  3595 | Reward:  0.062 | Value:  0.619 | VLoss: 0.0422 | PLoss: -0.0044 | Ent: 0.3103
+[ 71.8%] Iter  701/976 | SPS:  3595 | Reward:  0.063 | Value:  0.698 | VLoss: 0.0298 | PLoss: -0.0065 | Ent: 0.2813
+[ 71.9%] Iter  702/976 | SPS:  3595 | Reward:  0.074 | Value:  0.638 | VLoss: 0.0351 | PLoss: -0.0075 | Ent: 0.2966
+[ 72.0%] Iter  703/976 | SPS:  3595 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0335 | PLoss: -0.0035 | Ent: 0.3717
+[ 72.1%] Iter  704/976 | SPS:  3595 | Reward:  0.057 | Value:  0.645 | VLoss: 0.0411 | PLoss: -0.0035 | Ent: 0.2961
+[ 72.2%] Iter  705/976 | SPS:  3595 | Reward:  0.058 | Value:  0.635 | VLoss: 0.0314 | PLoss: -0.0056 | Ent: 0.3056
+[ 72.3%] Iter  706/976 | SPS:  3595 | Reward:  0.071 | Value:  0.669 | VLoss: 0.0409 | PLoss: -0.0054 | Ent: 0.3195
+[ 72.4%] Iter  707/976 | SPS:  3595 | Reward:  0.067 | Value:  0.602 | VLoss: 0.0395 | PLoss: -0.0090 | Ent: 0.3215
+[ 72.5%] Iter  708/976 | SPS:  3595 | Reward:  0.045 | Value:  0.647 | VLoss: 0.0351 | PLoss: -0.0104 | Ent: 0.3134
+[ 72.6%] Iter  709/976 | SPS:  3595 | Reward:  0.056 | Value:  0.560 | VLoss: 0.0427 | PLoss: -0.0028 | Ent: 0.2865
+[ 72.7%] Iter  710/976 | SPS:  3595 | Reward:  0.058 | Value:  0.611 | VLoss: 0.0283 | PLoss: -0.0011 | Ent: 0.3281
+[ 72.8%] Iter  711/976 | SPS:  3594 | Reward:  0.073 | Value:  0.602 | VLoss: 0.0432 | PLoss: -0.0091 | Ent: 0.2644
+[ 73.0%] Iter  712/976 | SPS:  3594 | Reward:  0.058 | Value:  0.638 | VLoss: 0.0294 | PLoss: -0.0022 | Ent: 0.3140
+[ 73.1%] Iter  713/976 | SPS:  3594 | Reward:  0.058 | Value:  0.661 | VLoss: 0.0340 | PLoss: -0.0010 | Ent: 0.3729
+[ 73.2%] Iter  714/976 | SPS:  3594 | Reward:  0.054 | Value:  0.628 | VLoss: 0.0371 | PLoss: -0.0018 | Ent: 0.2964
+[ 73.3%] Iter  715/976 | SPS:  3594 | Reward:  0.057 | Value:  0.583 | VLoss: 0.0357 | PLoss: -0.0040 | Ent: 0.3221
+[ 73.4%] Iter  716/976 | SPS:  3594 | Reward:  0.057 | Value:  0.607 | VLoss: 0.0405 | PLoss: -0.0057 | Ent: 0.2707
+[ 73.5%] Iter  717/976 | SPS:  3594 | Reward:  0.048 | Value:  0.641 | VLoss: 0.0271 | PLoss: -0.0018 | Ent: 0.2602
+[ 73.6%] Iter  718/976 | SPS:  3595 | Reward:  0.052 | Value:  0.667 | VLoss: 0.0293 | PLoss: -0.0056 | Ent: 0.2885
+[ 73.7%] Iter  719/976 | SPS:  3594 | Reward:  0.074 | Value:  0.610 | VLoss: 0.0463 | PLoss: -0.0009 | Ent: 0.2347
+[ 73.8%] Iter  720/976 | SPS:  3594 | Reward:  0.062 | Value:  0.638 | VLoss: 0.0383 | PLoss: -0.0100 | Ent: 0.3542
+[ 73.9%] Iter  721/976 | SPS:  3594 | Reward:  0.060 | Value:  0.661 | VLoss: 0.0429 | PLoss: -0.0053 | Ent: 0.2711
+[ 74.0%] Iter  722/976 | SPS:  3594 | Reward:  0.080 | Value:  0.674 | VLoss: 0.0392 | PLoss: -0.0007 | Ent: 0.2954
+[ 74.1%] Iter  723/976 | SPS:  3594 | Reward:  0.069 | Value:  0.654 | VLoss: 0.0409 | PLoss: -0.0046 | Ent: 0.2644
+[ 74.2%] Iter  724/976 | SPS:  3594 | Reward:  0.072 | Value:  0.635 | VLoss: 0.0421 | PLoss: -0.0090 | Ent: 0.3144
+[ 74.3%] Iter  725/976 | SPS:  3594 | Reward:  0.065 | Value:  0.640 | VLoss: 0.0361 | PLoss: 0.0007 | Ent: 0.3403
+[ 74.4%] Iter  726/976 | SPS:  3594 | Reward:  0.049 | Value:  0.623 | VLoss: 0.0344 | PLoss: -0.0127 | Ent: 0.4096
+[ 74.5%] Iter  727/976 | SPS:  3594 | Reward:  0.072 | Value:  0.688 | VLoss: 0.0382 | PLoss: -0.0035 | Ent: 0.2787
+[ 74.6%] Iter  728/976 | SPS:  3593 | Reward:  0.076 | Value:  0.679 | VLoss: 0.0431 | PLoss: -0.0085 | Ent: 0.2791
+[ 74.7%] Iter  729/976 | SPS:  3593 | Reward:  0.057 | Value:  0.662 | VLoss: 0.0226 | PLoss: -0.0013 | Ent: 0.3057
+[ 74.8%] Iter  730/976 | SPS:  3593 | Reward:  0.062 | Value:  0.616 | VLoss: 0.0468 | PLoss: -0.0096 | Ent: 0.3136
+[ 74.9%] Iter  731/976 | SPS:  3593 | Reward:  0.065 | Value:  0.611 | VLoss: 0.0345 | PLoss: -0.0076 | Ent: 0.3049
+[ 75.0%] Iter  732/976 | SPS:  3593 | Reward:  0.054 | Value:  0.678 | VLoss: 0.0342 | PLoss: -0.0045 | Ent: 0.3548
+[ 75.1%] Iter  733/976 | SPS:  3593 | Reward:  0.065 | Value:  0.543 | VLoss: 0.0343 | PLoss: -0.0036 | Ent: 0.3441
+[ 75.2%] Iter  734/976 | SPS:  3593 | Reward:  0.045 | Value:  0.539 | VLoss: 0.0310 | PLoss: -0.0045 | Ent: 0.4277
+[ 75.3%] Iter  735/976 | SPS:  3593 | Reward:  0.061 | Value:  0.571 | VLoss: 0.0403 | PLoss: -0.0086 | Ent: 0.3430
+[ 75.4%] Iter  736/976 | SPS:  3593 | Reward:  0.063 | Value:  0.677 | VLoss: 0.0371 | PLoss: -0.0037 | Ent: 0.3213
+[ 75.5%] Iter  737/976 | SPS:  3593 | Reward:  0.048 | Value:  0.690 | VLoss: 0.0314 | PLoss: -0.0015 | Ent: 0.2733
+[ 75.6%] Iter  738/976 | SPS:  3593 | Reward:  0.061 | Value:  0.618 | VLoss: 0.0430 | PLoss: -0.0080 | Ent: 0.3250
+[ 75.7%] Iter  739/976 | SPS:  3593 | Reward:  0.071 | Value:  0.646 | VLoss: 0.0371 | PLoss: -0.0094 | Ent: 0.2980
+[ 75.8%] Iter  740/976 | SPS:  3593 | Reward:  0.062 | Value:  0.651 | VLoss: 0.0295 | PLoss: -0.0125 | Ent: 0.3024
+[ 75.9%] Iter  741/976 | SPS:  3593 | Reward:  0.073 | Value:  0.640 | VLoss: 0.0318 | PLoss: -0.0052 | Ent: 0.3419
+[ 76.0%] Iter  742/976 | SPS:  3592 | Reward:  0.060 | Value:  0.694 | VLoss: 0.0388 | PLoss: -0.0023 | Ent: 0.2989
+[ 76.1%] Iter  743/976 | SPS:  3592 | Reward:  0.076 | Value:  0.642 | VLoss: 0.0414 | PLoss: -0.0072 | Ent: 0.3372
+[ 76.2%] Iter  744/976 | SPS:  3592 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0380 | PLoss: -0.0021 | Ent: 0.3233
+[ 76.3%] Iter  745/976 | SPS:  3592 | Reward:  0.069 | Value:  0.621 | VLoss: 0.0473 | PLoss: -0.0035 | Ent: 0.2739
+[ 76.4%] Iter  746/976 | SPS:  3592 | Reward:  0.085 | Value:  0.683 | VLoss: 0.0420 | PLoss: -0.0047 | Ent: 0.2761
+[ 76.5%] Iter  747/976 | SPS:  3591 | Reward:  0.081 | Value:  0.686 | VLoss: 0.0399 | PLoss: -0.0008 | Ent: 0.3185
+[ 76.6%] Iter  748/976 | SPS:  3591 | Reward:  0.063 | Value:  0.696 | VLoss: 0.0403 | PLoss: -0.0137 | Ent: 0.2945
+[ 76.7%] Iter  749/976 | SPS:  3591 | Reward:  0.060 | Value:  0.640 | VLoss: 0.0309 | PLoss: -0.0015 | Ent: 0.2655
+[ 76.8%] Iter  750/976 | SPS:  3591 | Reward:  0.057 | Value:  0.662 | VLoss: 0.0376 | PLoss: -0.0057 | Ent: 0.3317
+[ 76.9%] Iter  751/976 | SPS:  3591 | Reward:  0.060 | Value:  0.679 | VLoss: 0.0237 | PLoss: -0.0032 | Ent: 0.2937
+[ 77.0%] Iter  752/976 | SPS:  3591 | Reward:  0.066 | Value:  0.603 | VLoss: 0.0432 | PLoss: -0.0057 | Ent: 0.3890
+[ 77.2%] Iter  753/976 | SPS:  3591 | Reward:  0.069 | Value:  0.645 | VLoss: 0.0351 | PLoss: -0.0073 | Ent: 0.2963
+[ 77.3%] Iter  754/976 | SPS:  3591 | Reward:  0.053 | Value:  0.696 | VLoss: 0.0388 | PLoss: -0.0082 | Ent: 0.3142
+[ 77.4%] Iter  755/976 | SPS:  3591 | Reward:  0.054 | Value:  0.656 | VLoss: 0.0292 | PLoss: -0.0050 | Ent: 0.2877
+[ 77.5%] Iter  756/976 | SPS:  3591 | Reward:  0.053 | Value:  0.630 | VLoss: 0.0375 | PLoss: -0.0083 | Ent: 0.2921
+[ 77.6%] Iter  757/976 | SPS:  3591 | Reward:  0.069 | Value:  0.596 | VLoss: 0.0381 | PLoss: -0.0024 | Ent: 0.2988
+[ 77.7%] Iter  758/976 | SPS:  3591 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0447 | PLoss: -0.0056 | Ent: 0.3189
+[ 77.8%] Iter  759/976 | SPS:  3591 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0319 | PLoss: -0.0060 | Ent: 0.3106
+[ 77.9%] Iter  760/976 | SPS:  3591 | Reward:  0.051 | Value:  0.608 | VLoss: 0.0342 | PLoss: -0.0042 | Ent: 0.3696
+[ 78.0%] Iter  761/976 | SPS:  3591 | Reward:  0.052 | Value:  0.610 | VLoss: 0.0293 | PLoss: -0.0008 | Ent: 0.3363
+[ 78.1%] Iter  762/976 | SPS:  3591 | Reward:  0.060 | Value:  0.635 | VLoss: 0.0290 | PLoss: -0.0107 | Ent: 0.3126
+[ 78.2%] Iter  763/976 | SPS:  3591 | Reward:  0.060 | Value:  0.661 | VLoss: 0.0292 | PLoss: -0.0089 | Ent: 0.3166
+[ 78.3%] Iter  764/976 | SPS:  3591 | Reward:  0.057 | Value:  0.739 | VLoss: 0.0388 | PLoss: -0.0136 | Ent: 0.2930
+[ 78.4%] Iter  765/976 | SPS:  3591 | Reward:  0.069 | Value:  0.638 | VLoss: 0.0329 | PLoss: -0.0055 | Ent: 0.3814
+[ 78.5%] Iter  766/976 | SPS:  3591 | Reward:  0.064 | Value:  0.683 | VLoss: 0.0438 | PLoss: -0.0037 | Ent: 0.3164
+[ 78.6%] Iter  767/976 | SPS:  3591 | Reward:  0.077 | Value:  0.699 | VLoss: 0.0344 | PLoss: -0.0033 | Ent: 0.3156
+[ 78.7%] Iter  768/976 | SPS:  3591 | Reward:  0.069 | Value:  0.743 | VLoss: 0.0463 | PLoss: -0.0065 | Ent: 0.3145
+[ 78.8%] Iter  769/976 | SPS:  3591 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0350 | PLoss: -0.0037 | Ent: 0.3263
+[ 78.9%] Iter  770/976 | SPS:  3590 | Reward:  0.069 | Value:  0.599 | VLoss: 0.0409 | PLoss: -0.0044 | Ent: 0.3100
+[ 79.0%] Iter  771/976 | SPS:  3590 | Reward:  0.079 | Value:  0.684 | VLoss: 0.0446 | PLoss: -0.0025 | Ent: 0.3211
+[ 79.1%] Iter  772/976 | SPS:  3590 | Reward:  0.071 | Value:  0.709 | VLoss: 0.0404 | PLoss: -0.0048 | Ent: 0.3252
+[ 79.2%] Iter  773/976 | SPS:  3590 | Reward:  0.061 | Value:  0.612 | VLoss: 0.0383 | PLoss: -0.0073 | Ent: 0.3204
+[ 79.3%] Iter  774/976 | SPS:  3590 | Reward:  0.065 | Value:  0.608 | VLoss: 0.0260 | PLoss: -0.0033 | Ent: 0.3043
+[ 79.4%] Iter  775/976 | SPS:  3590 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0381 | PLoss: -0.0098 | Ent: 0.2952
+[ 79.5%] Iter  776/976 | SPS:  3589 | Reward:  0.077 | Value:  0.665 | VLoss: 0.0372 | PLoss: -0.0034 | Ent: 0.3185
+[ 79.6%] Iter  777/976 | SPS:  3589 | Reward:  0.073 | Value:  0.696 | VLoss: 0.0427 | PLoss: -0.0041 | Ent: 0.3112
+[ 79.7%] Iter  778/976 | SPS:  3589 | Reward:  0.072 | Value:  0.704 | VLoss: 0.0277 | PLoss: -0.0043 | Ent: 0.3059
+[ 79.8%] Iter  779/976 | SPS:  3589 | Reward:  0.071 | Value:  0.689 | VLoss: 0.0349 | PLoss: -0.0082 | Ent: 0.2740
+[ 79.9%] Iter  780/976 | SPS:  3589 | Reward:  0.072 | Value:  0.633 | VLoss: 0.0466 | PLoss: -0.0015 | Ent: 0.2933
+[ 80.0%] Iter  781/976 | SPS:  3589 | Reward:  0.047 | Value:  0.724 | VLoss: 0.0322 | PLoss: -0.0017 | Ent: 0.2634
+[ 80.1%] Iter  782/976 | SPS:  3589 | Reward:  0.071 | Value:  0.634 | VLoss: 0.0392 | PLoss: 0.0008 | Ent: 0.3579
+[ 80.2%] Iter  783/976 | SPS:  3589 | Reward:  0.056 | Value:  0.689 | VLoss: 0.0397 | PLoss: -0.0044 | Ent: 0.3528
+[ 80.3%] Iter  784/976 | SPS:  3589 | Reward:  0.062 | Value:  0.632 | VLoss: 0.0341 | PLoss: -0.0019 | Ent: 0.3226
+[ 80.4%] Iter  785/976 | SPS:  3589 | Reward:  0.061 | Value:  0.652 | VLoss: 0.0307 | PLoss: -0.0052 | Ent: 0.2988
+[ 80.5%] Iter  786/976 | SPS:  3589 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0318 | PLoss: -0.0038 | Ent: 0.2931
+[ 80.6%] Iter  787/976 | SPS:  3589 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0216 | PLoss: -0.0063 | Ent: 0.3139
+[ 80.7%] Iter  788/976 | SPS:  3589 | Reward:  0.064 | Value:  0.660 | VLoss: 0.0328 | PLoss: -0.0049 | Ent: 0.2967
+[ 80.8%] Iter  789/976 | SPS:  3588 | Reward:  0.066 | Value:  0.666 | VLoss: 0.0325 | PLoss: -0.0022 | Ent: 0.2481
+[ 80.9%] Iter  790/976 | SPS:  3589 | Reward:  0.066 | Value:  0.703 | VLoss: 0.0259 | PLoss: -0.0053 | Ent: 0.2254
+[ 81.0%] Iter  791/976 | SPS:  3588 | Reward:  0.058 | Value:  0.743 | VLoss: 0.0399 | PLoss: -0.0018 | Ent: 0.2895
+[ 81.1%] Iter  792/976 | SPS:  3588 | Reward:  0.059 | Value:  0.574 | VLoss: 0.0359 | PLoss: -0.0021 | Ent: 0.3522
+[ 81.2%] Iter  793/976 | SPS:  3588 | Reward:  0.062 | Value:  0.623 | VLoss: 0.0339 | PLoss: -0.0039 | Ent: 0.2865
+[ 81.4%] Iter  794/976 | SPS:  3588 | Reward:  0.062 | Value:  0.605 | VLoss: 0.0330 | PLoss: -0.0032 | Ent: 0.3667
+[ 81.5%] Iter  795/976 | SPS:  3588 | Reward:  0.064 | Value:  0.640 | VLoss: 0.0492 | PLoss: -0.0042 | Ent: 0.3373
+[ 81.6%] Iter  796/976 | SPS:  3588 | Reward:  0.054 | Value:  0.532 | VLoss: 0.0384 | PLoss: -0.0044 | Ent: 0.3702
+[ 81.7%] Iter  797/976 | SPS:  3588 | Reward:  0.059 | Value:  0.600 | VLoss: 0.0346 | PLoss: -0.0023 | Ent: 0.3023
+[ 81.8%] Iter  798/976 | SPS:  3588 | Reward:  0.075 | Value:  0.620 | VLoss: 0.0341 | PLoss: -0.0032 | Ent: 0.3521
+[ 81.9%] Iter  799/976 | SPS:  3588 | Reward:  0.064 | Value:  0.669 | VLoss: 0.0310 | PLoss: -0.0052 | Ent: 0.2960
+[ 82.0%] Iter  800/976 | SPS:  3588 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0415 | PLoss: -0.0018 | Ent: 0.2529
+[ 82.1%] Iter  801/976 | SPS:  3588 | Reward:  0.062 | Value:  0.622 | VLoss: 0.0326 | PLoss: -0.0033 | Ent: 0.3026
+[ 82.2%] Iter  802/976 | SPS:  3588 | Reward:  0.062 | Value:  0.639 | VLoss: 0.0337 | PLoss: -0.0041 | Ent: 0.2811
+[ 82.3%] Iter  803/976 | SPS:  3587 | Reward:  0.070 | Value:  0.605 | VLoss: 0.0387 | PLoss: -0.0035 | Ent: 0.3421
+[ 82.4%] Iter  804/976 | SPS:  3587 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0461 | PLoss: -0.0037 | Ent: 0.2903
+[ 82.5%] Iter  805/976 | SPS:  3587 | Reward:  0.059 | Value:  0.664 | VLoss: 0.0395 | PLoss: -0.0011 | Ent: 0.3028
+[ 82.6%] Iter  806/976 | SPS:  3587 | Reward:  0.073 | Value:  0.548 | VLoss: 0.0358 | PLoss: -0.0063 | Ent: 0.3328
+[ 82.7%] Iter  807/976 | SPS:  3587 | Reward:  0.067 | Value:  0.615 | VLoss: 0.0459 | PLoss: -0.0038 | Ent: 0.2821
+[ 82.8%] Iter  808/976 | SPS:  3587 | Reward:  0.066 | Value:  0.625 | VLoss: 0.0386 | PLoss: -0.0034 | Ent: 0.2598
+[ 82.9%] Iter  809/976 | SPS:  3587 | Reward:  0.066 | Value:  0.641 | VLoss: 0.0296 | PLoss: -0.0052 | Ent: 0.3367
+[ 83.0%] Iter  810/976 | SPS:  3586 | Reward:  0.060 | Value:  0.700 | VLoss: 0.0333 | PLoss: -0.0068 | Ent: 0.3535
+[ 83.1%] Iter  811/976 | SPS:  3586 | Reward:  0.063 | Value:  0.627 | VLoss: 0.0463 | PLoss: -0.0066 | Ent: 0.3271
+[ 83.2%] Iter  812/976 | SPS:  3586 | Reward:  0.069 | Value:  0.655 | VLoss: 0.0378 | PLoss: -0.0056 | Ent: 0.2465
+[ 83.3%] Iter  813/976 | SPS:  3585 | Reward:  0.064 | Value:  0.682 | VLoss: 0.0410 | PLoss: -0.0030 | Ent: 0.3205
+[ 83.4%] Iter  814/976 | SPS:  3585 | Reward:  0.059 | Value:  0.647 | VLoss: 0.0350 | PLoss: -0.0029 | Ent: 0.2840
+[ 83.5%] Iter  815/976 | SPS:  3585 | Reward:  0.063 | Value:  0.614 | VLoss: 0.0405 | PLoss: -0.0090 | Ent: 0.2958
+[ 83.6%] Iter  816/976 | SPS:  3585 | Reward:  0.062 | Value:  0.639 | VLoss: 0.0292 | PLoss: -0.0068 | Ent: 0.3159
+[ 83.7%] Iter  817/976 | SPS:  3585 | Reward:  0.043 | Value:  0.689 | VLoss: 0.0235 | PLoss: -0.0060 | Ent: 0.3425
+[ 83.8%] Iter  818/976 | SPS:  3585 | Reward:  0.051 | Value:  0.622 | VLoss: 0.0332 | PLoss: -0.0009 | Ent: 0.3385
+[ 83.9%] Iter  819/976 | SPS:  3585 | Reward:  0.061 | Value:  0.691 | VLoss: 0.0312 | PLoss: -0.0016 | Ent: 0.2586
+[ 84.0%] Iter  820/976 | SPS:  3585 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0248 | PLoss: -0.0040 | Ent: 0.2694
+[ 84.1%] Iter  821/976 | SPS:  3586 | Reward:  0.057 | Value:  0.657 | VLoss: 0.0322 | PLoss: -0.0011 | Ent: 0.2797
+[ 84.2%] Iter  822/976 | SPS:  3586 | Reward:  0.054 | Value:  0.675 | VLoss: 0.0335 | PLoss: -0.0035 | Ent: 0.2557
+[ 84.3%] Iter  823/976 | SPS:  3586 | Reward:  0.065 | Value:  0.614 | VLoss: 0.0393 | PLoss: -0.0024 | Ent: 0.3007
+[ 84.4%] Iter  824/976 | SPS:  3586 | Reward:  0.055 | Value:  0.686 | VLoss: 0.0461 | PLoss: -0.0056 | Ent: 0.3305
+[ 84.5%] Iter  825/976 | SPS:  3586 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0281 | PLoss: -0.0023 | Ent: 0.2818
+[ 84.6%] Iter  826/976 | SPS:  3585 | Reward:  0.066 | Value:  0.615 | VLoss: 0.0375 | PLoss: -0.0074 | Ent: 0.3385
+[ 84.7%] Iter  827/976 | SPS:  3585 | Reward:  0.069 | Value:  0.684 | VLoss: 0.0367 | PLoss: -0.0051 | Ent: 0.3342
+[ 84.8%] Iter  828/976 | SPS:  3585 | Reward:  0.079 | Value:  0.675 | VLoss: 0.0431 | PLoss: -0.0046 | Ent: 0.2577
+[ 84.9%] Iter  829/976 | SPS:  3585 | Reward:  0.080 | Value:  0.681 | VLoss: 0.0408 | PLoss: -0.0028 | Ent: 0.2953
+[ 85.0%] Iter  830/976 | SPS:  3585 | Reward:  0.064 | Value:  0.712 | VLoss: 0.0421 | PLoss: -0.0014 | Ent: 0.3098
+[ 85.1%] Iter  831/976 | SPS:  3585 | Reward:  0.054 | Value:  0.704 | VLoss: 0.0351 | PLoss: -0.0018 | Ent: 0.3168
+[ 85.2%] Iter  832/976 | SPS:  3585 | Reward:  0.062 | Value:  0.629 | VLoss: 0.0460 | PLoss: -0.0020 | Ent: 0.3136
+[ 85.3%] Iter  833/976 | SPS:  3585 | Reward:  0.058 | Value:  0.621 | VLoss: 0.0369 | PLoss: -0.0007 | Ent: 0.2797
+[ 85.5%] Iter  834/976 | SPS:  3585 | Reward:  0.065 | Value:  0.633 | VLoss: 0.0340 | PLoss: -0.0037 | Ent: 0.2504
+[ 85.6%] Iter  835/976 | SPS:  3585 | Reward:  0.056 | Value:  0.647 | VLoss: 0.0398 | PLoss: -0.0022 | Ent: 0.2687
+[ 85.7%] Iter  836/976 | SPS:  3585 | Reward:  0.070 | Value:  0.662 | VLoss: 0.0376 | PLoss: -0.0041 | Ent: 0.3370
+[ 85.8%] Iter  837/976 | SPS:  3585 | Reward:  0.057 | Value:  0.693 | VLoss: 0.0369 | PLoss: 0.0011 | Ent: 0.3263
+[ 85.9%] Iter  838/976 | SPS:  3585 | Reward:  0.063 | Value:  0.644 | VLoss: 0.0349 | PLoss: -0.0016 | Ent: 0.3208
+[ 86.0%] Iter  839/976 | SPS:  3585 | Reward:  0.058 | Value:  0.631 | VLoss: 0.0328 | PLoss: -0.0010 | Ent: 0.3102
+[ 86.1%] Iter  840/976 | SPS:  3585 | Reward:  0.060 | Value:  0.641 | VLoss: 0.0355 | PLoss: -0.0033 | Ent: 0.3600
+[ 86.2%] Iter  841/976 | SPS:  3585 | Reward:  0.054 | Value:  0.615 | VLoss: 0.0322 | PLoss: -0.0055 | Ent: 0.3104
+[ 86.3%] Iter  842/976 | SPS:  3585 | Reward:  0.057 | Value:  0.655 | VLoss: 0.0344 | PLoss: -0.0029 | Ent: 0.2922
+[ 86.4%] Iter  843/976 | SPS:  3585 | Reward:  0.061 | Value:  0.659 | VLoss: 0.0343 | PLoss: -0.0017 | Ent: 0.3057
+[ 86.5%] Iter  844/976 | SPS:  3584 | Reward:  0.082 | Value:  0.629 | VLoss: 0.0431 | PLoss: -0.0012 | Ent: 0.2886
+[ 86.6%] Iter  845/976 | SPS:  3584 | Reward:  0.069 | Value:  0.685 | VLoss: 0.0488 | PLoss: -0.0022 | Ent: 0.3097
+[ 86.7%] Iter  846/976 | SPS:  3584 | Reward:  0.069 | Value:  0.678 | VLoss: 0.0438 | PLoss: 0.0037 | Ent: 0.2855
+[ 86.8%] Iter  847/976 | SPS:  3584 | Reward:  0.072 | Value:  0.664 | VLoss: 0.0434 | PLoss: -0.0036 | Ent: 0.3393
+[ 86.9%] Iter  848/976 | SPS:  3584 | Reward:  0.076 | Value:  0.724 | VLoss: 0.0362 | PLoss: -0.0012 | Ent: 0.3050
+[ 87.0%] Iter  849/976 | SPS:  3583 | Reward:  0.061 | Value:  0.689 | VLoss: 0.0343 | PLoss: -0.0008 | Ent: 0.2709
+[ 87.1%] Iter  850/976 | SPS:  3584 | Reward:  0.067 | Value:  0.669 | VLoss: 0.0245 | PLoss: -0.0025 | Ent: 0.2930
+[ 87.2%] Iter  851/976 | SPS:  3583 | Reward:  0.069 | Value:  0.651 | VLoss: 0.0335 | PLoss: -0.0045 | Ent: 0.3652
+[ 87.3%] Iter  852/976 | SPS:  3583 | Reward:  0.053 | Value:  0.649 | VLoss: 0.0327 | PLoss: -0.0007 | Ent: 0.3397
+[ 87.4%] Iter  853/976 | SPS:  3584 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0301 | PLoss: -0.0020 | Ent: 0.2596
+[ 87.5%] Iter  854/976 | SPS:  3583 | Reward:  0.074 | Value:  0.704 | VLoss: 0.0325 | PLoss: -0.0027 | Ent: 0.2858
+[ 87.6%] Iter  855/976 | SPS:  3583 | Reward:  0.065 | Value:  0.718 | VLoss: 0.0363 | PLoss: 0.0028 | Ent: 0.2750
+[ 87.7%] Iter  856/976 | SPS:  3584 | Reward:  0.050 | Value:  0.721 | VLoss: 0.0284 | PLoss: -0.0029 | Ent: 0.2971
+[ 87.8%] Iter  857/976 | SPS:  3584 | Reward:  0.051 | Value:  0.665 | VLoss: 0.0377 | PLoss: -0.0005 | Ent: 0.2609
+[ 87.9%] Iter  858/976 | SPS:  3584 | Reward:  0.059 | Value:  0.610 | VLoss: 0.0327 | PLoss: -0.0035 | Ent: 0.3189
+[ 88.0%] Iter  859/976 | SPS:  3584 | Reward:  0.070 | Value:  0.661 | VLoss: 0.0367 | PLoss: -0.0029 | Ent: 0.2730
+[ 88.1%] Iter  860/976 | SPS:  3584 | Reward:  0.066 | Value:  0.677 | VLoss: 0.0347 | PLoss: -0.0023 | Ent: 0.2839
+[ 88.2%] Iter  861/976 | SPS:  3584 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0316 | PLoss: -0.0042 | Ent: 0.2815
+[ 88.3%] Iter  862/976 | SPS:  3583 | Reward:  0.074 | Value:  0.712 | VLoss: 0.0355 | PLoss: -0.0045 | Ent: 0.2945
+[ 88.4%] Iter  863/976 | SPS:  3583 | Reward:  0.070 | Value:  0.670 | VLoss: 0.0373 | PLoss: 0.0004 | Ent: 0.2472
+[ 88.5%] Iter  864/976 | SPS:  3583 | Reward:  0.059 | Value:  0.602 | VLoss: 0.0393 | PLoss: -0.0027 | Ent: 0.3363
+[ 88.6%] Iter  865/976 | SPS:  3583 | Reward:  0.072 | Value:  0.622 | VLoss: 0.0341 | PLoss: -0.0022 | Ent: 0.2844
+[ 88.7%] Iter  866/976 | SPS:  3583 | Reward:  0.073 | Value:  0.697 | VLoss: 0.0346 | PLoss: -0.0051 | Ent: 0.3035
+[ 88.8%] Iter  867/976 | SPS:  3583 | Reward:  0.070 | Value:  0.663 | VLoss: 0.0457 | PLoss: -0.0021 | Ent: 0.3303
+[ 88.9%] Iter  868/976 | SPS:  3583 | Reward:  0.068 | Value:  0.610 | VLoss: 0.0365 | PLoss: -0.0047 | Ent: 0.3146
+[ 89.0%] Iter  869/976 | SPS:  3583 | Reward:  0.069 | Value:  0.622 | VLoss: 0.0422 | PLoss: -0.0025 | Ent: 0.2925
+[ 89.1%] Iter  870/976 | SPS:  3582 | Reward:  0.066 | Value:  0.694 | VLoss: 0.0362 | PLoss: -0.0005 | Ent: 0.2763
+[ 89.2%] Iter  871/976 | SPS:  3582 | Reward:  0.066 | Value:  0.640 | VLoss: 0.0420 | PLoss: -0.0007 | Ent: 0.3386
+[ 89.3%] Iter  872/976 | SPS:  3582 | Reward:  0.062 | Value:  0.599 | VLoss: 0.0264 | PLoss: -0.0035 | Ent: 0.3267
+[ 89.4%] Iter  873/976 | SPS:  3582 | Reward:  0.077 | Value:  0.643 | VLoss: 0.0470 | PLoss: -0.0022 | Ent: 0.2902
+[ 89.5%] Iter  874/976 | SPS:  3582 | Reward:  0.059 | Value:  0.741 | VLoss: 0.0441 | PLoss: -0.0018 | Ent: 0.2664
+[ 89.7%] Iter  875/976 | SPS:  3582 | Reward:  0.070 | Value:  0.618 | VLoss: 0.0372 | PLoss: -0.0012 | Ent: 0.2543
+[ 89.8%] Iter  876/976 | SPS:  3582 | Reward:  0.057 | Value:  0.614 | VLoss: 0.0391 | PLoss: -0.0004 | Ent: 0.2408
+[ 89.9%] Iter  877/976 | SPS:  3582 | Reward:  0.075 | Value:  0.598 | VLoss: 0.0495 | PLoss: -0.0036 | Ent: 0.2938
+[ 90.0%] Iter  878/976 | SPS:  3581 | Reward:  0.064 | Value:  0.617 | VLoss: 0.0379 | PLoss: -0.0004 | Ent: 0.2713
+[ 90.1%] Iter  879/976 | SPS:  3581 | Reward:  0.061 | Value:  0.630 | VLoss: 0.0331 | PLoss: -0.0022 | Ent: 0.2723
+[ 90.2%] Iter  880/976 | SPS:  3581 | Reward:  0.077 | Value:  0.687 | VLoss: 0.0486 | PLoss: -0.0026 | Ent: 0.3143
+[ 90.3%] Iter  881/976 | SPS:  3581 | Reward:  0.064 | Value:  0.667 | VLoss: 0.0362 | PLoss: -0.0019 | Ent: 0.3132
+[ 90.4%] Iter  882/976 | SPS:  3581 | Reward:  0.060 | Value:  0.667 | VLoss: 0.0335 | PLoss: -0.0021 | Ent: 0.2723
+[ 90.5%] Iter  883/976 | SPS:  3581 | Reward:  0.067 | Value:  0.634 | VLoss: 0.0455 | PLoss: -0.0007 | Ent: 0.2873
+[ 90.6%] Iter  884/976 | SPS:  3580 | Reward:  0.086 | Value:  0.621 | VLoss: 0.0398 | PLoss: -0.0025 | Ent: 0.2974
+[ 90.7%] Iter  885/976 | SPS:  3580 | Reward:  0.048 | Value:  0.727 | VLoss: 0.0373 | PLoss: -0.0007 | Ent: 0.3182
+[ 90.8%] Iter  886/976 | SPS:  3580 | Reward:  0.071 | Value:  0.639 | VLoss: 0.0434 | PLoss: -0.0023 | Ent: 0.2770
+[ 90.9%] Iter  887/976 | SPS:  3580 | Reward:  0.070 | Value:  0.663 | VLoss: 0.0378 | PLoss: -0.0033 | Ent: 0.2906
+[ 91.0%] Iter  888/976 | SPS:  3580 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0352 | PLoss: -0.0027 | Ent: 0.2821
+[ 91.1%] Iter  889/976 | SPS:  3580 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0403 | PLoss: -0.0031 | Ent: 0.3255
+[ 91.2%] Iter  890/976 | SPS:  3580 | Reward:  0.057 | Value:  0.640 | VLoss: 0.0406 | PLoss: -0.0001 | Ent: 0.2915
+[ 91.3%] Iter  891/976 | SPS:  3580 | Reward:  0.067 | Value:  0.611 | VLoss: 0.0447 | PLoss: -0.0036 | Ent: 0.3068
+[ 91.4%] Iter  892/976 | SPS:  3580 | Reward:  0.061 | Value:  0.573 | VLoss: 0.0388 | PLoss: -0.0005 | Ent: 0.2808
+[ 91.5%] Iter  893/976 | SPS:  3580 | Reward:  0.084 | Value:  0.634 | VLoss: 0.0449 | PLoss: -0.0010 | Ent: 0.2776
+[ 91.6%] Iter  894/976 | SPS:  3580 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0280 | PLoss: -0.0003 | Ent: 0.3270
+[ 91.7%] Iter  895/976 | SPS:  3580 | Reward:  0.065 | Value:  0.657 | VLoss: 0.0393 | PLoss: -0.0005 | Ent: 0.2803
+[ 91.8%] Iter  896/976 | SPS:  3580 | Reward:  0.061 | Value:  0.624 | VLoss: 0.0317 | PLoss: -0.0006 | Ent: 0.3499
+[ 91.9%] Iter  897/976 | SPS:  3580 | Reward:  0.074 | Value:  0.631 | VLoss: 0.0428 | PLoss: -0.0014 | Ent: 0.3487
+[ 92.0%] Iter  898/976 | SPS:  3579 | Reward:  0.071 | Value:  0.660 | VLoss: 0.0501 | PLoss: -0.0022 | Ent: 0.3023
+[ 92.1%] Iter  899/976 | SPS:  3579 | Reward:  0.064 | Value:  0.689 | VLoss: 0.0471 | PLoss: -0.0029 | Ent: 0.3051
+[ 92.2%] Iter  900/976 | SPS:  3579 | Reward:  0.076 | Value:  0.648 | VLoss: 0.0427 | PLoss: -0.0009 | Ent: 0.2940
+[ 92.3%] Iter  901/976 | SPS:  3579 | Reward:  0.061 | Value:  0.622 | VLoss: 0.0342 | PLoss: -0.0002 | Ent: 0.3315
+[ 92.4%] Iter  902/976 | SPS:  3579 | Reward:  0.049 | Value:  0.591 | VLoss: 0.0372 | PLoss: -0.0008 | Ent: 0.3438
+[ 92.5%] Iter  903/976 | SPS:  3579 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0398 | PLoss: -0.0032 | Ent: 0.3343
+[ 92.6%] Iter  904/976 | SPS:  3579 | Reward:  0.059 | Value:  0.680 | VLoss: 0.0269 | PLoss: 0.0003 | Ent: 0.3172
+[ 92.7%] Iter  905/976 | SPS:  3579 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0353 | PLoss: -0.0015 | Ent: 0.3612
+[ 92.8%] Iter  906/976 | SPS:  3579 | Reward:  0.055 | Value:  0.610 | VLoss: 0.0417 | PLoss: -0.0023 | Ent: 0.3119
+[ 92.9%] Iter  907/976 | SPS:  3579 | Reward:  0.054 | Value:  0.673 | VLoss: 0.0332 | PLoss: -0.0014 | Ent: 0.3304
+[ 93.0%] Iter  908/976 | SPS:  3579 | Reward:  0.050 | Value:  0.667 | VLoss: 0.0429 | PLoss: -0.0002 | Ent: 0.2843
+[ 93.1%] Iter  909/976 | SPS:  3579 | Reward:  0.068 | Value:  0.654 | VLoss: 0.0252 | PLoss: -0.0006 | Ent: 0.3387
+[ 93.2%] Iter  910/976 | SPS:  3579 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0443 | PLoss: -0.0013 | Ent: 0.2526
+[ 93.3%] Iter  911/976 | SPS:  3579 | Reward:  0.066 | Value:  0.680 | VLoss: 0.0370 | PLoss: -0.0029 | Ent: 0.3019
+[ 93.4%] Iter  912/976 | SPS:  3579 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0314 | PLoss: -0.0006 | Ent: 0.2418
+[ 93.5%] Iter  913/976 | SPS:  3579 | Reward:  0.066 | Value:  0.674 | VLoss: 0.0301 | PLoss: -0.0022 | Ent: 0.3132
+[ 93.6%] Iter  914/976 | SPS:  3579 | Reward:  0.067 | Value:  0.648 | VLoss: 0.0424 | PLoss: -0.0017 | Ent: 0.2801
+[ 93.8%] Iter  915/976 | SPS:  3578 | Reward:  0.074 | Value:  0.669 | VLoss: 0.0489 | PLoss: -0.0021 | Ent: 0.2770
+[ 93.9%] Iter  916/976 | SPS:  3578 | Reward:  0.064 | Value:  0.647 | VLoss: 0.0388 | PLoss: -0.0013 | Ent: 0.3002
+[ 94.0%] Iter  917/976 | SPS:  3578 | Reward:  0.061 | Value:  0.654 | VLoss: 0.0319 | PLoss: -0.0003 | Ent: 0.3040
+[ 94.1%] Iter  918/976 | SPS:  3578 | Reward:  0.067 | Value:  0.652 | VLoss: 0.0390 | PLoss: -0.0008 | Ent: 0.3220
+[ 94.2%] Iter  919/976 | SPS:  3578 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0463 | PLoss: -0.0030 | Ent: 0.2905
+[ 94.3%] Iter  920/976 | SPS:  3578 | Reward:  0.071 | Value:  0.595 | VLoss: 0.0413 | PLoss: -0.0022 | Ent: 0.2591
+[ 94.4%] Iter  921/976 | SPS:  3578 | Reward:  0.057 | Value:  0.615 | VLoss: 0.0245 | PLoss: -0.0013 | Ent: 0.3218
+[ 94.5%] Iter  922/976 | SPS:  3578 | Reward:  0.069 | Value:  0.669 | VLoss: 0.0345 | PLoss: -0.0005 | Ent: 0.2778
+[ 94.6%] Iter  923/976 | SPS:  3578 | Reward:  0.063 | Value:  0.666 | VLoss: 0.0320 | PLoss: -0.0012 | Ent: 0.3115
+[ 94.7%] Iter  924/976 | SPS:  3578 | Reward:  0.065 | Value:  0.671 | VLoss: 0.0334 | PLoss: -0.0012 | Ent: 0.2765
+[ 94.8%] Iter  925/976 | SPS:  3578 | Reward:  0.069 | Value:  0.701 | VLoss: 0.0319 | PLoss: 0.0002 | Ent: 0.2322
+[ 94.9%] Iter  926/976 | SPS:  3577 | Reward:  0.069 | Value:  0.702 | VLoss: 0.0440 | PLoss: -0.0012 | Ent: 0.2855
+[ 95.0%] Iter  927/976 | SPS:  3577 | Reward:  0.069 | Value:  0.672 | VLoss: 0.0407 | PLoss: -0.0017 | Ent: 0.2807
+[ 95.1%] Iter  928/976 | SPS:  3577 | Reward:  0.067 | Value:  0.632 | VLoss: 0.0501 | PLoss: -0.0009 | Ent: 0.2866
+[ 95.2%] Iter  929/976 | SPS:  3577 | Reward:  0.070 | Value:  0.643 | VLoss: 0.0420 | PLoss: -0.0016 | Ent: 0.2829
+[ 95.3%] Iter  930/976 | SPS:  3577 | Reward:  0.070 | Value:  0.640 | VLoss: 0.0353 | PLoss: -0.0016 | Ent: 0.3500
+[ 95.4%] Iter  931/976 | SPS:  3577 | Reward:  0.066 | Value:  0.694 | VLoss: 0.0349 | PLoss: -0.0003 | Ent: 0.2561
+[ 95.5%] Iter  932/976 | SPS:  3577 | Reward:  0.064 | Value:  0.665 | VLoss: 0.0409 | PLoss: -0.0029 | Ent: 0.3236
+[ 95.6%] Iter  933/976 | SPS:  3576 | Reward:  0.063 | Value:  0.660 | VLoss: 0.0459 | PLoss: -0.0016 | Ent: 0.3004
+[ 95.7%] Iter  934/976 | SPS:  3576 | Reward:  0.073 | Value:  0.672 | VLoss: 0.0399 | PLoss: -0.0007 | Ent: 0.3199
+[ 95.8%] Iter  935/976 | SPS:  3576 | Reward:  0.067 | Value:  0.625 | VLoss: 0.0437 | PLoss: -0.0006 | Ent: 0.3141
+[ 95.9%] Iter  936/976 | SPS:  3576 | Reward:  0.075 | Value:  0.658 | VLoss: 0.0370 | PLoss: -0.0015 | Ent: 0.2738
+[ 96.0%] Iter  937/976 | SPS:  3576 | Reward:  0.068 | Value:  0.623 | VLoss: 0.0370 | PLoss: -0.0008 | Ent: 0.2698
+[ 96.1%] Iter  938/976 | SPS:  3576 | Reward:  0.082 | Value:  0.615 | VLoss: 0.0347 | PLoss: -0.0008 | Ent: 0.2929
+[ 96.2%] Iter  939/976 | SPS:  3576 | Reward:  0.072 | Value:  0.653 | VLoss: 0.0376 | PLoss: -0.0013 | Ent: 0.3108
+[ 96.3%] Iter  940/976 | SPS:  3575 | Reward:  0.078 | Value:  0.666 | VLoss: 0.0444 | PLoss: -0.0007 | Ent: 0.2980
+[ 96.4%] Iter  941/976 | SPS:  3575 | Reward:  0.071 | Value:  0.735 | VLoss: 0.0335 | PLoss: -0.0001 | Ent: 0.2662
+[ 96.5%] Iter  942/976 | SPS:  3575 | Reward:  0.079 | Value:  0.717 | VLoss: 0.0355 | PLoss: 0.0005 | Ent: 0.2585
+[ 96.6%] Iter  943/976 | SPS:  3575 | Reward:  0.064 | Value:  0.762 | VLoss: 0.0331 | PLoss: -0.0009 | Ent: 0.3013
+[ 96.7%] Iter  944/976 | SPS:  3575 | Reward:  0.076 | Value:  0.719 | VLoss: 0.0322 | PLoss: -0.0003 | Ent: 0.3107
+[ 96.8%] Iter  945/976 | SPS:  3575 | Reward:  0.074 | Value:  0.662 | VLoss: 0.0424 | PLoss: -0.0008 | Ent: 0.2868
+[ 96.9%] Iter  946/976 | SPS:  3575 | Reward:  0.055 | Value:  0.624 | VLoss: 0.0354 | PLoss: 0.0001 | Ent: 0.3187
+[ 97.0%] Iter  947/976 | SPS:  3575 | Reward:  0.065 | Value:  0.662 | VLoss: 0.0320 | PLoss: -0.0008 | Ent: 0.2789
+[ 97.1%] Iter  948/976 | SPS:  3574 | Reward:  0.073 | Value:  0.673 | VLoss: 0.0402 | PLoss: 0.0003 | Ent: 0.2780
+[ 97.2%] Iter  949/976 | SPS:  3574 | Reward:  0.067 | Value:  0.646 | VLoss: 0.0351 | PLoss: -0.0006 | Ent: 0.3313
+[ 97.3%] Iter  950/976 | SPS:  3574 | Reward:  0.070 | Value:  0.659 | VLoss: 0.0411 | PLoss: -0.0002 | Ent: 0.3114
+[ 97.4%] Iter  951/976 | SPS:  3574 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0349 | PLoss: -0.0006 | Ent: 0.3030
+[ 97.5%] Iter  952/976 | SPS:  3574 | Reward:  0.068 | Value:  0.686 | VLoss: 0.0289 | PLoss: 0.0000 | Ent: 0.3129
+[ 97.6%] Iter  953/976 | SPS:  3574 | Reward:  0.070 | Value:  0.668 | VLoss: 0.0379 | PLoss: -0.0004 | Ent: 0.2706
+[ 97.7%] Iter  954/976 | SPS:  3574 | Reward:  0.049 | Value:  0.698 | VLoss: 0.0431 | PLoss: -0.0002 | Ent: 0.3192
+[ 97.8%] Iter  955/976 | SPS:  3574 | Reward:  0.069 | Value:  0.659 | VLoss: 0.0450 | PLoss: -0.0003 | Ent: 0.3184
+[ 98.0%] Iter  956/976 | SPS:  3574 | Reward:  0.057 | Value:  0.637 | VLoss: 0.0357 | PLoss: -0.0004 | Ent: 0.3390
+[ 98.1%] Iter  957/976 | SPS:  3574 | Reward:  0.077 | Value:  0.622 | VLoss: 0.0371 | PLoss: -0.0008 | Ent: 0.3207
+[ 98.2%] Iter  958/976 | SPS:  3574 | Reward:  0.072 | Value:  0.650 | VLoss: 0.0363 | PLoss: -0.0002 | Ent: 0.2914
+[ 98.3%] Iter  959/976 | SPS:  3574 | Reward:  0.055 | Value:  0.650 | VLoss: 0.0323 | PLoss: -0.0006 | Ent: 0.3069
+[ 98.4%] Iter  960/976 | SPS:  3574 | Reward:  0.050 | Value:  0.627 | VLoss: 0.0332 | PLoss: -0.0005 | Ent: 0.3129
+[ 98.5%] Iter  961/976 | SPS:  3574 | Reward:  0.066 | Value:  0.627 | VLoss: 0.0371 | PLoss: -0.0004 | Ent: 0.2800
+[ 98.6%] Iter  962/976 | SPS:  3574 | Reward:  0.083 | Value:  0.684 | VLoss: 0.0420 | PLoss: -0.0007 | Ent: 0.3202
+[ 98.7%] Iter  963/976 | SPS:  3573 | Reward:  0.078 | Value:  0.635 | VLoss: 0.0378 | PLoss: -0.0000 | Ent: 0.3246
+[ 98.8%] Iter  964/976 | SPS:  3573 | Reward:  0.071 | Value:  0.630 | VLoss: 0.0417 | PLoss: -0.0003 | Ent: 0.3019
+[ 98.9%] Iter  965/976 | SPS:  3573 | Reward:  0.072 | Value:  0.668 | VLoss: 0.0453 | PLoss: -0.0003 | Ent: 0.2984
+[ 99.0%] Iter  966/976 | SPS:  3573 | Reward:  0.062 | Value:  0.602 | VLoss: 0.0426 | PLoss: -0.0004 | Ent: 0.3243
+[ 99.1%] Iter  967/976 | SPS:  3573 | Reward:  0.068 | Value:  0.651 | VLoss: 0.0489 | PLoss: -0.0002 | Ent: 0.3023
+[ 99.2%] Iter  968/976 | SPS:  3573 | Reward:  0.058 | Value:  0.631 | VLoss: 0.0433 | PLoss: -0.0002 | Ent: 0.2609
+[ 99.3%] Iter  969/976 | SPS:  3573 | Reward:  0.071 | Value:  0.680 | VLoss: 0.0407 | PLoss: -0.0003 | Ent: 0.2690
+[ 99.4%] Iter  970/976 | SPS:  3573 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0379 | PLoss: -0.0005 | Ent: 0.3153
+[ 99.5%] Iter  971/976 | SPS:  3573 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0388 | PLoss: -0.0000 | Ent: 0.2579
+[ 99.6%] Iter  972/976 | SPS:  3573 | Reward:  0.047 | Value:  0.672 | VLoss: 0.0315 | PLoss: -0.0001 | Ent: 0.3279
+[ 99.7%] Iter  973/976 | SPS:  3573 | Reward:  0.060 | Value:  0.634 | VLoss: 0.0364 | PLoss: -0.0001 | Ent: 0.3457
+[ 99.8%] Iter  974/976 | SPS:  3573 | Reward:  0.074 | Value:  0.678 | VLoss: 0.0488 | PLoss: 0.0000 | Ent: 0.3268
+[ 99.9%] Iter  975/976 | SPS:  3573 | Reward:  0.057 | Value:  0.648 | VLoss: 0.0326 | PLoss: -0.0000 | Ent: 0.2812
+[100.0%] Iter  976/976 | SPS:  3573 | Reward:  0.068 | Value:  0.637 | VLoss: 0.0442 | PLoss: -0.0000 | Ent: 0.2489
diff --git a/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/wandb-metadata.json b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..901165189143154575cbc69acb00b7e2b3f1ea2a
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_103331-e2e7wd2b/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2025-11-07T02:33:31.320641Z",
+  "args":  [
+    "--track",
+    "--wandb-project-name",
+    "ragen-bandit"
+  ],
+  "program":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_frozenlake.py",
+  "codePath":  "cleanrl/ppo_frozenlake.py",
+  "codePathLocal":  "ppo_frozenlake.py",
+  "git":  {
+    "remote":  "https://github.com/vwxyzjn/cleanrl.git",
+    "commit":  "004f8a086a892a2a180f4dd332b90d83a968aa7a"
+  },
+  "email":  "haoyu-wa22@mails.tsinghua.edu.cn",
+  "root":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl",
+  "host":  "pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0",
+  "executable":  "/root/local/miniconda3/envs/ragen/bin/python",
+  "cpu_count":  64,
+  "cpu_count_logical":  128,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "5153960755200",
+      "used":  "30509891584"
+    }
+  },
+  "memory":  {
+    "total":  "2163642122240"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-35e2d43d-4067-82ce-90d4-def9e389bf28"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-af4135e3-88f2-e9ac-518d-502c75a85429"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b7cf0ec6-7c29-1179-dceb-09565da51890"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "e26crksekquh3exjmxygxro2irogd3e1"
+}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/code/cleanrl/ppo_sokoban.py b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/code/cleanrl/ppo_sokoban.py
new file mode 100644
index 0000000000000000000000000000000000000000..97abbf5c5a939144f7ba252c631309660666fff9
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/code/cleanrl/ppo_sokoban.py
@@ -0,0 +1,352 @@
+# PPO implementation for RAGEN Sokoban environment
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.sokoban.env import SokobanEnv
+from ragen.env.sokoban.config import SokobanEnvConfig
+from ragen_wrappers import SokobanWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "Sokoban"
+    """the id of the environment"""
+    total_timesteps: int = 5000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 8
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.2
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.01
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+    
+    # Sokoban specific
+    dim_room: tuple = (6, 6)
+    """dimensions of the sokoban room"""
+    num_boxes: int = 1
+    """number of boxes in sokoban"""
+    max_steps: int = 100
+    """maximum steps per episode"""
+    search_depth: int = 100
+    """search depth for sokoban level generation"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, dim_room, num_boxes, max_steps, search_depth):
+    def thunk():
+        config = SokobanEnvConfig(
+            dim_room=dim_room,
+            num_boxes=num_boxes,
+            max_steps=max_steps,
+            search_depth=search_depth
+        )
+        env = SokobanEnv(config)
+        env = SokobanWrapper(env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = np.array(envs.single_observation_space.shape).prod()
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 256)),
+            nn.Tanh(),
+            layer_init(nn.Linear(256, 256)),
+            nn.Tanh(),
+            layer_init(nn.Linear(256, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 256)),
+            nn.Tanh(),
+            layer_init(nn.Linear(256, 256)),
+            nn.Tanh(),
+            layer_init(nn.Linear(256, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None):
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed, 
+                  args.dim_room, args.num_boxes, args.max_steps, args.search_depth) 
+         for i in range(args.num_envs)],
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                        writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                        writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped)
+                    v_loss = 0.5 * v_loss_max.mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        
+        # Additional useful metrics
+        writer.add_scalar("charts/avg_reward", rewards.mean().item(), global_step)
+        writer.add_scalar("charts/avg_value", values.mean().item(), global_step)
+        writer.add_scalar("charts/max_reward", rewards.max().item(), global_step)
+        writer.add_scalar("charts/min_reward", rewards.min().item(), global_step)
+        
+        # Console output with key metrics
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Value: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        writer.add_scalar("charts/SPS", sps, global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..9c3dd3491df4dce673a07d78e8696333f27cc19f
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/config.yaml
@@ -0,0 +1,163 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/ppo_sokoban.py
+        e:
+            z96hq3pus2sxc7xhydj92rmnpv0sy2t0:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-sokoban
+                codePath: cleanrl/ppo_sokoban.py
+                codePathLocal: ppo_sokoban.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "31672418304"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_sokoban.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T03:26:47.139012Z"
+                writerId: z96hq3pus2sxc7xhydj92rmnpv0sy2t0
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+anneal_lr:
+    value: true
+batch_size:
+    value: 1024
+capture_video:
+    value: false
+clip_coef:
+    value: 0.2
+clip_vloss:
+    value: true
+cuda:
+    value: true
+dim_room:
+    value:
+        - 6
+        - 6
+ent_coef:
+    value: 0.01
+env_id:
+    value: Sokoban
+exp_name:
+    value: ppo_sokoban
+gae_lambda:
+    value: 0.95
+gamma:
+    value: 0.99
+learning_rate:
+    value: 0.00025
+max_grad_norm:
+    value: 0.5
+max_steps:
+    value: 100
+minibatch_size:
+    value: 256
+norm_adv:
+    value: true
+num_boxes:
+    value: 1
+num_envs:
+    value: 8
+num_iterations:
+    value: 4882
+num_minibatches:
+    value: 4
+num_steps:
+    value: 128
+search_depth:
+    value: 100
+seed:
+    value: 1
+target_kl:
+    value: null
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 5000000
+track:
+    value: true
+update_epochs:
+    value: 4
+vf_coef:
+    value: 0.5
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-sokoban
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/output.log b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..1dcb420a0899478e5173173c30d2ce1f31005ce0
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/output.log
@@ -0,0 +1,16 @@
+Traceback (most recent call last):
+  File "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_sokoban.py", line 184, in <module>
+    envs = gym.vector.SyncVectorEnv(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/gymnasium/vector/sync_vector_env.py", line 97, in __init__
+    self.envs = [env_fn() for env_fn in env_fns]
+                 ^^^^^^^^
+  File "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_sokoban.py", line 108, in thunk
+    env = gym.wrappers.RecordEpisodeStatistics(env)
+          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/gymnasium/wrappers/common.py", line 496, in __init__
+    gym.Wrapper.__init__(self, env)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/gymnasium/core.py", line 313, in __init__
+    assert isinstance(
+           ^^^^^^^^^^^
+AssertionError: Expected env to be a `gymnasium.Env` but got <class 'ragen_wrappers.SokobanWrapper'>
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/requirements.txt b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..5d592ae1a79f938781a73a558d1be3b4aa58e912
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/requirements.txt
@@ -0,0 +1,305 @@
+ragen==0.1
+setuptools==80.9.0
+wheel==0.45.1
+pip==25.2
+zipp==3.23.0
+verl==0.2.0.dev0
+ragen==0.1
+triton==3.2.0
+nvidia-cusparselt-cu12==0.6.2
+mpmath==1.3.0
+typing_extensions==4.15.0
+sympy==1.13.1
+nvidia-nvtx-cu12==12.4.127
+nvidia-nvjitlink-cu12==12.4.127
+nvidia-nccl-cu12==2.21.5
+nvidia-curand-cu12==10.3.5.147
+nvidia-cufft-cu12==11.2.1.3
+nvidia-cuda-runtime-cu12==12.4.127
+nvidia-cuda-nvrtc-cu12==12.4.127
+nvidia-cuda-cupti-cu12==12.4.127
+nvidia-cublas-cu12==12.4.5.8
+networkx==3.5
+MarkupSafe==2.1.5
+fsspec==2025.9.0
+filelock==3.19.1
+nvidia-cusparse-cu12==12.3.1.170
+nvidia-cudnn-cu12==9.1.0.70
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.6.1.9
+torch==2.6.0+cu124
+einops==0.8.1
+flash_attn==2.7.4.post1
+pytz==2025.2
+pyperclip==1.11.0
+pylatexenc==2.10
+pyjnius==1.7.0
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+gym-notices==0.1.0
+flatbuffers==25.9.23
+fastrlock==0.8.3
+Farama-Notifications==0.0.4
+cymem==2.0.11
+antlr4-python3-runtime==4.9.3
+xxhash==3.6.0
+wrapt==2.0.0
+Werkzeug==3.1.3
+websockets==15.0.1
+wcwidth==0.2.14
+wasabi==1.1.3
+uvloop==0.22.1
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.2
+traitlets==5.14.3
+tqdm==4.67.1
+threadpoolctl==3.6.0
+tabulate==0.9.0
+spacy-loggers==1.0.5
+spacy-legacy==3.0.12
+soupsieve==2.8
+sniffio==1.3.1
+smmap==5.0.2
+six==1.17.0
+shellingham==1.5.4
+sentencepiece==0.2.1
+safetensors==0.6.2
+rpds-py==0.28.0
+rignore==0.7.6
+regex==2025.11.3
+RapidFuzz==3.14.3
+pyzmq==27.1.0
+PyYAML==6.0.3
+pytokens==0.3.0
+python-multipart==0.0.20
+python-json-logger==4.0.0
+python-dotenv==1.2.1
+PySocks==1.7.1
+pyparsing==3.2.5
+PyJWT==2.10.1
+Pygments==2.19.2
+pygame==2.6.1
+pydantic_core==2.41.5
+pycparser==2.23
+pycountry==24.6.1
+pybind11==3.0.1
+pyarrow==22.0.0
+psutil==7.1.3
+protobuf==6.33.0
+propcache==0.4.1
+prometheus_client==0.23.1
+platformdirs==4.5.0
+pillow==11.3.0
+pexpect==4.9.0
+pathvalidate==3.3.1
+pathspec==0.12.1
+pathable==0.4.4
+partial-json-parser==0.2.1.1.post6
+parso==0.8.5
+packaging==25.0
+orjson==3.11.4
+numpy==1.26.4
+ninja==1.13.0
+nest-asyncio==1.6.0
+mypy_extensions==1.1.0
+murmurhash==1.0.13
+multidict==6.7.0
+msgspec==0.19.0
+msgpack==1.1.2
+more-itertools==10.8.0
+mdurl==0.1.2
+marisa-trie==1.3.1
+llvmlite==0.43.0
+llguidance==0.7.30
+lark==1.2.2
+kiwisolver==1.4.9
+joblib==1.5.2
+jiter==0.11.1
+jeepney==0.9.0
+jaraco.context==6.0.1
+itsdangerous==2.2.0
+interegular==0.3.3
+idna==3.11
+humanfriendly==10.0
+httpx-sse==0.4.3
+httptools==0.7.1
+html2text==2025.4.15
+hf-xet==1.2.0
+h11==0.16.0
+frozenlist==1.8.0
+fonttools==4.60.1
+executing==2.2.1
+exceptiongroup==1.3.0
+eval_type_backport==0.2.2
+docutils==0.22.3
+docstring_parser==0.17.0
+dnspython==2.8.0
+distro==1.9.0
+diskcache==5.6.3
+dill==0.4.0
+decorator==5.2.1
+debugpy==1.8.17
+Cython==3.2.0
+cycler==0.12.1
+codetiming==1.4.0
+cloudpickle==3.1.2
+cloudpathlib==0.23.0
+click==8.2.1
+charset-normalizer==3.4.4
+certifi==2025.10.5
+catalogue==2.0.10
+cachetools==6.2.1
+blinker==1.9.0
+blake3==1.0.8
+beartype==0.22.5
+attrs==25.4.0
+asttokens==3.0.0
+astor==0.8.1
+annotated-types==0.7.0
+annotated-doc==0.0.3
+airportsdata==20250909
+aiohappyeyeballs==2.6.1
+yarl==1.22.0
+uvicorn==0.38.0
+typer-slim==0.20.0
+thefuzz==0.22.1
+stack-data==0.6.3
+srsly==2.5.1
+smart_open==7.4.4
+sentry-sdk==2.43.0
+scipy==1.16.3
+requests==2.32.5
+referencing==0.36.2
+rank-bm25==0.2.2
+python-dateutil==2.9.0.post0
+pydantic==2.12.4
+py-key-value-shared==0.2.8
+prompt_toolkit==3.0.52
+preshed==3.0.10
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+numba==0.60.0
+nltk==3.9.2
+multiprocess==0.70.18
+matplotlib-inline==0.2.1
+markdown-it-py==4.0.0
+language_data==1.3.0
+jedi==0.19.2
+jaraco.functools==4.3.0
+jaraco.classes==3.4.0
+ipython_pygments_lexers==1.1.1
+importlib_metadata==8.7.0
+ImageIO==2.37.2
+httpcore==1.0.9
+gymnasium==1.2.2
+gym==0.26.2
+gitdb==4.0.12
+gguf==0.10.0
+Flask==3.1.2
+faiss-cpu==1.12.0
+email-validator==2.3.0
+depyf==0.18.0
+cupy-cuda12x==13.6.0
+contourpy==1.3.3
+coloredlogs==15.0.1
+cffi==2.0.0
+blis==1.3.0
+black==25.9.0
+beautifulsoup4==4.14.2
+anyio==4.11.0
+aiosignal==1.4.0
+watchfiles==1.1.1
+tiktoken==0.12.0
+starlette==0.49.3
+sse-starlette==3.0.3
+scikit-learn==1.7.2
+rich==14.2.0
+pydantic-settings==2.11.0
+pydantic-extra-types==2.10.6
+py-key-value-aio==0.2.8
+pandas==2.3.3
+openapi-pydantic==0.5.1
+onnxruntime==1.23.2
+matplotlib==3.10.7
+lm-format-enforcer==0.10.12
+langcodes==3.5.0
+jsonschema-specifications==2025.9.1
+jsonschema-path==0.3.4
+ipython==9.7.0
+hydra-core==1.3.2
+huggingface-hub==0.36.0
+httpx==0.28.1
+gym-sokoban==0.0.6
+GitPython==3.1.45
+cryptography==46.0.3
+confection==0.1.5
+cleantext==1.1.4
+aiohttp==3.13.2
+xformers==0.0.29.post2
+weasel==0.4.2
+wandb==0.22.3
+typer==0.19.2
+torchvision==0.21.0
+torchdata==0.11.0
+torchaudio==2.6.0
+tokenizers==0.22.1
+thinc==8.3.7
+tensordict==0.8.3
+SecretStorage==3.4.0
+rich-toolkit==0.15.1
+rich-rst==1.3.2
+prometheus-fastapi-instrumentator==7.1.0
+openai==2.7.1
+jsonschema==4.25.1
+gdown==5.2.0
+fastapi==0.121.0
+Authlib==1.6.5
+anthropic==0.72.0
+accelerate==1.11.0
+transformers==4.57.1
+together==1.5.30
+spacy==3.8.7
+ray==2.51.1
+outlines_core==0.1.26
+mistral_common==1.8.5
+mcp==1.20.0
+keyring==25.6.0
+fastapi-cloud-cli==0.3.1
+fastapi-cli==0.0.14
+datasets==4.4.1
+cyclopts==4.2.1
+xgrammar==0.1.16
+peft==0.17.1
+outlines==0.1.11
+compressed-tensors==0.9.2
+fastmcp==2.13.0.2
+vllm==0.8.2
+pyserini==1.3.0
+typeguard==4.4.4
+shtab==1.7.2
+tyro==0.9.35
+tensorboard-data-server==0.7.2
+Markdown==3.10
+grpcio==1.76.0
+absl-py==2.3.1
+tensorboard==2.20.0
+ragen==0.1
+verl==0.2.0.dev0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.0.0
+inflect==7.3.1
+jaraco.collections==5.1.0
+jaraco.context==5.3.0
+jaraco.functools==4.0.1
+jaraco.text==3.12.1
+more-itertools==10.3.0
+packaging==24.2
+platformdirs==4.2.2
+tomli==2.0.1
+typeguard==4.3.0
+typing_extensions==4.12.2
+wheel==0.45.1
+zipp==3.19.2
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/wandb-metadata.json b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..a2b19b189db48f5a90a660b8aa265963f5e89740
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2025-11-07T03:26:47.139012Z",
+  "args":  [
+    "--track",
+    "--wandb-project-name",
+    "ragen-sokoban"
+  ],
+  "program":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_sokoban.py",
+  "codePath":  "cleanrl/ppo_sokoban.py",
+  "codePathLocal":  "ppo_sokoban.py",
+  "git":  {
+    "remote":  "https://github.com/vwxyzjn/cleanrl.git",
+    "commit":  "004f8a086a892a2a180f4dd332b90d83a968aa7a"
+  },
+  "email":  "haoyu-wa22@mails.tsinghua.edu.cn",
+  "root":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl",
+  "host":  "pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0",
+  "executable":  "/root/local/miniconda3/envs/ragen/bin/python",
+  "cpu_count":  64,
+  "cpu_count_logical":  128,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "5153960755200",
+      "used":  "31672418304"
+    }
+  },
+  "memory":  {
+    "total":  "2163642122240"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-35e2d43d-4067-82ce-90d4-def9e389bf28"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-af4135e3-88f2-e9ac-518d-502c75a85429"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b7cf0ec6-7c29-1179-dceb-09565da51890"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "z96hq3pus2sxc7xhydj92rmnpv0sy2t0"
+}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/logs/debug.log b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..fa9443c3a539b4145b2c853a146bddc47b9564c6
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/logs/debug.log
@@ -0,0 +1,24 @@
+2025-11-07 11:26:47,186 INFO    MainThread:212707 [wandb_setup.py:_flush():81] Current SDK version is 0.22.3
+2025-11-07 11:26:47,186 INFO    MainThread:212707 [wandb_setup.py:_flush():81] Configure stats pid to 212707
+2025-11-07 11:26:47,186 INFO    MainThread:212707 [wandb_setup.py:_flush():81] Loading settings from /root/.config/wandb/settings
+2025-11-07 11:26:47,186 INFO    MainThread:212707 [wandb_setup.py:_flush():81] Loading settings from /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/settings
+2025-11-07 11:26:47,187 INFO    MainThread:212707 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2025-11-07 11:26:47,187 INFO    MainThread:212707 [wandb_init.py:setup_run_log_directory():706] Logging user logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/logs/debug.log
+2025-11-07 11:26:47,187 INFO    MainThread:212707 [wandb_init.py:setup_run_log_directory():707] Logging internal logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/logs/debug-internal.log
+2025-11-07 11:26:47,189 INFO    MainThread:212707 [wandb_init.py:init():833] calling init triggers
+2025-11-07 11:26:47,189 INFO    MainThread:212707 [wandb_init.py:init():838] wandb.init called with sweep_config: {}
+config: {'exp_name': 'ppo_sokoban', 'seed': 1, 'torch_deterministic': True, 'cuda': True, 'track': True, 'wandb_project_name': 'ragen-sokoban', 'wandb_entity': None, 'capture_video': False, 'env_id': 'Sokoban', 'total_timesteps': 5000000, 'learning_rate': 0.00025, 'num_envs': 8, 'num_steps': 128, 'anneal_lr': True, 'gamma': 0.99, 'gae_lambda': 0.95, 'num_minibatches': 4, 'update_epochs': 4, 'norm_adv': True, 'clip_coef': 0.2, 'clip_vloss': True, 'ent_coef': 0.01, 'vf_coef': 0.5, 'max_grad_norm': 0.5, 'target_kl': None, 'dim_room': (6, 6), 'num_boxes': 1, 'max_steps': 100, 'search_depth': 100, 'batch_size': 1024, 'minibatch_size': 256, 'num_iterations': 4882, '_wandb': {'code_path': 'code/cleanrl/ppo_sokoban.py'}}
+2025-11-07 11:26:47,189 INFO    MainThread:212707 [wandb_init.py:init():881] starting backend
+2025-11-07 11:26:47,395 INFO    MainThread:212707 [wandb_init.py:init():884] sending inform_init request
+2025-11-07 11:26:47,406 INFO    MainThread:212707 [wandb_init.py:init():892] backend started and connected
+2025-11-07 11:26:47,407 INFO    MainThread:212707 [wandb_init.py:init():962] updated telemetry
+2025-11-07 11:26:47,441 INFO    MainThread:212707 [wandb_init.py:init():986] communicating run to backend with 90.0 second timeout
+2025-11-07 11:26:48,271 INFO    MainThread:212707 [wandb_init.py:init():1033] starting run threads in backend
+2025-11-07 11:26:48,414 INFO    MainThread:212707 [wandb_run.py:_console_start():2506] atexit reg
+2025-11-07 11:26:48,414 INFO    MainThread:212707 [wandb_run.py:_redirect():2354] redirect: wrap_raw
+2025-11-07 11:26:48,415 INFO    MainThread:212707 [wandb_run.py:_redirect():2423] Wrapping output streams.
+2025-11-07 11:26:48,415 INFO    MainThread:212707 [wandb_run.py:_redirect():2446] Redirects installed.
+2025-11-07 11:26:48,417 INFO    MainThread:212707 [wandb_init.py:init():1073] run started, returning control to user process
+2025-11-07 11:26:48,417 INFO    MainThread:212707 [wandb_run.py:_tensorboard_callback():1598] tensorboard callback: runs/Sokoban__ppo_sokoban__1__1762486000, True
+2025-11-07 11:26:48,527 INFO    wandb-AsyncioManager-main:212707 [service_client.py:_forward_responses():80] Reached EOF.
+2025-11-07 11:26:48,527 INFO    wandb-AsyncioManager-main:212707 [mailbox.py:close():137] Closing mailbox, abandoning 2 handles.
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/run-9utis4xx.wandb b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/run-9utis4xx.wandb
new file mode 100644
index 0000000000000000000000000000000000000000..dba253ccdaf54a7be1be6f2b1a7a7ce80a109d4a
Binary files /dev/null and b/cleanrl/cleanrl/wandb/run-20251107_112647-9utis4xx/run-9utis4xx.wandb differ
diff --git a/cleanrl/cleanrl/wandb/run-20251107_112903-8xop3upl/files/wandb-metadata.json b/cleanrl/cleanrl/wandb/run-20251107_112903-8xop3upl/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c1cf7d30d01058072938b09b3473449735747b4
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_112903-8xop3upl/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2025-11-07T03:29:03.539841Z",
+  "args":  [
+    "--track",
+    "--wandb-project-name",
+    "ragen-bandit"
+  ],
+  "program":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_frozenlake.py",
+  "codePath":  "cleanrl/ppo_frozenlake.py",
+  "codePathLocal":  "ppo_frozenlake.py",
+  "git":  {
+    "remote":  "https://github.com/vwxyzjn/cleanrl.git",
+    "commit":  "004f8a086a892a2a180f4dd332b90d83a968aa7a"
+  },
+  "email":  "haoyu-wa22@mails.tsinghua.edu.cn",
+  "root":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl",
+  "host":  "pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0",
+  "executable":  "/root/local/miniconda3/envs/ragen/bin/python",
+  "cpu_count":  64,
+  "cpu_count_logical":  128,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "5153960755200",
+      "used":  "31673995264"
+    }
+  },
+  "memory":  {
+    "total":  "2163642122240"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-35e2d43d-4067-82ce-90d4-def9e389bf28"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-af4135e3-88f2-e9ac-518d-502c75a85429"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b7cf0ec6-7c29-1179-dceb-09565da51890"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "27m01iw4c1djia38oq0povle1209kj1y"
+}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/code/cleanrl/ppo_frozenlake.py b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/code/cleanrl/ppo_frozenlake.py
new file mode 100644
index 0000000000000000000000000000000000000000..64039047d5927adf5575525c449370af9dfa4e71
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/code/cleanrl/ppo_frozenlake.py
@@ -0,0 +1,347 @@
+# PPO implementation for RAGEN FrozenLake environment
+import os
+import random
+import time
+from dataclasses import dataclass
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import tyro
+from torch.distributions.categorical import Categorical
+from torch.utils.tensorboard import SummaryWriter
+
+import sys
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '../../'))
+
+from ragen.env.frozen_lake.env import FrozenLakeEnv
+from ragen.env.frozen_lake.config import FrozenLakeEnvConfig
+from ragen_wrappers import FrozenLakeWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+
+    # Algorithm specific arguments
+    env_id: str = "FrozenLake"
+    """the id of the environment"""
+    total_timesteps: int = 10000000
+    """total timesteps of the experiments"""
+    learning_rate: float = 2.5e-4
+    """the learning rate of the optimizer"""
+    num_envs: int = 8
+    """the number of parallel game environments"""
+    num_steps: int = 128
+    """the number of steps to run in each environment per policy rollout"""
+    anneal_lr: bool = True
+    """Toggle learning rate annealing for policy and value networks"""
+    gamma: float = 0.99
+    """the discount factor gamma"""
+    gae_lambda: float = 0.95
+    """the lambda for the general advantage estimation"""
+    num_minibatches: int = 4
+    """the number of mini-batches"""
+    update_epochs: int = 4
+    """the K epochs to update the policy"""
+    norm_adv: bool = True
+    """Toggles advantages normalization"""
+    clip_coef: float = 0.2
+    """the surrogate clipping coefficient"""
+    clip_vloss: bool = True
+    """Toggles whether or not to use a clipped loss for the value function, as per the paper."""
+    ent_coef: float = 0.01
+    """coefficient of the entropy"""
+    vf_coef: float = 0.5
+    """coefficient of the value function"""
+    max_grad_norm: float = 0.5
+    """the maximum norm for the gradient clipping"""
+    target_kl: float = None
+    """the target KL divergence threshold"""
+    
+    # FrozenLake specific
+    grid_size: int = 4
+    """size of the frozen lake grid"""
+    is_slippery: bool = True
+    """whether the ice is slippery"""
+
+    # to be filled in runtime
+    batch_size: int = 0
+    """the batch size (computed in runtime)"""
+    minibatch_size: int = 0
+    """the mini-batch size (computed in runtime)"""
+    num_iterations: int = 0
+    """the number of iterations (computed in runtime)"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, grid_size, is_slippery):
+    def thunk():
+        config = FrozenLakeEnvConfig(
+            size=grid_size,
+            p=0.8,
+            is_slippery=is_slippery,
+            map_seed=seed + idx
+        )
+        env = FrozenLakeEnv(config)
+        env = FrozenLakeWrapper(env)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+def layer_init(layer, std=np.sqrt(2), bias_const=0.0):
+    torch.nn.init.orthogonal_(layer.weight, std)
+    torch.nn.init.constant_(layer.bias, bias_const)
+    return layer
+
+
+class Agent(nn.Module):
+    def __init__(self, envs):
+        super().__init__()
+        obs_shape = np.array(envs.single_observation_space.shape).prod()
+        self.critic = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 1), std=1.0),
+        )
+        self.actor = nn.Sequential(
+            layer_init(nn.Linear(obs_shape, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, 128)),
+            nn.Tanh(),
+            layer_init(nn.Linear(128, envs.single_action_space.n), std=0.01),
+        )
+
+    def get_value(self, x):
+        return self.critic(x)
+
+    def get_action_and_value(self, x, action=None):
+        logits = self.actor(x)
+        probs = Categorical(logits=logits)
+        if action is None:
+            action = probs.sample()
+        return action, probs.log_prob(action), probs.entropy(), self.critic(x)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    args.batch_size = int(args.num_envs * args.num_steps)
+    args.minibatch_size = int(args.batch_size // args.num_minibatches)
+    args.num_iterations = args.total_timesteps // args.batch_size
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed, args.grid_size, args.is_slippery) 
+         for i in range(args.num_envs)],
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+
+    # ALGO Logic: Storage setup
+    obs = torch.zeros((args.num_steps, args.num_envs) + envs.single_observation_space.shape).to(device)
+    actions = torch.zeros((args.num_steps, args.num_envs) + envs.single_action_space.shape).to(device)
+    logprobs = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    rewards = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    dones = torch.zeros((args.num_steps, args.num_envs)).to(device)
+    values = torch.zeros((args.num_steps, args.num_envs)).to(device)
+
+    # TRY NOT TO MODIFY: start the game
+    global_step = 0
+    start_time = time.time()
+    next_obs, _ = envs.reset(seed=args.seed)
+    next_obs = torch.Tensor(next_obs).to(device)
+    next_done = torch.zeros(args.num_envs).to(device)
+
+    for iteration in range(1, args.num_iterations + 1):
+        # Annealing the rate if instructed to do so.
+        if args.anneal_lr:
+            frac = 1.0 - (iteration - 1.0) / args.num_iterations
+            lrnow = frac * args.learning_rate
+            optimizer.param_groups[0]["lr"] = lrnow
+
+        for step in range(0, args.num_steps):
+            global_step += args.num_envs
+            obs[step] = next_obs
+            dones[step] = next_done
+
+            # ALGO LOGIC: action logic
+            with torch.no_grad():
+                action, logprob, _, value = agent.get_action_and_value(next_obs)
+                values[step] = value.flatten()
+            actions[step] = action
+            logprobs[step] = logprob
+
+            # TRY NOT TO MODIFY: execute the game and log data.
+            next_obs, reward, terminations, truncations, infos = envs.step(action.cpu().numpy())
+            next_done = np.logical_or(terminations, truncations)
+            rewards[step] = torch.tensor(reward).to(device).view(-1)
+            next_obs, next_done = torch.Tensor(next_obs).to(device), torch.Tensor(next_done).to(device)
+
+            if "final_info" in infos:
+                for info in infos["final_info"]:
+                    if info and "episode" in info:
+                        print(f"global_step={global_step}, episodic_return={info['episode']['r']}")
+                        writer.add_scalar("charts/episodic_return", info["episode"]["r"], global_step)
+                        writer.add_scalar("charts/episodic_length", info["episode"]["l"], global_step)
+
+        # bootstrap value if not done
+        with torch.no_grad():
+            next_value = agent.get_value(next_obs).reshape(1, -1)
+            advantages = torch.zeros_like(rewards).to(device)
+            lastgaelam = 0
+            for t in reversed(range(args.num_steps)):
+                if t == args.num_steps - 1:
+                    nextnonterminal = 1.0 - next_done
+                    nextvalues = next_value
+                else:
+                    nextnonterminal = 1.0 - dones[t + 1]
+                    nextvalues = values[t + 1]
+                delta = rewards[t] + args.gamma * nextvalues * nextnonterminal - values[t]
+                advantages[t] = lastgaelam = delta + args.gamma * args.gae_lambda * nextnonterminal * lastgaelam
+            returns = advantages + values
+
+        # flatten the batch
+        b_obs = obs.reshape((-1,) + envs.single_observation_space.shape)
+        b_logprobs = logprobs.reshape(-1)
+        b_actions = actions.reshape((-1,) + envs.single_action_space.shape)
+        b_advantages = advantages.reshape(-1)
+        b_returns = returns.reshape(-1)
+        b_values = values.reshape(-1)
+
+        # Optimizing the policy and value network
+        b_inds = np.arange(args.batch_size)
+        clipfracs = []
+        for epoch in range(args.update_epochs):
+            np.random.shuffle(b_inds)
+            for start in range(0, args.batch_size, args.minibatch_size):
+                end = start + args.minibatch_size
+                mb_inds = b_inds[start:end]
+
+                _, newlogprob, entropy, newvalue = agent.get_action_and_value(b_obs[mb_inds], b_actions.long()[mb_inds])
+                logratio = newlogprob - b_logprobs[mb_inds]
+                ratio = logratio.exp()
+
+                with torch.no_grad():
+                    # calculate approx_kl http://joschu.net/blog/kl-approx.html
+                    old_approx_kl = (-logratio).mean()
+                    approx_kl = ((ratio - 1) - logratio).mean()
+                    clipfracs += [((ratio - 1.0).abs() > args.clip_coef).float().mean().item()]
+
+                mb_advantages = b_advantages[mb_inds]
+                if args.norm_adv:
+                    mb_advantages = (mb_advantages - mb_advantages.mean()) / (mb_advantages.std() + 1e-8)
+
+                # Policy loss
+                pg_loss1 = -mb_advantages * ratio
+                pg_loss2 = -mb_advantages * torch.clamp(ratio, 1 - args.clip_coef, 1 + args.clip_coef)
+                pg_loss = torch.max(pg_loss1, pg_loss2).mean()
+
+                # Value loss
+                newvalue = newvalue.view(-1)
+                if args.clip_vloss:
+                    v_loss_unclipped = (newvalue - b_returns[mb_inds]) ** 2
+                    v_clipped = b_values[mb_inds] + torch.clamp(
+                        newvalue - b_values[mb_inds],
+                        -args.clip_coef,
+                        args.clip_coef,
+                    )
+                    v_loss_clipped = (v_clipped - b_returns[mb_inds]) ** 2
+                    v_loss_max = torch.max(v_loss_unclipped, v_loss_clipped)
+                    v_loss = 0.5 * v_loss_max.mean()
+                else:
+                    v_loss = 0.5 * ((newvalue - b_returns[mb_inds]) ** 2).mean()
+
+                entropy_loss = entropy.mean()
+                loss = pg_loss - args.ent_coef * entropy_loss + v_loss * args.vf_coef
+
+                optimizer.zero_grad()
+                loss.backward()
+                nn.utils.clip_grad_norm_(agent.parameters(), args.max_grad_norm)
+                optimizer.step()
+
+            if args.target_kl is not None and approx_kl > args.target_kl:
+                break
+
+        y_pred, y_true = b_values.cpu().numpy(), b_returns.cpu().numpy()
+        var_y = np.var(y_true)
+        explained_var = np.nan if var_y == 0 else 1 - np.var(y_true - y_pred) / var_y
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        writer.add_scalar("charts/learning_rate", optimizer.param_groups[0]["lr"], global_step)
+        writer.add_scalar("losses/value_loss", v_loss.item(), global_step)
+        writer.add_scalar("losses/policy_loss", pg_loss.item(), global_step)
+        writer.add_scalar("losses/entropy", entropy_loss.item(), global_step)
+        writer.add_scalar("losses/old_approx_kl", old_approx_kl.item(), global_step)
+        writer.add_scalar("losses/approx_kl", approx_kl.item(), global_step)
+        writer.add_scalar("losses/clipfrac", np.mean(clipfracs), global_step)
+        writer.add_scalar("losses/explained_variance", explained_var, global_step)
+        
+        # Additional useful metrics
+        writer.add_scalar("charts/avg_reward", rewards.mean().item(), global_step)
+        writer.add_scalar("charts/avg_value", values.mean().item(), global_step)
+        writer.add_scalar("charts/max_reward", rewards.max().item(), global_step)
+        writer.add_scalar("charts/min_reward", rewards.min().item(), global_step)
+        
+        # Console output with key metrics
+        sps = int(global_step / (time.time() - start_time))
+        progress = 100 * iteration / args.num_iterations
+        print(f"[{progress:5.1f}%] Iter {iteration:4d}/{args.num_iterations} | "
+              f"SPS: {sps:5d} | "
+              f"Reward: {rewards.mean().item():6.3f} | "
+              f"Value: {values.mean().item():6.3f} | "
+              f"VLoss: {v_loss.item():.4f} | "
+              f"PLoss: {pg_loss.item():.4f} | "
+              f"Ent: {entropy_loss.item():.4f}")
+        writer.add_scalar("charts/SPS", sps, global_step)
+
+    envs.close()
+    writer.close()
diff --git a/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/output.log b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..5762a7ce1f6c6cbe3a152518da5f80dc9c3fafce
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/output.log
@@ -0,0 +1,9765 @@
+[  0.0%] Iter    1/9765 | SPS:   375 | Reward:  0.031 | Value:  0.466 | VLoss: 0.0530 | PLoss: -0.0083 | Ent: 1.3842
+[  0.0%] Iter    2/9765 | SPS:   406 | Reward:  0.027 | Value:  0.393 | VLoss: 0.0417 | PLoss: -0.0030 | Ent: 1.3797
+[  0.0%] Iter    3/9765 | SPS:   417 | Reward:  0.036 | Value:  0.423 | VLoss: 0.0515 | PLoss: -0.0018 | Ent: 1.3743
+[  0.0%] Iter    4/9765 | SPS:   423 | Reward:  0.030 | Value:  0.396 | VLoss: 0.0423 | PLoss: -0.0048 | Ent: 1.3754
+[  0.1%] Iter    5/9765 | SPS:   427 | Reward:  0.039 | Value:  0.354 | VLoss: 0.0493 | PLoss: -0.0031 | Ent: 1.3778
+[  0.1%] Iter    6/9765 | SPS:   430 | Reward:  0.027 | Value:  0.423 | VLoss: 0.0325 | PLoss: -0.0070 | Ent: 1.3749
+[  0.1%] Iter    7/9765 | SPS:   431 | Reward:  0.034 | Value:  0.393 | VLoss: 0.0481 | PLoss: -0.0044 | Ent: 1.3717
+[  0.1%] Iter    8/9765 | SPS:   432 | Reward:  0.031 | Value:  0.378 | VLoss: 0.0426 | PLoss: -0.0019 | Ent: 1.3785
+[  0.1%] Iter    9/9765 | SPS:   433 | Reward:  0.028 | Value:  0.375 | VLoss: 0.0411 | PLoss: 0.0001 | Ent: 1.3808
+[  0.1%] Iter   10/9765 | SPS:   434 | Reward:  0.021 | Value:  0.374 | VLoss: 0.0297 | PLoss: -0.0067 | Ent: 1.3827
+[  0.1%] Iter   11/9765 | SPS:   435 | Reward:  0.027 | Value:  0.324 | VLoss: 0.0420 | PLoss: -0.0026 | Ent: 1.3840
+[  0.1%] Iter   12/9765 | SPS:   436 | Reward:  0.035 | Value:  0.350 | VLoss: 0.0454 | PLoss: -0.0039 | Ent: 1.3829
+[  0.1%] Iter   13/9765 | SPS:   436 | Reward:  0.022 | Value:  0.388 | VLoss: 0.0374 | PLoss: -0.0004 | Ent: 1.3839
+[  0.1%] Iter   14/9765 | SPS:   437 | Reward:  0.029 | Value:  0.337 | VLoss: 0.0362 | PLoss: 0.0024 | Ent: 1.3821
+[  0.2%] Iter   15/9765 | SPS:   437 | Reward:  0.030 | Value:  0.365 | VLoss: 0.0437 | PLoss: -0.0028 | Ent: 1.3767
+[  0.2%] Iter   16/9765 | SPS:   437 | Reward:  0.034 | Value:  0.468 | VLoss: 0.0382 | PLoss: -0.0029 | Ent: 1.3751
+[  0.2%] Iter   17/9765 | SPS:   441 | Reward:  0.030 | Value:  0.397 | VLoss: 0.0366 | PLoss: -0.0135 | Ent: 1.3729
+[  0.2%] Iter   18/9765 | SPS:   463 | Reward:  0.032 | Value:  0.378 | VLoss: 0.0401 | PLoss: -0.0134 | Ent: 1.3652
+[  0.2%] Iter   19/9765 | SPS:   486 | Reward:  0.026 | Value:  0.359 | VLoss: 0.0337 | PLoss: 0.0006 | Ent: 1.3418
+[  0.2%] Iter   20/9765 | SPS:   508 | Reward:  0.028 | Value:  0.381 | VLoss: 0.0392 | PLoss: -0.0037 | Ent: 1.3056
+[  0.2%] Iter   21/9765 | SPS:   530 | Reward:  0.027 | Value:  0.340 | VLoss: 0.0359 | PLoss: -0.0053 | Ent: 1.2999
+[  0.2%] Iter   22/9765 | SPS:   547 | Reward:  0.028 | Value:  0.309 | VLoss: 0.0432 | PLoss: -0.0148 | Ent: 1.3370
+[  0.2%] Iter   23/9765 | SPS:   554 | Reward:  0.032 | Value:  0.322 | VLoss: 0.0385 | PLoss: -0.0163 | Ent: 1.3688
+[  0.2%] Iter   24/9765 | SPS:   571 | Reward:  0.024 | Value:  0.373 | VLoss: 0.0355 | PLoss: -0.0056 | Ent: 1.3665
+[  0.3%] Iter   25/9765 | SPS:   591 | Reward:  0.035 | Value:  0.311 | VLoss: 0.0398 | PLoss: -0.0056 | Ent: 1.3586
+[  0.3%] Iter   26/9765 | SPS:   610 | Reward:  0.028 | Value:  0.330 | VLoss: 0.0426 | PLoss: -0.0149 | Ent: 1.3632
+[  0.3%] Iter   27/9765 | SPS:   630 | Reward:  0.018 | Value:  0.370 | VLoss: 0.0301 | PLoss: -0.0070 | Ent: 1.3577
+[  0.3%] Iter   28/9765 | SPS:   649 | Reward:  0.028 | Value:  0.280 | VLoss: 0.0460 | PLoss: -0.0050 | Ent: 1.3577
+[  0.3%] Iter   29/9765 | SPS:   662 | Reward:  0.024 | Value:  0.240 | VLoss: 0.0358 | PLoss: -0.0100 | Ent: 1.3486
+[  0.3%] Iter   30/9765 | SPS:   669 | Reward:  0.028 | Value:  0.304 | VLoss: 0.0295 | PLoss: -0.0045 | Ent: 1.3526
+[  0.3%] Iter   31/9765 | SPS:   669 | Reward:  0.028 | Value:  0.327 | VLoss: 0.0384 | PLoss: -0.0059 | Ent: 1.3353
+[  0.3%] Iter   32/9765 | SPS:   669 | Reward:  0.028 | Value:  0.389 | VLoss: 0.0336 | PLoss: -0.0050 | Ent: 1.3334
+[  0.3%] Iter   33/9765 | SPS:   667 | Reward:  0.032 | Value:  0.392 | VLoss: 0.0318 | PLoss: -0.0062 | Ent: 1.3558
+[  0.3%] Iter   34/9765 | SPS:   666 | Reward:  0.028 | Value:  0.384 | VLoss: 0.0382 | PLoss: -0.0083 | Ent: 1.3474
+[  0.4%] Iter   35/9765 | SPS:   667 | Reward:  0.037 | Value:  0.313 | VLoss: 0.0431 | PLoss: -0.0080 | Ent: 1.3287
+[  0.4%] Iter   36/9765 | SPS:   678 | Reward:  0.027 | Value:  0.350 | VLoss: 0.0372 | PLoss: -0.0062 | Ent: 1.3030
+[  0.4%] Iter   37/9765 | SPS:   693 | Reward:  0.029 | Value:  0.381 | VLoss: 0.0306 | PLoss: -0.0033 | Ent: 1.3150
+[  0.4%] Iter   38/9765 | SPS:   706 | Reward:  0.036 | Value:  0.402 | VLoss: 0.0456 | PLoss: -0.0063 | Ent: 1.2943
+[  0.4%] Iter   39/9765 | SPS:   721 | Reward:  0.021 | Value:  0.400 | VLoss: 0.0290 | PLoss: -0.0188 | Ent: 1.2727
+[  0.4%] Iter   40/9765 | SPS:   736 | Reward:  0.025 | Value:  0.405 | VLoss: 0.0304 | PLoss: -0.0088 | Ent: 1.2773
+[  0.4%] Iter   41/9765 | SPS:   751 | Reward:  0.031 | Value:  0.371 | VLoss: 0.0372 | PLoss: -0.0007 | Ent: 1.2959
+[  0.4%] Iter   42/9765 | SPS:   760 | Reward:  0.039 | Value:  0.352 | VLoss: 0.0523 | PLoss: -0.0246 | Ent: 1.3054
+[  0.4%] Iter   43/9765 | SPS:   772 | Reward:  0.031 | Value:  0.422 | VLoss: 0.0375 | PLoss: -0.0069 | Ent: 1.3234
+[  0.5%] Iter   44/9765 | SPS:   786 | Reward:  0.030 | Value:  0.378 | VLoss: 0.0403 | PLoss: -0.0057 | Ent: 1.3324
+[  0.5%] Iter   45/9765 | SPS:   799 | Reward:  0.032 | Value:  0.287 | VLoss: 0.0393 | PLoss: -0.0099 | Ent: 1.3315
+[  0.5%] Iter   46/9765 | SPS:   803 | Reward:  0.030 | Value:  0.383 | VLoss: 0.0339 | PLoss: -0.0071 | Ent: 1.3113
+[  0.5%] Iter   47/9765 | SPS:   811 | Reward:  0.043 | Value:  0.353 | VLoss: 0.0418 | PLoss: -0.0051 | Ent: 1.2765
+[  0.5%] Iter   48/9765 | SPS:   820 | Reward:  0.033 | Value:  0.428 | VLoss: 0.0366 | PLoss: -0.0055 | Ent: 1.2941
+[  0.5%] Iter   49/9765 | SPS:   828 | Reward:  0.031 | Value:  0.365 | VLoss: 0.0376 | PLoss: 0.0015 | Ent: 1.3067
+[  0.5%] Iter   50/9765 | SPS:   836 | Reward:  0.036 | Value:  0.366 | VLoss: 0.0435 | PLoss: -0.0107 | Ent: 1.3396
+[  0.5%] Iter   51/9765 | SPS:   843 | Reward:  0.035 | Value:  0.320 | VLoss: 0.0416 | PLoss: -0.0171 | Ent: 1.3486
+[  0.5%] Iter   52/9765 | SPS:   851 | Reward:  0.033 | Value:  0.344 | VLoss: 0.0424 | PLoss: -0.0034 | Ent: 1.3479
+[  0.5%] Iter   53/9765 | SPS:   858 | Reward:  0.033 | Value:  0.348 | VLoss: 0.0438 | PLoss: -0.0070 | Ent: 1.3412
+[  0.6%] Iter   54/9765 | SPS:   865 | Reward:  0.032 | Value:  0.438 | VLoss: 0.0435 | PLoss: -0.0128 | Ent: 1.3335
+[  0.6%] Iter   55/9765 | SPS:   872 | Reward:  0.034 | Value:  0.361 | VLoss: 0.0271 | PLoss: -0.0100 | Ent: 1.3243
+[  0.6%] Iter   56/9765 | SPS:   879 | Reward:  0.035 | Value:  0.399 | VLoss: 0.0388 | PLoss: -0.0062 | Ent: 1.2528
+[  0.6%] Iter   57/9765 | SPS:   885 | Reward:  0.027 | Value:  0.403 | VLoss: 0.0326 | PLoss: -0.0114 | Ent: 1.2199
+[  0.6%] Iter   58/9765 | SPS:   891 | Reward:  0.034 | Value:  0.401 | VLoss: 0.0444 | PLoss: -0.0009 | Ent: 1.2590
+[  0.6%] Iter   59/9765 | SPS:   897 | Reward:  0.031 | Value:  0.453 | VLoss: 0.0371 | PLoss: -0.0107 | Ent: 1.2937
+[  0.6%] Iter   60/9765 | SPS:   902 | Reward:  0.030 | Value:  0.472 | VLoss: 0.0449 | PLoss: -0.0020 | Ent: 1.2907
+[  0.6%] Iter   61/9765 | SPS:   907 | Reward:  0.040 | Value:  0.387 | VLoss: 0.0462 | PLoss: -0.0023 | Ent: 1.2669
+[  0.6%] Iter   62/9765 | SPS:   911 | Reward:  0.034 | Value:  0.421 | VLoss: 0.0471 | PLoss: -0.0068 | Ent: 1.2756
+[  0.6%] Iter   63/9765 | SPS:   915 | Reward:  0.041 | Value:  0.396 | VLoss: 0.0344 | PLoss: -0.0027 | Ent: 1.2795
+[  0.7%] Iter   64/9765 | SPS:   918 | Reward:  0.036 | Value:  0.446 | VLoss: 0.0312 | PLoss: -0.0075 | Ent: 1.2620
+[  0.7%] Iter   65/9765 | SPS:   920 | Reward:  0.036 | Value:  0.538 | VLoss: 0.0395 | PLoss: -0.0201 | Ent: 1.2811
+[  0.7%] Iter   66/9765 | SPS:   922 | Reward:  0.025 | Value:  0.437 | VLoss: 0.0304 | PLoss: -0.0060 | Ent: 1.2989
+[  0.7%] Iter   67/9765 | SPS:   923 | Reward:  0.035 | Value:  0.366 | VLoss: 0.0309 | PLoss: -0.0235 | Ent: 1.3138
+[  0.7%] Iter   68/9765 | SPS:   924 | Reward:  0.037 | Value:  0.440 | VLoss: 0.0373 | PLoss: 0.0056 | Ent: 1.3050
+[  0.7%] Iter   69/9765 | SPS:   923 | Reward:  0.040 | Value:  0.488 | VLoss: 0.0463 | PLoss: -0.0210 | Ent: 1.2829
+[  0.7%] Iter   70/9765 | SPS:   921 | Reward:  0.041 | Value:  0.440 | VLoss: 0.0485 | PLoss: -0.0164 | Ent: 1.2863
+[  0.7%] Iter   71/9765 | SPS:   919 | Reward:  0.039 | Value:  0.448 | VLoss: 0.0330 | PLoss: -0.0114 | Ent: 1.2324
+[  0.7%] Iter   72/9765 | SPS:   917 | Reward:  0.045 | Value:  0.416 | VLoss: 0.0458 | PLoss: -0.0059 | Ent: 1.1997
+[  0.7%] Iter   73/9765 | SPS:   914 | Reward:  0.034 | Value:  0.523 | VLoss: 0.0382 | PLoss: -0.0110 | Ent: 1.2219
+[  0.8%] Iter   74/9765 | SPS:   912 | Reward:  0.042 | Value:  0.441 | VLoss: 0.0430 | PLoss: -0.0070 | Ent: 1.2393
+[  0.8%] Iter   75/9765 | SPS:   910 | Reward:  0.027 | Value:  0.424 | VLoss: 0.0441 | PLoss: -0.0147 | Ent: 1.2361
+[  0.8%] Iter   76/9765 | SPS:   908 | Reward:  0.044 | Value:  0.427 | VLoss: 0.0315 | PLoss: -0.0147 | Ent: 1.1996
+[  0.8%] Iter   77/9765 | SPS:   906 | Reward:  0.042 | Value:  0.436 | VLoss: 0.0528 | PLoss: -0.0228 | Ent: 1.2141
+[  0.8%] Iter   78/9765 | SPS:   904 | Reward:  0.031 | Value:  0.510 | VLoss: 0.0338 | PLoss: -0.0108 | Ent: 1.2145
+[  0.8%] Iter   79/9765 | SPS:   902 | Reward:  0.030 | Value:  0.464 | VLoss: 0.0379 | PLoss: -0.0138 | Ent: 1.1870
+[  0.8%] Iter   80/9765 | SPS:   901 | Reward:  0.034 | Value:  0.458 | VLoss: 0.0353 | PLoss: -0.0229 | Ent: 1.1987
+[  0.8%] Iter   81/9765 | SPS:   899 | Reward:  0.043 | Value:  0.406 | VLoss: 0.0446 | PLoss: -0.0050 | Ent: 1.1450
+[  0.8%] Iter   82/9765 | SPS:   897 | Reward:  0.043 | Value:  0.423 | VLoss: 0.0497 | PLoss: -0.0120 | Ent: 1.1233
+[  0.8%] Iter   83/9765 | SPS:   896 | Reward:  0.029 | Value:  0.500 | VLoss: 0.0432 | PLoss: -0.0107 | Ent: 1.1751
+[  0.9%] Iter   84/9765 | SPS:   894 | Reward:  0.030 | Value:  0.483 | VLoss: 0.0381 | PLoss: -0.0255 | Ent: 1.1693
+[  0.9%] Iter   85/9765 | SPS:   893 | Reward:  0.044 | Value:  0.467 | VLoss: 0.0413 | PLoss: -0.0088 | Ent: 1.1573
+[  0.9%] Iter   86/9765 | SPS:   892 | Reward:  0.024 | Value:  0.533 | VLoss: 0.0295 | PLoss: -0.0141 | Ent: 1.1566
+[  0.9%] Iter   87/9765 | SPS:   900 | Reward:  0.037 | Value:  0.474 | VLoss: 0.0429 | PLoss: -0.0096 | Ent: 1.1542
+[  0.9%] Iter   88/9765 | SPS:   908 | Reward:  0.041 | Value:  0.468 | VLoss: 0.0444 | PLoss: -0.0246 | Ent: 1.1646
+[  0.9%] Iter   89/9765 | SPS:   916 | Reward:  0.041 | Value:  0.485 | VLoss: 0.0405 | PLoss: -0.0201 | Ent: 1.1034
+[  0.9%] Iter   90/9765 | SPS:   923 | Reward:  0.035 | Value:  0.513 | VLoss: 0.0425 | PLoss: -0.0218 | Ent: 1.1453
+[  0.9%] Iter   91/9765 | SPS:   931 | Reward:  0.042 | Value:  0.502 | VLoss: 0.0363 | PLoss: -0.0203 | Ent: 1.1466
+[  0.9%] Iter   92/9765 | SPS:   938 | Reward:  0.042 | Value:  0.451 | VLoss: 0.0413 | PLoss: -0.0033 | Ent: 1.0917
+[  1.0%] Iter   93/9765 | SPS:   946 | Reward:  0.037 | Value:  0.488 | VLoss: 0.0345 | PLoss: -0.0142 | Ent: 1.0513
+[  1.0%] Iter   94/9765 | SPS:   950 | Reward:  0.051 | Value:  0.426 | VLoss: 0.0549 | PLoss: -0.0137 | Ent: 1.0898
+[  1.0%] Iter   95/9765 | SPS:   948 | Reward:  0.034 | Value:  0.424 | VLoss: 0.0374 | PLoss: -0.0067 | Ent: 0.9986
+[  1.0%] Iter   96/9765 | SPS:   955 | Reward:  0.042 | Value:  0.464 | VLoss: 0.0384 | PLoss: -0.0171 | Ent: 0.9802
+[  1.0%] Iter   97/9765 | SPS:   962 | Reward:  0.028 | Value:  0.507 | VLoss: 0.0352 | PLoss: -0.0039 | Ent: 0.9926
+[  1.0%] Iter   98/9765 | SPS:   969 | Reward:  0.040 | Value:  0.433 | VLoss: 0.0391 | PLoss: -0.0034 | Ent: 1.0790
+[  1.0%] Iter   99/9765 | SPS:   977 | Reward:  0.036 | Value:  0.451 | VLoss: 0.0324 | PLoss: -0.0221 | Ent: 1.0553
+[  1.0%] Iter  100/9765 | SPS:   984 | Reward:  0.042 | Value:  0.507 | VLoss: 0.0372 | PLoss: -0.0144 | Ent: 1.0529
+[  1.0%] Iter  101/9765 | SPS:   985 | Reward:  0.034 | Value:  0.475 | VLoss: 0.0375 | PLoss: -0.0114 | Ent: 1.0030
+[  1.0%] Iter  102/9765 | SPS:   983 | Reward:  0.056 | Value:  0.532 | VLoss: 0.0497 | PLoss: -0.0053 | Ent: 1.0501
+[  1.1%] Iter  103/9765 | SPS:   980 | Reward:  0.048 | Value:  0.537 | VLoss: 0.0414 | PLoss: -0.0158 | Ent: 1.0170
+[  1.1%] Iter  104/9765 | SPS:   975 | Reward:  0.026 | Value:  0.494 | VLoss: 0.0264 | PLoss: -0.0140 | Ent: 1.0054
+[  1.1%] Iter  105/9765 | SPS:   971 | Reward:  0.026 | Value:  0.439 | VLoss: 0.0314 | PLoss: -0.0115 | Ent: 0.9745
+[  1.1%] Iter  106/9765 | SPS:   967 | Reward:  0.039 | Value:  0.431 | VLoss: 0.0466 | PLoss: -0.0075 | Ent: 0.9913
+[  1.1%] Iter  107/9765 | SPS:   964 | Reward:  0.046 | Value:  0.477 | VLoss: 0.0345 | PLoss: -0.0021 | Ent: 1.0250
+[  1.1%] Iter  108/9765 | SPS:   970 | Reward:  0.033 | Value:  0.460 | VLoss: 0.0305 | PLoss: -0.0009 | Ent: 0.9960
+[  1.1%] Iter  109/9765 | SPS:   976 | Reward:  0.033 | Value:  0.476 | VLoss: 0.0314 | PLoss: -0.0197 | Ent: 0.9579
+[  1.1%] Iter  110/9765 | SPS:   982 | Reward:  0.033 | Value:  0.460 | VLoss: 0.0319 | PLoss: -0.0080 | Ent: 0.9467
+[  1.1%] Iter  111/9765 | SPS:   989 | Reward:  0.038 | Value:  0.447 | VLoss: 0.0445 | PLoss: -0.0058 | Ent: 0.9781
+[  1.1%] Iter  112/9765 | SPS:   995 | Reward:  0.044 | Value:  0.513 | VLoss: 0.0300 | PLoss: -0.0152 | Ent: 0.9995
+[  1.2%] Iter  113/9765 | SPS:  1001 | Reward:  0.044 | Value:  0.511 | VLoss: 0.0388 | PLoss: -0.0152 | Ent: 0.9791
+[  1.2%] Iter  114/9765 | SPS:  1008 | Reward:  0.038 | Value:  0.457 | VLoss: 0.0399 | PLoss: -0.0084 | Ent: 0.8957
+[  1.2%] Iter  115/9765 | SPS:  1014 | Reward:  0.034 | Value:  0.462 | VLoss: 0.0423 | PLoss: -0.0146 | Ent: 0.9025
+[  1.2%] Iter  116/9765 | SPS:  1021 | Reward:  0.033 | Value:  0.526 | VLoss: 0.0365 | PLoss: -0.0123 | Ent: 0.9104
+[  1.2%] Iter  117/9765 | SPS:  1027 | Reward:  0.037 | Value:  0.446 | VLoss: 0.0367 | PLoss: -0.0034 | Ent: 0.9011
+[  1.2%] Iter  118/9765 | SPS:  1033 | Reward:  0.050 | Value:  0.468 | VLoss: 0.0473 | PLoss: -0.0060 | Ent: 0.8826
+[  1.2%] Iter  119/9765 | SPS:  1038 | Reward:  0.053 | Value:  0.540 | VLoss: 0.0423 | PLoss: -0.0099 | Ent: 0.9037
+[  1.2%] Iter  120/9765 | SPS:  1043 | Reward:  0.040 | Value:  0.529 | VLoss: 0.0332 | PLoss: -0.0123 | Ent: 0.9141
+[  1.2%] Iter  121/9765 | SPS:  1049 | Reward:  0.045 | Value:  0.532 | VLoss: 0.0359 | PLoss: -0.0238 | Ent: 0.8948
+[  1.2%] Iter  122/9765 | SPS:  1055 | Reward:  0.039 | Value:  0.551 | VLoss: 0.0371 | PLoss: -0.0177 | Ent: 0.9078
+[  1.3%] Iter  123/9765 | SPS:  1061 | Reward:  0.051 | Value:  0.503 | VLoss: 0.0502 | PLoss: -0.0100 | Ent: 0.9172
+[  1.3%] Iter  124/9765 | SPS:  1061 | Reward:  0.052 | Value:  0.477 | VLoss: 0.0394 | PLoss: -0.0082 | Ent: 0.8769
+[  1.3%] Iter  125/9765 | SPS:  1065 | Reward:  0.035 | Value:  0.563 | VLoss: 0.0278 | PLoss: -0.0110 | Ent: 0.9246
+[  1.3%] Iter  126/9765 | SPS:  1068 | Reward:  0.034 | Value:  0.503 | VLoss: 0.0379 | PLoss: -0.0165 | Ent: 0.8693
+[  1.3%] Iter  127/9765 | SPS:  1071 | Reward:  0.042 | Value:  0.497 | VLoss: 0.0340 | PLoss: -0.0116 | Ent: 0.8824
+[  1.3%] Iter  128/9765 | SPS:  1075 | Reward:  0.039 | Value:  0.487 | VLoss: 0.0425 | PLoss: -0.0129 | Ent: 0.9090
+[  1.3%] Iter  129/9765 | SPS:  1078 | Reward:  0.042 | Value:  0.502 | VLoss: 0.0352 | PLoss: -0.0124 | Ent: 0.8512
+[  1.3%] Iter  130/9765 | SPS:  1081 | Reward:  0.039 | Value:  0.524 | VLoss: 0.0352 | PLoss: -0.0017 | Ent: 0.8078
+[  1.3%] Iter  131/9765 | SPS:  1084 | Reward:  0.031 | Value:  0.480 | VLoss: 0.0319 | PLoss: -0.0227 | Ent: 0.8341
+[  1.4%] Iter  132/9765 | SPS:  1087 | Reward:  0.049 | Value:  0.445 | VLoss: 0.0454 | PLoss: -0.0234 | Ent: 0.8294
+[  1.4%] Iter  133/9765 | SPS:  1090 | Reward:  0.034 | Value:  0.491 | VLoss: 0.0281 | PLoss: -0.0057 | Ent: 0.7795
+[  1.4%] Iter  134/9765 | SPS:  1093 | Reward:  0.038 | Value:  0.555 | VLoss: 0.0396 | PLoss: -0.0038 | Ent: 0.8513
+[  1.4%] Iter  135/9765 | SPS:  1096 | Reward:  0.056 | Value:  0.493 | VLoss: 0.0456 | PLoss: -0.0125 | Ent: 0.8636
+[  1.4%] Iter  136/9765 | SPS:  1098 | Reward:  0.035 | Value:  0.511 | VLoss: 0.0380 | PLoss: -0.0076 | Ent: 0.7887
+[  1.4%] Iter  137/9765 | SPS:  1101 | Reward:  0.043 | Value:  0.456 | VLoss: 0.0390 | PLoss: -0.0142 | Ent: 0.8064
+[  1.4%] Iter  138/9765 | SPS:  1104 | Reward:  0.049 | Value:  0.505 | VLoss: 0.0347 | PLoss: -0.0104 | Ent: 0.8701
+[  1.4%] Iter  139/9765 | SPS:  1107 | Reward:  0.046 | Value:  0.539 | VLoss: 0.0389 | PLoss: -0.0108 | Ent: 0.7671
+[  1.4%] Iter  140/9765 | SPS:  1110 | Reward:  0.058 | Value:  0.560 | VLoss: 0.0334 | PLoss: -0.0103 | Ent: 0.7845
+[  1.4%] Iter  141/9765 | SPS:  1112 | Reward:  0.050 | Value:  0.628 | VLoss: 0.0379 | PLoss: -0.0120 | Ent: 0.7761
+[  1.5%] Iter  142/9765 | SPS:  1114 | Reward:  0.052 | Value:  0.659 | VLoss: 0.0358 | PLoss: -0.0180 | Ent: 0.7334
+[  1.5%] Iter  143/9765 | SPS:  1117 | Reward:  0.038 | Value:  0.586 | VLoss: 0.0400 | PLoss: -0.0235 | Ent: 0.8207
+[  1.5%] Iter  144/9765 | SPS:  1119 | Reward:  0.040 | Value:  0.548 | VLoss: 0.0346 | PLoss: -0.0103 | Ent: 0.8206
+[  1.5%] Iter  145/9765 | SPS:  1121 | Reward:  0.041 | Value:  0.557 | VLoss: 0.0301 | PLoss: -0.0112 | Ent: 0.6800
+[  1.5%] Iter  146/9765 | SPS:  1123 | Reward:  0.052 | Value:  0.577 | VLoss: 0.0380 | PLoss: -0.0041 | Ent: 0.8023
+[  1.5%] Iter  147/9765 | SPS:  1125 | Reward:  0.046 | Value:  0.534 | VLoss: 0.0432 | PLoss: -0.0083 | Ent: 0.8077
+[  1.5%] Iter  148/9765 | SPS:  1127 | Reward:  0.039 | Value:  0.518 | VLoss: 0.0310 | PLoss: -0.0091 | Ent: 0.7877
+[  1.5%] Iter  149/9765 | SPS:  1129 | Reward:  0.036 | Value:  0.518 | VLoss: 0.0325 | PLoss: -0.0123 | Ent: 0.7451
+[  1.5%] Iter  150/9765 | SPS:  1130 | Reward:  0.051 | Value:  0.555 | VLoss: 0.0288 | PLoss: -0.0125 | Ent: 0.6729
+[  1.5%] Iter  151/9765 | SPS:  1136 | Reward:  0.043 | Value:  0.510 | VLoss: 0.0371 | PLoss: -0.0173 | Ent: 0.6994
+[  1.6%] Iter  152/9765 | SPS:  1139 | Reward:  0.055 | Value:  0.571 | VLoss: 0.0345 | PLoss: -0.0112 | Ent: 0.7023
+[  1.6%] Iter  153/9765 | SPS:  1140 | Reward:  0.047 | Value:  0.612 | VLoss: 0.0357 | PLoss: -0.0142 | Ent: 0.7364
+[  1.6%] Iter  154/9765 | SPS:  1140 | Reward:  0.043 | Value:  0.612 | VLoss: 0.0279 | PLoss: -0.0181 | Ent: 0.7060
+[  1.6%] Iter  155/9765 | SPS:  1140 | Reward:  0.031 | Value:  0.549 | VLoss: 0.0271 | PLoss: -0.0076 | Ent: 0.7330
+[  1.6%] Iter  156/9765 | SPS:  1140 | Reward:  0.037 | Value:  0.531 | VLoss: 0.0297 | PLoss: -0.0082 | Ent: 0.6087
+[  1.6%] Iter  157/9765 | SPS:  1139 | Reward:  0.045 | Value:  0.488 | VLoss: 0.0374 | PLoss: -0.0129 | Ent: 0.7061
+[  1.6%] Iter  158/9765 | SPS:  1137 | Reward:  0.044 | Value:  0.526 | VLoss: 0.0313 | PLoss: -0.0120 | Ent: 0.7759
+[  1.6%] Iter  159/9765 | SPS:  1136 | Reward:  0.047 | Value:  0.503 | VLoss: 0.0351 | PLoss: -0.0153 | Ent: 0.7827
+[  1.6%] Iter  160/9765 | SPS:  1134 | Reward:  0.052 | Value:  0.587 | VLoss: 0.0385 | PLoss: -0.0201 | Ent: 0.7429
+[  1.6%] Iter  161/9765 | SPS:  1133 | Reward:  0.046 | Value:  0.615 | VLoss: 0.0364 | PLoss: -0.0092 | Ent: 0.6633
+[  1.7%] Iter  162/9765 | SPS:  1132 | Reward:  0.045 | Value:  0.520 | VLoss: 0.0310 | PLoss: -0.0047 | Ent: 0.7111
+[  1.7%] Iter  163/9765 | SPS:  1130 | Reward:  0.052 | Value:  0.503 | VLoss: 0.0408 | PLoss: -0.0239 | Ent: 0.6668
+[  1.7%] Iter  164/9765 | SPS:  1128 | Reward:  0.046 | Value:  0.523 | VLoss: 0.0533 | PLoss: -0.0104 | Ent: 0.6736
+[  1.7%] Iter  165/9765 | SPS:  1126 | Reward:  0.067 | Value:  0.601 | VLoss: 0.0385 | PLoss: -0.0087 | Ent: 0.6508
+[  1.7%] Iter  166/9765 | SPS:  1124 | Reward:  0.043 | Value:  0.569 | VLoss: 0.0337 | PLoss: -0.0119 | Ent: 0.7380
+[  1.7%] Iter  167/9765 | SPS:  1122 | Reward:  0.045 | Value:  0.587 | VLoss: 0.0386 | PLoss: -0.0078 | Ent: 0.6354
+[  1.7%] Iter  168/9765 | SPS:  1120 | Reward:  0.044 | Value:  0.568 | VLoss: 0.0419 | PLoss: -0.0128 | Ent: 0.6511
+[  1.7%] Iter  169/9765 | SPS:  1119 | Reward:  0.031 | Value:  0.486 | VLoss: 0.0327 | PLoss: -0.0185 | Ent: 0.6512
+[  1.7%] Iter  170/9765 | SPS:  1116 | Reward:  0.054 | Value:  0.494 | VLoss: 0.0464 | PLoss: -0.0116 | Ent: 0.6427
+[  1.8%] Iter  171/9765 | SPS:  1113 | Reward:  0.050 | Value:  0.528 | VLoss: 0.0381 | PLoss: -0.0066 | Ent: 0.6912
+[  1.8%] Iter  172/9765 | SPS:  1115 | Reward:  0.061 | Value:  0.547 | VLoss: 0.0443 | PLoss: -0.0123 | Ent: 0.7168
+[  1.8%] Iter  173/9765 | SPS:  1119 | Reward:  0.045 | Value:  0.572 | VLoss: 0.0318 | PLoss: -0.0138 | Ent: 0.6622
+[  1.8%] Iter  174/9765 | SPS:  1124 | Reward:  0.053 | Value:  0.569 | VLoss: 0.0306 | PLoss: -0.0118 | Ent: 0.6160
+[  1.8%] Iter  175/9765 | SPS:  1128 | Reward:  0.051 | Value:  0.605 | VLoss: 0.0343 | PLoss: -0.0124 | Ent: 0.7111
+[  1.8%] Iter  176/9765 | SPS:  1125 | Reward:  0.054 | Value:  0.582 | VLoss: 0.0377 | PLoss: 0.0058 | Ent: 0.6562
+[  1.8%] Iter  177/9765 | SPS:  1130 | Reward:  0.052 | Value:  0.585 | VLoss: 0.0270 | PLoss: -0.0087 | Ent: 0.6619
+[  1.8%] Iter  178/9765 | SPS:  1134 | Reward:  0.045 | Value:  0.610 | VLoss: 0.0390 | PLoss: -0.0096 | Ent: 0.6521
+[  1.8%] Iter  179/9765 | SPS:  1138 | Reward:  0.034 | Value:  0.553 | VLoss: 0.0312 | PLoss: -0.0120 | Ent: 0.6953
+[  1.8%] Iter  180/9765 | SPS:  1143 | Reward:  0.032 | Value:  0.520 | VLoss: 0.0350 | PLoss: -0.0134 | Ent: 0.6962
+[  1.9%] Iter  181/9765 | SPS:  1147 | Reward:  0.053 | Value:  0.539 | VLoss: 0.0478 | PLoss: -0.0049 | Ent: 0.6262
+[  1.9%] Iter  182/9765 | SPS:  1152 | Reward:  0.038 | Value:  0.536 | VLoss: 0.0331 | PLoss: -0.0118 | Ent: 0.5963
+[  1.9%] Iter  183/9765 | SPS:  1156 | Reward:  0.056 | Value:  0.616 | VLoss: 0.0384 | PLoss: -0.0033 | Ent: 0.5767
+[  1.9%] Iter  184/9765 | SPS:  1161 | Reward:  0.040 | Value:  0.639 | VLoss: 0.0250 | PLoss: -0.0052 | Ent: 0.6316
+[  1.9%] Iter  185/9765 | SPS:  1165 | Reward:  0.047 | Value:  0.550 | VLoss: 0.0368 | PLoss: -0.0118 | Ent: 0.6463
+[  1.9%] Iter  186/9765 | SPS:  1168 | Reward:  0.050 | Value:  0.481 | VLoss: 0.0416 | PLoss: -0.0158 | Ent: 0.6228
+[  1.9%] Iter  187/9765 | SPS:  1171 | Reward:  0.058 | Value:  0.564 | VLoss: 0.0354 | PLoss: -0.0076 | Ent: 0.5501
+[  1.9%] Iter  188/9765 | SPS:  1176 | Reward:  0.048 | Value:  0.580 | VLoss: 0.0340 | PLoss: -0.0059 | Ent: 0.5909
+[  1.9%] Iter  189/9765 | SPS:  1179 | Reward:  0.048 | Value:  0.592 | VLoss: 0.0254 | PLoss: -0.0062 | Ent: 0.6446
+[  1.9%] Iter  190/9765 | SPS:  1183 | Reward:  0.041 | Value:  0.646 | VLoss: 0.0362 | PLoss: -0.0175 | Ent: 0.6774
+[  2.0%] Iter  191/9765 | SPS:  1184 | Reward:  0.054 | Value:  0.625 | VLoss: 0.0268 | PLoss: -0.0150 | Ent: 0.6336
+[  2.0%] Iter  192/9765 | SPS:  1186 | Reward:  0.034 | Value:  0.634 | VLoss: 0.0272 | PLoss: 0.0026 | Ent: 0.6669
+[  2.0%] Iter  193/9765 | SPS:  1188 | Reward:  0.059 | Value:  0.636 | VLoss: 0.0328 | PLoss: 0.0005 | Ent: 0.6300
+[  2.0%] Iter  194/9765 | SPS:  1189 | Reward:  0.055 | Value:  0.584 | VLoss: 0.0369 | PLoss: -0.0070 | Ent: 0.6641
+[  2.0%] Iter  195/9765 | SPS:  1191 | Reward:  0.051 | Value:  0.614 | VLoss: 0.0366 | PLoss: -0.0146 | Ent: 0.6700
+[  2.0%] Iter  196/9765 | SPS:  1193 | Reward:  0.042 | Value:  0.531 | VLoss: 0.0348 | PLoss: -0.0183 | Ent: 0.5487
+[  2.0%] Iter  197/9765 | SPS:  1194 | Reward:  0.033 | Value:  0.556 | VLoss: 0.0253 | PLoss: -0.0166 | Ent: 0.5540
+[  2.0%] Iter  198/9765 | SPS:  1196 | Reward:  0.036 | Value:  0.470 | VLoss: 0.0287 | PLoss: -0.0065 | Ent: 0.6018
+[  2.0%] Iter  199/9765 | SPS:  1198 | Reward:  0.047 | Value:  0.448 | VLoss: 0.0376 | PLoss: -0.0017 | Ent: 0.6527
+[  2.0%] Iter  200/9765 | SPS:  1199 | Reward:  0.058 | Value:  0.503 | VLoss: 0.0417 | PLoss: -0.0111 | Ent: 0.5806
+[  2.1%] Iter  201/9765 | SPS:  1200 | Reward:  0.062 | Value:  0.593 | VLoss: 0.0400 | PLoss: -0.0130 | Ent: 0.6128
+[  2.1%] Iter  202/9765 | SPS:  1202 | Reward:  0.042 | Value:  0.608 | VLoss: 0.0371 | PLoss: -0.0065 | Ent: 0.6219
+[  2.1%] Iter  203/9765 | SPS:  1203 | Reward:  0.057 | Value:  0.507 | VLoss: 0.0415 | PLoss: -0.0155 | Ent: 0.6239
+[  2.1%] Iter  204/9765 | SPS:  1204 | Reward:  0.054 | Value:  0.525 | VLoss: 0.0401 | PLoss: -0.0132 | Ent: 0.6133
+[  2.1%] Iter  205/9765 | SPS:  1205 | Reward:  0.060 | Value:  0.559 | VLoss: 0.0477 | PLoss: -0.0151 | Ent: 0.5862
+[  2.1%] Iter  206/9765 | SPS:  1206 | Reward:  0.051 | Value:  0.552 | VLoss: 0.0441 | PLoss: -0.0075 | Ent: 0.6115
+[  2.1%] Iter  207/9765 | SPS:  1207 | Reward:  0.047 | Value:  0.573 | VLoss: 0.0307 | PLoss: -0.0057 | Ent: 0.6028
+[  2.1%] Iter  208/9765 | SPS:  1208 | Reward:  0.038 | Value:  0.632 | VLoss: 0.0246 | PLoss: -0.0025 | Ent: 0.5672
+[  2.1%] Iter  209/9765 | SPS:  1208 | Reward:  0.041 | Value:  0.525 | VLoss: 0.0380 | PLoss: -0.0118 | Ent: 0.5818
+[  2.2%] Iter  210/9765 | SPS:  1208 | Reward:  0.041 | Value:  0.567 | VLoss: 0.0264 | PLoss: -0.0156 | Ent: 0.5550
+[  2.2%] Iter  211/9765 | SPS:  1208 | Reward:  0.039 | Value:  0.592 | VLoss: 0.0339 | PLoss: -0.0096 | Ent: 0.6281
+[  2.2%] Iter  212/9765 | SPS:  1207 | Reward:  0.054 | Value:  0.560 | VLoss: 0.0415 | PLoss: -0.0025 | Ent: 0.6226
+[  2.2%] Iter  213/9765 | SPS:  1206 | Reward:  0.058 | Value:  0.562 | VLoss: 0.0481 | PLoss: -0.0183 | Ent: 0.6205
+[  2.2%] Iter  214/9765 | SPS:  1205 | Reward:  0.062 | Value:  0.568 | VLoss: 0.0395 | PLoss: -0.0081 | Ent: 0.5471
+[  2.2%] Iter  215/9765 | SPS:  1204 | Reward:  0.053 | Value:  0.588 | VLoss: 0.0428 | PLoss: -0.0114 | Ent: 0.5354
+[  2.2%] Iter  216/9765 | SPS:  1202 | Reward:  0.042 | Value:  0.609 | VLoss: 0.0331 | PLoss: -0.0081 | Ent: 0.6119
+[  2.2%] Iter  217/9765 | SPS:  1201 | Reward:  0.045 | Value:  0.589 | VLoss: 0.0226 | PLoss: -0.0195 | Ent: 0.6061
+[  2.2%] Iter  218/9765 | SPS:  1199 | Reward:  0.056 | Value:  0.571 | VLoss: 0.0340 | PLoss: -0.0114 | Ent: 0.6418
+[  2.2%] Iter  219/9765 | SPS:  1197 | Reward:  0.036 | Value:  0.640 | VLoss: 0.0253 | PLoss: -0.0027 | Ent: 0.5947
+[  2.3%] Iter  220/9765 | SPS:  1195 | Reward:  0.037 | Value:  0.547 | VLoss: 0.0246 | PLoss: -0.0141 | Ent: 0.6563
+[  2.3%] Iter  221/9765 | SPS:  1192 | Reward:  0.051 | Value:  0.560 | VLoss: 0.0286 | PLoss: -0.0093 | Ent: 0.5652
+[  2.3%] Iter  222/9765 | SPS:  1195 | Reward:  0.057 | Value:  0.627 | VLoss: 0.0354 | PLoss: -0.0164 | Ent: 0.5108
+[  2.3%] Iter  223/9765 | SPS:  1198 | Reward:  0.053 | Value:  0.613 | VLoss: 0.0314 | PLoss: -0.0129 | Ent: 0.6281
+[  2.3%] Iter  224/9765 | SPS:  1202 | Reward:  0.052 | Value:  0.659 | VLoss: 0.0341 | PLoss: -0.0030 | Ent: 0.5212
+[  2.3%] Iter  225/9765 | SPS:  1202 | Reward:  0.047 | Value:  0.637 | VLoss: 0.0364 | PLoss: -0.0048 | Ent: 0.5818
+[  2.3%] Iter  226/9765 | SPS:  1193 | Reward:  0.056 | Value:  0.589 | VLoss: 0.0344 | PLoss: -0.0138 | Ent: 0.5828
+[  2.3%] Iter  227/9765 | SPS:  1184 | Reward:  0.050 | Value:  0.584 | VLoss: 0.0357 | PLoss: -0.0129 | Ent: 0.6537
+[  2.3%] Iter  228/9765 | SPS:  1176 | Reward:  0.048 | Value:  0.599 | VLoss: 0.0407 | PLoss: -0.0049 | Ent: 0.6600
+[  2.3%] Iter  229/9765 | SPS:  1167 | Reward:  0.055 | Value:  0.577 | VLoss: 0.0251 | PLoss: -0.0146 | Ent: 0.5585
+[  2.4%] Iter  230/9765 | SPS:  1159 | Reward:  0.052 | Value:  0.586 | VLoss: 0.0345 | PLoss: -0.0059 | Ent: 0.6434
+[  2.4%] Iter  231/9765 | SPS:  1151 | Reward:  0.060 | Value:  0.568 | VLoss: 0.0380 | PLoss: -0.0026 | Ent: 0.5618
+[  2.4%] Iter  232/9765 | SPS:  1149 | Reward:  0.050 | Value:  0.576 | VLoss: 0.0364 | PLoss: -0.0063 | Ent: 0.6303
+[  2.4%] Iter  233/9765 | SPS:  1152 | Reward:  0.037 | Value:  0.550 | VLoss: 0.0250 | PLoss: -0.0041 | Ent: 0.5539
+[  2.4%] Iter  234/9765 | SPS:  1156 | Reward:  0.046 | Value:  0.616 | VLoss: 0.0394 | PLoss: -0.0128 | Ent: 0.6187
+[  2.4%] Iter  235/9765 | SPS:  1159 | Reward:  0.046 | Value:  0.583 | VLoss: 0.0273 | PLoss: -0.0153 | Ent: 0.5876
+[  2.4%] Iter  236/9765 | SPS:  1161 | Reward:  0.048 | Value:  0.544 | VLoss: 0.0289 | PLoss: -0.0146 | Ent: 0.6355
+[  2.4%] Iter  237/9765 | SPS:  1160 | Reward:  0.037 | Value:  0.591 | VLoss: 0.0250 | PLoss: -0.0032 | Ent: 0.6189
+[  2.4%] Iter  238/9765 | SPS:  1161 | Reward:  0.063 | Value:  0.557 | VLoss: 0.0498 | PLoss: -0.0059 | Ent: 0.6090
+[  2.4%] Iter  239/9765 | SPS:  1164 | Reward:  0.055 | Value:  0.542 | VLoss: 0.0393 | PLoss: -0.0076 | Ent: 0.5942
+[  2.5%] Iter  240/9765 | SPS:  1168 | Reward:  0.057 | Value:  0.552 | VLoss: 0.0368 | PLoss: -0.0211 | Ent: 0.5995
+[  2.5%] Iter  241/9765 | SPS:  1171 | Reward:  0.058 | Value:  0.550 | VLoss: 0.0319 | PLoss: -0.0070 | Ent: 0.5640
+[  2.5%] Iter  242/9765 | SPS:  1173 | Reward:  0.062 | Value:  0.607 | VLoss: 0.0393 | PLoss: -0.0118 | Ent: 0.5815
+[  2.5%] Iter  243/9765 | SPS:  1174 | Reward:  0.053 | Value:  0.669 | VLoss: 0.0435 | PLoss: -0.0123 | Ent: 0.6046
+[  2.5%] Iter  244/9765 | SPS:  1174 | Reward:  0.053 | Value:  0.626 | VLoss: 0.0349 | PLoss: -0.0186 | Ent: 0.5824
+[  2.5%] Iter  245/9765 | SPS:  1170 | Reward:  0.040 | Value:  0.503 | VLoss: 0.0325 | PLoss: -0.0144 | Ent: 0.5765
+[  2.5%] Iter  246/9765 | SPS:  1167 | Reward:  0.054 | Value:  0.553 | VLoss: 0.0380 | PLoss: -0.0078 | Ent: 0.5885
+[  2.5%] Iter  247/9765 | SPS:  1164 | Reward:  0.046 | Value:  0.558 | VLoss: 0.0440 | PLoss: -0.0121 | Ent: 0.5644
+[  2.5%] Iter  248/9765 | SPS:  1160 | Reward:  0.062 | Value:  0.543 | VLoss: 0.0475 | PLoss: -0.0104 | Ent: 0.5922
+[  2.5%] Iter  249/9765 | SPS:  1157 | Reward:  0.053 | Value:  0.560 | VLoss: 0.0378 | PLoss: -0.0062 | Ent: 0.5079
+[  2.6%] Iter  250/9765 | SPS:  1158 | Reward:  0.047 | Value:  0.573 | VLoss: 0.0408 | PLoss: -0.0079 | Ent: 0.5586
+[  2.6%] Iter  251/9765 | SPS:  1161 | Reward:  0.056 | Value:  0.545 | VLoss: 0.0331 | PLoss: -0.0032 | Ent: 0.5727
+[  2.6%] Iter  252/9765 | SPS:  1163 | Reward:  0.043 | Value:  0.570 | VLoss: 0.0283 | PLoss: -0.0086 | Ent: 0.5433
+[  2.6%] Iter  253/9765 | SPS:  1167 | Reward:  0.049 | Value:  0.521 | VLoss: 0.0340 | PLoss: -0.0219 | Ent: 0.5870
+[  2.6%] Iter  254/9765 | SPS:  1170 | Reward:  0.047 | Value:  0.592 | VLoss: 0.0297 | PLoss: -0.0063 | Ent: 0.5292
+[  2.6%] Iter  255/9765 | SPS:  1173 | Reward:  0.062 | Value:  0.619 | VLoss: 0.0400 | PLoss: -0.0135 | Ent: 0.4985
+[  2.6%] Iter  256/9765 | SPS:  1176 | Reward:  0.044 | Value:  0.565 | VLoss: 0.0398 | PLoss: -0.0104 | Ent: 0.5670
+[  2.6%] Iter  257/9765 | SPS:  1177 | Reward:  0.056 | Value:  0.563 | VLoss: 0.0335 | PLoss: -0.0161 | Ent: 0.5117
+[  2.6%] Iter  258/9765 | SPS:  1179 | Reward:  0.067 | Value:  0.608 | VLoss: 0.0363 | PLoss: -0.0004 | Ent: 0.5266
+[  2.7%] Iter  259/9765 | SPS:  1182 | Reward:  0.055 | Value:  0.565 | VLoss: 0.0392 | PLoss: -0.0069 | Ent: 0.5607
+[  2.7%] Iter  260/9765 | SPS:  1185 | Reward:  0.034 | Value:  0.557 | VLoss: 0.0265 | PLoss: -0.0071 | Ent: 0.5085
+[  2.7%] Iter  261/9765 | SPS:  1186 | Reward:  0.049 | Value:  0.500 | VLoss: 0.0375 | PLoss: -0.0025 | Ent: 0.5178
+[  2.7%] Iter  262/9765 | SPS:  1187 | Reward:  0.044 | Value:  0.499 | VLoss: 0.0408 | PLoss: -0.0073 | Ent: 0.5728
+[  2.7%] Iter  263/9765 | SPS:  1188 | Reward:  0.048 | Value:  0.548 | VLoss: 0.0387 | PLoss: -0.0060 | Ent: 0.5147
+[  2.7%] Iter  264/9765 | SPS:  1190 | Reward:  0.055 | Value:  0.570 | VLoss: 0.0276 | PLoss: -0.0061 | Ent: 0.5094
+[  2.7%] Iter  265/9765 | SPS:  1191 | Reward:  0.037 | Value:  0.559 | VLoss: 0.0230 | PLoss: -0.0197 | Ent: 0.4806
+[  2.7%] Iter  266/9765 | SPS:  1192 | Reward:  0.064 | Value:  0.660 | VLoss: 0.0381 | PLoss: -0.0086 | Ent: 0.5051
+[  2.7%] Iter  267/9765 | SPS:  1193 | Reward:  0.061 | Value:  0.642 | VLoss: 0.0352 | PLoss: -0.0095 | Ent: 0.5160
+[  2.7%] Iter  268/9765 | SPS:  1194 | Reward:  0.051 | Value:  0.595 | VLoss: 0.0374 | PLoss: -0.0094 | Ent: 0.4799
+[  2.8%] Iter  269/9765 | SPS:  1196 | Reward:  0.059 | Value:  0.597 | VLoss: 0.0353 | PLoss: -0.0171 | Ent: 0.4435
+[  2.8%] Iter  270/9765 | SPS:  1197 | Reward:  0.060 | Value:  0.635 | VLoss: 0.0325 | PLoss: -0.0138 | Ent: 0.4914
+[  2.8%] Iter  271/9765 | SPS:  1198 | Reward:  0.055 | Value:  0.598 | VLoss: 0.0422 | PLoss: -0.0053 | Ent: 0.5286
+[  2.8%] Iter  272/9765 | SPS:  1198 | Reward:  0.067 | Value:  0.553 | VLoss: 0.0399 | PLoss: -0.0033 | Ent: 0.5145
+[  2.8%] Iter  273/9765 | SPS:  1199 | Reward:  0.055 | Value:  0.638 | VLoss: 0.0401 | PLoss: -0.0139 | Ent: 0.5298
+[  2.8%] Iter  274/9765 | SPS:  1200 | Reward:  0.062 | Value:  0.631 | VLoss: 0.0314 | PLoss: -0.0061 | Ent: 0.5120
+[  2.8%] Iter  275/9765 | SPS:  1201 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0463 | PLoss: -0.0129 | Ent: 0.5282
+[  2.8%] Iter  276/9765 | SPS:  1202 | Reward:  0.051 | Value:  0.630 | VLoss: 0.0433 | PLoss: -0.0200 | Ent: 0.4987
+[  2.8%] Iter  277/9765 | SPS:  1203 | Reward:  0.044 | Value:  0.606 | VLoss: 0.0390 | PLoss: -0.0141 | Ent: 0.5019
+[  2.8%] Iter  278/9765 | SPS:  1203 | Reward:  0.054 | Value:  0.560 | VLoss: 0.0334 | PLoss: -0.0099 | Ent: 0.5038
+[  2.9%] Iter  279/9765 | SPS:  1203 | Reward:  0.064 | Value:  0.584 | VLoss: 0.0359 | PLoss: -0.0148 | Ent: 0.4330
+[  2.9%] Iter  280/9765 | SPS:  1204 | Reward:  0.066 | Value:  0.627 | VLoss: 0.0371 | PLoss: -0.0085 | Ent: 0.4414
+[  2.9%] Iter  281/9765 | SPS:  1203 | Reward:  0.061 | Value:  0.683 | VLoss: 0.0390 | PLoss: -0.0141 | Ent: 0.4649
+[  2.9%] Iter  282/9765 | SPS:  1203 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0372 | PLoss: -0.0034 | Ent: 0.4703
+[  2.9%] Iter  283/9765 | SPS:  1202 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0454 | PLoss: -0.0023 | Ent: 0.4586
+[  2.9%] Iter  284/9765 | SPS:  1201 | Reward:  0.054 | Value:  0.607 | VLoss: 0.0384 | PLoss: -0.0141 | Ent: 0.4558
+[  2.9%] Iter  285/9765 | SPS:  1200 | Reward:  0.058 | Value:  0.646 | VLoss: 0.0382 | PLoss: -0.0171 | Ent: 0.4713
+[  2.9%] Iter  286/9765 | SPS:  1198 | Reward:  0.058 | Value:  0.563 | VLoss: 0.0421 | PLoss: -0.0102 | Ent: 0.4994
+[  2.9%] Iter  287/9765 | SPS:  1196 | Reward:  0.054 | Value:  0.556 | VLoss: 0.0354 | PLoss: -0.0175 | Ent: 0.4851
+[  2.9%] Iter  288/9765 | SPS:  1194 | Reward:  0.072 | Value:  0.578 | VLoss: 0.0417 | PLoss: -0.0125 | Ent: 0.4722
+[  3.0%] Iter  289/9765 | SPS:  1192 | Reward:  0.057 | Value:  0.615 | VLoss: 0.0390 | PLoss: -0.0059 | Ent: 0.5127
+[  3.0%] Iter  290/9765 | SPS:  1189 | Reward:  0.075 | Value:  0.594 | VLoss: 0.0455 | PLoss: -0.0131 | Ent: 0.4200
+[  3.0%] Iter  291/9765 | SPS:  1187 | Reward:  0.058 | Value:  0.642 | VLoss: 0.0369 | PLoss: -0.0092 | Ent: 0.4804
+[  3.0%] Iter  292/9765 | SPS:  1185 | Reward:  0.056 | Value:  0.661 | VLoss: 0.0291 | PLoss: -0.0091 | Ent: 0.5019
+[  3.0%] Iter  293/9765 | SPS:  1183 | Reward:  0.051 | Value:  0.631 | VLoss: 0.0301 | PLoss: -0.0074 | Ent: 0.4523
+[  3.0%] Iter  294/9765 | SPS:  1181 | Reward:  0.060 | Value:  0.629 | VLoss: 0.0395 | PLoss: -0.0094 | Ent: 0.4053
+[  3.0%] Iter  295/9765 | SPS:  1179 | Reward:  0.051 | Value:  0.657 | VLoss: 0.0309 | PLoss: -0.0036 | Ent: 0.4024
+[  3.0%] Iter  296/9765 | SPS:  1176 | Reward:  0.057 | Value:  0.605 | VLoss: 0.0403 | PLoss: -0.0061 | Ent: 0.4270
+[  3.0%] Iter  297/9765 | SPS:  1174 | Reward:  0.055 | Value:  0.596 | VLoss: 0.0459 | PLoss: -0.0096 | Ent: 0.4920
+[  3.1%] Iter  298/9765 | SPS:  1172 | Reward:  0.061 | Value:  0.582 | VLoss: 0.0408 | PLoss: -0.0090 | Ent: 0.4519
+[  3.1%] Iter  299/9765 | SPS:  1170 | Reward:  0.064 | Value:  0.540 | VLoss: 0.0384 | PLoss: -0.0123 | Ent: 0.4458
+[  3.1%] Iter  300/9765 | SPS:  1168 | Reward:  0.064 | Value:  0.612 | VLoss: 0.0399 | PLoss: -0.0153 | Ent: 0.4284
+[  3.1%] Iter  301/9765 | SPS:  1166 | Reward:  0.048 | Value:  0.629 | VLoss: 0.0369 | PLoss: -0.0159 | Ent: 0.3916
+[  3.1%] Iter  302/9765 | SPS:  1164 | Reward:  0.031 | Value:  0.552 | VLoss: 0.0287 | PLoss: -0.0242 | Ent: 0.4483
+[  3.1%] Iter  303/9765 | SPS:  1162 | Reward:  0.039 | Value:  0.527 | VLoss: 0.0267 | PLoss: -0.0127 | Ent: 0.4771
+[  3.1%] Iter  304/9765 | SPS:  1160 | Reward:  0.048 | Value:  0.567 | VLoss: 0.0238 | PLoss: -0.0101 | Ent: 0.3992
+[  3.1%] Iter  305/9765 | SPS:  1158 | Reward:  0.041 | Value:  0.561 | VLoss: 0.0340 | PLoss: -0.0101 | Ent: 0.4899
+[  3.1%] Iter  306/9765 | SPS:  1157 | Reward:  0.046 | Value:  0.531 | VLoss: 0.0283 | PLoss: -0.0077 | Ent: 0.5083
+[  3.1%] Iter  307/9765 | SPS:  1155 | Reward:  0.048 | Value:  0.534 | VLoss: 0.0274 | PLoss: 0.0002 | Ent: 0.4459
+[  3.2%] Iter  308/9765 | SPS:  1153 | Reward:  0.037 | Value:  0.555 | VLoss: 0.0219 | PLoss: -0.0127 | Ent: 0.4389
+[  3.2%] Iter  309/9765 | SPS:  1152 | Reward:  0.028 | Value:  0.536 | VLoss: 0.0189 | PLoss: -0.0112 | Ent: 0.5048
+[  3.2%] Iter  310/9765 | SPS:  1150 | Reward:  0.052 | Value:  0.509 | VLoss: 0.0356 | PLoss: -0.0043 | Ent: 0.5165
+[  3.2%] Iter  311/9765 | SPS:  1148 | Reward:  0.066 | Value:  0.564 | VLoss: 0.0373 | PLoss: -0.0088 | Ent: 0.5010
+[  3.2%] Iter  312/9765 | SPS:  1147 | Reward:  0.059 | Value:  0.595 | VLoss: 0.0401 | PLoss: -0.0033 | Ent: 0.4197
+[  3.2%] Iter  313/9765 | SPS:  1145 | Reward:  0.047 | Value:  0.616 | VLoss: 0.0284 | PLoss: -0.0052 | Ent: 0.4056
+[  3.2%] Iter  314/9765 | SPS:  1143 | Reward:  0.056 | Value:  0.584 | VLoss: 0.0479 | PLoss: -0.0088 | Ent: 0.4423
+[  3.2%] Iter  315/9765 | SPS:  1141 | Reward:  0.062 | Value:  0.613 | VLoss: 0.0408 | PLoss: -0.0145 | Ent: 0.4291
+[  3.2%] Iter  316/9765 | SPS:  1139 | Reward:  0.053 | Value:  0.604 | VLoss: 0.0411 | PLoss: -0.0114 | Ent: 0.4563
+[  3.2%] Iter  317/9765 | SPS:  1138 | Reward:  0.060 | Value:  0.609 | VLoss: 0.0328 | PLoss: -0.0121 | Ent: 0.4120
+[  3.3%] Iter  318/9765 | SPS:  1136 | Reward:  0.060 | Value:  0.575 | VLoss: 0.0333 | PLoss: -0.0027 | Ent: 0.4068
+[  3.3%] Iter  319/9765 | SPS:  1134 | Reward:  0.057 | Value:  0.615 | VLoss: 0.0351 | PLoss: -0.0197 | Ent: 0.4632
+[  3.3%] Iter  320/9765 | SPS:  1132 | Reward:  0.062 | Value:  0.600 | VLoss: 0.0377 | PLoss: -0.0106 | Ent: 0.4234
+[  3.3%] Iter  321/9765 | SPS:  1131 | Reward:  0.064 | Value:  0.647 | VLoss: 0.0386 | PLoss: -0.0081 | Ent: 0.4295
+[  3.3%] Iter  322/9765 | SPS:  1129 | Reward:  0.057 | Value:  0.626 | VLoss: 0.0289 | PLoss: -0.0065 | Ent: 0.4014
+[  3.3%] Iter  323/9765 | SPS:  1128 | Reward:  0.059 | Value:  0.678 | VLoss: 0.0332 | PLoss: -0.0099 | Ent: 0.4432
+[  3.3%] Iter  324/9765 | SPS:  1126 | Reward:  0.052 | Value:  0.595 | VLoss: 0.0475 | PLoss: -0.0123 | Ent: 0.3921
+[  3.3%] Iter  325/9765 | SPS:  1124 | Reward:  0.070 | Value:  0.597 | VLoss: 0.0442 | PLoss: -0.0176 | Ent: 0.4587
+[  3.3%] Iter  326/9765 | SPS:  1122 | Reward:  0.056 | Value:  0.592 | VLoss: 0.0376 | PLoss: -0.0074 | Ent: 0.4568
+[  3.3%] Iter  327/9765 | SPS:  1121 | Reward:  0.059 | Value:  0.608 | VLoss: 0.0378 | PLoss: -0.0101 | Ent: 0.4501
+[  3.4%] Iter  328/9765 | SPS:  1119 | Reward:  0.053 | Value:  0.548 | VLoss: 0.0295 | PLoss: -0.0105 | Ent: 0.5101
+[  3.4%] Iter  329/9765 | SPS:  1118 | Reward:  0.054 | Value:  0.659 | VLoss: 0.0274 | PLoss: -0.0079 | Ent: 0.3855
+[  3.4%] Iter  330/9765 | SPS:  1116 | Reward:  0.043 | Value:  0.606 | VLoss: 0.0295 | PLoss: -0.0173 | Ent: 0.4070
+[  3.4%] Iter  331/9765 | SPS:  1115 | Reward:  0.054 | Value:  0.659 | VLoss: 0.0240 | PLoss: -0.0138 | Ent: 0.4373
+[  3.4%] Iter  332/9765 | SPS:  1113 | Reward:  0.049 | Value:  0.608 | VLoss: 0.0370 | PLoss: -0.0125 | Ent: 0.4219
+[  3.4%] Iter  333/9765 | SPS:  1112 | Reward:  0.072 | Value:  0.632 | VLoss: 0.0370 | PLoss: -0.0057 | Ent: 0.4235
+[  3.4%] Iter  334/9765 | SPS:  1111 | Reward:  0.070 | Value:  0.670 | VLoss: 0.0408 | PLoss: -0.0100 | Ent: 0.3706
+[  3.4%] Iter  335/9765 | SPS:  1111 | Reward:  0.057 | Value:  0.703 | VLoss: 0.0339 | PLoss: -0.0112 | Ent: 0.3841
+[  3.4%] Iter  336/9765 | SPS:  1110 | Reward:  0.059 | Value:  0.571 | VLoss: 0.0367 | PLoss: -0.0190 | Ent: 0.4484
+[  3.5%] Iter  337/9765 | SPS:  1110 | Reward:  0.067 | Value:  0.627 | VLoss: 0.0353 | PLoss: -0.0089 | Ent: 0.4249
+[  3.5%] Iter  338/9765 | SPS:  1110 | Reward:  0.065 | Value:  0.642 | VLoss: 0.0405 | PLoss: -0.0175 | Ent: 0.3824
+[  3.5%] Iter  339/9765 | SPS:  1110 | Reward:  0.053 | Value:  0.641 | VLoss: 0.0311 | PLoss: 0.0030 | Ent: 0.3440
+[  3.5%] Iter  340/9765 | SPS:  1110 | Reward:  0.064 | Value:  0.588 | VLoss: 0.0397 | PLoss: -0.0126 | Ent: 0.4212
+[  3.5%] Iter  341/9765 | SPS:  1110 | Reward:  0.060 | Value:  0.632 | VLoss: 0.0386 | PLoss: -0.0134 | Ent: 0.4537
+[  3.5%] Iter  342/9765 | SPS:  1110 | Reward:  0.048 | Value:  0.643 | VLoss: 0.0336 | PLoss: -0.0078 | Ent: 0.4018
+[  3.5%] Iter  343/9765 | SPS:  1110 | Reward:  0.072 | Value:  0.600 | VLoss: 0.0416 | PLoss: -0.0087 | Ent: 0.4031
+[  3.5%] Iter  344/9765 | SPS:  1110 | Reward:  0.064 | Value:  0.651 | VLoss: 0.0417 | PLoss: -0.0043 | Ent: 0.3445
+[  3.5%] Iter  345/9765 | SPS:  1110 | Reward:  0.063 | Value:  0.616 | VLoss: 0.0361 | PLoss: -0.0068 | Ent: 0.4552
+[  3.5%] Iter  346/9765 | SPS:  1110 | Reward:  0.064 | Value:  0.619 | VLoss: 0.0421 | PLoss: -0.0128 | Ent: 0.4028
+[  3.6%] Iter  347/9765 | SPS:  1110 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0250 | PLoss: -0.0071 | Ent: 0.3113
+[  3.6%] Iter  348/9765 | SPS:  1110 | Reward:  0.049 | Value:  0.691 | VLoss: 0.0241 | PLoss: -0.0123 | Ent: 0.4099
+[  3.6%] Iter  349/9765 | SPS:  1110 | Reward:  0.055 | Value:  0.619 | VLoss: 0.0399 | PLoss: -0.0123 | Ent: 0.4551
+[  3.6%] Iter  350/9765 | SPS:  1110 | Reward:  0.061 | Value:  0.622 | VLoss: 0.0370 | PLoss: -0.0112 | Ent: 0.4264
+[  3.6%] Iter  351/9765 | SPS:  1110 | Reward:  0.069 | Value:  0.658 | VLoss: 0.0410 | PLoss: -0.0064 | Ent: 0.4314
+[  3.6%] Iter  352/9765 | SPS:  1110 | Reward:  0.050 | Value:  0.661 | VLoss: 0.0315 | PLoss: -0.0084 | Ent: 0.4233
+[  3.6%] Iter  353/9765 | SPS:  1110 | Reward:  0.046 | Value:  0.615 | VLoss: 0.0324 | PLoss: -0.0171 | Ent: 0.4563
+[  3.6%] Iter  354/9765 | SPS:  1110 | Reward:  0.056 | Value:  0.597 | VLoss: 0.0359 | PLoss: -0.0149 | Ent: 0.4314
+[  3.6%] Iter  355/9765 | SPS:  1110 | Reward:  0.048 | Value:  0.599 | VLoss: 0.0328 | PLoss: -0.0102 | Ent: 0.3822
+[  3.6%] Iter  356/9765 | SPS:  1110 | Reward:  0.042 | Value:  0.609 | VLoss: 0.0259 | PLoss: -0.0086 | Ent: 0.4117
+[  3.7%] Iter  357/9765 | SPS:  1110 | Reward:  0.054 | Value:  0.608 | VLoss: 0.0337 | PLoss: -0.0089 | Ent: 0.4394
+[  3.7%] Iter  358/9765 | SPS:  1109 | Reward:  0.065 | Value:  0.594 | VLoss: 0.0366 | PLoss: -0.0162 | Ent: 0.4057
+[  3.7%] Iter  359/9765 | SPS:  1109 | Reward:  0.054 | Value:  0.623 | VLoss: 0.0319 | PLoss: -0.0124 | Ent: 0.4293
+[  3.7%] Iter  360/9765 | SPS:  1109 | Reward:  0.070 | Value:  0.673 | VLoss: 0.0432 | PLoss: -0.0059 | Ent: 0.3327
+[  3.7%] Iter  361/9765 | SPS:  1109 | Reward:  0.053 | Value:  0.679 | VLoss: 0.0313 | PLoss: -0.0125 | Ent: 0.4336
+[  3.7%] Iter  362/9765 | SPS:  1109 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0415 | PLoss: -0.0081 | Ent: 0.4244
+[  3.7%] Iter  363/9765 | SPS:  1109 | Reward:  0.053 | Value:  0.601 | VLoss: 0.0273 | PLoss: -0.0132 | Ent: 0.4340
+[  3.7%] Iter  364/9765 | SPS:  1109 | Reward:  0.064 | Value:  0.612 | VLoss: 0.0459 | PLoss: -0.0192 | Ent: 0.4409
+[  3.7%] Iter  365/9765 | SPS:  1109 | Reward:  0.043 | Value:  0.619 | VLoss: 0.0365 | PLoss: -0.0122 | Ent: 0.4357
+[  3.7%] Iter  366/9765 | SPS:  1109 | Reward:  0.058 | Value:  0.560 | VLoss: 0.0369 | PLoss: -0.0088 | Ent: 0.3930
+[  3.8%] Iter  367/9765 | SPS:  1109 | Reward:  0.072 | Value:  0.613 | VLoss: 0.0453 | PLoss: -0.0084 | Ent: 0.4027
+[  3.8%] Iter  368/9765 | SPS:  1109 | Reward:  0.059 | Value:  0.631 | VLoss: 0.0297 | PLoss: -0.0051 | Ent: 0.4329
+[  3.8%] Iter  369/9765 | SPS:  1109 | Reward:  0.059 | Value:  0.642 | VLoss: 0.0279 | PLoss: 0.0003 | Ent: 0.3497
+[  3.8%] Iter  370/9765 | SPS:  1109 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0380 | PLoss: -0.0057 | Ent: 0.3423
+[  3.8%] Iter  371/9765 | SPS:  1109 | Reward:  0.050 | Value:  0.592 | VLoss: 0.0282 | PLoss: -0.0067 | Ent: 0.4081
+[  3.8%] Iter  372/9765 | SPS:  1109 | Reward:  0.058 | Value:  0.664 | VLoss: 0.0308 | PLoss: -0.0137 | Ent: 0.3596
+[  3.8%] Iter  373/9765 | SPS:  1109 | Reward:  0.061 | Value:  0.637 | VLoss: 0.0335 | PLoss: -0.0118 | Ent: 0.3621
+[  3.8%] Iter  374/9765 | SPS:  1108 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0319 | PLoss: -0.0143 | Ent: 0.4011
+[  3.8%] Iter  375/9765 | SPS:  1108 | Reward:  0.052 | Value:  0.630 | VLoss: 0.0377 | PLoss: -0.0129 | Ent: 0.4380
+[  3.9%] Iter  376/9765 | SPS:  1108 | Reward:  0.064 | Value:  0.612 | VLoss: 0.0351 | PLoss: -0.0141 | Ent: 0.4239
+[  3.9%] Iter  377/9765 | SPS:  1108 | Reward:  0.060 | Value:  0.632 | VLoss: 0.0418 | PLoss: -0.0079 | Ent: 0.4230
+[  3.9%] Iter  378/9765 | SPS:  1108 | Reward:  0.047 | Value:  0.641 | VLoss: 0.0289 | PLoss: -0.0119 | Ent: 0.4050
+[  3.9%] Iter  379/9765 | SPS:  1108 | Reward:  0.063 | Value:  0.594 | VLoss: 0.0405 | PLoss: -0.0075 | Ent: 0.4141
+[  3.9%] Iter  380/9765 | SPS:  1108 | Reward:  0.053 | Value:  0.595 | VLoss: 0.0294 | PLoss: -0.0228 | Ent: 0.3776
+[  3.9%] Iter  381/9765 | SPS:  1108 | Reward:  0.075 | Value:  0.645 | VLoss: 0.0392 | PLoss: -0.0072 | Ent: 0.4173
+[  3.9%] Iter  382/9765 | SPS:  1108 | Reward:  0.067 | Value:  0.644 | VLoss: 0.0478 | PLoss: -0.0159 | Ent: 0.3666
+[  3.9%] Iter  383/9765 | SPS:  1108 | Reward:  0.068 | Value:  0.606 | VLoss: 0.0353 | PLoss: -0.0116 | Ent: 0.3857
+[  3.9%] Iter  384/9765 | SPS:  1108 | Reward:  0.052 | Value:  0.590 | VLoss: 0.0359 | PLoss: -0.0105 | Ent: 0.3993
+[  3.9%] Iter  385/9765 | SPS:  1108 | Reward:  0.063 | Value:  0.619 | VLoss: 0.0363 | PLoss: -0.0045 | Ent: 0.4513
+[  4.0%] Iter  386/9765 | SPS:  1108 | Reward:  0.047 | Value:  0.581 | VLoss: 0.0378 | PLoss: 0.0011 | Ent: 0.4581
+[  4.0%] Iter  387/9765 | SPS:  1108 | Reward:  0.062 | Value:  0.569 | VLoss: 0.0369 | PLoss: -0.0145 | Ent: 0.4212
+[  4.0%] Iter  388/9765 | SPS:  1108 | Reward:  0.055 | Value:  0.573 | VLoss: 0.0389 | PLoss: -0.0075 | Ent: 0.3601
+[  4.0%] Iter  389/9765 | SPS:  1108 | Reward:  0.045 | Value:  0.558 | VLoss: 0.0314 | PLoss: -0.0041 | Ent: 0.4152
+[  4.0%] Iter  390/9765 | SPS:  1107 | Reward:  0.055 | Value:  0.632 | VLoss: 0.0364 | PLoss: -0.0124 | Ent: 0.3701
+[  4.0%] Iter  391/9765 | SPS:  1107 | Reward:  0.066 | Value:  0.597 | VLoss: 0.0360 | PLoss: -0.0137 | Ent: 0.3997
+[  4.0%] Iter  392/9765 | SPS:  1107 | Reward:  0.058 | Value:  0.654 | VLoss: 0.0287 | PLoss: -0.0099 | Ent: 0.3747
+[  4.0%] Iter  393/9765 | SPS:  1107 | Reward:  0.064 | Value:  0.655 | VLoss: 0.0356 | PLoss: -0.0090 | Ent: 0.4039
+[  4.0%] Iter  394/9765 | SPS:  1107 | Reward:  0.058 | Value:  0.649 | VLoss: 0.0339 | PLoss: -0.0029 | Ent: 0.3950
+[  4.0%] Iter  395/9765 | SPS:  1107 | Reward:  0.057 | Value:  0.665 | VLoss: 0.0485 | PLoss: -0.0086 | Ent: 0.4028
+[  4.1%] Iter  396/9765 | SPS:  1107 | Reward:  0.067 | Value:  0.613 | VLoss: 0.0365 | PLoss: -0.0113 | Ent: 0.3204
+[  4.1%] Iter  397/9765 | SPS:  1106 | Reward:  0.073 | Value:  0.627 | VLoss: 0.0451 | PLoss: 0.0006 | Ent: 0.3788
+[  4.1%] Iter  398/9765 | SPS:  1106 | Reward:  0.058 | Value:  0.648 | VLoss: 0.0342 | PLoss: -0.0162 | Ent: 0.4103
+[  4.1%] Iter  399/9765 | SPS:  1106 | Reward:  0.057 | Value:  0.669 | VLoss: 0.0269 | PLoss: -0.0171 | Ent: 0.4074
+[  4.1%] Iter  400/9765 | SPS:  1106 | Reward:  0.069 | Value:  0.657 | VLoss: 0.0411 | PLoss: -0.0110 | Ent: 0.3242
+[  4.1%] Iter  401/9765 | SPS:  1106 | Reward:  0.049 | Value:  0.663 | VLoss: 0.0263 | PLoss: -0.0075 | Ent: 0.3493
+[  4.1%] Iter  402/9765 | SPS:  1106 | Reward:  0.064 | Value:  0.649 | VLoss: 0.0264 | PLoss: -0.0104 | Ent: 0.3976
+[  4.1%] Iter  403/9765 | SPS:  1106 | Reward:  0.062 | Value:  0.613 | VLoss: 0.0422 | PLoss: -0.0146 | Ent: 0.3693
+[  4.1%] Iter  404/9765 | SPS:  1106 | Reward:  0.065 | Value:  0.622 | VLoss: 0.0325 | PLoss: -0.0073 | Ent: 0.4538
+[  4.1%] Iter  405/9765 | SPS:  1106 | Reward:  0.065 | Value:  0.663 | VLoss: 0.0388 | PLoss: -0.0107 | Ent: 0.3366
+[  4.2%] Iter  406/9765 | SPS:  1105 | Reward:  0.058 | Value:  0.607 | VLoss: 0.0383 | PLoss: -0.0077 | Ent: 0.4437
+[  4.2%] Iter  407/9765 | SPS:  1105 | Reward:  0.064 | Value:  0.631 | VLoss: 0.0412 | PLoss: -0.0063 | Ent: 0.3602
+[  4.2%] Iter  408/9765 | SPS:  1105 | Reward:  0.059 | Value:  0.662 | VLoss: 0.0276 | PLoss: -0.0117 | Ent: 0.3142
+[  4.2%] Iter  409/9765 | SPS:  1105 | Reward:  0.065 | Value:  0.676 | VLoss: 0.0384 | PLoss: -0.0072 | Ent: 0.3751
+[  4.2%] Iter  410/9765 | SPS:  1105 | Reward:  0.054 | Value:  0.632 | VLoss: 0.0402 | PLoss: -0.0161 | Ent: 0.4355
+[  4.2%] Iter  411/9765 | SPS:  1105 | Reward:  0.068 | Value:  0.631 | VLoss: 0.0366 | PLoss: -0.0039 | Ent: 0.4145
+[  4.2%] Iter  412/9765 | SPS:  1105 | Reward:  0.077 | Value:  0.671 | VLoss: 0.0304 | PLoss: -0.0079 | Ent: 0.3696
+[  4.2%] Iter  413/9765 | SPS:  1104 | Reward:  0.056 | Value:  0.692 | VLoss: 0.0331 | PLoss: -0.0192 | Ent: 0.3281
+[  4.2%] Iter  414/9765 | SPS:  1104 | Reward:  0.059 | Value:  0.584 | VLoss: 0.0312 | PLoss: -0.0109 | Ent: 0.3907
+[  4.2%] Iter  415/9765 | SPS:  1104 | Reward:  0.058 | Value:  0.673 | VLoss: 0.0368 | PLoss: -0.0134 | Ent: 0.3491
+[  4.3%] Iter  416/9765 | SPS:  1104 | Reward:  0.073 | Value:  0.699 | VLoss: 0.0376 | PLoss: -0.0127 | Ent: 0.3411
+[  4.3%] Iter  417/9765 | SPS:  1104 | Reward:  0.055 | Value:  0.656 | VLoss: 0.0327 | PLoss: -0.0112 | Ent: 0.4223
+[  4.3%] Iter  418/9765 | SPS:  1104 | Reward:  0.072 | Value:  0.713 | VLoss: 0.0323 | PLoss: -0.0090 | Ent: 0.3295
+[  4.3%] Iter  419/9765 | SPS:  1104 | Reward:  0.053 | Value:  0.685 | VLoss: 0.0253 | PLoss: -0.0068 | Ent: 0.4182
+[  4.3%] Iter  420/9765 | SPS:  1104 | Reward:  0.050 | Value:  0.607 | VLoss: 0.0296 | PLoss: -0.0025 | Ent: 0.4036
+[  4.3%] Iter  421/9765 | SPS:  1104 | Reward:  0.065 | Value:  0.630 | VLoss: 0.0458 | PLoss: -0.0152 | Ent: 0.3179
+[  4.3%] Iter  422/9765 | SPS:  1103 | Reward:  0.067 | Value:  0.630 | VLoss: 0.0335 | PLoss: -0.0047 | Ent: 0.2966
+[  4.3%] Iter  423/9765 | SPS:  1103 | Reward:  0.034 | Value:  0.578 | VLoss: 0.0267 | PLoss: -0.0032 | Ent: 0.3078
+[  4.3%] Iter  424/9765 | SPS:  1103 | Reward:  0.042 | Value:  0.623 | VLoss: 0.0234 | PLoss: -0.0159 | Ent: 0.4057
+[  4.4%] Iter  425/9765 | SPS:  1103 | Reward:  0.053 | Value:  0.628 | VLoss: 0.0374 | PLoss: -0.0075 | Ent: 0.3926
+[  4.4%] Iter  426/9765 | SPS:  1103 | Reward:  0.048 | Value:  0.669 | VLoss: 0.0286 | PLoss: -0.0105 | Ent: 0.4505
+[  4.4%] Iter  427/9765 | SPS:  1103 | Reward:  0.059 | Value:  0.650 | VLoss: 0.0390 | PLoss: -0.0120 | Ent: 0.3666
+[  4.4%] Iter  428/9765 | SPS:  1103 | Reward:  0.067 | Value:  0.618 | VLoss: 0.0340 | PLoss: -0.0058 | Ent: 0.3718
+[  4.4%] Iter  429/9765 | SPS:  1103 | Reward:  0.057 | Value:  0.614 | VLoss: 0.0311 | PLoss: -0.0116 | Ent: 0.3077
+[  4.4%] Iter  430/9765 | SPS:  1102 | Reward:  0.071 | Value:  0.641 | VLoss: 0.0372 | PLoss: -0.0134 | Ent: 0.3384
+[  4.4%] Iter  431/9765 | SPS:  1102 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0281 | PLoss: 0.0021 | Ent: 0.3451
+[  4.4%] Iter  432/9765 | SPS:  1102 | Reward:  0.061 | Value:  0.688 | VLoss: 0.0408 | PLoss: -0.0088 | Ent: 0.3335
+[  4.4%] Iter  433/9765 | SPS:  1102 | Reward:  0.053 | Value:  0.755 | VLoss: 0.0310 | PLoss: -0.0132 | Ent: 0.3722
+[  4.4%] Iter  434/9765 | SPS:  1102 | Reward:  0.048 | Value:  0.672 | VLoss: 0.0353 | PLoss: -0.0150 | Ent: 0.3584
+[  4.5%] Iter  435/9765 | SPS:  1102 | Reward:  0.067 | Value:  0.606 | VLoss: 0.0358 | PLoss: -0.0058 | Ent: 0.3418
+[  4.5%] Iter  436/9765 | SPS:  1101 | Reward:  0.049 | Value:  0.633 | VLoss: 0.0276 | PLoss: -0.0091 | Ent: 0.3210
+[  4.5%] Iter  437/9765 | SPS:  1101 | Reward:  0.062 | Value:  0.628 | VLoss: 0.0297 | PLoss: -0.0144 | Ent: 0.3451
+[  4.5%] Iter  438/9765 | SPS:  1101 | Reward:  0.058 | Value:  0.640 | VLoss: 0.0309 | PLoss: -0.0063 | Ent: 0.3191
+[  4.5%] Iter  439/9765 | SPS:  1101 | Reward:  0.068 | Value:  0.630 | VLoss: 0.0364 | PLoss: -0.0109 | Ent: 0.3920
+[  4.5%] Iter  440/9765 | SPS:  1101 | Reward:  0.067 | Value:  0.645 | VLoss: 0.0392 | PLoss: -0.0114 | Ent: 0.3776
+[  4.5%] Iter  441/9765 | SPS:  1101 | Reward:  0.056 | Value:  0.612 | VLoss: 0.0407 | PLoss: -0.0173 | Ent: 0.3864
+[  4.5%] Iter  442/9765 | SPS:  1101 | Reward:  0.064 | Value:  0.633 | VLoss: 0.0282 | PLoss: -0.0118 | Ent: 0.4008
+[  4.5%] Iter  443/9765 | SPS:  1100 | Reward:  0.046 | Value:  0.628 | VLoss: 0.0307 | PLoss: -0.0119 | Ent: 0.3273
+[  4.5%] Iter  444/9765 | SPS:  1100 | Reward:  0.051 | Value:  0.585 | VLoss: 0.0345 | PLoss: -0.0069 | Ent: 0.4078
+[  4.6%] Iter  445/9765 | SPS:  1100 | Reward:  0.055 | Value:  0.652 | VLoss: 0.0306 | PLoss: -0.0123 | Ent: 0.3944
+[  4.6%] Iter  446/9765 | SPS:  1100 | Reward:  0.064 | Value:  0.681 | VLoss: 0.0307 | PLoss: -0.0112 | Ent: 0.3396
+[  4.6%] Iter  447/9765 | SPS:  1100 | Reward:  0.046 | Value:  0.656 | VLoss: 0.0348 | PLoss: -0.0195 | Ent: 0.4216
+[  4.6%] Iter  448/9765 | SPS:  1099 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0336 | PLoss: -0.0089 | Ent: 0.3319
+[  4.6%] Iter  449/9765 | SPS:  1099 | Reward:  0.052 | Value:  0.644 | VLoss: 0.0351 | PLoss: -0.0140 | Ent: 0.3848
+[  4.6%] Iter  450/9765 | SPS:  1099 | Reward:  0.069 | Value:  0.633 | VLoss: 0.0396 | PLoss: -0.0075 | Ent: 0.3513
+[  4.6%] Iter  451/9765 | SPS:  1099 | Reward:  0.066 | Value:  0.646 | VLoss: 0.0383 | PLoss: -0.0062 | Ent: 0.3752
+[  4.6%] Iter  452/9765 | SPS:  1098 | Reward:  0.062 | Value:  0.645 | VLoss: 0.0303 | PLoss: -0.0068 | Ent: 0.3547
+[  4.6%] Iter  453/9765 | SPS:  1098 | Reward:  0.052 | Value:  0.646 | VLoss: 0.0328 | PLoss: -0.0118 | Ent: 0.3719
+[  4.6%] Iter  454/9765 | SPS:  1098 | Reward:  0.039 | Value:  0.586 | VLoss: 0.0265 | PLoss: -0.0081 | Ent: 0.3569
+[  4.7%] Iter  455/9765 | SPS:  1098 | Reward:  0.061 | Value:  0.610 | VLoss: 0.0313 | PLoss: -0.0168 | Ent: 0.3718
+[  4.7%] Iter  456/9765 | SPS:  1098 | Reward:  0.070 | Value:  0.661 | VLoss: 0.0349 | PLoss: -0.0100 | Ent: 0.3605
+[  4.7%] Iter  457/9765 | SPS:  1097 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0290 | PLoss: -0.0083 | Ent: 0.3492
+[  4.7%] Iter  458/9765 | SPS:  1097 | Reward:  0.054 | Value:  0.715 | VLoss: 0.0358 | PLoss: -0.0085 | Ent: 0.3371
+[  4.7%] Iter  459/9765 | SPS:  1097 | Reward:  0.056 | Value:  0.689 | VLoss: 0.0368 | PLoss: -0.0106 | Ent: 0.3275
+[  4.7%] Iter  460/9765 | SPS:  1097 | Reward:  0.048 | Value:  0.662 | VLoss: 0.0316 | PLoss: -0.0158 | Ent: 0.3234
+[  4.7%] Iter  461/9765 | SPS:  1097 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0307 | PLoss: -0.0122 | Ent: 0.3045
+[  4.7%] Iter  462/9765 | SPS:  1097 | Reward:  0.057 | Value:  0.654 | VLoss: 0.0355 | PLoss: -0.0066 | Ent: 0.3393
+[  4.7%] Iter  463/9765 | SPS:  1097 | Reward:  0.052 | Value:  0.584 | VLoss: 0.0388 | PLoss: -0.0117 | Ent: 0.3702
+[  4.8%] Iter  464/9765 | SPS:  1097 | Reward:  0.054 | Value:  0.631 | VLoss: 0.0357 | PLoss: -0.0081 | Ent: 0.3033
+[  4.8%] Iter  465/9765 | SPS:  1097 | Reward:  0.057 | Value:  0.602 | VLoss: 0.0464 | PLoss: -0.0102 | Ent: 0.3148
+[  4.8%] Iter  466/9765 | SPS:  1097 | Reward:  0.072 | Value:  0.635 | VLoss: 0.0391 | PLoss: -0.0053 | Ent: 0.2922
+[  4.8%] Iter  467/9765 | SPS:  1096 | Reward:  0.059 | Value:  0.697 | VLoss: 0.0294 | PLoss: -0.0119 | Ent: 0.2956
+[  4.8%] Iter  468/9765 | SPS:  1096 | Reward:  0.061 | Value:  0.632 | VLoss: 0.0390 | PLoss: -0.0183 | Ent: 0.3488
+[  4.8%] Iter  469/9765 | SPS:  1096 | Reward:  0.081 | Value:  0.636 | VLoss: 0.0450 | PLoss: -0.0066 | Ent: 0.3530
+[  4.8%] Iter  470/9765 | SPS:  1096 | Reward:  0.070 | Value:  0.680 | VLoss: 0.0268 | PLoss: -0.0058 | Ent: 0.3054
+[  4.8%] Iter  471/9765 | SPS:  1096 | Reward:  0.053 | Value:  0.682 | VLoss: 0.0340 | PLoss: -0.0153 | Ent: 0.4012
+[  4.8%] Iter  472/9765 | SPS:  1096 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0330 | PLoss: -0.0055 | Ent: 0.3255
+[  4.8%] Iter  473/9765 | SPS:  1095 | Reward:  0.071 | Value:  0.677 | VLoss: 0.0362 | PLoss: -0.0089 | Ent: 0.3071
+[  4.9%] Iter  474/9765 | SPS:  1095 | Reward:  0.075 | Value:  0.685 | VLoss: 0.0403 | PLoss: -0.0115 | Ent: 0.3207
+[  4.9%] Iter  475/9765 | SPS:  1095 | Reward:  0.053 | Value:  0.730 | VLoss: 0.0283 | PLoss: -0.0123 | Ent: 0.3813
+[  4.9%] Iter  476/9765 | SPS:  1095 | Reward:  0.073 | Value:  0.696 | VLoss: 0.0374 | PLoss: -0.0111 | Ent: 0.3518
+[  4.9%] Iter  477/9765 | SPS:  1094 | Reward:  0.059 | Value:  0.683 | VLoss: 0.0363 | PLoss: -0.0099 | Ent: 0.3557
+[  4.9%] Iter  478/9765 | SPS:  1094 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0338 | PLoss: -0.0132 | Ent: 0.3834
+[  4.9%] Iter  479/9765 | SPS:  1094 | Reward:  0.055 | Value:  0.676 | VLoss: 0.0331 | PLoss: -0.0010 | Ent: 0.3102
+[  4.9%] Iter  480/9765 | SPS:  1094 | Reward:  0.060 | Value:  0.601 | VLoss: 0.0279 | PLoss: -0.0091 | Ent: 0.3402
+[  4.9%] Iter  481/9765 | SPS:  1094 | Reward:  0.055 | Value:  0.649 | VLoss: 0.0306 | PLoss: -0.0084 | Ent: 0.3653
+[  4.9%] Iter  482/9765 | SPS:  1093 | Reward:  0.068 | Value:  0.681 | VLoss: 0.0370 | PLoss: -0.0129 | Ent: 0.2626
+[  4.9%] Iter  483/9765 | SPS:  1093 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0339 | PLoss: -0.0096 | Ent: 0.3168
+[  5.0%] Iter  484/9765 | SPS:  1093 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0350 | PLoss: -0.0077 | Ent: 0.3049
+[  5.0%] Iter  485/9765 | SPS:  1093 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0361 | PLoss: -0.0154 | Ent: 0.3242
+[  5.0%] Iter  486/9765 | SPS:  1095 | Reward:  0.059 | Value:  0.648 | VLoss: 0.0323 | PLoss: -0.0038 | Ent: 0.3218
+[  5.0%] Iter  487/9765 | SPS:  1096 | Reward:  0.057 | Value:  0.658 | VLoss: 0.0391 | PLoss: -0.0141 | Ent: 0.3027
+[  5.0%] Iter  488/9765 | SPS:  1098 | Reward:  0.058 | Value:  0.612 | VLoss: 0.0335 | PLoss: -0.0119 | Ent: 0.3134
+[  5.0%] Iter  489/9765 | SPS:  1100 | Reward:  0.062 | Value:  0.590 | VLoss: 0.0283 | PLoss: -0.0091 | Ent: 0.4135
+[  5.0%] Iter  490/9765 | SPS:  1101 | Reward:  0.046 | Value:  0.647 | VLoss: 0.0308 | PLoss: -0.0108 | Ent: 0.3303
+[  5.0%] Iter  491/9765 | SPS:  1103 | Reward:  0.057 | Value:  0.596 | VLoss: 0.0417 | PLoss: -0.0101 | Ent: 0.3207
+[  5.0%] Iter  492/9765 | SPS:  1104 | Reward:  0.060 | Value:  0.639 | VLoss: 0.0334 | PLoss: -0.0108 | Ent: 0.3460
+[  5.0%] Iter  493/9765 | SPS:  1105 | Reward:  0.054 | Value:  0.643 | VLoss: 0.0410 | PLoss: -0.0134 | Ent: 0.2905
+[  5.1%] Iter  494/9765 | SPS:  1103 | Reward:  0.052 | Value:  0.640 | VLoss: 0.0346 | PLoss: -0.0097 | Ent: 0.3841
+[  5.1%] Iter  495/9765 | SPS:  1105 | Reward:  0.062 | Value:  0.587 | VLoss: 0.0344 | PLoss: -0.0105 | Ent: 0.3007
+[  5.1%] Iter  496/9765 | SPS:  1106 | Reward:  0.066 | Value:  0.613 | VLoss: 0.0419 | PLoss: 0.0009 | Ent: 0.3695
+[  5.1%] Iter  497/9765 | SPS:  1108 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0391 | PLoss: -0.0092 | Ent: 0.3120
+[  5.1%] Iter  498/9765 | SPS:  1109 | Reward:  0.070 | Value:  0.666 | VLoss: 0.0333 | PLoss: -0.0105 | Ent: 0.3063
+[  5.1%] Iter  499/9765 | SPS:  1111 | Reward:  0.059 | Value:  0.653 | VLoss: 0.0268 | PLoss: -0.0141 | Ent: 0.3520
+[  5.1%] Iter  500/9765 | SPS:  1111 | Reward:  0.070 | Value:  0.659 | VLoss: 0.0341 | PLoss: -0.0097 | Ent: 0.2804
+[  5.1%] Iter  501/9765 | SPS:  1109 | Reward:  0.073 | Value:  0.646 | VLoss: 0.0396 | PLoss: -0.0095 | Ent: 0.3192
+[  5.1%] Iter  502/9765 | SPS:  1108 | Reward:  0.061 | Value:  0.655 | VLoss: 0.0392 | PLoss: -0.0113 | Ent: 0.2874
+[  5.2%] Iter  503/9765 | SPS:  1106 | Reward:  0.064 | Value:  0.663 | VLoss: 0.0273 | PLoss: -0.0079 | Ent: 0.3597
+[  5.2%] Iter  504/9765 | SPS:  1105 | Reward:  0.061 | Value:  0.598 | VLoss: 0.0435 | PLoss: -0.0100 | Ent: 0.3424
+[  5.2%] Iter  505/9765 | SPS:  1103 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0311 | PLoss: -0.0147 | Ent: 0.3575
+[  5.2%] Iter  506/9765 | SPS:  1102 | Reward:  0.051 | Value:  0.680 | VLoss: 0.0225 | PLoss: -0.0090 | Ent: 0.3393
+[  5.2%] Iter  507/9765 | SPS:  1102 | Reward:  0.056 | Value:  0.668 | VLoss: 0.0362 | PLoss: -0.0086 | Ent: 0.3361
+[  5.2%] Iter  508/9765 | SPS:  1103 | Reward:  0.068 | Value:  0.616 | VLoss: 0.0421 | PLoss: -0.0013 | Ent: 0.3280
+[  5.2%] Iter  509/9765 | SPS:  1105 | Reward:  0.052 | Value:  0.653 | VLoss: 0.0275 | PLoss: -0.0104 | Ent: 0.3059
+[  5.2%] Iter  510/9765 | SPS:  1106 | Reward:  0.060 | Value:  0.624 | VLoss: 0.0344 | PLoss: -0.0089 | Ent: 0.2667
+[  5.2%] Iter  511/9765 | SPS:  1107 | Reward:  0.053 | Value:  0.596 | VLoss: 0.0270 | PLoss: -0.0164 | Ent: 0.3832
+[  5.2%] Iter  512/9765 | SPS:  1109 | Reward:  0.068 | Value:  0.605 | VLoss: 0.0362 | PLoss: -0.0078 | Ent: 0.2649
+[  5.3%] Iter  513/9765 | SPS:  1109 | Reward:  0.068 | Value:  0.576 | VLoss: 0.0409 | PLoss: 0.0006 | Ent: 0.3055
+[  5.3%] Iter  514/9765 | SPS:  1110 | Reward:  0.061 | Value:  0.610 | VLoss: 0.0367 | PLoss: -0.0156 | Ent: 0.3065
+[  5.3%] Iter  515/9765 | SPS:  1112 | Reward:  0.042 | Value:  0.666 | VLoss: 0.0256 | PLoss: -0.0047 | Ent: 0.2679
+[  5.3%] Iter  516/9765 | SPS:  1113 | Reward:  0.061 | Value:  0.631 | VLoss: 0.0395 | PLoss: -0.0114 | Ent: 0.3533
+[  5.3%] Iter  517/9765 | SPS:  1113 | Reward:  0.071 | Value:  0.648 | VLoss: 0.0322 | PLoss: -0.0048 | Ent: 0.3430
+[  5.3%] Iter  518/9765 | SPS:  1114 | Reward:  0.062 | Value:  0.699 | VLoss: 0.0352 | PLoss: -0.0166 | Ent: 0.3139
+[  5.3%] Iter  519/9765 | SPS:  1114 | Reward:  0.075 | Value:  0.677 | VLoss: 0.0394 | PLoss: -0.0100 | Ent: 0.3409
+[  5.3%] Iter  520/9765 | SPS:  1115 | Reward:  0.049 | Value:  0.645 | VLoss: 0.0297 | PLoss: -0.0150 | Ent: 0.2784
+[  5.3%] Iter  521/9765 | SPS:  1116 | Reward:  0.048 | Value:  0.623 | VLoss: 0.0280 | PLoss: -0.0048 | Ent: 0.2945
+[  5.3%] Iter  522/9765 | SPS:  1116 | Reward:  0.056 | Value:  0.655 | VLoss: 0.0250 | PLoss: -0.0150 | Ent: 0.2500
+[  5.4%] Iter  523/9765 | SPS:  1117 | Reward:  0.059 | Value:  0.711 | VLoss: 0.0291 | PLoss: -0.0099 | Ent: 0.3516
+[  5.4%] Iter  524/9765 | SPS:  1117 | Reward:  0.054 | Value:  0.628 | VLoss: 0.0311 | PLoss: -0.0100 | Ent: 0.2854
+[  5.4%] Iter  525/9765 | SPS:  1118 | Reward:  0.057 | Value:  0.617 | VLoss: 0.0309 | PLoss: -0.0083 | Ent: 0.2394
+[  5.4%] Iter  526/9765 | SPS:  1119 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0398 | PLoss: -0.0033 | Ent: 0.2891
+[  5.4%] Iter  527/9765 | SPS:  1119 | Reward:  0.060 | Value:  0.653 | VLoss: 0.0367 | PLoss: -0.0092 | Ent: 0.2962
+[  5.4%] Iter  528/9765 | SPS:  1120 | Reward:  0.066 | Value:  0.619 | VLoss: 0.0299 | PLoss: -0.0099 | Ent: 0.3050
+[  5.4%] Iter  529/9765 | SPS:  1120 | Reward:  0.075 | Value:  0.656 | VLoss: 0.0333 | PLoss: -0.0069 | Ent: 0.3236
+[  5.4%] Iter  530/9765 | SPS:  1121 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0315 | PLoss: -0.0094 | Ent: 0.2701
+[  5.4%] Iter  531/9765 | SPS:  1121 | Reward:  0.052 | Value:  0.579 | VLoss: 0.0392 | PLoss: -0.0104 | Ent: 0.3209
+[  5.4%] Iter  532/9765 | SPS:  1121 | Reward:  0.064 | Value:  0.592 | VLoss: 0.0323 | PLoss: -0.0031 | Ent: 0.2796
+[  5.5%] Iter  533/9765 | SPS:  1122 | Reward:  0.050 | Value:  0.572 | VLoss: 0.0293 | PLoss: -0.0052 | Ent: 0.3143
+[  5.5%] Iter  534/9765 | SPS:  1122 | Reward:  0.059 | Value:  0.666 | VLoss: 0.0358 | PLoss: -0.0119 | Ent: 0.2657
+[  5.5%] Iter  535/9765 | SPS:  1122 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0297 | PLoss: -0.0096 | Ent: 0.2952
+[  5.5%] Iter  536/9765 | SPS:  1122 | Reward:  0.064 | Value:  0.649 | VLoss: 0.0365 | PLoss: -0.0031 | Ent: 0.2825
+[  5.5%] Iter  537/9765 | SPS:  1123 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0219 | PLoss: -0.0018 | Ent: 0.3166
+[  5.5%] Iter  538/9765 | SPS:  1123 | Reward:  0.066 | Value:  0.722 | VLoss: 0.0304 | PLoss: -0.0107 | Ent: 0.2990
+[  5.5%] Iter  539/9765 | SPS:  1123 | Reward:  0.063 | Value:  0.725 | VLoss: 0.0358 | PLoss: -0.0100 | Ent: 0.3046
+[  5.5%] Iter  540/9765 | SPS:  1123 | Reward:  0.056 | Value:  0.680 | VLoss: 0.0392 | PLoss: -0.0164 | Ent: 0.2849
+[  5.5%] Iter  541/9765 | SPS:  1122 | Reward:  0.072 | Value:  0.653 | VLoss: 0.0460 | PLoss: -0.0160 | Ent: 0.2827
+[  5.6%] Iter  542/9765 | SPS:  1122 | Reward:  0.051 | Value:  0.629 | VLoss: 0.0325 | PLoss: -0.0161 | Ent: 0.3125
+[  5.6%] Iter  543/9765 | SPS:  1122 | Reward:  0.061 | Value:  0.610 | VLoss: 0.0334 | PLoss: -0.0160 | Ent: 0.2710
+[  5.6%] Iter  544/9765 | SPS:  1122 | Reward:  0.065 | Value:  0.624 | VLoss: 0.0389 | PLoss: -0.0120 | Ent: 0.2416
+[  5.6%] Iter  545/9765 | SPS:  1121 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0233 | PLoss: -0.0076 | Ent: 0.3513
+[  5.6%] Iter  546/9765 | SPS:  1121 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0328 | PLoss: -0.0118 | Ent: 0.2899
+[  5.6%] Iter  547/9765 | SPS:  1120 | Reward:  0.062 | Value:  0.671 | VLoss: 0.0322 | PLoss: -0.0170 | Ent: 0.2795
+[  5.6%] Iter  548/9765 | SPS:  1120 | Reward:  0.060 | Value:  0.651 | VLoss: 0.0378 | PLoss: -0.0118 | Ent: 0.2628
+[  5.6%] Iter  549/9765 | SPS:  1119 | Reward:  0.058 | Value:  0.675 | VLoss: 0.0314 | PLoss: -0.0123 | Ent: 0.3067
+[  5.6%] Iter  550/9765 | SPS:  1118 | Reward:  0.066 | Value:  0.628 | VLoss: 0.0348 | PLoss: -0.0044 | Ent: 0.2807
+[  5.6%] Iter  551/9765 | SPS:  1117 | Reward:  0.070 | Value:  0.667 | VLoss: 0.0419 | PLoss: -0.0079 | Ent: 0.2982
+[  5.7%] Iter  552/9765 | SPS:  1117 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0324 | PLoss: -0.0116 | Ent: 0.3271
+[  5.7%] Iter  553/9765 | SPS:  1116 | Reward:  0.070 | Value:  0.621 | VLoss: 0.0356 | PLoss: -0.0028 | Ent: 0.2681
+[  5.7%] Iter  554/9765 | SPS:  1115 | Reward:  0.060 | Value:  0.650 | VLoss: 0.0336 | PLoss: -0.0191 | Ent: 0.2745
+[  5.7%] Iter  555/9765 | SPS:  1114 | Reward:  0.054 | Value:  0.581 | VLoss: 0.0455 | PLoss: -0.0153 | Ent: 0.3507
+[  5.7%] Iter  556/9765 | SPS:  1113 | Reward:  0.065 | Value:  0.544 | VLoss: 0.0467 | PLoss: -0.0075 | Ent: 0.3381
+[  5.7%] Iter  557/9765 | SPS:  1113 | Reward:  0.055 | Value:  0.562 | VLoss: 0.0329 | PLoss: -0.0149 | Ent: 0.3421
+[  5.7%] Iter  558/9765 | SPS:  1112 | Reward:  0.072 | Value:  0.600 | VLoss: 0.0329 | PLoss: -0.0162 | Ent: 0.3132
+[  5.7%] Iter  559/9765 | SPS:  1111 | Reward:  0.079 | Value:  0.645 | VLoss: 0.0395 | PLoss: -0.0078 | Ent: 0.2523
+[  5.7%] Iter  560/9765 | SPS:  1110 | Reward:  0.066 | Value:  0.601 | VLoss: 0.0356 | PLoss: -0.0085 | Ent: 0.2970
+[  5.7%] Iter  561/9765 | SPS:  1110 | Reward:  0.055 | Value:  0.581 | VLoss: 0.0318 | PLoss: -0.0069 | Ent: 0.3098
+[  5.8%] Iter  562/9765 | SPS:  1109 | Reward:  0.050 | Value:  0.654 | VLoss: 0.0342 | PLoss: -0.0064 | Ent: 0.2698
+[  5.8%] Iter  563/9765 | SPS:  1108 | Reward:  0.062 | Value:  0.649 | VLoss: 0.0354 | PLoss: -0.0107 | Ent: 0.2981
+[  5.8%] Iter  564/9765 | SPS:  1107 | Reward:  0.055 | Value:  0.629 | VLoss: 0.0330 | PLoss: -0.0116 | Ent: 0.3126
+[  5.8%] Iter  565/9765 | SPS:  1107 | Reward:  0.051 | Value:  0.606 | VLoss: 0.0243 | PLoss: -0.0085 | Ent: 0.3252
+[  5.8%] Iter  566/9765 | SPS:  1106 | Reward:  0.042 | Value:  0.614 | VLoss: 0.0178 | PLoss: -0.0048 | Ent: 0.2655
+[  5.8%] Iter  567/9765 | SPS:  1105 | Reward:  0.037 | Value:  0.545 | VLoss: 0.0167 | PLoss: -0.0116 | Ent: 0.3082
+[  5.8%] Iter  568/9765 | SPS:  1104 | Reward:  0.062 | Value:  0.630 | VLoss: 0.0314 | PLoss: -0.0151 | Ent: 0.2527
+[  5.8%] Iter  569/9765 | SPS:  1104 | Reward:  0.048 | Value:  0.620 | VLoss: 0.0304 | PLoss: -0.0133 | Ent: 0.3016
+[  5.8%] Iter  570/9765 | SPS:  1103 | Reward:  0.050 | Value:  0.607 | VLoss: 0.0268 | PLoss: -0.0046 | Ent: 0.2764
+[  5.8%] Iter  571/9765 | SPS:  1102 | Reward:  0.050 | Value:  0.564 | VLoss: 0.0252 | PLoss: -0.0037 | Ent: 0.3008
+[  5.9%] Iter  572/9765 | SPS:  1102 | Reward:  0.057 | Value:  0.615 | VLoss: 0.0324 | PLoss: -0.0039 | Ent: 0.2364
+[  5.9%] Iter  573/9765 | SPS:  1103 | Reward:  0.056 | Value:  0.701 | VLoss: 0.0208 | PLoss: -0.0053 | Ent: 0.2201
+[  5.9%] Iter  574/9765 | SPS:  1105 | Reward:  0.067 | Value:  0.693 | VLoss: 0.0333 | PLoss: -0.0232 | Ent: 0.2498
+[  5.9%] Iter  575/9765 | SPS:  1106 | Reward:  0.063 | Value:  0.637 | VLoss: 0.0356 | PLoss: -0.0091 | Ent: 0.2122
+[  5.9%] Iter  576/9765 | SPS:  1107 | Reward:  0.061 | Value:  0.592 | VLoss: 0.0351 | PLoss: -0.0089 | Ent: 0.3206
+[  5.9%] Iter  577/9765 | SPS:  1109 | Reward:  0.051 | Value:  0.624 | VLoss: 0.0342 | PLoss: -0.0020 | Ent: 0.2876
+[  5.9%] Iter  578/9765 | SPS:  1110 | Reward:  0.055 | Value:  0.610 | VLoss: 0.0347 | PLoss: -0.0148 | Ent: 0.2593
+[  5.9%] Iter  579/9765 | SPS:  1111 | Reward:  0.054 | Value:  0.643 | VLoss: 0.0417 | PLoss: -0.0051 | Ent: 0.2743
+[  5.9%] Iter  580/9765 | SPS:  1111 | Reward:  0.056 | Value:  0.654 | VLoss: 0.0323 | PLoss: -0.0149 | Ent: 0.2375
+[  5.9%] Iter  581/9765 | SPS:  1111 | Reward:  0.066 | Value:  0.626 | VLoss: 0.0353 | PLoss: -0.0087 | Ent: 0.2613
+[  6.0%] Iter  582/9765 | SPS:  1112 | Reward:  0.060 | Value:  0.644 | VLoss: 0.0389 | PLoss: -0.0110 | Ent: 0.3249
+[  6.0%] Iter  583/9765 | SPS:  1114 | Reward:  0.064 | Value:  0.647 | VLoss: 0.0320 | PLoss: -0.0178 | Ent: 0.3086
+[  6.0%] Iter  584/9765 | SPS:  1115 | Reward:  0.053 | Value:  0.627 | VLoss: 0.0335 | PLoss: 0.0004 | Ent: 0.2468
+[  6.0%] Iter  585/9765 | SPS:  1116 | Reward:  0.076 | Value:  0.619 | VLoss: 0.0444 | PLoss: -0.0102 | Ent: 0.2718
+[  6.0%] Iter  586/9765 | SPS:  1116 | Reward:  0.071 | Value:  0.657 | VLoss: 0.0337 | PLoss: -0.0111 | Ent: 0.2880
+[  6.0%] Iter  587/9765 | SPS:  1116 | Reward:  0.064 | Value:  0.622 | VLoss: 0.0360 | PLoss: -0.0024 | Ent: 0.2811
+[  6.0%] Iter  588/9765 | SPS:  1115 | Reward:  0.072 | Value:  0.616 | VLoss: 0.0373 | PLoss: -0.0136 | Ent: 0.2641
+[  6.0%] Iter  589/9765 | SPS:  1113 | Reward:  0.067 | Value:  0.672 | VLoss: 0.0345 | PLoss: -0.0186 | Ent: 0.3127
+[  6.0%] Iter  590/9765 | SPS:  1112 | Reward:  0.071 | Value:  0.619 | VLoss: 0.0372 | PLoss: -0.0114 | Ent: 0.2746
+[  6.1%] Iter  591/9765 | SPS:  1111 | Reward:  0.062 | Value:  0.667 | VLoss: 0.0358 | PLoss: -0.0074 | Ent: 0.2018
+[  6.1%] Iter  592/9765 | SPS:  1109 | Reward:  0.082 | Value:  0.645 | VLoss: 0.0397 | PLoss: -0.0099 | Ent: 0.2669
+[  6.1%] Iter  593/9765 | SPS:  1110 | Reward:  0.074 | Value:  0.676 | VLoss: 0.0492 | PLoss: -0.0114 | Ent: 0.2908
+[  6.1%] Iter  594/9765 | SPS:  1112 | Reward:  0.077 | Value:  0.683 | VLoss: 0.0267 | PLoss: -0.0070 | Ent: 0.2443
+[  6.1%] Iter  595/9765 | SPS:  1113 | Reward:  0.065 | Value:  0.744 | VLoss: 0.0383 | PLoss: -0.0171 | Ent: 0.2924
+[  6.1%] Iter  596/9765 | SPS:  1114 | Reward:  0.073 | Value:  0.719 | VLoss: 0.0387 | PLoss: -0.0129 | Ent: 0.2851
+[  6.1%] Iter  597/9765 | SPS:  1115 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0347 | PLoss: -0.0119 | Ent: 0.2927
+[  6.1%] Iter  598/9765 | SPS:  1116 | Reward:  0.064 | Value:  0.637 | VLoss: 0.0315 | PLoss: -0.0127 | Ent: 0.3038
+[  6.1%] Iter  599/9765 | SPS:  1117 | Reward:  0.062 | Value:  0.588 | VLoss: 0.0479 | PLoss: -0.0127 | Ent: 0.2491
+[  6.1%] Iter  600/9765 | SPS:  1118 | Reward:  0.075 | Value:  0.664 | VLoss: 0.0295 | PLoss: -0.0113 | Ent: 0.2444
+[  6.2%] Iter  601/9765 | SPS:  1118 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0358 | PLoss: -0.0014 | Ent: 0.2028
+[  6.2%] Iter  602/9765 | SPS:  1119 | Reward:  0.073 | Value:  0.660 | VLoss: 0.0398 | PLoss: -0.0158 | Ent: 0.2476
+[  6.2%] Iter  603/9765 | SPS:  1119 | Reward:  0.058 | Value:  0.706 | VLoss: 0.0290 | PLoss: -0.0100 | Ent: 0.2587
+[  6.2%] Iter  604/9765 | SPS:  1120 | Reward:  0.072 | Value:  0.695 | VLoss: 0.0382 | PLoss: -0.0027 | Ent: 0.2742
+[  6.2%] Iter  605/9765 | SPS:  1120 | Reward:  0.067 | Value:  0.651 | VLoss: 0.0414 | PLoss: -0.0118 | Ent: 0.2491
+[  6.2%] Iter  606/9765 | SPS:  1121 | Reward:  0.072 | Value:  0.658 | VLoss: 0.0435 | PLoss: -0.0110 | Ent: 0.2885
+[  6.2%] Iter  607/9765 | SPS:  1121 | Reward:  0.069 | Value:  0.654 | VLoss: 0.0331 | PLoss: -0.0099 | Ent: 0.2408
+[  6.2%] Iter  608/9765 | SPS:  1121 | Reward:  0.066 | Value:  0.647 | VLoss: 0.0365 | PLoss: -0.0027 | Ent: 0.2700
+[  6.2%] Iter  609/9765 | SPS:  1122 | Reward:  0.067 | Value:  0.597 | VLoss: 0.0342 | PLoss: -0.0247 | Ent: 0.2753
+[  6.2%] Iter  610/9765 | SPS:  1122 | Reward:  0.059 | Value:  0.662 | VLoss: 0.0327 | PLoss: -0.0082 | Ent: 0.2388
+[  6.3%] Iter  611/9765 | SPS:  1123 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0442 | PLoss: -0.0101 | Ent: 0.2823
+[  6.3%] Iter  612/9765 | SPS:  1123 | Reward:  0.059 | Value:  0.640 | VLoss: 0.0388 | PLoss: -0.0035 | Ent: 0.2738
+[  6.3%] Iter  613/9765 | SPS:  1123 | Reward:  0.062 | Value:  0.601 | VLoss: 0.0379 | PLoss: -0.0092 | Ent: 0.2957
+[  6.3%] Iter  614/9765 | SPS:  1123 | Reward:  0.055 | Value:  0.627 | VLoss: 0.0260 | PLoss: -0.0116 | Ent: 0.2104
+[  6.3%] Iter  615/9765 | SPS:  1124 | Reward:  0.055 | Value:  0.599 | VLoss: 0.0279 | PLoss: -0.0116 | Ent: 0.2549
+[  6.3%] Iter  616/9765 | SPS:  1124 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0450 | PLoss: -0.0132 | Ent: 0.2721
+[  6.3%] Iter  617/9765 | SPS:  1124 | Reward:  0.060 | Value:  0.670 | VLoss: 0.0347 | PLoss: 0.0009 | Ent: 0.2101
+[  6.3%] Iter  618/9765 | SPS:  1124 | Reward:  0.068 | Value:  0.664 | VLoss: 0.0322 | PLoss: -0.0013 | Ent: 0.2331
+[  6.3%] Iter  619/9765 | SPS:  1123 | Reward:  0.061 | Value:  0.588 | VLoss: 0.0356 | PLoss: -0.0086 | Ent: 0.2399
+[  6.3%] Iter  620/9765 | SPS:  1123 | Reward:  0.049 | Value:  0.658 | VLoss: 0.0250 | PLoss: -0.0148 | Ent: 0.2079
+[  6.4%] Iter  621/9765 | SPS:  1122 | Reward:  0.046 | Value:  0.655 | VLoss: 0.0247 | PLoss: -0.0082 | Ent: 0.2485
+[  6.4%] Iter  622/9765 | SPS:  1122 | Reward:  0.068 | Value:  0.613 | VLoss: 0.0342 | PLoss: -0.0043 | Ent: 0.2024
+[  6.4%] Iter  623/9765 | SPS:  1121 | Reward:  0.064 | Value:  0.662 | VLoss: 0.0341 | PLoss: -0.0085 | Ent: 0.3117
+[  6.4%] Iter  624/9765 | SPS:  1120 | Reward:  0.068 | Value:  0.643 | VLoss: 0.0346 | PLoss: -0.0096 | Ent: 0.2126
+[  6.4%] Iter  625/9765 | SPS:  1121 | Reward:  0.069 | Value:  0.688 | VLoss: 0.0333 | PLoss: -0.0033 | Ent: 0.2356
+[  6.4%] Iter  626/9765 | SPS:  1122 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0257 | PLoss: -0.0139 | Ent: 0.2337
+[  6.4%] Iter  627/9765 | SPS:  1123 | Reward:  0.057 | Value:  0.687 | VLoss: 0.0273 | PLoss: -0.0082 | Ent: 0.2788
+[  6.4%] Iter  628/9765 | SPS:  1122 | Reward:  0.067 | Value:  0.717 | VLoss: 0.0339 | PLoss: -0.0148 | Ent: 0.2315
+[  6.4%] Iter  629/9765 | SPS:  1120 | Reward:  0.067 | Value:  0.681 | VLoss: 0.0372 | PLoss: -0.0087 | Ent: 0.2676
+[  6.5%] Iter  630/9765 | SPS:  1117 | Reward:  0.053 | Value:  0.632 | VLoss: 0.0341 | PLoss: -0.0106 | Ent: 0.2665
+[  6.5%] Iter  631/9765 | SPS:  1114 | Reward:  0.068 | Value:  0.651 | VLoss: 0.0332 | PLoss: -0.0067 | Ent: 0.2681
+[  6.5%] Iter  632/9765 | SPS:  1112 | Reward:  0.061 | Value:  0.717 | VLoss: 0.0347 | PLoss: -0.0129 | Ent: 0.2900
+[  6.5%] Iter  633/9765 | SPS:  1110 | Reward:  0.068 | Value:  0.669 | VLoss: 0.0362 | PLoss: -0.0243 | Ent: 0.3011
+[  6.5%] Iter  634/9765 | SPS:  1111 | Reward:  0.056 | Value:  0.640 | VLoss: 0.0367 | PLoss: -0.0095 | Ent: 0.3525
+[  6.5%] Iter  635/9765 | SPS:  1112 | Reward:  0.060 | Value:  0.621 | VLoss: 0.0427 | PLoss: -0.0102 | Ent: 0.3177
+[  6.5%] Iter  636/9765 | SPS:  1113 | Reward:  0.047 | Value:  0.592 | VLoss: 0.0314 | PLoss: -0.0177 | Ent: 0.2940
+[  6.5%] Iter  637/9765 | SPS:  1114 | Reward:  0.063 | Value:  0.606 | VLoss: 0.0318 | PLoss: -0.0142 | Ent: 0.2431
+[  6.5%] Iter  638/9765 | SPS:  1114 | Reward:  0.074 | Value:  0.620 | VLoss: 0.0265 | PLoss: -0.0065 | Ent: 0.3210
+[  6.5%] Iter  639/9765 | SPS:  1114 | Reward:  0.060 | Value:  0.651 | VLoss: 0.0390 | PLoss: -0.0023 | Ent: 0.2418
+[  6.6%] Iter  640/9765 | SPS:  1115 | Reward:  0.067 | Value:  0.664 | VLoss: 0.0364 | PLoss: -0.0139 | Ent: 0.2941
+[  6.6%] Iter  641/9765 | SPS:  1116 | Reward:  0.065 | Value:  0.677 | VLoss: 0.0329 | PLoss: -0.0063 | Ent: 0.2633
+[  6.6%] Iter  642/9765 | SPS:  1118 | Reward:  0.056 | Value:  0.636 | VLoss: 0.0425 | PLoss: 0.0003 | Ent: 0.3136
+[  6.6%] Iter  643/9765 | SPS:  1119 | Reward:  0.059 | Value:  0.572 | VLoss: 0.0428 | PLoss: -0.0123 | Ent: 0.2766
+[  6.6%] Iter  644/9765 | SPS:  1119 | Reward:  0.062 | Value:  0.591 | VLoss: 0.0303 | PLoss: -0.0144 | Ent: 0.2443
+[  6.6%] Iter  645/9765 | SPS:  1119 | Reward:  0.059 | Value:  0.618 | VLoss: 0.0303 | PLoss: -0.0079 | Ent: 0.2713
+[  6.6%] Iter  646/9765 | SPS:  1118 | Reward:  0.060 | Value:  0.655 | VLoss: 0.0304 | PLoss: -0.0105 | Ent: 0.2551
+[  6.6%] Iter  647/9765 | SPS:  1117 | Reward:  0.068 | Value:  0.621 | VLoss: 0.0335 | PLoss: -0.0044 | Ent: 0.2334
+[  6.6%] Iter  648/9765 | SPS:  1116 | Reward:  0.043 | Value:  0.590 | VLoss: 0.0332 | PLoss: -0.0098 | Ent: 0.1767
+[  6.6%] Iter  649/9765 | SPS:  1115 | Reward:  0.053 | Value:  0.587 | VLoss: 0.0298 | PLoss: -0.0041 | Ent: 0.2213
+[  6.7%] Iter  650/9765 | SPS:  1114 | Reward:  0.058 | Value:  0.581 | VLoss: 0.0289 | PLoss: -0.0045 | Ent: 0.3027
+[  6.7%] Iter  651/9765 | SPS:  1114 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0206 | PLoss: -0.0079 | Ent: 0.2302
+[  6.7%] Iter  652/9765 | SPS:  1116 | Reward:  0.054 | Value:  0.686 | VLoss: 0.0326 | PLoss: -0.0141 | Ent: 0.2208
+[  6.7%] Iter  653/9765 | SPS:  1117 | Reward:  0.042 | Value:  0.673 | VLoss: 0.0271 | PLoss: -0.0112 | Ent: 0.2413
+[  6.7%] Iter  654/9765 | SPS:  1118 | Reward:  0.047 | Value:  0.616 | VLoss: 0.0250 | PLoss: -0.0268 | Ent: 0.2453
+[  6.7%] Iter  655/9765 | SPS:  1119 | Reward:  0.055 | Value:  0.598 | VLoss: 0.0341 | PLoss: -0.0070 | Ent: 0.1828
+[  6.7%] Iter  656/9765 | SPS:  1120 | Reward:  0.057 | Value:  0.642 | VLoss: 0.0296 | PLoss: -0.0104 | Ent: 0.2289
+[  6.7%] Iter  657/9765 | SPS:  1121 | Reward:  0.062 | Value:  0.618 | VLoss: 0.0414 | PLoss: -0.0062 | Ent: 0.2588
+[  6.7%] Iter  658/9765 | SPS:  1122 | Reward:  0.058 | Value:  0.655 | VLoss: 0.0351 | PLoss: -0.0099 | Ent: 0.2849
+[  6.7%] Iter  659/9765 | SPS:  1123 | Reward:  0.064 | Value:  0.628 | VLoss: 0.0306 | PLoss: -0.0120 | Ent: 0.2970
+[  6.8%] Iter  660/9765 | SPS:  1124 | Reward:  0.070 | Value:  0.658 | VLoss: 0.0313 | PLoss: -0.0025 | Ent: 0.2198
+[  6.8%] Iter  661/9765 | SPS:  1124 | Reward:  0.063 | Value:  0.643 | VLoss: 0.0388 | PLoss: -0.0086 | Ent: 0.2670
+[  6.8%] Iter  662/9765 | SPS:  1124 | Reward:  0.063 | Value:  0.668 | VLoss: 0.0383 | PLoss: -0.0063 | Ent: 0.2120
+[  6.8%] Iter  663/9765 | SPS:  1125 | Reward:  0.063 | Value:  0.691 | VLoss: 0.0289 | PLoss: -0.0050 | Ent: 0.2323
+[  6.8%] Iter  664/9765 | SPS:  1125 | Reward:  0.067 | Value:  0.712 | VLoss: 0.0398 | PLoss: -0.0157 | Ent: 0.2361
+[  6.8%] Iter  665/9765 | SPS:  1126 | Reward:  0.056 | Value:  0.715 | VLoss: 0.0329 | PLoss: -0.0103 | Ent: 0.2346
+[  6.8%] Iter  666/9765 | SPS:  1127 | Reward:  0.037 | Value:  0.711 | VLoss: 0.0212 | PLoss: -0.0100 | Ent: 0.2542
+[  6.8%] Iter  667/9765 | SPS:  1127 | Reward:  0.051 | Value:  0.627 | VLoss: 0.0237 | PLoss: -0.0065 | Ent: 0.2291
+[  6.8%] Iter  668/9765 | SPS:  1128 | Reward:  0.067 | Value:  0.656 | VLoss: 0.0317 | PLoss: -0.0022 | Ent: 0.2467
+[  6.9%] Iter  669/9765 | SPS:  1128 | Reward:  0.055 | Value:  0.667 | VLoss: 0.0260 | PLoss: -0.0090 | Ent: 0.2851
+[  6.9%] Iter  670/9765 | SPS:  1129 | Reward:  0.053 | Value:  0.659 | VLoss: 0.0183 | PLoss: -0.0047 | Ent: 0.2468
+[  6.9%] Iter  671/9765 | SPS:  1129 | Reward:  0.063 | Value:  0.631 | VLoss: 0.0317 | PLoss: -0.0079 | Ent: 0.2306
+[  6.9%] Iter  672/9765 | SPS:  1129 | Reward:  0.065 | Value:  0.587 | VLoss: 0.0308 | PLoss: -0.0079 | Ent: 0.2462
+[  6.9%] Iter  673/9765 | SPS:  1130 | Reward:  0.052 | Value:  0.656 | VLoss: 0.0266 | PLoss: -0.0085 | Ent: 0.2200
+[  6.9%] Iter  674/9765 | SPS:  1130 | Reward:  0.051 | Value:  0.642 | VLoss: 0.0228 | PLoss: -0.0050 | Ent: 0.2245
+[  6.9%] Iter  675/9765 | SPS:  1131 | Reward:  0.066 | Value:  0.637 | VLoss: 0.0311 | PLoss: -0.0041 | Ent: 0.2554
+[  6.9%] Iter  676/9765 | SPS:  1131 | Reward:  0.070 | Value:  0.627 | VLoss: 0.0352 | PLoss: -0.0093 | Ent: 0.2595
+[  6.9%] Iter  677/9765 | SPS:  1131 | Reward:  0.068 | Value:  0.616 | VLoss: 0.0324 | PLoss: -0.0101 | Ent: 0.2791
+[  6.9%] Iter  678/9765 | SPS:  1131 | Reward:  0.065 | Value:  0.650 | VLoss: 0.0371 | PLoss: -0.0102 | Ent: 0.2733
+[  7.0%] Iter  679/9765 | SPS:  1131 | Reward:  0.056 | Value:  0.662 | VLoss: 0.0337 | PLoss: -0.0160 | Ent: 0.2487
+[  7.0%] Iter  680/9765 | SPS:  1131 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0353 | PLoss: -0.0134 | Ent: 0.3059
+[  7.0%] Iter  681/9765 | SPS:  1131 | Reward:  0.055 | Value:  0.592 | VLoss: 0.0337 | PLoss: -0.0106 | Ent: 0.2504
+[  7.0%] Iter  682/9765 | SPS:  1131 | Reward:  0.057 | Value:  0.657 | VLoss: 0.0346 | PLoss: -0.0109 | Ent: 0.2024
+[  7.0%] Iter  683/9765 | SPS:  1130 | Reward:  0.055 | Value:  0.630 | VLoss: 0.0221 | PLoss: -0.0007 | Ent: 0.2378
+[  7.0%] Iter  684/9765 | SPS:  1129 | Reward:  0.048 | Value:  0.599 | VLoss: 0.0239 | PLoss: -0.0062 | Ent: 0.2032
+[  7.0%] Iter  685/9765 | SPS:  1130 | Reward:  0.059 | Value:  0.591 | VLoss: 0.0234 | PLoss: -0.0093 | Ent: 0.2433
+[  7.0%] Iter  686/9765 | SPS:  1131 | Reward:  0.059 | Value:  0.617 | VLoss: 0.0293 | PLoss: -0.0116 | Ent: 0.3042
+[  7.0%] Iter  687/9765 | SPS:  1132 | Reward:  0.051 | Value:  0.621 | VLoss: 0.0299 | PLoss: -0.0186 | Ent: 0.2160
+[  7.0%] Iter  688/9765 | SPS:  1132 | Reward:  0.055 | Value:  0.607 | VLoss: 0.0268 | PLoss: -0.0069 | Ent: 0.1824
+[  7.1%] Iter  689/9765 | SPS:  1130 | Reward:  0.048 | Value:  0.530 | VLoss: 0.0237 | PLoss: -0.0060 | Ent: 0.2056
+[  7.1%] Iter  690/9765 | SPS:  1127 | Reward:  0.054 | Value:  0.585 | VLoss: 0.0299 | PLoss: -0.0146 | Ent: 0.2496
+[  7.1%] Iter  691/9765 | SPS:  1125 | Reward:  0.055 | Value:  0.539 | VLoss: 0.0303 | PLoss: -0.0062 | Ent: 0.2355
+[  7.1%] Iter  692/9765 | SPS:  1122 | Reward:  0.054 | Value:  0.575 | VLoss: 0.0298 | PLoss: -0.0056 | Ent: 0.2058
+[  7.1%] Iter  693/9765 | SPS:  1122 | Reward:  0.050 | Value:  0.535 | VLoss: 0.0311 | PLoss: -0.0202 | Ent: 0.2255
+[  7.1%] Iter  694/9765 | SPS:  1123 | Reward:  0.051 | Value:  0.530 | VLoss: 0.0364 | PLoss: -0.0052 | Ent: 0.2018
+[  7.1%] Iter  695/9765 | SPS:  1124 | Reward:  0.056 | Value:  0.611 | VLoss: 0.0269 | PLoss: -0.0036 | Ent: 0.2205
+[  7.1%] Iter  696/9765 | SPS:  1125 | Reward:  0.063 | Value:  0.597 | VLoss: 0.0299 | PLoss: -0.0005 | Ent: 0.1929
+[  7.1%] Iter  697/9765 | SPS:  1126 | Reward:  0.060 | Value:  0.617 | VLoss: 0.0368 | PLoss: -0.0060 | Ent: 0.2302
+[  7.1%] Iter  698/9765 | SPS:  1125 | Reward:  0.059 | Value:  0.645 | VLoss: 0.0369 | PLoss: -0.0057 | Ent: 0.1973
+[  7.2%] Iter  699/9765 | SPS:  1126 | Reward:  0.076 | Value:  0.642 | VLoss: 0.0313 | PLoss: -0.0103 | Ent: 0.2477
+[  7.2%] Iter  700/9765 | SPS:  1127 | Reward:  0.056 | Value:  0.684 | VLoss: 0.0314 | PLoss: -0.0076 | Ent: 0.2636
+[  7.2%] Iter  701/9765 | SPS:  1129 | Reward:  0.062 | Value:  0.765 | VLoss: 0.0255 | PLoss: -0.0078 | Ent: 0.2259
+[  7.2%] Iter  702/9765 | SPS:  1130 | Reward:  0.062 | Value:  0.643 | VLoss: 0.0394 | PLoss: -0.0094 | Ent: 0.2213
+[  7.2%] Iter  703/9765 | SPS:  1130 | Reward:  0.052 | Value:  0.694 | VLoss: 0.0299 | PLoss: -0.0130 | Ent: 0.2327
+[  7.2%] Iter  704/9765 | SPS:  1131 | Reward:  0.051 | Value:  0.628 | VLoss: 0.0323 | PLoss: -0.0104 | Ent: 0.2102
+[  7.2%] Iter  705/9765 | SPS:  1130 | Reward:  0.072 | Value:  0.600 | VLoss: 0.0353 | PLoss: -0.0129 | Ent: 0.2500
+[  7.2%] Iter  706/9765 | SPS:  1129 | Reward:  0.061 | Value:  0.613 | VLoss: 0.0416 | PLoss: -0.0098 | Ent: 0.2215
+[  7.2%] Iter  707/9765 | SPS:  1128 | Reward:  0.050 | Value:  0.654 | VLoss: 0.0292 | PLoss: -0.0047 | Ent: 0.1850
+[  7.3%] Iter  708/9765 | SPS:  1127 | Reward:  0.052 | Value:  0.649 | VLoss: 0.0253 | PLoss: -0.0104 | Ent: 0.2236
+[  7.3%] Iter  709/9765 | SPS:  1126 | Reward:  0.059 | Value:  0.607 | VLoss: 0.0365 | PLoss: -0.0028 | Ent: 0.2176
+[  7.3%] Iter  710/9765 | SPS:  1126 | Reward:  0.050 | Value:  0.625 | VLoss: 0.0210 | PLoss: -0.0078 | Ent: 0.2038
+[  7.3%] Iter  711/9765 | SPS:  1127 | Reward:  0.053 | Value:  0.650 | VLoss: 0.0257 | PLoss: -0.0067 | Ent: 0.1818
+[  7.3%] Iter  712/9765 | SPS:  1128 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0255 | PLoss: -0.0075 | Ent: 0.1772
+[  7.3%] Iter  713/9765 | SPS:  1129 | Reward:  0.067 | Value:  0.649 | VLoss: 0.0323 | PLoss: -0.0078 | Ent: 0.2070
+[  7.3%] Iter  714/9765 | SPS:  1130 | Reward:  0.052 | Value:  0.645 | VLoss: 0.0305 | PLoss: -0.0115 | Ent: 0.2006
+[  7.3%] Iter  715/9765 | SPS:  1131 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0340 | PLoss: -0.0058 | Ent: 0.1674
+[  7.3%] Iter  716/9765 | SPS:  1132 | Reward:  0.075 | Value:  0.623 | VLoss: 0.0341 | PLoss: -0.0066 | Ent: 0.2318
+[  7.3%] Iter  717/9765 | SPS:  1133 | Reward:  0.069 | Value:  0.662 | VLoss: 0.0426 | PLoss: -0.0114 | Ent: 0.2466
+[  7.4%] Iter  718/9765 | SPS:  1134 | Reward:  0.053 | Value:  0.711 | VLoss: 0.0300 | PLoss: -0.0129 | Ent: 0.2382
+[  7.4%] Iter  719/9765 | SPS:  1135 | Reward:  0.071 | Value:  0.596 | VLoss: 0.0434 | PLoss: -0.0017 | Ent: 0.2241
+[  7.4%] Iter  720/9765 | SPS:  1136 | Reward:  0.057 | Value:  0.638 | VLoss: 0.0347 | PLoss: -0.0120 | Ent: 0.1851
+[  7.4%] Iter  721/9765 | SPS:  1137 | Reward:  0.057 | Value:  0.563 | VLoss: 0.0357 | PLoss: -0.0145 | Ent: 0.2521
+[  7.4%] Iter  722/9765 | SPS:  1138 | Reward:  0.067 | Value:  0.571 | VLoss: 0.0287 | PLoss: -0.0042 | Ent: 0.1992
+[  7.4%] Iter  723/9765 | SPS:  1139 | Reward:  0.055 | Value:  0.579 | VLoss: 0.0369 | PLoss: -0.0080 | Ent: 0.2380
+[  7.4%] Iter  724/9765 | SPS:  1139 | Reward:  0.069 | Value:  0.556 | VLoss: 0.0400 | PLoss: -0.0108 | Ent: 0.2676
+[  7.4%] Iter  725/9765 | SPS:  1140 | Reward:  0.057 | Value:  0.631 | VLoss: 0.0305 | PLoss: -0.0118 | Ent: 0.2045
+[  7.4%] Iter  726/9765 | SPS:  1140 | Reward:  0.067 | Value:  0.636 | VLoss: 0.0316 | PLoss: -0.0081 | Ent: 0.2741
+[  7.4%] Iter  727/9765 | SPS:  1141 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0334 | PLoss: -0.0195 | Ent: 0.2522
+[  7.5%] Iter  728/9765 | SPS:  1141 | Reward:  0.051 | Value:  0.659 | VLoss: 0.0425 | PLoss: -0.0147 | Ent: 0.2800
+[  7.5%] Iter  729/9765 | SPS:  1142 | Reward:  0.084 | Value:  0.631 | VLoss: 0.0410 | PLoss: -0.0158 | Ent: 0.2254
+[  7.5%] Iter  730/9765 | SPS:  1142 | Reward:  0.062 | Value:  0.642 | VLoss: 0.0358 | PLoss: -0.0118 | Ent: 0.2001
+[  7.5%] Iter  731/9765 | SPS:  1143 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0361 | PLoss: -0.0025 | Ent: 0.2392
+[  7.5%] Iter  732/9765 | SPS:  1143 | Reward:  0.058 | Value:  0.651 | VLoss: 0.0248 | PLoss: -0.0102 | Ent: 0.2087
+[  7.5%] Iter  733/9765 | SPS:  1144 | Reward:  0.069 | Value:  0.684 | VLoss: 0.0238 | PLoss: -0.0117 | Ent: 0.2030
+[  7.5%] Iter  734/9765 | SPS:  1144 | Reward:  0.063 | Value:  0.691 | VLoss: 0.0275 | PLoss: -0.0020 | Ent: 0.2613
+[  7.5%] Iter  735/9765 | SPS:  1145 | Reward:  0.068 | Value:  0.675 | VLoss: 0.0363 | PLoss: -0.0083 | Ent: 0.1985
+[  7.5%] Iter  736/9765 | SPS:  1145 | Reward:  0.053 | Value:  0.670 | VLoss: 0.0328 | PLoss: -0.0069 | Ent: 0.2353
+[  7.5%] Iter  737/9765 | SPS:  1146 | Reward:  0.074 | Value:  0.651 | VLoss: 0.0379 | PLoss: -0.0043 | Ent: 0.1964
+[  7.6%] Iter  738/9765 | SPS:  1146 | Reward:  0.058 | Value:  0.643 | VLoss: 0.0294 | PLoss: -0.0052 | Ent: 0.2421
+[  7.6%] Iter  739/9765 | SPS:  1146 | Reward:  0.059 | Value:  0.620 | VLoss: 0.0381 | PLoss: 0.0003 | Ent: 0.1685
+[  7.6%] Iter  740/9765 | SPS:  1147 | Reward:  0.065 | Value:  0.592 | VLoss: 0.0327 | PLoss: -0.0076 | Ent: 0.2400
+[  7.6%] Iter  741/9765 | SPS:  1147 | Reward:  0.063 | Value:  0.565 | VLoss: 0.0377 | PLoss: -0.0096 | Ent: 0.2187
+[  7.6%] Iter  742/9765 | SPS:  1148 | Reward:  0.066 | Value:  0.653 | VLoss: 0.0381 | PLoss: -0.0067 | Ent: 0.2007
+[  7.6%] Iter  743/9765 | SPS:  1148 | Reward:  0.067 | Value:  0.677 | VLoss: 0.0359 | PLoss: -0.0060 | Ent: 0.1611
+[  7.6%] Iter  744/9765 | SPS:  1148 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0348 | PLoss: -0.0022 | Ent: 0.2345
+[  7.6%] Iter  745/9765 | SPS:  1149 | Reward:  0.075 | Value:  0.705 | VLoss: 0.0317 | PLoss: -0.0029 | Ent: 0.1547
+[  7.6%] Iter  746/9765 | SPS:  1149 | Reward:  0.063 | Value:  0.653 | VLoss: 0.0363 | PLoss: -0.0048 | Ent: 0.2176
+[  7.6%] Iter  747/9765 | SPS:  1149 | Reward:  0.068 | Value:  0.651 | VLoss: 0.0407 | PLoss: -0.0070 | Ent: 0.2173
+[  7.7%] Iter  748/9765 | SPS:  1149 | Reward:  0.076 | Value:  0.606 | VLoss: 0.0400 | PLoss: -0.0113 | Ent: 0.2106
+[  7.7%] Iter  749/9765 | SPS:  1149 | Reward:  0.063 | Value:  0.662 | VLoss: 0.0320 | PLoss: -0.0049 | Ent: 0.2399
+[  7.7%] Iter  750/9765 | SPS:  1149 | Reward:  0.054 | Value:  0.657 | VLoss: 0.0249 | PLoss: -0.0123 | Ent: 0.2205
+[  7.7%] Iter  751/9765 | SPS:  1149 | Reward:  0.065 | Value:  0.649 | VLoss: 0.0281 | PLoss: -0.0053 | Ent: 0.2166
+[  7.7%] Iter  752/9765 | SPS:  1149 | Reward:  0.054 | Value:  0.660 | VLoss: 0.0329 | PLoss: -0.0082 | Ent: 0.2311
+[  7.7%] Iter  753/9765 | SPS:  1149 | Reward:  0.052 | Value:  0.675 | VLoss: 0.0331 | PLoss: -0.0076 | Ent: 0.2282
+[  7.7%] Iter  754/9765 | SPS:  1148 | Reward:  0.058 | Value:  0.632 | VLoss: 0.0301 | PLoss: -0.0038 | Ent: 0.2676
+[  7.7%] Iter  755/9765 | SPS:  1148 | Reward:  0.051 | Value:  0.652 | VLoss: 0.0266 | PLoss: -0.0041 | Ent: 0.2450
+[  7.7%] Iter  756/9765 | SPS:  1148 | Reward:  0.053 | Value:  0.657 | VLoss: 0.0268 | PLoss: -0.0063 | Ent: 0.2141
+[  7.8%] Iter  757/9765 | SPS:  1147 | Reward:  0.046 | Value:  0.623 | VLoss: 0.0266 | PLoss: -0.0012 | Ent: 0.2035
+[  7.8%] Iter  758/9765 | SPS:  1147 | Reward:  0.058 | Value:  0.647 | VLoss: 0.0311 | PLoss: -0.0081 | Ent: 0.3009
+[  7.8%] Iter  759/9765 | SPS:  1148 | Reward:  0.074 | Value:  0.652 | VLoss: 0.0416 | PLoss: -0.0076 | Ent: 0.2068
+[  7.8%] Iter  760/9765 | SPS:  1148 | Reward:  0.068 | Value:  0.649 | VLoss: 0.0295 | PLoss: -0.0090 | Ent: 0.2128
+[  7.8%] Iter  761/9765 | SPS:  1148 | Reward:  0.062 | Value:  0.643 | VLoss: 0.0377 | PLoss: -0.0133 | Ent: 0.2198
+[  7.8%] Iter  762/9765 | SPS:  1147 | Reward:  0.066 | Value:  0.677 | VLoss: 0.0368 | PLoss: -0.0069 | Ent: 0.1578
+[  7.8%] Iter  763/9765 | SPS:  1146 | Reward:  0.055 | Value:  0.721 | VLoss: 0.0356 | PLoss: -0.0024 | Ent: 0.2015
+[  7.8%] Iter  764/9765 | SPS:  1146 | Reward:  0.050 | Value:  0.666 | VLoss: 0.0330 | PLoss: -0.0063 | Ent: 0.2318
+[  7.8%] Iter  765/9765 | SPS:  1145 | Reward:  0.063 | Value:  0.637 | VLoss: 0.0265 | PLoss: -0.0086 | Ent: 0.2509
+[  7.8%] Iter  766/9765 | SPS:  1144 | Reward:  0.070 | Value:  0.682 | VLoss: 0.0261 | PLoss: -0.0066 | Ent: 0.1998
+[  7.9%] Iter  767/9765 | SPS:  1144 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0210 | PLoss: -0.0097 | Ent: 0.1841
+[  7.9%] Iter  768/9765 | SPS:  1143 | Reward:  0.074 | Value:  0.698 | VLoss: 0.0332 | PLoss: -0.0117 | Ent: 0.2505
+[  7.9%] Iter  769/9765 | SPS:  1142 | Reward:  0.067 | Value:  0.726 | VLoss: 0.0294 | PLoss: -0.0081 | Ent: 0.2223
+[  7.9%] Iter  770/9765 | SPS:  1142 | Reward:  0.059 | Value:  0.673 | VLoss: 0.0291 | PLoss: -0.0078 | Ent: 0.2342
+[  7.9%] Iter  771/9765 | SPS:  1141 | Reward:  0.061 | Value:  0.659 | VLoss: 0.0419 | PLoss: -0.0062 | Ent: 0.2234
+[  7.9%] Iter  772/9765 | SPS:  1141 | Reward:  0.053 | Value:  0.631 | VLoss: 0.0248 | PLoss: -0.0051 | Ent: 0.2109
+[  7.9%] Iter  773/9765 | SPS:  1142 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0281 | PLoss: -0.0064 | Ent: 0.1926
+[  7.9%] Iter  774/9765 | SPS:  1143 | Reward:  0.062 | Value:  0.633 | VLoss: 0.0401 | PLoss: -0.0110 | Ent: 0.2081
+[  7.9%] Iter  775/9765 | SPS:  1144 | Reward:  0.072 | Value:  0.661 | VLoss: 0.0355 | PLoss: -0.0136 | Ent: 0.2313
+[  7.9%] Iter  776/9765 | SPS:  1142 | Reward:  0.062 | Value:  0.648 | VLoss: 0.0407 | PLoss: -0.0095 | Ent: 0.2347
+[  8.0%] Iter  777/9765 | SPS:  1140 | Reward:  0.070 | Value:  0.626 | VLoss: 0.0360 | PLoss: -0.0102 | Ent: 0.2074
+[  8.0%] Iter  778/9765 | SPS:  1137 | Reward:  0.054 | Value:  0.647 | VLoss: 0.0311 | PLoss: -0.0070 | Ent: 0.1617
+[  8.0%] Iter  779/9765 | SPS:  1135 | Reward:  0.058 | Value:  0.640 | VLoss: 0.0292 | PLoss: -0.0139 | Ent: 0.2324
+[  8.0%] Iter  780/9765 | SPS:  1133 | Reward:  0.060 | Value:  0.619 | VLoss: 0.0259 | PLoss: -0.0138 | Ent: 0.2446
+[  8.0%] Iter  781/9765 | SPS:  1130 | Reward:  0.055 | Value:  0.713 | VLoss: 0.0281 | PLoss: -0.0108 | Ent: 0.1962
+[  8.0%] Iter  782/9765 | SPS:  1128 | Reward:  0.072 | Value:  0.743 | VLoss: 0.0321 | PLoss: -0.0081 | Ent: 0.2762
+[  8.0%] Iter  783/9765 | SPS:  1126 | Reward:  0.078 | Value:  0.708 | VLoss: 0.0401 | PLoss: -0.0043 | Ent: 0.2123
+[  8.0%] Iter  784/9765 | SPS:  1124 | Reward:  0.063 | Value:  0.731 | VLoss: 0.0366 | PLoss: -0.0103 | Ent: 0.2310
+[  8.0%] Iter  785/9765 | SPS:  1122 | Reward:  0.066 | Value:  0.673 | VLoss: 0.0343 | PLoss: -0.0153 | Ent: 0.2134
+[  8.0%] Iter  786/9765 | SPS:  1119 | Reward:  0.074 | Value:  0.701 | VLoss: 0.0328 | PLoss: -0.0078 | Ent: 0.2244
+[  8.1%] Iter  787/9765 | SPS:  1117 | Reward:  0.083 | Value:  0.724 | VLoss: 0.0303 | PLoss: -0.0127 | Ent: 0.2265
+[  8.1%] Iter  788/9765 | SPS:  1115 | Reward:  0.067 | Value:  0.697 | VLoss: 0.0288 | PLoss: -0.0133 | Ent: 0.2491
+[  8.1%] Iter  789/9765 | SPS:  1113 | Reward:  0.072 | Value:  0.682 | VLoss: 0.0310 | PLoss: -0.0100 | Ent: 0.2114
+[  8.1%] Iter  790/9765 | SPS:  1111 | Reward:  0.069 | Value:  0.737 | VLoss: 0.0285 | PLoss: -0.0127 | Ent: 0.1839
+[  8.1%] Iter  791/9765 | SPS:  1109 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0352 | PLoss: -0.0067 | Ent: 0.2321
+[  8.1%] Iter  792/9765 | SPS:  1107 | Reward:  0.069 | Value:  0.610 | VLoss: 0.0414 | PLoss: -0.0044 | Ent: 0.2021
+[  8.1%] Iter  793/9765 | SPS:  1104 | Reward:  0.068 | Value:  0.638 | VLoss: 0.0385 | PLoss: -0.0034 | Ent: 0.2024
+[  8.1%] Iter  794/9765 | SPS:  1102 | Reward:  0.071 | Value:  0.604 | VLoss: 0.0356 | PLoss: -0.0051 | Ent: 0.1886
+[  8.1%] Iter  795/9765 | SPS:  1100 | Reward:  0.051 | Value:  0.672 | VLoss: 0.0368 | PLoss: -0.0120 | Ent: 0.2193
+[  8.2%] Iter  796/9765 | SPS:  1098 | Reward:  0.069 | Value:  0.622 | VLoss: 0.0285 | PLoss: -0.0111 | Ent: 0.1445
+[  8.2%] Iter  797/9765 | SPS:  1096 | Reward:  0.054 | Value:  0.648 | VLoss: 0.0315 | PLoss: -0.0129 | Ent: 0.1989
+[  8.2%] Iter  798/9765 | SPS:  1094 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0317 | PLoss: -0.0038 | Ent: 0.2100
+[  8.2%] Iter  799/9765 | SPS:  1092 | Reward:  0.040 | Value:  0.662 | VLoss: 0.0280 | PLoss: -0.0140 | Ent: 0.2416
+[  8.2%] Iter  800/9765 | SPS:  1090 | Reward:  0.073 | Value:  0.687 | VLoss: 0.0355 | PLoss: -0.0065 | Ent: 0.1490
+[  8.2%] Iter  801/9765 | SPS:  1088 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0336 | PLoss: -0.0028 | Ent: 0.2277
+[  8.2%] Iter  802/9765 | SPS:  1086 | Reward:  0.055 | Value:  0.661 | VLoss: 0.0270 | PLoss: -0.0152 | Ent: 0.2372
+[  8.2%] Iter  803/9765 | SPS:  1084 | Reward:  0.075 | Value:  0.651 | VLoss: 0.0402 | PLoss: -0.0038 | Ent: 0.1766
+[  8.2%] Iter  804/9765 | SPS:  1082 | Reward:  0.066 | Value:  0.626 | VLoss: 0.0323 | PLoss: -0.0174 | Ent: 0.1825
+[  8.2%] Iter  805/9765 | SPS:  1080 | Reward:  0.060 | Value:  0.628 | VLoss: 0.0386 | PLoss: -0.0044 | Ent: 0.1952
+[  8.3%] Iter  806/9765 | SPS:  1078 | Reward:  0.061 | Value:  0.670 | VLoss: 0.0337 | PLoss: -0.0150 | Ent: 0.2329
+[  8.3%] Iter  807/9765 | SPS:  1077 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0330 | PLoss: -0.0090 | Ent: 0.2125
+[  8.3%] Iter  808/9765 | SPS:  1075 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0350 | PLoss: -0.0037 | Ent: 0.2223
+[  8.3%] Iter  809/9765 | SPS:  1073 | Reward:  0.064 | Value:  0.670 | VLoss: 0.0357 | PLoss: -0.0053 | Ent: 0.2206
+[  8.3%] Iter  810/9765 | SPS:  1071 | Reward:  0.063 | Value:  0.641 | VLoss: 0.0394 | PLoss: -0.0098 | Ent: 0.2133
+[  8.3%] Iter  811/9765 | SPS:  1069 | Reward:  0.072 | Value:  0.644 | VLoss: 0.0385 | PLoss: -0.0085 | Ent: 0.1881
+[  8.3%] Iter  812/9765 | SPS:  1067 | Reward:  0.061 | Value:  0.655 | VLoss: 0.0353 | PLoss: -0.0091 | Ent: 0.2015
+[  8.3%] Iter  813/9765 | SPS:  1065 | Reward:  0.078 | Value:  0.658 | VLoss: 0.0346 | PLoss: -0.0064 | Ent: 0.2022
+[  8.3%] Iter  814/9765 | SPS:  1063 | Reward:  0.078 | Value:  0.644 | VLoss: 0.0386 | PLoss: -0.0115 | Ent: 0.1849
+[  8.3%] Iter  815/9765 | SPS:  1062 | Reward:  0.073 | Value:  0.695 | VLoss: 0.0393 | PLoss: 0.0002 | Ent: 0.2163
+[  8.4%] Iter  816/9765 | SPS:  1060 | Reward:  0.070 | Value:  0.660 | VLoss: 0.0411 | PLoss: -0.0093 | Ent: 0.1995
+[  8.4%] Iter  817/9765 | SPS:  1058 | Reward:  0.066 | Value:  0.645 | VLoss: 0.0382 | PLoss: -0.0089 | Ent: 0.1940
+[  8.4%] Iter  818/9765 | SPS:  1056 | Reward:  0.083 | Value:  0.635 | VLoss: 0.0328 | PLoss: -0.0068 | Ent: 0.2046
+[  8.4%] Iter  819/9765 | SPS:  1054 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0250 | PLoss: -0.0120 | Ent: 0.1901
+[  8.4%] Iter  820/9765 | SPS:  1053 | Reward:  0.073 | Value:  0.645 | VLoss: 0.0354 | PLoss: -0.0104 | Ent: 0.1914
+[  8.4%] Iter  821/9765 | SPS:  1051 | Reward:  0.068 | Value:  0.676 | VLoss: 0.0370 | PLoss: -0.0029 | Ent: 0.2052
+[  8.4%] Iter  822/9765 | SPS:  1049 | Reward:  0.054 | Value:  0.672 | VLoss: 0.0277 | PLoss: -0.0057 | Ent: 0.2006
+[  8.4%] Iter  823/9765 | SPS:  1047 | Reward:  0.073 | Value:  0.685 | VLoss: 0.0298 | PLoss: -0.0132 | Ent: 0.2272
+[  8.4%] Iter  824/9765 | SPS:  1046 | Reward:  0.078 | Value:  0.683 | VLoss: 0.0351 | PLoss: -0.0087 | Ent: 0.2298
+[  8.4%] Iter  825/9765 | SPS:  1044 | Reward:  0.071 | Value:  0.703 | VLoss: 0.0309 | PLoss: -0.0072 | Ent: 0.1812
+[  8.5%] Iter  826/9765 | SPS:  1042 | Reward:  0.080 | Value:  0.701 | VLoss: 0.0368 | PLoss: -0.0057 | Ent: 0.2051
+[  8.5%] Iter  827/9765 | SPS:  1040 | Reward:  0.070 | Value:  0.720 | VLoss: 0.0401 | PLoss: -0.0056 | Ent: 0.2451
+[  8.5%] Iter  828/9765 | SPS:  1039 | Reward:  0.068 | Value:  0.625 | VLoss: 0.0421 | PLoss: -0.0094 | Ent: 0.2292
+[  8.5%] Iter  829/9765 | SPS:  1037 | Reward:  0.079 | Value:  0.668 | VLoss: 0.0342 | PLoss: -0.0051 | Ent: 0.1844
+[  8.5%] Iter  830/9765 | SPS:  1035 | Reward:  0.053 | Value:  0.679 | VLoss: 0.0307 | PLoss: -0.0056 | Ent: 0.1778
+[  8.5%] Iter  831/9765 | SPS:  1034 | Reward:  0.068 | Value:  0.617 | VLoss: 0.0450 | PLoss: -0.0139 | Ent: 0.2570
+[  8.5%] Iter  832/9765 | SPS:  1032 | Reward:  0.076 | Value:  0.640 | VLoss: 0.0406 | PLoss: -0.0155 | Ent: 0.2452
+[  8.5%] Iter  833/9765 | SPS:  1030 | Reward:  0.066 | Value:  0.672 | VLoss: 0.0347 | PLoss: -0.0084 | Ent: 0.2413
+[  8.5%] Iter  834/9765 | SPS:  1029 | Reward:  0.064 | Value:  0.670 | VLoss: 0.0415 | PLoss: -0.0064 | Ent: 0.2830
+[  8.6%] Iter  835/9765 | SPS:  1027 | Reward:  0.066 | Value:  0.610 | VLoss: 0.0304 | PLoss: -0.0074 | Ent: 0.1868
+[  8.6%] Iter  836/9765 | SPS:  1026 | Reward:  0.064 | Value:  0.600 | VLoss: 0.0391 | PLoss: -0.0092 | Ent: 0.1957
+[  8.6%] Iter  837/9765 | SPS:  1024 | Reward:  0.063 | Value:  0.649 | VLoss: 0.0298 | PLoss: -0.0045 | Ent: 0.2354
+[  8.6%] Iter  838/9765 | SPS:  1022 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0309 | PLoss: -0.0046 | Ent: 0.2395
+[  8.6%] Iter  839/9765 | SPS:  1021 | Reward:  0.072 | Value:  0.660 | VLoss: 0.0410 | PLoss: -0.0081 | Ent: 0.2085
+[  8.6%] Iter  840/9765 | SPS:  1019 | Reward:  0.076 | Value:  0.703 | VLoss: 0.0285 | PLoss: 0.0013 | Ent: 0.2227
+[  8.6%] Iter  841/9765 | SPS:  1018 | Reward:  0.064 | Value:  0.692 | VLoss: 0.0334 | PLoss: -0.0082 | Ent: 0.2457
+[  8.6%] Iter  842/9765 | SPS:  1016 | Reward:  0.058 | Value:  0.668 | VLoss: 0.0333 | PLoss: -0.0024 | Ent: 0.2587
+[  8.6%] Iter  843/9765 | SPS:  1015 | Reward:  0.058 | Value:  0.661 | VLoss: 0.0331 | PLoss: -0.0065 | Ent: 0.1906
+[  8.6%] Iter  844/9765 | SPS:  1016 | Reward:  0.067 | Value:  0.633 | VLoss: 0.0379 | PLoss: -0.0093 | Ent: 0.2259
+[  8.7%] Iter  845/9765 | SPS:  1017 | Reward:  0.068 | Value:  0.638 | VLoss: 0.0254 | PLoss: -0.0129 | Ent: 0.1908
+[  8.7%] Iter  846/9765 | SPS:  1018 | Reward:  0.083 | Value:  0.659 | VLoss: 0.0337 | PLoss: -0.0125 | Ent: 0.2169
+[  8.7%] Iter  847/9765 | SPS:  1018 | Reward:  0.066 | Value:  0.683 | VLoss: 0.0315 | PLoss: -0.0032 | Ent: 0.1988
+[  8.7%] Iter  848/9765 | SPS:  1019 | Reward:  0.068 | Value:  0.726 | VLoss: 0.0338 | PLoss: -0.0143 | Ent: 0.1972
+[  8.7%] Iter  849/9765 | SPS:  1020 | Reward:  0.063 | Value:  0.721 | VLoss: 0.0347 | PLoss: -0.0055 | Ent: 0.2130
+[  8.7%] Iter  850/9765 | SPS:  1021 | Reward:  0.051 | Value:  0.685 | VLoss: 0.0293 | PLoss: -0.0125 | Ent: 0.2779
+[  8.7%] Iter  851/9765 | SPS:  1022 | Reward:  0.049 | Value:  0.637 | VLoss: 0.0321 | PLoss: 0.0005 | Ent: 0.2412
+[  8.7%] Iter  852/9765 | SPS:  1023 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0355 | PLoss: -0.0091 | Ent: 0.1725
+[  8.7%] Iter  853/9765 | SPS:  1024 | Reward:  0.062 | Value:  0.631 | VLoss: 0.0289 | PLoss: -0.0122 | Ent: 0.2423
+[  8.7%] Iter  854/9765 | SPS:  1024 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0371 | PLoss: -0.0134 | Ent: 0.1877
+[  8.8%] Iter  855/9765 | SPS:  1025 | Reward:  0.063 | Value:  0.633 | VLoss: 0.0361 | PLoss: -0.0041 | Ent: 0.1994
+[  8.8%] Iter  856/9765 | SPS:  1026 | Reward:  0.067 | Value:  0.664 | VLoss: 0.0249 | PLoss: -0.0088 | Ent: 0.1887
+[  8.8%] Iter  857/9765 | SPS:  1027 | Reward:  0.054 | Value:  0.682 | VLoss: 0.0250 | PLoss: -0.0032 | Ent: 0.1834
+[  8.8%] Iter  858/9765 | SPS:  1027 | Reward:  0.052 | Value:  0.655 | VLoss: 0.0275 | PLoss: -0.0163 | Ent: 0.1940
+[  8.8%] Iter  859/9765 | SPS:  1027 | Reward:  0.055 | Value:  0.637 | VLoss: 0.0261 | PLoss: -0.0128 | Ent: 0.1730
+[  8.8%] Iter  860/9765 | SPS:  1028 | Reward:  0.061 | Value:  0.645 | VLoss: 0.0273 | PLoss: -0.0058 | Ent: 0.1743
+[  8.8%] Iter  861/9765 | SPS:  1028 | Reward:  0.050 | Value:  0.613 | VLoss: 0.0293 | PLoss: -0.0077 | Ent: 0.2078
+[  8.8%] Iter  862/9765 | SPS:  1029 | Reward:  0.058 | Value:  0.627 | VLoss: 0.0376 | PLoss: -0.0172 | Ent: 0.2088
+[  8.8%] Iter  863/9765 | SPS:  1029 | Reward:  0.073 | Value:  0.678 | VLoss: 0.0347 | PLoss: -0.0039 | Ent: 0.1825
+[  8.8%] Iter  864/9765 | SPS:  1029 | Reward:  0.066 | Value:  0.655 | VLoss: 0.0360 | PLoss: -0.0069 | Ent: 0.2084
+[  8.9%] Iter  865/9765 | SPS:  1030 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0326 | PLoss: -0.0090 | Ent: 0.1985
+[  8.9%] Iter  866/9765 | SPS:  1030 | Reward:  0.060 | Value:  0.630 | VLoss: 0.0252 | PLoss: -0.0054 | Ent: 0.2096
+[  8.9%] Iter  867/9765 | SPS:  1031 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0378 | PLoss: -0.0175 | Ent: 0.1928
+[  8.9%] Iter  868/9765 | SPS:  1031 | Reward:  0.059 | Value:  0.679 | VLoss: 0.0260 | PLoss: -0.0107 | Ent: 0.2354
+[  8.9%] Iter  869/9765 | SPS:  1031 | Reward:  0.080 | Value:  0.652 | VLoss: 0.0414 | PLoss: -0.0093 | Ent: 0.1893
+[  8.9%] Iter  870/9765 | SPS:  1032 | Reward:  0.071 | Value:  0.641 | VLoss: 0.0370 | PLoss: -0.0108 | Ent: 0.1975
+[  8.9%] Iter  871/9765 | SPS:  1032 | Reward:  0.054 | Value:  0.666 | VLoss: 0.0321 | PLoss: -0.0077 | Ent: 0.1753
+[  8.9%] Iter  872/9765 | SPS:  1033 | Reward:  0.062 | Value:  0.622 | VLoss: 0.0329 | PLoss: -0.0089 | Ent: 0.1925
+[  8.9%] Iter  873/9765 | SPS:  1033 | Reward:  0.059 | Value:  0.664 | VLoss: 0.0301 | PLoss: -0.0050 | Ent: 0.1985
+[  9.0%] Iter  874/9765 | SPS:  1033 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0321 | PLoss: -0.0103 | Ent: 0.1741
+[  9.0%] Iter  875/9765 | SPS:  1033 | Reward:  0.064 | Value:  0.672 | VLoss: 0.0430 | PLoss: -0.0062 | Ent: 0.2024
+[  9.0%] Iter  876/9765 | SPS:  1033 | Reward:  0.072 | Value:  0.707 | VLoss: 0.0285 | PLoss: -0.0150 | Ent: 0.2157
+[  9.0%] Iter  877/9765 | SPS:  1034 | Reward:  0.054 | Value:  0.684 | VLoss: 0.0408 | PLoss: -0.0118 | Ent: 0.1997
+[  9.0%] Iter  878/9765 | SPS:  1034 | Reward:  0.064 | Value:  0.716 | VLoss: 0.0259 | PLoss: -0.0098 | Ent: 0.2344
+[  9.0%] Iter  879/9765 | SPS:  1034 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0367 | PLoss: 0.0016 | Ent: 0.1840
+[  9.0%] Iter  880/9765 | SPS:  1033 | Reward:  0.070 | Value:  0.639 | VLoss: 0.0337 | PLoss: -0.0103 | Ent: 0.2108
+[  9.0%] Iter  881/9765 | SPS:  1033 | Reward:  0.069 | Value:  0.719 | VLoss: 0.0237 | PLoss: -0.0107 | Ent: 0.2156
+[  9.0%] Iter  882/9765 | SPS:  1033 | Reward:  0.081 | Value:  0.687 | VLoss: 0.0411 | PLoss: -0.0052 | Ent: 0.2158
+[  9.0%] Iter  883/9765 | SPS:  1033 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0288 | PLoss: -0.0044 | Ent: 0.1803
+[  9.1%] Iter  884/9765 | SPS:  1032 | Reward:  0.075 | Value:  0.670 | VLoss: 0.0406 | PLoss: -0.0096 | Ent: 0.1936
+[  9.1%] Iter  885/9765 | SPS:  1033 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0309 | PLoss: -0.0045 | Ent: 0.2093
+[  9.1%] Iter  886/9765 | SPS:  1034 | Reward:  0.052 | Value:  0.631 | VLoss: 0.0267 | PLoss: -0.0067 | Ent: 0.2503
+[  9.1%] Iter  887/9765 | SPS:  1035 | Reward:  0.061 | Value:  0.622 | VLoss: 0.0376 | PLoss: -0.0090 | Ent: 0.2014
+[  9.1%] Iter  888/9765 | SPS:  1034 | Reward:  0.066 | Value:  0.656 | VLoss: 0.0330 | PLoss: -0.0062 | Ent: 0.2104
+[  9.1%] Iter  889/9765 | SPS:  1032 | Reward:  0.067 | Value:  0.650 | VLoss: 0.0360 | PLoss: -0.0111 | Ent: 0.2382
+[  9.1%] Iter  890/9765 | SPS:  1031 | Reward:  0.060 | Value:  0.619 | VLoss: 0.0369 | PLoss: -0.0082 | Ent: 0.1932
+[  9.1%] Iter  891/9765 | SPS:  1029 | Reward:  0.057 | Value:  0.629 | VLoss: 0.0268 | PLoss: -0.0022 | Ent: 0.1874
+[  9.1%] Iter  892/9765 | SPS:  1028 | Reward:  0.063 | Value:  0.702 | VLoss: 0.0267 | PLoss: -0.0092 | Ent: 0.1961
+[  9.1%] Iter  893/9765 | SPS:  1026 | Reward:  0.063 | Value:  0.664 | VLoss: 0.0374 | PLoss: -0.0089 | Ent: 0.2167
+[  9.2%] Iter  894/9765 | SPS:  1025 | Reward:  0.060 | Value:  0.672 | VLoss: 0.0300 | PLoss: -0.0098 | Ent: 0.2159
+[  9.2%] Iter  895/9765 | SPS:  1023 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0319 | PLoss: -0.0056 | Ent: 0.2269
+[  9.2%] Iter  896/9765 | SPS:  1022 | Reward:  0.070 | Value:  0.679 | VLoss: 0.0319 | PLoss: -0.0157 | Ent: 0.1927
+[  9.2%] Iter  897/9765 | SPS:  1020 | Reward:  0.065 | Value:  0.657 | VLoss: 0.0370 | PLoss: -0.0107 | Ent: 0.1842
+[  9.2%] Iter  898/9765 | SPS:  1019 | Reward:  0.070 | Value:  0.616 | VLoss: 0.0442 | PLoss: -0.0053 | Ent: 0.1877
+[  9.2%] Iter  899/9765 | SPS:  1017 | Reward:  0.072 | Value:  0.643 | VLoss: 0.0355 | PLoss: -0.0054 | Ent: 0.1680
+[  9.2%] Iter  900/9765 | SPS:  1016 | Reward:  0.056 | Value:  0.648 | VLoss: 0.0338 | PLoss: -0.0078 | Ent: 0.2151
+[  9.2%] Iter  901/9765 | SPS:  1014 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0352 | PLoss: -0.0086 | Ent: 0.2051
+[  9.2%] Iter  902/9765 | SPS:  1013 | Reward:  0.081 | Value:  0.624 | VLoss: 0.0353 | PLoss: 0.0011 | Ent: 0.2083
+[  9.2%] Iter  903/9765 | SPS:  1011 | Reward:  0.060 | Value:  0.656 | VLoss: 0.0325 | PLoss: -0.0093 | Ent: 0.2243
+[  9.3%] Iter  904/9765 | SPS:  1010 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0307 | PLoss: -0.0136 | Ent: 0.1939
+[  9.3%] Iter  905/9765 | SPS:  1009 | Reward:  0.060 | Value:  0.611 | VLoss: 0.0292 | PLoss: -0.0057 | Ent: 0.2458
+[  9.3%] Iter  906/9765 | SPS:  1007 | Reward:  0.060 | Value:  0.755 | VLoss: 0.0294 | PLoss: -0.0169 | Ent: 0.1869
+[  9.3%] Iter  907/9765 | SPS:  1006 | Reward:  0.077 | Value:  0.663 | VLoss: 0.0290 | PLoss: -0.0085 | Ent: 0.2002
+[  9.3%] Iter  908/9765 | SPS:  1004 | Reward:  0.070 | Value:  0.674 | VLoss: 0.0280 | PLoss: -0.0073 | Ent: 0.1635
+[  9.3%] Iter  909/9765 | SPS:  1003 | Reward:  0.043 | Value:  0.666 | VLoss: 0.0245 | PLoss: -0.0037 | Ent: 0.2007
+[  9.3%] Iter  910/9765 | SPS:  1002 | Reward:  0.058 | Value:  0.625 | VLoss: 0.0347 | PLoss: -0.0057 | Ent: 0.1949
+[  9.3%] Iter  911/9765 | SPS:  1000 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0323 | PLoss: -0.0067 | Ent: 0.1635
+[  9.3%] Iter  912/9765 | SPS:   999 | Reward:  0.071 | Value:  0.681 | VLoss: 0.0337 | PLoss: -0.0107 | Ent: 0.1565
+[  9.3%] Iter  913/9765 | SPS:   997 | Reward:  0.059 | Value:  0.706 | VLoss: 0.0292 | PLoss: -0.0081 | Ent: 0.1988
+[  9.4%] Iter  914/9765 | SPS:   996 | Reward:  0.054 | Value:  0.694 | VLoss: 0.0265 | PLoss: -0.0063 | Ent: 0.1699
+[  9.4%] Iter  915/9765 | SPS:   995 | Reward:  0.061 | Value:  0.675 | VLoss: 0.0301 | PLoss: -0.0029 | Ent: 0.1980
+[  9.4%] Iter  916/9765 | SPS:   993 | Reward:  0.068 | Value:  0.609 | VLoss: 0.0364 | PLoss: -0.0056 | Ent: 0.1770
+[  9.4%] Iter  917/9765 | SPS:   992 | Reward:  0.055 | Value:  0.644 | VLoss: 0.0301 | PLoss: -0.0077 | Ent: 0.2375
+[  9.4%] Iter  918/9765 | SPS:   991 | Reward:  0.066 | Value:  0.650 | VLoss: 0.0346 | PLoss: -0.0058 | Ent: 0.2318
+[  9.4%] Iter  919/9765 | SPS:   990 | Reward:  0.051 | Value:  0.661 | VLoss: 0.0340 | PLoss: 0.0013 | Ent: 0.1785
+[  9.4%] Iter  920/9765 | SPS:   990 | Reward:  0.064 | Value:  0.624 | VLoss: 0.0346 | PLoss: -0.0099 | Ent: 0.2034
+[  9.4%] Iter  921/9765 | SPS:   991 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0389 | PLoss: -0.0100 | Ent: 0.1962
+[  9.4%] Iter  922/9765 | SPS:   992 | Reward:  0.064 | Value:  0.563 | VLoss: 0.0380 | PLoss: -0.0181 | Ent: 0.2079
+[  9.5%] Iter  923/9765 | SPS:   993 | Reward:  0.066 | Value:  0.648 | VLoss: 0.0331 | PLoss: -0.0058 | Ent: 0.2081
+[  9.5%] Iter  924/9765 | SPS:   993 | Reward:  0.074 | Value:  0.714 | VLoss: 0.0295 | PLoss: -0.0026 | Ent: 0.1589
+[  9.5%] Iter  925/9765 | SPS:   993 | Reward:  0.068 | Value:  0.655 | VLoss: 0.0260 | PLoss: -0.0054 | Ent: 0.2311
+[  9.5%] Iter  926/9765 | SPS:   994 | Reward:  0.067 | Value:  0.688 | VLoss: 0.0288 | PLoss: -0.0060 | Ent: 0.1910
+[  9.5%] Iter  927/9765 | SPS:   994 | Reward:  0.078 | Value:  0.673 | VLoss: 0.0399 | PLoss: 0.0003 | Ent: 0.1619
+[  9.5%] Iter  928/9765 | SPS:   995 | Reward:  0.065 | Value:  0.667 | VLoss: 0.0356 | PLoss: -0.0053 | Ent: 0.1928
+[  9.5%] Iter  929/9765 | SPS:   996 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0256 | PLoss: -0.0125 | Ent: 0.2013
+[  9.5%] Iter  930/9765 | SPS:   996 | Reward:  0.058 | Value:  0.685 | VLoss: 0.0305 | PLoss: -0.0074 | Ent: 0.1392
+[  9.5%] Iter  931/9765 | SPS:   996 | Reward:  0.056 | Value:  0.631 | VLoss: 0.0336 | PLoss: -0.0060 | Ent: 0.1688
+[  9.5%] Iter  932/9765 | SPS:   995 | Reward:  0.071 | Value:  0.692 | VLoss: 0.0345 | PLoss: -0.0107 | Ent: 0.2114
+[  9.6%] Iter  933/9765 | SPS:   995 | Reward:  0.065 | Value:  0.653 | VLoss: 0.0378 | PLoss: -0.0077 | Ent: 0.1974
+[  9.6%] Iter  934/9765 | SPS:   994 | Reward:  0.068 | Value:  0.671 | VLoss: 0.0342 | PLoss: -0.0069 | Ent: 0.2006
+[  9.6%] Iter  935/9765 | SPS:   994 | Reward:  0.060 | Value:  0.716 | VLoss: 0.0231 | PLoss: -0.0039 | Ent: 0.1978
+[  9.6%] Iter  936/9765 | SPS:   993 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0367 | PLoss: -0.0077 | Ent: 0.2103
+[  9.6%] Iter  937/9765 | SPS:   994 | Reward:  0.065 | Value:  0.668 | VLoss: 0.0312 | PLoss: -0.0050 | Ent: 0.1642
+[  9.6%] Iter  938/9765 | SPS:   995 | Reward:  0.055 | Value:  0.638 | VLoss: 0.0390 | PLoss: -0.0081 | Ent: 0.1990
+[  9.6%] Iter  939/9765 | SPS:   995 | Reward:  0.075 | Value:  0.602 | VLoss: 0.0422 | PLoss: -0.0119 | Ent: 0.1900
+[  9.6%] Iter  940/9765 | SPS:   996 | Reward:  0.052 | Value:  0.628 | VLoss: 0.0236 | PLoss: -0.0114 | Ent: 0.1637
+[  9.6%] Iter  941/9765 | SPS:   997 | Reward:  0.065 | Value:  0.662 | VLoss: 0.0348 | PLoss: -0.0052 | Ent: 0.2461
+[  9.6%] Iter  942/9765 | SPS:   997 | Reward:  0.064 | Value:  0.662 | VLoss: 0.0274 | PLoss: -0.0060 | Ent: 0.1940
+[  9.7%] Iter  943/9765 | SPS:   998 | Reward:  0.066 | Value:  0.694 | VLoss: 0.0269 | PLoss: -0.0102 | Ent: 0.2231
+[  9.7%] Iter  944/9765 | SPS:   999 | Reward:  0.066 | Value:  0.682 | VLoss: 0.0358 | PLoss: -0.0055 | Ent: 0.1869
+[  9.7%] Iter  945/9765 | SPS:   999 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0247 | PLoss: -0.0092 | Ent: 0.2341
+[  9.7%] Iter  946/9765 | SPS:  1000 | Reward:  0.069 | Value:  0.658 | VLoss: 0.0298 | PLoss: -0.0036 | Ent: 0.1938
+[  9.7%] Iter  947/9765 | SPS:  1000 | Reward:  0.072 | Value:  0.705 | VLoss: 0.0400 | PLoss: -0.0113 | Ent: 0.1924
+[  9.7%] Iter  948/9765 | SPS:  1001 | Reward:  0.078 | Value:  0.650 | VLoss: 0.0401 | PLoss: -0.0065 | Ent: 0.2092
+[  9.7%] Iter  949/9765 | SPS:  1001 | Reward:  0.075 | Value:  0.676 | VLoss: 0.0309 | PLoss: -0.0104 | Ent: 0.1455
+[  9.7%] Iter  950/9765 | SPS:  1001 | Reward:  0.064 | Value:  0.678 | VLoss: 0.0305 | PLoss: -0.0020 | Ent: 0.1546
+[  9.7%] Iter  951/9765 | SPS:  1002 | Reward:  0.068 | Value:  0.684 | VLoss: 0.0265 | PLoss: -0.0064 | Ent: 0.2064
+[  9.7%] Iter  952/9765 | SPS:  1002 | Reward:  0.058 | Value:  0.701 | VLoss: 0.0287 | PLoss: -0.0105 | Ent: 0.2075
+[  9.8%] Iter  953/9765 | SPS:  1003 | Reward:  0.088 | Value:  0.643 | VLoss: 0.0330 | PLoss: -0.0016 | Ent: 0.2066
+[  9.8%] Iter  954/9765 | SPS:  1003 | Reward:  0.062 | Value:  0.692 | VLoss: 0.0431 | PLoss: -0.0065 | Ent: 0.2306
+[  9.8%] Iter  955/9765 | SPS:  1003 | Reward:  0.051 | Value:  0.629 | VLoss: 0.0293 | PLoss: -0.0070 | Ent: 0.2606
+[  9.8%] Iter  956/9765 | SPS:  1004 | Reward:  0.075 | Value:  0.674 | VLoss: 0.0311 | PLoss: -0.0144 | Ent: 0.2105
+[  9.8%] Iter  957/9765 | SPS:  1004 | Reward:  0.063 | Value:  0.699 | VLoss: 0.0406 | PLoss: -0.0082 | Ent: 0.2308
+[  9.8%] Iter  958/9765 | SPS:  1004 | Reward:  0.061 | Value:  0.645 | VLoss: 0.0295 | PLoss: -0.0145 | Ent: 0.2312
+[  9.8%] Iter  959/9765 | SPS:  1005 | Reward:  0.065 | Value:  0.661 | VLoss: 0.0415 | PLoss: -0.0109 | Ent: 0.1766
+[  9.8%] Iter  960/9765 | SPS:  1005 | Reward:  0.071 | Value:  0.686 | VLoss: 0.0415 | PLoss: -0.0089 | Ent: 0.1934
+[  9.8%] Iter  961/9765 | SPS:  1005 | Reward:  0.073 | Value:  0.680 | VLoss: 0.0246 | PLoss: -0.0049 | Ent: 0.1773
+[  9.9%] Iter  962/9765 | SPS:  1006 | Reward:  0.068 | Value:  0.659 | VLoss: 0.0470 | PLoss: -0.0055 | Ent: 0.2745
+[  9.9%] Iter  963/9765 | SPS:  1006 | Reward:  0.065 | Value:  0.629 | VLoss: 0.0278 | PLoss: -0.0101 | Ent: 0.1856
+[  9.9%] Iter  964/9765 | SPS:  1006 | Reward:  0.055 | Value:  0.722 | VLoss: 0.0306 | PLoss: -0.0018 | Ent: 0.1982
+[  9.9%] Iter  965/9765 | SPS:  1007 | Reward:  0.046 | Value:  0.695 | VLoss: 0.0241 | PLoss: -0.0053 | Ent: 0.2071
+[  9.9%] Iter  966/9765 | SPS:  1007 | Reward:  0.067 | Value:  0.613 | VLoss: 0.0302 | PLoss: -0.0021 | Ent: 0.1945
+[  9.9%] Iter  967/9765 | SPS:  1007 | Reward:  0.063 | Value:  0.659 | VLoss: 0.0373 | PLoss: -0.0060 | Ent: 0.1893
+[  9.9%] Iter  968/9765 | SPS:  1007 | Reward:  0.066 | Value:  0.654 | VLoss: 0.0406 | PLoss: -0.0082 | Ent: 0.1866
+[  9.9%] Iter  969/9765 | SPS:  1007 | Reward:  0.072 | Value:  0.689 | VLoss: 0.0308 | PLoss: -0.0076 | Ent: 0.2119
+[  9.9%] Iter  970/9765 | SPS:  1007 | Reward:  0.061 | Value:  0.649 | VLoss: 0.0263 | PLoss: -0.0145 | Ent: 0.1969
+[  9.9%] Iter  971/9765 | SPS:  1007 | Reward:  0.073 | Value:  0.690 | VLoss: 0.0317 | PLoss: -0.0112 | Ent: 0.2179
+[ 10.0%] Iter  972/9765 | SPS:  1007 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0358 | PLoss: -0.0124 | Ent: 0.1812
+[ 10.0%] Iter  973/9765 | SPS:  1007 | Reward:  0.050 | Value:  0.648 | VLoss: 0.0302 | PLoss: -0.0089 | Ent: 0.2094
+[ 10.0%] Iter  974/9765 | SPS:  1007 | Reward:  0.069 | Value:  0.619 | VLoss: 0.0381 | PLoss: -0.0101 | Ent: 0.1996
+[ 10.0%] Iter  975/9765 | SPS:  1007 | Reward:  0.065 | Value:  0.637 | VLoss: 0.0332 | PLoss: -0.0017 | Ent: 0.2516
+[ 10.0%] Iter  976/9765 | SPS:  1007 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0250 | PLoss: -0.0085 | Ent: 0.1947
+[ 10.0%] Iter  977/9765 | SPS:  1006 | Reward:  0.075 | Value:  0.652 | VLoss: 0.0353 | PLoss: -0.0017 | Ent: 0.1893
+[ 10.0%] Iter  978/9765 | SPS:  1006 | Reward:  0.068 | Value:  0.673 | VLoss: 0.0254 | PLoss: -0.0098 | Ent: 0.1861
+[ 10.0%] Iter  979/9765 | SPS:  1006 | Reward:  0.065 | Value:  0.692 | VLoss: 0.0289 | PLoss: -0.0017 | Ent: 0.1824
+[ 10.0%] Iter  980/9765 | SPS:  1007 | Reward:  0.059 | Value:  0.717 | VLoss: 0.0267 | PLoss: -0.0109 | Ent: 0.2022
+[ 10.0%] Iter  981/9765 | SPS:  1007 | Reward:  0.029 | Value:  0.698 | VLoss: 0.0191 | PLoss: -0.0056 | Ent: 0.2044
+[ 10.1%] Iter  982/9765 | SPS:  1008 | Reward:  0.039 | Value:  0.678 | VLoss: 0.0174 | PLoss: -0.0060 | Ent: 0.1961
+[ 10.1%] Iter  983/9765 | SPS:  1007 | Reward:  0.053 | Value:  0.619 | VLoss: 0.0200 | PLoss: -0.0073 | Ent: 0.2183
+[ 10.1%] Iter  984/9765 | SPS:  1006 | Reward:  0.055 | Value:  0.600 | VLoss: 0.0256 | PLoss: -0.0141 | Ent: 0.2187
+[ 10.1%] Iter  985/9765 | SPS:  1005 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0287 | PLoss: -0.0076 | Ent: 0.2258
+[ 10.1%] Iter  986/9765 | SPS:  1004 | Reward:  0.067 | Value:  0.736 | VLoss: 0.0292 | PLoss: -0.0062 | Ent: 0.2231
+[ 10.1%] Iter  987/9765 | SPS:  1002 | Reward:  0.081 | Value:  0.700 | VLoss: 0.0280 | PLoss: -0.0039 | Ent: 0.1855
+[ 10.1%] Iter  988/9765 | SPS:  1001 | Reward:  0.078 | Value:  0.705 | VLoss: 0.0335 | PLoss: -0.0034 | Ent: 0.2241
+[ 10.1%] Iter  989/9765 | SPS:  1002 | Reward:  0.066 | Value:  0.636 | VLoss: 0.0325 | PLoss: -0.0113 | Ent: 0.2046
+[ 10.1%] Iter  990/9765 | SPS:  1003 | Reward:  0.078 | Value:  0.702 | VLoss: 0.0370 | PLoss: -0.0089 | Ent: 0.2147
+[ 10.1%] Iter  991/9765 | SPS:  1004 | Reward:  0.063 | Value:  0.660 | VLoss: 0.0331 | PLoss: -0.0086 | Ent: 0.1732
+[ 10.2%] Iter  992/9765 | SPS:  1004 | Reward:  0.063 | Value:  0.693 | VLoss: 0.0297 | PLoss: -0.0119 | Ent: 0.2561
+[ 10.2%] Iter  993/9765 | SPS:  1004 | Reward:  0.069 | Value:  0.611 | VLoss: 0.0277 | PLoss: -0.0106 | Ent: 0.2630
+[ 10.2%] Iter  994/9765 | SPS:  1004 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0279 | PLoss: -0.0107 | Ent: 0.2885
+[ 10.2%] Iter  995/9765 | SPS:  1005 | Reward:  0.083 | Value:  0.689 | VLoss: 0.0291 | PLoss: -0.0029 | Ent: 0.1750
+[ 10.2%] Iter  996/9765 | SPS:  1006 | Reward:  0.065 | Value:  0.632 | VLoss: 0.0296 | PLoss: -0.0019 | Ent: 0.1696
+[ 10.2%] Iter  997/9765 | SPS:  1007 | Reward:  0.060 | Value:  0.642 | VLoss: 0.0264 | PLoss: -0.0084 | Ent: 0.1784
+[ 10.2%] Iter  998/9765 | SPS:  1007 | Reward:  0.079 | Value:  0.620 | VLoss: 0.0306 | PLoss: -0.0091 | Ent: 0.2019
+[ 10.2%] Iter  999/9765 | SPS:  1007 | Reward:  0.068 | Value:  0.623 | VLoss: 0.0374 | PLoss: -0.0050 | Ent: 0.1479
+[ 10.2%] Iter 1000/9765 | SPS:  1008 | Reward:  0.075 | Value:  0.624 | VLoss: 0.0333 | PLoss: -0.0071 | Ent: 0.2085
+[ 10.3%] Iter 1001/9765 | SPS:  1007 | Reward:  0.064 | Value:  0.640 | VLoss: 0.0344 | PLoss: 0.0005 | Ent: 0.1758
+[ 10.3%] Iter 1002/9765 | SPS:  1006 | Reward:  0.061 | Value:  0.702 | VLoss: 0.0285 | PLoss: -0.0097 | Ent: 0.1540
+[ 10.3%] Iter 1003/9765 | SPS:  1006 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0304 | PLoss: -0.0052 | Ent: 0.2207
+[ 10.3%] Iter 1004/9765 | SPS:  1006 | Reward:  0.055 | Value:  0.694 | VLoss: 0.0303 | PLoss: -0.0064 | Ent: 0.1572
+[ 10.3%] Iter 1005/9765 | SPS:  1005 | Reward:  0.058 | Value:  0.615 | VLoss: 0.0394 | PLoss: -0.0103 | Ent: 0.2560
+[ 10.3%] Iter 1006/9765 | SPS:  1006 | Reward:  0.065 | Value:  0.626 | VLoss: 0.0288 | PLoss: -0.0011 | Ent: 0.2161
+[ 10.3%] Iter 1007/9765 | SPS:  1006 | Reward:  0.066 | Value:  0.646 | VLoss: 0.0310 | PLoss: -0.0115 | Ent: 0.1690
+[ 10.3%] Iter 1008/9765 | SPS:  1007 | Reward:  0.073 | Value:  0.662 | VLoss: 0.0307 | PLoss: -0.0072 | Ent: 0.1796
+[ 10.3%] Iter 1009/9765 | SPS:  1008 | Reward:  0.071 | Value:  0.656 | VLoss: 0.0279 | PLoss: -0.0116 | Ent: 0.1877
+[ 10.3%] Iter 1010/9765 | SPS:  1008 | Reward:  0.062 | Value:  0.658 | VLoss: 0.0387 | PLoss: -0.0048 | Ent: 0.2186
+[ 10.4%] Iter 1011/9765 | SPS:  1009 | Reward:  0.064 | Value:  0.699 | VLoss: 0.0345 | PLoss: -0.0160 | Ent: 0.2143
+[ 10.4%] Iter 1012/9765 | SPS:  1010 | Reward:  0.057 | Value:  0.729 | VLoss: 0.0251 | PLoss: -0.0020 | Ent: 0.1945
+[ 10.4%] Iter 1013/9765 | SPS:  1010 | Reward:  0.063 | Value:  0.655 | VLoss: 0.0289 | PLoss: -0.0086 | Ent: 0.2030
+[ 10.4%] Iter 1014/9765 | SPS:  1011 | Reward:  0.059 | Value:  0.629 | VLoss: 0.0301 | PLoss: -0.0111 | Ent: 0.1809
+[ 10.4%] Iter 1015/9765 | SPS:  1012 | Reward:  0.067 | Value:  0.699 | VLoss: 0.0306 | PLoss: -0.0075 | Ent: 0.1902
+[ 10.4%] Iter 1016/9765 | SPS:  1012 | Reward:  0.064 | Value:  0.709 | VLoss: 0.0243 | PLoss: -0.0083 | Ent: 0.2022
+[ 10.4%] Iter 1017/9765 | SPS:  1013 | Reward:  0.066 | Value:  0.708 | VLoss: 0.0261 | PLoss: -0.0095 | Ent: 0.2343
+[ 10.4%] Iter 1018/9765 | SPS:  1013 | Reward:  0.073 | Value:  0.650 | VLoss: 0.0338 | PLoss: -0.0089 | Ent: 0.1969
+[ 10.4%] Iter 1019/9765 | SPS:  1013 | Reward:  0.060 | Value:  0.675 | VLoss: 0.0360 | PLoss: -0.0146 | Ent: 0.1815
+[ 10.4%] Iter 1020/9765 | SPS:  1014 | Reward:  0.068 | Value:  0.661 | VLoss: 0.0424 | PLoss: -0.0086 | Ent: 0.1784
+[ 10.5%] Iter 1021/9765 | SPS:  1014 | Reward:  0.057 | Value:  0.638 | VLoss: 0.0333 | PLoss: -0.0091 | Ent: 0.1479
+[ 10.5%] Iter 1022/9765 | SPS:  1014 | Reward:  0.059 | Value:  0.645 | VLoss: 0.0297 | PLoss: -0.0075 | Ent: 0.2232
+[ 10.5%] Iter 1023/9765 | SPS:  1015 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0272 | PLoss: -0.0029 | Ent: 0.1982
+[ 10.5%] Iter 1024/9765 | SPS:  1015 | Reward:  0.054 | Value:  0.684 | VLoss: 0.0330 | PLoss: -0.0041 | Ent: 0.1885
+[ 10.5%] Iter 1025/9765 | SPS:  1015 | Reward:  0.067 | Value:  0.685 | VLoss: 0.0297 | PLoss: -0.0112 | Ent: 0.1822
+[ 10.5%] Iter 1026/9765 | SPS:  1016 | Reward:  0.069 | Value:  0.730 | VLoss: 0.0271 | PLoss: -0.0071 | Ent: 0.1784
+[ 10.5%] Iter 1027/9765 | SPS:  1016 | Reward:  0.061 | Value:  0.648 | VLoss: 0.0386 | PLoss: -0.0050 | Ent: 0.1922
+[ 10.5%] Iter 1028/9765 | SPS:  1016 | Reward:  0.072 | Value:  0.695 | VLoss: 0.0265 | PLoss: -0.0148 | Ent: 0.1933
+[ 10.5%] Iter 1029/9765 | SPS:  1017 | Reward:  0.050 | Value:  0.670 | VLoss: 0.0213 | PLoss: -0.0054 | Ent: 0.2158
+[ 10.5%] Iter 1030/9765 | SPS:  1017 | Reward:  0.044 | Value:  0.643 | VLoss: 0.0275 | PLoss: -0.0085 | Ent: 0.2387
+[ 10.6%] Iter 1031/9765 | SPS:  1017 | Reward:  0.055 | Value:  0.694 | VLoss: 0.0301 | PLoss: -0.0182 | Ent: 0.2285
+[ 10.6%] Iter 1032/9765 | SPS:  1017 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0303 | PLoss: -0.0077 | Ent: 0.2256
+[ 10.6%] Iter 1033/9765 | SPS:  1018 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0355 | PLoss: -0.0103 | Ent: 0.1716
+[ 10.6%] Iter 1034/9765 | SPS:  1018 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0303 | PLoss: -0.0035 | Ent: 0.1288
+[ 10.6%] Iter 1035/9765 | SPS:  1018 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0339 | PLoss: -0.0098 | Ent: 0.2010
+[ 10.6%] Iter 1036/9765 | SPS:  1017 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0339 | PLoss: -0.0099 | Ent: 0.2098
+[ 10.6%] Iter 1037/9765 | SPS:  1017 | Reward:  0.061 | Value:  0.654 | VLoss: 0.0331 | PLoss: -0.0110 | Ent: 0.1723
+[ 10.6%] Iter 1038/9765 | SPS:  1017 | Reward:  0.055 | Value:  0.675 | VLoss: 0.0286 | PLoss: -0.0121 | Ent: 0.1910
+[ 10.6%] Iter 1039/9765 | SPS:  1018 | Reward:  0.067 | Value:  0.679 | VLoss: 0.0260 | PLoss: -0.0106 | Ent: 0.2354
+[ 10.7%] Iter 1040/9765 | SPS:  1018 | Reward:  0.049 | Value:  0.756 | VLoss: 0.0268 | PLoss: -0.0041 | Ent: 0.2136
+[ 10.7%] Iter 1041/9765 | SPS:  1019 | Reward:  0.057 | Value:  0.691 | VLoss: 0.0287 | PLoss: -0.0113 | Ent: 0.1841
+[ 10.7%] Iter 1042/9765 | SPS:  1019 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0279 | PLoss: -0.0081 | Ent: 0.1890
+[ 10.7%] Iter 1043/9765 | SPS:  1018 | Reward:  0.049 | Value:  0.585 | VLoss: 0.0289 | PLoss: -0.0035 | Ent: 0.1854
+[ 10.7%] Iter 1044/9765 | SPS:  1017 | Reward:  0.053 | Value:  0.641 | VLoss: 0.0261 | PLoss: -0.0026 | Ent: 0.1436
+[ 10.7%] Iter 1045/9765 | SPS:  1016 | Reward:  0.049 | Value:  0.681 | VLoss: 0.0314 | PLoss: -0.0047 | Ent: 0.1853
+[ 10.7%] Iter 1046/9765 | SPS:  1014 | Reward:  0.056 | Value:  0.654 | VLoss: 0.0199 | PLoss: -0.0097 | Ent: 0.1804
+[ 10.7%] Iter 1047/9765 | SPS:  1013 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0351 | PLoss: -0.0052 | Ent: 0.1771
+[ 10.7%] Iter 1048/9765 | SPS:  1012 | Reward:  0.067 | Value:  0.639 | VLoss: 0.0396 | PLoss: -0.0051 | Ent: 0.1442
+[ 10.7%] Iter 1049/9765 | SPS:  1011 | Reward:  0.076 | Value:  0.656 | VLoss: 0.0305 | PLoss: -0.0043 | Ent: 0.1596
+[ 10.8%] Iter 1050/9765 | SPS:  1009 | Reward:  0.057 | Value:  0.666 | VLoss: 0.0358 | PLoss: -0.0128 | Ent: 0.1847
+[ 10.8%] Iter 1051/9765 | SPS:  1008 | Reward:  0.054 | Value:  0.639 | VLoss: 0.0335 | PLoss: -0.0096 | Ent: 0.2003
+[ 10.8%] Iter 1052/9765 | SPS:  1007 | Reward:  0.076 | Value:  0.713 | VLoss: 0.0330 | PLoss: -0.0076 | Ent: 0.1660
+[ 10.8%] Iter 1053/9765 | SPS:  1006 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0280 | PLoss: -0.0054 | Ent: 0.1743
+[ 10.8%] Iter 1054/9765 | SPS:  1005 | Reward:  0.084 | Value:  0.704 | VLoss: 0.0307 | PLoss: -0.0068 | Ent: 0.1751
+[ 10.8%] Iter 1055/9765 | SPS:  1005 | Reward:  0.060 | Value:  0.698 | VLoss: 0.0333 | PLoss: -0.0050 | Ent: 0.1601
+[ 10.8%] Iter 1056/9765 | SPS:  1005 | Reward:  0.052 | Value:  0.741 | VLoss: 0.0194 | PLoss: -0.0113 | Ent: 0.2047
+[ 10.8%] Iter 1057/9765 | SPS:  1006 | Reward:  0.058 | Value:  0.690 | VLoss: 0.0290 | PLoss: -0.0093 | Ent: 0.2564
+[ 10.8%] Iter 1058/9765 | SPS:  1007 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0307 | PLoss: -0.0006 | Ent: 0.1921
+[ 10.8%] Iter 1059/9765 | SPS:  1007 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0274 | PLoss: -0.0086 | Ent: 0.1626
+[ 10.9%] Iter 1060/9765 | SPS:  1007 | Reward:  0.065 | Value:  0.743 | VLoss: 0.0315 | PLoss: -0.0022 | Ent: 0.1618
+[ 10.9%] Iter 1061/9765 | SPS:  1008 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0340 | PLoss: -0.0035 | Ent: 0.1481
+[ 10.9%] Iter 1062/9765 | SPS:  1008 | Reward:  0.060 | Value:  0.671 | VLoss: 0.0303 | PLoss: -0.0083 | Ent: 0.1375
+[ 10.9%] Iter 1063/9765 | SPS:  1009 | Reward:  0.062 | Value:  0.646 | VLoss: 0.0231 | PLoss: -0.0039 | Ent: 0.2127
+[ 10.9%] Iter 1064/9765 | SPS:  1010 | Reward:  0.059 | Value:  0.598 | VLoss: 0.0197 | PLoss: 0.0002 | Ent: 0.1794
+[ 10.9%] Iter 1065/9765 | SPS:  1010 | Reward:  0.068 | Value:  0.655 | VLoss: 0.0280 | PLoss: -0.0058 | Ent: 0.1727
+[ 10.9%] Iter 1066/9765 | SPS:  1011 | Reward:  0.043 | Value:  0.598 | VLoss: 0.0293 | PLoss: -0.0096 | Ent: 0.2321
+[ 10.9%] Iter 1067/9765 | SPS:  1010 | Reward:  0.052 | Value:  0.565 | VLoss: 0.0289 | PLoss: -0.0055 | Ent: 0.1554
+[ 10.9%] Iter 1068/9765 | SPS:  1010 | Reward:  0.052 | Value:  0.722 | VLoss: 0.0180 | PLoss: -0.0140 | Ent: 0.1941
+[ 10.9%] Iter 1069/9765 | SPS:  1009 | Reward:  0.061 | Value:  0.684 | VLoss: 0.0369 | PLoss: -0.0140 | Ent: 0.2245
+[ 11.0%] Iter 1070/9765 | SPS:  1009 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0230 | PLoss: -0.0056 | Ent: 0.1581
+[ 11.0%] Iter 1071/9765 | SPS:  1008 | Reward:  0.065 | Value:  0.648 | VLoss: 0.0307 | PLoss: -0.0120 | Ent: 0.1698
+[ 11.0%] Iter 1072/9765 | SPS:  1008 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0334 | PLoss: -0.0112 | Ent: 0.2303
+[ 11.0%] Iter 1073/9765 | SPS:  1009 | Reward:  0.083 | Value:  0.694 | VLoss: 0.0319 | PLoss: -0.0064 | Ent: 0.1679
+[ 11.0%] Iter 1074/9765 | SPS:  1009 | Reward:  0.071 | Value:  0.690 | VLoss: 0.0326 | PLoss: -0.0098 | Ent: 0.1996
+[ 11.0%] Iter 1075/9765 | SPS:  1010 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0399 | PLoss: -0.0065 | Ent: 0.1844
+[ 11.0%] Iter 1076/9765 | SPS:  1011 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0393 | PLoss: -0.0061 | Ent: 0.1469
+[ 11.0%] Iter 1077/9765 | SPS:  1011 | Reward:  0.065 | Value:  0.610 | VLoss: 0.0362 | PLoss: -0.0050 | Ent: 0.1660
+[ 11.0%] Iter 1078/9765 | SPS:  1012 | Reward:  0.054 | Value:  0.612 | VLoss: 0.0211 | PLoss: -0.0099 | Ent: 0.2058
+[ 11.0%] Iter 1079/9765 | SPS:  1012 | Reward:  0.058 | Value:  0.671 | VLoss: 0.0250 | PLoss: -0.0093 | Ent: 0.1750
+[ 11.1%] Iter 1080/9765 | SPS:  1013 | Reward:  0.050 | Value:  0.626 | VLoss: 0.0325 | PLoss: -0.0019 | Ent: 0.2301
+[ 11.1%] Iter 1081/9765 | SPS:  1014 | Reward:  0.063 | Value:  0.688 | VLoss: 0.0287 | PLoss: -0.0040 | Ent: 0.1640
+[ 11.1%] Iter 1082/9765 | SPS:  1014 | Reward:  0.061 | Value:  0.614 | VLoss: 0.0358 | PLoss: -0.0113 | Ent: 0.1567
+[ 11.1%] Iter 1083/9765 | SPS:  1014 | Reward:  0.068 | Value:  0.640 | VLoss: 0.0394 | PLoss: -0.0138 | Ent: 0.1896
+[ 11.1%] Iter 1084/9765 | SPS:  1015 | Reward:  0.063 | Value:  0.667 | VLoss: 0.0225 | PLoss: -0.0046 | Ent: 0.1481
+[ 11.1%] Iter 1085/9765 | SPS:  1015 | Reward:  0.063 | Value:  0.634 | VLoss: 0.0334 | PLoss: -0.0086 | Ent: 0.2121
+[ 11.1%] Iter 1086/9765 | SPS:  1015 | Reward:  0.065 | Value:  0.645 | VLoss: 0.0300 | PLoss: -0.0046 | Ent: 0.2227
+[ 11.1%] Iter 1087/9765 | SPS:  1016 | Reward:  0.060 | Value:  0.671 | VLoss: 0.0369 | PLoss: -0.0120 | Ent: 0.1894
+[ 11.1%] Iter 1088/9765 | SPS:  1016 | Reward:  0.080 | Value:  0.664 | VLoss: 0.0285 | PLoss: -0.0022 | Ent: 0.1602
+[ 11.2%] Iter 1089/9765 | SPS:  1016 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0356 | PLoss: -0.0085 | Ent: 0.1733
+[ 11.2%] Iter 1090/9765 | SPS:  1016 | Reward:  0.067 | Value:  0.712 | VLoss: 0.0316 | PLoss: -0.0066 | Ent: 0.1926
+[ 11.2%] Iter 1091/9765 | SPS:  1017 | Reward:  0.060 | Value:  0.717 | VLoss: 0.0311 | PLoss: -0.0020 | Ent: 0.1604
+[ 11.2%] Iter 1092/9765 | SPS:  1017 | Reward:  0.062 | Value:  0.698 | VLoss: 0.0257 | PLoss: -0.0221 | Ent: 0.2190
+[ 11.2%] Iter 1093/9765 | SPS:  1017 | Reward:  0.059 | Value:  0.692 | VLoss: 0.0232 | PLoss: -0.0034 | Ent: 0.1950
+[ 11.2%] Iter 1094/9765 | SPS:  1018 | Reward:  0.058 | Value:  0.651 | VLoss: 0.0393 | PLoss: -0.0126 | Ent: 0.1564
+[ 11.2%] Iter 1095/9765 | SPS:  1018 | Reward:  0.062 | Value:  0.623 | VLoss: 0.0358 | PLoss: -0.0081 | Ent: 0.1561
+[ 11.2%] Iter 1096/9765 | SPS:  1018 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0282 | PLoss: -0.0090 | Ent: 0.1874
+[ 11.2%] Iter 1097/9765 | SPS:  1019 | Reward:  0.059 | Value:  0.608 | VLoss: 0.0373 | PLoss: -0.0036 | Ent: 0.1814
+[ 11.2%] Iter 1098/9765 | SPS:  1019 | Reward:  0.061 | Value:  0.615 | VLoss: 0.0249 | PLoss: -0.0048 | Ent: 0.1926
+[ 11.3%] Iter 1099/9765 | SPS:  1019 | Reward:  0.077 | Value:  0.670 | VLoss: 0.0267 | PLoss: -0.0112 | Ent: 0.1739
+[ 11.3%] Iter 1100/9765 | SPS:  1019 | Reward:  0.073 | Value:  0.654 | VLoss: 0.0332 | PLoss: -0.0046 | Ent: 0.1725
+[ 11.3%] Iter 1101/9765 | SPS:  1019 | Reward:  0.061 | Value:  0.624 | VLoss: 0.0326 | PLoss: -0.0023 | Ent: 0.1762
+[ 11.3%] Iter 1102/9765 | SPS:  1018 | Reward:  0.064 | Value:  0.700 | VLoss: 0.0280 | PLoss: -0.0095 | Ent: 0.1458
+[ 11.3%] Iter 1103/9765 | SPS:  1018 | Reward:  0.083 | Value:  0.605 | VLoss: 0.0418 | PLoss: -0.0066 | Ent: 0.1732
+[ 11.3%] Iter 1104/9765 | SPS:  1018 | Reward:  0.055 | Value:  0.592 | VLoss: 0.0256 | PLoss: -0.0027 | Ent: 0.1182
+[ 11.3%] Iter 1105/9765 | SPS:  1019 | Reward:  0.070 | Value:  0.649 | VLoss: 0.0339 | PLoss: -0.0065 | Ent: 0.1469
+[ 11.3%] Iter 1106/9765 | SPS:  1020 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0357 | PLoss: -0.0092 | Ent: 0.1659
+[ 11.3%] Iter 1107/9765 | SPS:  1020 | Reward:  0.072 | Value:  0.657 | VLoss: 0.0301 | PLoss: -0.0071 | Ent: 0.1906
+[ 11.3%] Iter 1108/9765 | SPS:  1019 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0317 | PLoss: -0.0173 | Ent: 0.1832
+[ 11.4%] Iter 1109/9765 | SPS:  1018 | Reward:  0.061 | Value:  0.622 | VLoss: 0.0273 | PLoss: -0.0056 | Ent: 0.1992
+[ 11.4%] Iter 1110/9765 | SPS:  1018 | Reward:  0.041 | Value:  0.633 | VLoss: 0.0209 | PLoss: -0.0065 | Ent: 0.1853
+[ 11.4%] Iter 1111/9765 | SPS:  1018 | Reward:  0.064 | Value:  0.624 | VLoss: 0.0282 | PLoss: -0.0105 | Ent: 0.1442
+[ 11.4%] Iter 1112/9765 | SPS:  1019 | Reward:  0.059 | Value:  0.679 | VLoss: 0.0358 | PLoss: -0.0049 | Ent: 0.1547
+[ 11.4%] Iter 1113/9765 | SPS:  1020 | Reward:  0.061 | Value:  0.657 | VLoss: 0.0400 | PLoss: -0.0107 | Ent: 0.1616
+[ 11.4%] Iter 1114/9765 | SPS:  1020 | Reward:  0.055 | Value:  0.609 | VLoss: 0.0296 | PLoss: -0.0117 | Ent: 0.2086
+[ 11.4%] Iter 1115/9765 | SPS:  1020 | Reward:  0.065 | Value:  0.625 | VLoss: 0.0325 | PLoss: -0.0091 | Ent: 0.1742
+[ 11.4%] Iter 1116/9765 | SPS:  1021 | Reward:  0.056 | Value:  0.670 | VLoss: 0.0249 | PLoss: -0.0116 | Ent: 0.1881
+[ 11.4%] Iter 1117/9765 | SPS:  1021 | Reward:  0.072 | Value:  0.660 | VLoss: 0.0357 | PLoss: -0.0075 | Ent: 0.1758
+[ 11.4%] Iter 1118/9765 | SPS:  1022 | Reward:  0.061 | Value:  0.669 | VLoss: 0.0320 | PLoss: -0.0084 | Ent: 0.1756
+[ 11.5%] Iter 1119/9765 | SPS:  1022 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0322 | PLoss: -0.0007 | Ent: 0.1602
+[ 11.5%] Iter 1120/9765 | SPS:  1023 | Reward:  0.068 | Value:  0.603 | VLoss: 0.0357 | PLoss: -0.0151 | Ent: 0.1851
+[ 11.5%] Iter 1121/9765 | SPS:  1023 | Reward:  0.052 | Value:  0.609 | VLoss: 0.0292 | PLoss: -0.0088 | Ent: 0.1364
+[ 11.5%] Iter 1122/9765 | SPS:  1023 | Reward:  0.068 | Value:  0.652 | VLoss: 0.0299 | PLoss: -0.0086 | Ent: 0.1758
+[ 11.5%] Iter 1123/9765 | SPS:  1023 | Reward:  0.079 | Value:  0.682 | VLoss: 0.0323 | PLoss: -0.0033 | Ent: 0.1473
+[ 11.5%] Iter 1124/9765 | SPS:  1022 | Reward:  0.068 | Value:  0.673 | VLoss: 0.0259 | PLoss: -0.0069 | Ent: 0.1978
+[ 11.5%] Iter 1125/9765 | SPS:  1022 | Reward:  0.072 | Value:  0.700 | VLoss: 0.0345 | PLoss: -0.0101 | Ent: 0.1785
+[ 11.5%] Iter 1126/9765 | SPS:  1022 | Reward:  0.077 | Value:  0.686 | VLoss: 0.0287 | PLoss: -0.0031 | Ent: 0.2098
+[ 11.5%] Iter 1127/9765 | SPS:  1021 | Reward:  0.057 | Value:  0.704 | VLoss: 0.0339 | PLoss: -0.0070 | Ent: 0.1508
+[ 11.6%] Iter 1128/9765 | SPS:  1022 | Reward:  0.055 | Value:  0.661 | VLoss: 0.0214 | PLoss: -0.0092 | Ent: 0.1878
+[ 11.6%] Iter 1129/9765 | SPS:  1022 | Reward:  0.065 | Value:  0.698 | VLoss: 0.0319 | PLoss: -0.0056 | Ent: 0.1674
+[ 11.6%] Iter 1130/9765 | SPS:  1023 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0293 | PLoss: 0.0029 | Ent: 0.1954
+[ 11.6%] Iter 1131/9765 | SPS:  1024 | Reward:  0.070 | Value:  0.654 | VLoss: 0.0384 | PLoss: -0.0002 | Ent: 0.1564
+[ 11.6%] Iter 1132/9765 | SPS:  1024 | Reward:  0.066 | Value:  0.686 | VLoss: 0.0286 | PLoss: -0.0109 | Ent: 0.1482
+[ 11.6%] Iter 1133/9765 | SPS:  1025 | Reward:  0.054 | Value:  0.626 | VLoss: 0.0313 | PLoss: -0.0140 | Ent: 0.1953
+[ 11.6%] Iter 1134/9765 | SPS:  1025 | Reward:  0.052 | Value:  0.631 | VLoss: 0.0212 | PLoss: -0.0082 | Ent: 0.1649
+[ 11.6%] Iter 1135/9765 | SPS:  1026 | Reward:  0.054 | Value:  0.638 | VLoss: 0.0331 | PLoss: -0.0136 | Ent: 0.1462
+[ 11.6%] Iter 1136/9765 | SPS:  1026 | Reward:  0.055 | Value:  0.567 | VLoss: 0.0313 | PLoss: -0.0107 | Ent: 0.1914
+[ 11.6%] Iter 1137/9765 | SPS:  1027 | Reward:  0.060 | Value:  0.624 | VLoss: 0.0331 | PLoss: -0.0076 | Ent: 0.1425
+[ 11.7%] Iter 1138/9765 | SPS:  1027 | Reward:  0.074 | Value:  0.681 | VLoss: 0.0237 | PLoss: -0.0077 | Ent: 0.1399
+[ 11.7%] Iter 1139/9765 | SPS:  1027 | Reward:  0.069 | Value:  0.641 | VLoss: 0.0335 | PLoss: -0.0157 | Ent: 0.1911
+[ 11.7%] Iter 1140/9765 | SPS:  1028 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0312 | PLoss: -0.0060 | Ent: 0.2200
+[ 11.7%] Iter 1141/9765 | SPS:  1028 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0294 | PLoss: -0.0108 | Ent: 0.1949
+[ 11.7%] Iter 1142/9765 | SPS:  1028 | Reward:  0.062 | Value:  0.636 | VLoss: 0.0297 | PLoss: -0.0128 | Ent: 0.2188
+[ 11.7%] Iter 1143/9765 | SPS:  1028 | Reward:  0.070 | Value:  0.687 | VLoss: 0.0308 | PLoss: -0.0146 | Ent: 0.1752
+[ 11.7%] Iter 1144/9765 | SPS:  1029 | Reward:  0.058 | Value:  0.668 | VLoss: 0.0323 | PLoss: -0.0060 | Ent: 0.2225
+[ 11.7%] Iter 1145/9765 | SPS:  1029 | Reward:  0.072 | Value:  0.649 | VLoss: 0.0291 | PLoss: -0.0042 | Ent: 0.1154
+[ 11.7%] Iter 1146/9765 | SPS:  1029 | Reward:  0.060 | Value:  0.663 | VLoss: 0.0253 | PLoss: -0.0016 | Ent: 0.1805
+[ 11.7%] Iter 1147/9765 | SPS:  1030 | Reward:  0.040 | Value:  0.672 | VLoss: 0.0253 | PLoss: -0.0068 | Ent: 0.1718
+[ 11.8%] Iter 1148/9765 | SPS:  1030 | Reward:  0.041 | Value:  0.642 | VLoss: 0.0193 | PLoss: -0.0063 | Ent: 0.1449
+[ 11.8%] Iter 1149/9765 | SPS:  1030 | Reward:  0.059 | Value:  0.678 | VLoss: 0.0259 | PLoss: -0.0127 | Ent: 0.1579
+[ 11.8%] Iter 1150/9765 | SPS:  1031 | Reward:  0.076 | Value:  0.740 | VLoss: 0.0327 | PLoss: -0.0035 | Ent: 0.1528
+[ 11.8%] Iter 1151/9765 | SPS:  1031 | Reward:  0.073 | Value:  0.682 | VLoss: 0.0278 | PLoss: -0.0076 | Ent: 0.1810
+[ 11.8%] Iter 1152/9765 | SPS:  1031 | Reward:  0.064 | Value:  0.733 | VLoss: 0.0272 | PLoss: -0.0088 | Ent: 0.1353
+[ 11.8%] Iter 1153/9765 | SPS:  1031 | Reward:  0.050 | Value:  0.667 | VLoss: 0.0220 | PLoss: -0.0130 | Ent: 0.1650
+[ 11.8%] Iter 1154/9765 | SPS:  1031 | Reward:  0.052 | Value:  0.668 | VLoss: 0.0283 | PLoss: -0.0025 | Ent: 0.2037
+[ 11.8%] Iter 1155/9765 | SPS:  1031 | Reward:  0.041 | Value:  0.650 | VLoss: 0.0277 | PLoss: 0.0000 | Ent: 0.1627
+[ 11.8%] Iter 1156/9765 | SPS:  1031 | Reward:  0.062 | Value:  0.652 | VLoss: 0.0328 | PLoss: -0.0115 | Ent: 0.1689
+[ 11.8%] Iter 1157/9765 | SPS:  1031 | Reward:  0.075 | Value:  0.656 | VLoss: 0.0305 | PLoss: -0.0065 | Ent: 0.1646
+[ 11.9%] Iter 1158/9765 | SPS:  1031 | Reward:  0.067 | Value:  0.731 | VLoss: 0.0239 | PLoss: -0.0139 | Ent: 0.1646
+[ 11.9%] Iter 1159/9765 | SPS:  1031 | Reward:  0.049 | Value:  0.702 | VLoss: 0.0219 | PLoss: -0.0137 | Ent: 0.1363
+[ 11.9%] Iter 1160/9765 | SPS:  1030 | Reward:  0.064 | Value:  0.679 | VLoss: 0.0368 | PLoss: -0.0076 | Ent: 0.2003
+[ 11.9%] Iter 1161/9765 | SPS:  1031 | Reward:  0.062 | Value:  0.620 | VLoss: 0.0290 | PLoss: -0.0054 | Ent: 0.1837
+[ 11.9%] Iter 1162/9765 | SPS:  1031 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0454 | PLoss: -0.0117 | Ent: 0.2219
+[ 11.9%] Iter 1163/9765 | SPS:  1032 | Reward:  0.066 | Value:  0.617 | VLoss: 0.0318 | PLoss: -0.0062 | Ent: 0.2132
+[ 11.9%] Iter 1164/9765 | SPS:  1032 | Reward:  0.068 | Value:  0.683 | VLoss: 0.0268 | PLoss: -0.0054 | Ent: 0.1892
+[ 11.9%] Iter 1165/9765 | SPS:  1030 | Reward:  0.065 | Value:  0.630 | VLoss: 0.0407 | PLoss: -0.0110 | Ent: 0.1961
+[ 11.9%] Iter 1166/9765 | SPS:  1029 | Reward:  0.090 | Value:  0.678 | VLoss: 0.0346 | PLoss: -0.0034 | Ent: 0.1437
+[ 12.0%] Iter 1167/9765 | SPS:  1028 | Reward:  0.059 | Value:  0.713 | VLoss: 0.0296 | PLoss: -0.0063 | Ent: 0.1921
+[ 12.0%] Iter 1168/9765 | SPS:  1027 | Reward:  0.069 | Value:  0.711 | VLoss: 0.0310 | PLoss: -0.0106 | Ent: 0.1524
+[ 12.0%] Iter 1169/9765 | SPS:  1026 | Reward:  0.065 | Value:  0.688 | VLoss: 0.0331 | PLoss: -0.0133 | Ent: 0.1764
+[ 12.0%] Iter 1170/9765 | SPS:  1025 | Reward:  0.069 | Value:  0.653 | VLoss: 0.0307 | PLoss: -0.0071 | Ent: 0.1604
+[ 12.0%] Iter 1171/9765 | SPS:  1023 | Reward:  0.064 | Value:  0.708 | VLoss: 0.0314 | PLoss: -0.0082 | Ent: 0.1932
+[ 12.0%] Iter 1172/9765 | SPS:  1022 | Reward:  0.073 | Value:  0.670 | VLoss: 0.0388 | PLoss: -0.0076 | Ent: 0.1630
+[ 12.0%] Iter 1173/9765 | SPS:  1023 | Reward:  0.068 | Value:  0.710 | VLoss: 0.0352 | PLoss: -0.0063 | Ent: 0.1803
+[ 12.0%] Iter 1174/9765 | SPS:  1024 | Reward:  0.064 | Value:  0.668 | VLoss: 0.0321 | PLoss: -0.0112 | Ent: 0.1772
+[ 12.0%] Iter 1175/9765 | SPS:  1024 | Reward:  0.062 | Value:  0.608 | VLoss: 0.0379 | PLoss: -0.0020 | Ent: 0.1166
+[ 12.0%] Iter 1176/9765 | SPS:  1025 | Reward:  0.068 | Value:  0.646 | VLoss: 0.0250 | PLoss: -0.0066 | Ent: 0.2168
+[ 12.1%] Iter 1177/9765 | SPS:  1025 | Reward:  0.081 | Value:  0.692 | VLoss: 0.0331 | PLoss: -0.0068 | Ent: 0.1433
+[ 12.1%] Iter 1178/9765 | SPS:  1025 | Reward:  0.056 | Value:  0.709 | VLoss: 0.0303 | PLoss: -0.0114 | Ent: 0.1846
+[ 12.1%] Iter 1179/9765 | SPS:  1025 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0266 | PLoss: -0.0086 | Ent: 0.1830
+[ 12.1%] Iter 1180/9765 | SPS:  1026 | Reward:  0.074 | Value:  0.693 | VLoss: 0.0327 | PLoss: -0.0113 | Ent: 0.1646
+[ 12.1%] Iter 1181/9765 | SPS:  1027 | Reward:  0.070 | Value:  0.686 | VLoss: 0.0331 | PLoss: -0.0145 | Ent: 0.1932
+[ 12.1%] Iter 1182/9765 | SPS:  1027 | Reward:  0.066 | Value:  0.695 | VLoss: 0.0251 | PLoss: -0.0024 | Ent: 0.1428
+[ 12.1%] Iter 1183/9765 | SPS:  1028 | Reward:  0.079 | Value:  0.598 | VLoss: 0.0335 | PLoss: -0.0083 | Ent: 0.1811
+[ 12.1%] Iter 1184/9765 | SPS:  1028 | Reward:  0.067 | Value:  0.654 | VLoss: 0.0398 | PLoss: -0.0011 | Ent: 0.1473
+[ 12.1%] Iter 1185/9765 | SPS:  1028 | Reward:  0.066 | Value:  0.657 | VLoss: 0.0340 | PLoss: -0.0043 | Ent: 0.1814
+[ 12.1%] Iter 1186/9765 | SPS:  1028 | Reward:  0.063 | Value:  0.690 | VLoss: 0.0290 | PLoss: -0.0079 | Ent: 0.1651
+[ 12.2%] Iter 1187/9765 | SPS:  1027 | Reward:  0.054 | Value:  0.742 | VLoss: 0.0210 | PLoss: -0.0048 | Ent: 0.1322
+[ 12.2%] Iter 1188/9765 | SPS:  1027 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0325 | PLoss: -0.0044 | Ent: 0.1606
+[ 12.2%] Iter 1189/9765 | SPS:  1026 | Reward:  0.068 | Value:  0.647 | VLoss: 0.0286 | PLoss: -0.0039 | Ent: 0.1673
+[ 12.2%] Iter 1190/9765 | SPS:  1026 | Reward:  0.071 | Value:  0.654 | VLoss: 0.0403 | PLoss: -0.0061 | Ent: 0.1479
+[ 12.2%] Iter 1191/9765 | SPS:  1026 | Reward:  0.061 | Value:  0.666 | VLoss: 0.0389 | PLoss: -0.0091 | Ent: 0.1924
+[ 12.2%] Iter 1192/9765 | SPS:  1027 | Reward:  0.053 | Value:  0.580 | VLoss: 0.0373 | PLoss: -0.0108 | Ent: 0.1826
+[ 12.2%] Iter 1193/9765 | SPS:  1027 | Reward:  0.067 | Value:  0.581 | VLoss: 0.0308 | PLoss: -0.0021 | Ent: 0.1310
+[ 12.2%] Iter 1194/9765 | SPS:  1028 | Reward:  0.061 | Value:  0.680 | VLoss: 0.0275 | PLoss: -0.0015 | Ent: 0.1484
+[ 12.2%] Iter 1195/9765 | SPS:  1028 | Reward:  0.062 | Value:  0.648 | VLoss: 0.0335 | PLoss: -0.0111 | Ent: 0.2149
+[ 12.2%] Iter 1196/9765 | SPS:  1029 | Reward:  0.054 | Value:  0.648 | VLoss: 0.0285 | PLoss: 0.0055 | Ent: 0.1770
+[ 12.3%] Iter 1197/9765 | SPS:  1030 | Reward:  0.067 | Value:  0.605 | VLoss: 0.0368 | PLoss: -0.0095 | Ent: 0.1899
+[ 12.3%] Iter 1198/9765 | SPS:  1030 | Reward:  0.061 | Value:  0.679 | VLoss: 0.0236 | PLoss: -0.0090 | Ent: 0.2032
+[ 12.3%] Iter 1199/9765 | SPS:  1031 | Reward:  0.081 | Value:  0.649 | VLoss: 0.0300 | PLoss: -0.0021 | Ent: 0.1379
+[ 12.3%] Iter 1200/9765 | SPS:  1032 | Reward:  0.049 | Value:  0.613 | VLoss: 0.0209 | PLoss: -0.0123 | Ent: 0.1684
+[ 12.3%] Iter 1201/9765 | SPS:  1032 | Reward:  0.061 | Value:  0.687 | VLoss: 0.0378 | PLoss: -0.0057 | Ent: 0.2021
+[ 12.3%] Iter 1202/9765 | SPS:  1033 | Reward:  0.066 | Value:  0.666 | VLoss: 0.0330 | PLoss: -0.0088 | Ent: 0.1780
+[ 12.3%] Iter 1203/9765 | SPS:  1033 | Reward:  0.066 | Value:  0.669 | VLoss: 0.0303 | PLoss: -0.0110 | Ent: 0.1504
+[ 12.3%] Iter 1204/9765 | SPS:  1033 | Reward:  0.062 | Value:  0.666 | VLoss: 0.0272 | PLoss: -0.0136 | Ent: 0.1905
+[ 12.3%] Iter 1205/9765 | SPS:  1034 | Reward:  0.053 | Value:  0.703 | VLoss: 0.0274 | PLoss: -0.0017 | Ent: 0.1526
+[ 12.4%] Iter 1206/9765 | SPS:  1034 | Reward:  0.053 | Value:  0.669 | VLoss: 0.0329 | PLoss: -0.0085 | Ent: 0.1428
+[ 12.4%] Iter 1207/9765 | SPS:  1035 | Reward:  0.076 | Value:  0.700 | VLoss: 0.0363 | PLoss: -0.0170 | Ent: 0.1906
+[ 12.4%] Iter 1208/9765 | SPS:  1035 | Reward:  0.049 | Value:  0.688 | VLoss: 0.0208 | PLoss: -0.0056 | Ent: 0.1631
+[ 12.4%] Iter 1209/9765 | SPS:  1035 | Reward:  0.067 | Value:  0.651 | VLoss: 0.0337 | PLoss: -0.0014 | Ent: 0.2074
+[ 12.4%] Iter 1210/9765 | SPS:  1036 | Reward:  0.061 | Value:  0.652 | VLoss: 0.0345 | PLoss: -0.0060 | Ent: 0.2174
+[ 12.4%] Iter 1211/9765 | SPS:  1036 | Reward:  0.075 | Value:  0.675 | VLoss: 0.0374 | PLoss: -0.0064 | Ent: 0.1693
+[ 12.4%] Iter 1212/9765 | SPS:  1036 | Reward:  0.071 | Value:  0.640 | VLoss: 0.0404 | PLoss: -0.0120 | Ent: 0.1744
+[ 12.4%] Iter 1213/9765 | SPS:  1036 | Reward:  0.071 | Value:  0.718 | VLoss: 0.0323 | PLoss: -0.0110 | Ent: 0.1653
+[ 12.4%] Iter 1214/9765 | SPS:  1037 | Reward:  0.054 | Value:  0.638 | VLoss: 0.0296 | PLoss: -0.0101 | Ent: 0.2067
+[ 12.4%] Iter 1215/9765 | SPS:  1037 | Reward:  0.052 | Value:  0.651 | VLoss: 0.0316 | PLoss: -0.0098 | Ent: 0.1756
+[ 12.5%] Iter 1216/9765 | SPS:  1037 | Reward:  0.058 | Value:  0.685 | VLoss: 0.0352 | PLoss: -0.0078 | Ent: 0.1830
+[ 12.5%] Iter 1217/9765 | SPS:  1038 | Reward:  0.064 | Value:  0.652 | VLoss: 0.0339 | PLoss: -0.0068 | Ent: 0.1616
+[ 12.5%] Iter 1218/9765 | SPS:  1038 | Reward:  0.073 | Value:  0.655 | VLoss: 0.0296 | PLoss: -0.0087 | Ent: 0.1562
+[ 12.5%] Iter 1219/9765 | SPS:  1038 | Reward:  0.075 | Value:  0.636 | VLoss: 0.0354 | PLoss: -0.0107 | Ent: 0.1832
+[ 12.5%] Iter 1220/9765 | SPS:  1039 | Reward:  0.072 | Value:  0.673 | VLoss: 0.0300 | PLoss: -0.0057 | Ent: 0.1378
+[ 12.5%] Iter 1221/9765 | SPS:  1039 | Reward:  0.076 | Value:  0.625 | VLoss: 0.0298 | PLoss: -0.0028 | Ent: 0.1978
+[ 12.5%] Iter 1222/9765 | SPS:  1039 | Reward:  0.055 | Value:  0.677 | VLoss: 0.0383 | PLoss: -0.0060 | Ent: 0.2000
+[ 12.5%] Iter 1223/9765 | SPS:  1040 | Reward:  0.072 | Value:  0.679 | VLoss: 0.0417 | PLoss: -0.0150 | Ent: 0.2207
+[ 12.5%] Iter 1224/9765 | SPS:  1040 | Reward:  0.057 | Value:  0.665 | VLoss: 0.0307 | PLoss: -0.0088 | Ent: 0.1640
+[ 12.5%] Iter 1225/9765 | SPS:  1040 | Reward:  0.061 | Value:  0.631 | VLoss: 0.0306 | PLoss: -0.0055 | Ent: 0.1492
+[ 12.6%] Iter 1226/9765 | SPS:  1040 | Reward:  0.068 | Value:  0.655 | VLoss: 0.0341 | PLoss: -0.0066 | Ent: 0.1815
+[ 12.6%] Iter 1227/9765 | SPS:  1040 | Reward:  0.046 | Value:  0.586 | VLoss: 0.0250 | PLoss: -0.0077 | Ent: 0.1665
+[ 12.6%] Iter 1228/9765 | SPS:  1041 | Reward:  0.068 | Value:  0.672 | VLoss: 0.0285 | PLoss: -0.0072 | Ent: 0.1380
+[ 12.6%] Iter 1229/9765 | SPS:  1041 | Reward:  0.066 | Value:  0.613 | VLoss: 0.0351 | PLoss: -0.0144 | Ent: 0.1816
+[ 12.6%] Iter 1230/9765 | SPS:  1041 | Reward:  0.061 | Value:  0.592 | VLoss: 0.0353 | PLoss: -0.0054 | Ent: 0.2060
+[ 12.6%] Iter 1231/9765 | SPS:  1041 | Reward:  0.055 | Value:  0.619 | VLoss: 0.0314 | PLoss: -0.0082 | Ent: 0.1703
+[ 12.6%] Iter 1232/9765 | SPS:  1042 | Reward:  0.050 | Value:  0.610 | VLoss: 0.0306 | PLoss: -0.0106 | Ent: 0.2012
+[ 12.6%] Iter 1233/9765 | SPS:  1042 | Reward:  0.055 | Value:  0.631 | VLoss: 0.0265 | PLoss: -0.0037 | Ent: 0.2090
+[ 12.6%] Iter 1234/9765 | SPS:  1042 | Reward:  0.049 | Value:  0.668 | VLoss: 0.0235 | PLoss: -0.0042 | Ent: 0.1713
+[ 12.6%] Iter 1235/9765 | SPS:  1042 | Reward:  0.058 | Value:  0.641 | VLoss: 0.0296 | PLoss: -0.0044 | Ent: 0.1932
+[ 12.7%] Iter 1236/9765 | SPS:  1042 | Reward:  0.058 | Value:  0.649 | VLoss: 0.0436 | PLoss: -0.0024 | Ent: 0.1810
+[ 12.7%] Iter 1237/9765 | SPS:  1042 | Reward:  0.056 | Value:  0.663 | VLoss: 0.0256 | PLoss: -0.0056 | Ent: 0.1998
+[ 12.7%] Iter 1238/9765 | SPS:  1042 | Reward:  0.060 | Value:  0.672 | VLoss: 0.0349 | PLoss: -0.0023 | Ent: 0.1896
+[ 12.7%] Iter 1239/9765 | SPS:  1041 | Reward:  0.070 | Value:  0.620 | VLoss: 0.0387 | PLoss: -0.0051 | Ent: 0.1796
+[ 12.7%] Iter 1240/9765 | SPS:  1041 | Reward:  0.054 | Value:  0.664 | VLoss: 0.0319 | PLoss: -0.0101 | Ent: 0.1749
+[ 12.7%] Iter 1241/9765 | SPS:  1041 | Reward:  0.059 | Value:  0.725 | VLoss: 0.0317 | PLoss: -0.0096 | Ent: 0.1648
+[ 12.7%] Iter 1242/9765 | SPS:  1041 | Reward:  0.050 | Value:  0.737 | VLoss: 0.0203 | PLoss: -0.0069 | Ent: 0.1472
+[ 12.7%] Iter 1243/9765 | SPS:  1041 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0284 | PLoss: -0.0074 | Ent: 0.1666
+[ 12.7%] Iter 1244/9765 | SPS:  1040 | Reward:  0.053 | Value:  0.652 | VLoss: 0.0315 | PLoss: -0.0083 | Ent: 0.1829
+[ 12.7%] Iter 1245/9765 | SPS:  1040 | Reward:  0.060 | Value:  0.641 | VLoss: 0.0314 | PLoss: -0.0035 | Ent: 0.1682
+[ 12.8%] Iter 1246/9765 | SPS:  1040 | Reward:  0.080 | Value:  0.655 | VLoss: 0.0312 | PLoss: -0.0106 | Ent: 0.1673
+[ 12.8%] Iter 1247/9765 | SPS:  1040 | Reward:  0.065 | Value:  0.706 | VLoss: 0.0255 | PLoss: 0.0015 | Ent: 0.1512
+[ 12.8%] Iter 1248/9765 | SPS:  1039 | Reward:  0.053 | Value:  0.690 | VLoss: 0.0325 | PLoss: -0.0115 | Ent: 0.1320
+[ 12.8%] Iter 1249/9765 | SPS:  1039 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0320 | PLoss: -0.0141 | Ent: 0.1532
+[ 12.8%] Iter 1250/9765 | SPS:  1039 | Reward:  0.054 | Value:  0.616 | VLoss: 0.0348 | PLoss: -0.0127 | Ent: 0.2064
+[ 12.8%] Iter 1251/9765 | SPS:  1039 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0323 | PLoss: -0.0053 | Ent: 0.1915
+[ 12.8%] Iter 1252/9765 | SPS:  1038 | Reward:  0.063 | Value:  0.647 | VLoss: 0.0331 | PLoss: -0.0110 | Ent: 0.1650
+[ 12.8%] Iter 1253/9765 | SPS:  1038 | Reward:  0.067 | Value:  0.695 | VLoss: 0.0227 | PLoss: -0.0056 | Ent: 0.1901
+[ 12.8%] Iter 1254/9765 | SPS:  1038 | Reward:  0.070 | Value:  0.667 | VLoss: 0.0298 | PLoss: -0.0084 | Ent: 0.1864
+[ 12.9%] Iter 1255/9765 | SPS:  1038 | Reward:  0.056 | Value:  0.719 | VLoss: 0.0395 | PLoss: -0.0038 | Ent: 0.2091
+[ 12.9%] Iter 1256/9765 | SPS:  1037 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0423 | PLoss: -0.0163 | Ent: 0.1554
+[ 12.9%] Iter 1257/9765 | SPS:  1037 | Reward:  0.065 | Value:  0.674 | VLoss: 0.0273 | PLoss: -0.0078 | Ent: 0.1587
+[ 12.9%] Iter 1258/9765 | SPS:  1037 | Reward:  0.060 | Value:  0.702 | VLoss: 0.0278 | PLoss: -0.0087 | Ent: 0.1194
+[ 12.9%] Iter 1259/9765 | SPS:  1038 | Reward:  0.056 | Value:  0.638 | VLoss: 0.0216 | PLoss: -0.0060 | Ent: 0.1912
+[ 12.9%] Iter 1260/9765 | SPS:  1038 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0219 | PLoss: -0.0054 | Ent: 0.1628
+[ 12.9%] Iter 1261/9765 | SPS:  1039 | Reward:  0.066 | Value:  0.672 | VLoss: 0.0411 | PLoss: -0.0155 | Ent: 0.1544
+[ 12.9%] Iter 1262/9765 | SPS:  1039 | Reward:  0.064 | Value:  0.716 | VLoss: 0.0245 | PLoss: -0.0105 | Ent: 0.1918
+[ 12.9%] Iter 1263/9765 | SPS:  1040 | Reward:  0.055 | Value:  0.749 | VLoss: 0.0257 | PLoss: -0.0129 | Ent: 0.1857
+[ 12.9%] Iter 1264/9765 | SPS:  1041 | Reward:  0.047 | Value:  0.752 | VLoss: 0.0262 | PLoss: -0.0025 | Ent: 0.2040
+[ 13.0%] Iter 1265/9765 | SPS:  1041 | Reward:  0.060 | Value:  0.708 | VLoss: 0.0281 | PLoss: -0.0164 | Ent: 0.1697
+[ 13.0%] Iter 1266/9765 | SPS:  1042 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0256 | PLoss: -0.0087 | Ent: 0.1889
+[ 13.0%] Iter 1267/9765 | SPS:  1042 | Reward:  0.059 | Value:  0.694 | VLoss: 0.0258 | PLoss: -0.0136 | Ent: 0.1668
+[ 13.0%] Iter 1268/9765 | SPS:  1043 | Reward:  0.045 | Value:  0.702 | VLoss: 0.0267 | PLoss: -0.0079 | Ent: 0.1311
+[ 13.0%] Iter 1269/9765 | SPS:  1044 | Reward:  0.052 | Value:  0.679 | VLoss: 0.0229 | PLoss: -0.0112 | Ent: 0.1543
+[ 13.0%] Iter 1270/9765 | SPS:  1044 | Reward:  0.049 | Value:  0.645 | VLoss: 0.0274 | PLoss: -0.0074 | Ent: 0.2001
+[ 13.0%] Iter 1271/9765 | SPS:  1045 | Reward:  0.073 | Value:  0.678 | VLoss: 0.0268 | PLoss: -0.0066 | Ent: 0.1662
+[ 13.0%] Iter 1272/9765 | SPS:  1045 | Reward:  0.061 | Value:  0.668 | VLoss: 0.0267 | PLoss: -0.0035 | Ent: 0.1618
+[ 13.0%] Iter 1273/9765 | SPS:  1046 | Reward:  0.052 | Value:  0.729 | VLoss: 0.0216 | PLoss: -0.0018 | Ent: 0.1635
+[ 13.0%] Iter 1274/9765 | SPS:  1046 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0250 | PLoss: -0.0048 | Ent: 0.1789
+[ 13.1%] Iter 1275/9765 | SPS:  1047 | Reward:  0.058 | Value:  0.704 | VLoss: 0.0378 | PLoss: -0.0081 | Ent: 0.1463
+[ 13.1%] Iter 1276/9765 | SPS:  1048 | Reward:  0.049 | Value:  0.622 | VLoss: 0.0271 | PLoss: -0.0119 | Ent: 0.1740
+[ 13.1%] Iter 1277/9765 | SPS:  1048 | Reward:  0.061 | Value:  0.627 | VLoss: 0.0274 | PLoss: -0.0009 | Ent: 0.1863
+[ 13.1%] Iter 1278/9765 | SPS:  1049 | Reward:  0.066 | Value:  0.698 | VLoss: 0.0313 | PLoss: -0.0117 | Ent: 0.2081
+[ 13.1%] Iter 1279/9765 | SPS:  1049 | Reward:  0.071 | Value:  0.677 | VLoss: 0.0301 | PLoss: -0.0157 | Ent: 0.1901
+[ 13.1%] Iter 1280/9765 | SPS:  1050 | Reward:  0.069 | Value:  0.659 | VLoss: 0.0334 | PLoss: -0.0046 | Ent: 0.2026
+[ 13.1%] Iter 1281/9765 | SPS:  1050 | Reward:  0.066 | Value:  0.695 | VLoss: 0.0366 | PLoss: -0.0123 | Ent: 0.1746
+[ 13.1%] Iter 1282/9765 | SPS:  1051 | Reward:  0.047 | Value:  0.665 | VLoss: 0.0232 | PLoss: -0.0105 | Ent: 0.1897
+[ 13.1%] Iter 1283/9765 | SPS:  1052 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0279 | PLoss: -0.0063 | Ent: 0.1788
+[ 13.1%] Iter 1284/9765 | SPS:  1052 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0289 | PLoss: -0.0054 | Ent: 0.1680
+[ 13.2%] Iter 1285/9765 | SPS:  1053 | Reward:  0.062 | Value:  0.738 | VLoss: 0.0267 | PLoss: -0.0058 | Ent: 0.1687
+[ 13.2%] Iter 1286/9765 | SPS:  1053 | Reward:  0.071 | Value:  0.700 | VLoss: 0.0308 | PLoss: -0.0054 | Ent: 0.1563
+[ 13.2%] Iter 1287/9765 | SPS:  1054 | Reward:  0.053 | Value:  0.674 | VLoss: 0.0231 | PLoss: -0.0116 | Ent: 0.1968
+[ 13.2%] Iter 1288/9765 | SPS:  1055 | Reward:  0.054 | Value:  0.738 | VLoss: 0.0210 | PLoss: -0.0033 | Ent: 0.1779
+[ 13.2%] Iter 1289/9765 | SPS:  1055 | Reward:  0.062 | Value:  0.648 | VLoss: 0.0269 | PLoss: -0.0071 | Ent: 0.1713
+[ 13.2%] Iter 1290/9765 | SPS:  1056 | Reward:  0.060 | Value:  0.659 | VLoss: 0.0283 | PLoss: -0.0092 | Ent: 0.1903
+[ 13.2%] Iter 1291/9765 | SPS:  1056 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0271 | PLoss: -0.0063 | Ent: 0.1431
+[ 13.2%] Iter 1292/9765 | SPS:  1057 | Reward:  0.061 | Value:  0.679 | VLoss: 0.0345 | PLoss: -0.0062 | Ent: 0.1413
+[ 13.2%] Iter 1293/9765 | SPS:  1057 | Reward:  0.063 | Value:  0.642 | VLoss: 0.0349 | PLoss: -0.0031 | Ent: 0.1483
+[ 13.3%] Iter 1294/9765 | SPS:  1058 | Reward:  0.059 | Value:  0.661 | VLoss: 0.0277 | PLoss: -0.0066 | Ent: 0.1758
+[ 13.3%] Iter 1295/9765 | SPS:  1059 | Reward:  0.071 | Value:  0.668 | VLoss: 0.0294 | PLoss: -0.0132 | Ent: 0.1681
+[ 13.3%] Iter 1296/9765 | SPS:  1059 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0291 | PLoss: -0.0067 | Ent: 0.1547
+[ 13.3%] Iter 1297/9765 | SPS:  1060 | Reward:  0.079 | Value:  0.673 | VLoss: 0.0322 | PLoss: -0.0086 | Ent: 0.1596
+[ 13.3%] Iter 1298/9765 | SPS:  1060 | Reward:  0.059 | Value:  0.716 | VLoss: 0.0292 | PLoss: -0.0127 | Ent: 0.1888
+[ 13.3%] Iter 1299/9765 | SPS:  1061 | Reward:  0.052 | Value:  0.696 | VLoss: 0.0262 | PLoss: -0.0049 | Ent: 0.1869
+[ 13.3%] Iter 1300/9765 | SPS:  1062 | Reward:  0.067 | Value:  0.720 | VLoss: 0.0385 | PLoss: -0.0074 | Ent: 0.1399
+[ 13.3%] Iter 1301/9765 | SPS:  1062 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0250 | PLoss: -0.0039 | Ent: 0.1931
+[ 13.3%] Iter 1302/9765 | SPS:  1063 | Reward:  0.059 | Value:  0.660 | VLoss: 0.0263 | PLoss: -0.0053 | Ent: 0.1673
+[ 13.3%] Iter 1303/9765 | SPS:  1063 | Reward:  0.073 | Value:  0.675 | VLoss: 0.0328 | PLoss: -0.0099 | Ent: 0.1794
+[ 13.4%] Iter 1304/9765 | SPS:  1064 | Reward:  0.064 | Value:  0.638 | VLoss: 0.0328 | PLoss: -0.0135 | Ent: 0.1490
+[ 13.4%] Iter 1305/9765 | SPS:  1064 | Reward:  0.066 | Value:  0.721 | VLoss: 0.0287 | PLoss: -0.0034 | Ent: 0.1111
+[ 13.4%] Iter 1306/9765 | SPS:  1065 | Reward:  0.054 | Value:  0.711 | VLoss: 0.0354 | PLoss: -0.0109 | Ent: 0.1398
+[ 13.4%] Iter 1307/9765 | SPS:  1066 | Reward:  0.054 | Value:  0.711 | VLoss: 0.0237 | PLoss: -0.0060 | Ent: 0.2154
+[ 13.4%] Iter 1308/9765 | SPS:  1066 | Reward:  0.078 | Value:  0.686 | VLoss: 0.0293 | PLoss: -0.0076 | Ent: 0.1647
+[ 13.4%] Iter 1309/9765 | SPS:  1067 | Reward:  0.059 | Value:  0.739 | VLoss: 0.0210 | PLoss: -0.0032 | Ent: 0.1463
+[ 13.4%] Iter 1310/9765 | SPS:  1067 | Reward:  0.063 | Value:  0.695 | VLoss: 0.0304 | PLoss: -0.0099 | Ent: 0.1650
+[ 13.4%] Iter 1311/9765 | SPS:  1068 | Reward:  0.059 | Value:  0.649 | VLoss: 0.0347 | PLoss: -0.0101 | Ent: 0.1717
+[ 13.4%] Iter 1312/9765 | SPS:  1068 | Reward:  0.064 | Value:  0.690 | VLoss: 0.0298 | PLoss: -0.0042 | Ent: 0.1574
+[ 13.4%] Iter 1313/9765 | SPS:  1069 | Reward:  0.050 | Value:  0.701 | VLoss: 0.0245 | PLoss: -0.0095 | Ent: 0.2025
+[ 13.5%] Iter 1314/9765 | SPS:  1070 | Reward:  0.060 | Value:  0.688 | VLoss: 0.0340 | PLoss: -0.0090 | Ent: 0.1994
+[ 13.5%] Iter 1315/9765 | SPS:  1070 | Reward:  0.056 | Value:  0.701 | VLoss: 0.0216 | PLoss: -0.0033 | Ent: 0.1660
+[ 13.5%] Iter 1316/9765 | SPS:  1071 | Reward:  0.051 | Value:  0.740 | VLoss: 0.0205 | PLoss: -0.0125 | Ent: 0.1408
+[ 13.5%] Iter 1317/9765 | SPS:  1071 | Reward:  0.042 | Value:  0.712 | VLoss: 0.0265 | PLoss: -0.0079 | Ent: 0.1924
+[ 13.5%] Iter 1318/9765 | SPS:  1072 | Reward:  0.053 | Value:  0.669 | VLoss: 0.0207 | PLoss: -0.0060 | Ent: 0.1281
+[ 13.5%] Iter 1319/9765 | SPS:  1073 | Reward:  0.072 | Value:  0.627 | VLoss: 0.0285 | PLoss: -0.0067 | Ent: 0.1475
+[ 13.5%] Iter 1320/9765 | SPS:  1073 | Reward:  0.065 | Value:  0.689 | VLoss: 0.0297 | PLoss: -0.0087 | Ent: 0.1568
+[ 13.5%] Iter 1321/9765 | SPS:  1074 | Reward:  0.052 | Value:  0.615 | VLoss: 0.0261 | PLoss: -0.0078 | Ent: 0.2000
+[ 13.5%] Iter 1322/9765 | SPS:  1074 | Reward:  0.057 | Value:  0.648 | VLoss: 0.0263 | PLoss: -0.0116 | Ent: 0.1510
+[ 13.5%] Iter 1323/9765 | SPS:  1075 | Reward:  0.056 | Value:  0.679 | VLoss: 0.0230 | PLoss: -0.0123 | Ent: 0.2033
+[ 13.6%] Iter 1324/9765 | SPS:  1075 | Reward:  0.059 | Value:  0.649 | VLoss: 0.0269 | PLoss: -0.0093 | Ent: 0.1817
+[ 13.6%] Iter 1325/9765 | SPS:  1076 | Reward:  0.065 | Value:  0.681 | VLoss: 0.0289 | PLoss: -0.0050 | Ent: 0.1832
+[ 13.6%] Iter 1326/9765 | SPS:  1077 | Reward:  0.059 | Value:  0.686 | VLoss: 0.0265 | PLoss: -0.0104 | Ent: 0.1523
+[ 13.6%] Iter 1327/9765 | SPS:  1077 | Reward:  0.066 | Value:  0.684 | VLoss: 0.0483 | PLoss: -0.0127 | Ent: 0.1672
+[ 13.6%] Iter 1328/9765 | SPS:  1078 | Reward:  0.065 | Value:  0.698 | VLoss: 0.0303 | PLoss: -0.0143 | Ent: 0.1720
+[ 13.6%] Iter 1329/9765 | SPS:  1078 | Reward:  0.058 | Value:  0.672 | VLoss: 0.0262 | PLoss: -0.0001 | Ent: 0.1665
+[ 13.6%] Iter 1330/9765 | SPS:  1079 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0260 | PLoss: -0.0121 | Ent: 0.2004
+[ 13.6%] Iter 1331/9765 | SPS:  1079 | Reward:  0.061 | Value:  0.697 | VLoss: 0.0270 | PLoss: -0.0110 | Ent: 0.1508
+[ 13.6%] Iter 1332/9765 | SPS:  1080 | Reward:  0.077 | Value:  0.693 | VLoss: 0.0334 | PLoss: 0.0041 | Ent: 0.1849
+[ 13.7%] Iter 1333/9765 | SPS:  1080 | Reward:  0.060 | Value:  0.646 | VLoss: 0.0330 | PLoss: -0.0080 | Ent: 0.1852
+[ 13.7%] Iter 1334/9765 | SPS:  1081 | Reward:  0.063 | Value:  0.657 | VLoss: 0.0265 | PLoss: -0.0156 | Ent: 0.1662
+[ 13.7%] Iter 1335/9765 | SPS:  1082 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0260 | PLoss: -0.0097 | Ent: 0.1250
+[ 13.7%] Iter 1336/9765 | SPS:  1082 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0379 | PLoss: -0.0095 | Ent: 0.1939
+[ 13.7%] Iter 1337/9765 | SPS:  1083 | Reward:  0.055 | Value:  0.651 | VLoss: 0.0248 | PLoss: -0.0081 | Ent: 0.1380
+[ 13.7%] Iter 1338/9765 | SPS:  1083 | Reward:  0.053 | Value:  0.657 | VLoss: 0.0252 | PLoss: -0.0089 | Ent: 0.1714
+[ 13.7%] Iter 1339/9765 | SPS:  1084 | Reward:  0.048 | Value:  0.640 | VLoss: 0.0268 | PLoss: -0.0109 | Ent: 0.1486
+[ 13.7%] Iter 1340/9765 | SPS:  1084 | Reward:  0.052 | Value:  0.642 | VLoss: 0.0190 | PLoss: -0.0092 | Ent: 0.1937
+[ 13.7%] Iter 1341/9765 | SPS:  1085 | Reward:  0.073 | Value:  0.637 | VLoss: 0.0404 | PLoss: -0.0090 | Ent: 0.1753
+[ 13.7%] Iter 1342/9765 | SPS:  1086 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0291 | PLoss: -0.0085 | Ent: 0.1728
+[ 13.8%] Iter 1343/9765 | SPS:  1086 | Reward:  0.075 | Value:  0.656 | VLoss: 0.0319 | PLoss: -0.0159 | Ent: 0.2164
+[ 13.8%] Iter 1344/9765 | SPS:  1087 | Reward:  0.063 | Value:  0.709 | VLoss: 0.0308 | PLoss: -0.0041 | Ent: 0.1986
+[ 13.8%] Iter 1345/9765 | SPS:  1087 | Reward:  0.048 | Value:  0.690 | VLoss: 0.0314 | PLoss: -0.0057 | Ent: 0.1789
+[ 13.8%] Iter 1346/9765 | SPS:  1088 | Reward:  0.067 | Value:  0.687 | VLoss: 0.0302 | PLoss: -0.0066 | Ent: 0.1857
+[ 13.8%] Iter 1347/9765 | SPS:  1088 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0285 | PLoss: -0.0187 | Ent: 0.1970
+[ 13.8%] Iter 1348/9765 | SPS:  1089 | Reward:  0.055 | Value:  0.614 | VLoss: 0.0321 | PLoss: -0.0040 | Ent: 0.1655
+[ 13.8%] Iter 1349/9765 | SPS:  1090 | Reward:  0.070 | Value:  0.746 | VLoss: 0.0239 | PLoss: -0.0064 | Ent: 0.1590
+[ 13.8%] Iter 1350/9765 | SPS:  1090 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0282 | PLoss: -0.0056 | Ent: 0.1649
+[ 13.8%] Iter 1351/9765 | SPS:  1091 | Reward:  0.060 | Value:  0.681 | VLoss: 0.0346 | PLoss: -0.0139 | Ent: 0.1417
+[ 13.8%] Iter 1352/9765 | SPS:  1091 | Reward:  0.064 | Value:  0.632 | VLoss: 0.0368 | PLoss: -0.0131 | Ent: 0.1674
+[ 13.9%] Iter 1353/9765 | SPS:  1092 | Reward:  0.069 | Value:  0.652 | VLoss: 0.0351 | PLoss: -0.0100 | Ent: 0.1437
+[ 13.9%] Iter 1354/9765 | SPS:  1092 | Reward:  0.065 | Value:  0.632 | VLoss: 0.0349 | PLoss: -0.0083 | Ent: 0.1711
+[ 13.9%] Iter 1355/9765 | SPS:  1093 | Reward:  0.065 | Value:  0.683 | VLoss: 0.0349 | PLoss: -0.0112 | Ent: 0.2024
+[ 13.9%] Iter 1356/9765 | SPS:  1094 | Reward:  0.059 | Value:  0.629 | VLoss: 0.0285 | PLoss: -0.0055 | Ent: 0.1522
+[ 13.9%] Iter 1357/9765 | SPS:  1094 | Reward:  0.053 | Value:  0.622 | VLoss: 0.0361 | PLoss: -0.0063 | Ent: 0.1651
+[ 13.9%] Iter 1358/9765 | SPS:  1095 | Reward:  0.070 | Value:  0.663 | VLoss: 0.0308 | PLoss: -0.0047 | Ent: 0.1248
+[ 13.9%] Iter 1359/9765 | SPS:  1095 | Reward:  0.056 | Value:  0.698 | VLoss: 0.0208 | PLoss: -0.0066 | Ent: 0.1535
+[ 13.9%] Iter 1360/9765 | SPS:  1096 | Reward:  0.053 | Value:  0.712 | VLoss: 0.0288 | PLoss: -0.0112 | Ent: 0.1841
+[ 13.9%] Iter 1361/9765 | SPS:  1096 | Reward:  0.061 | Value:  0.672 | VLoss: 0.0309 | PLoss: -0.0168 | Ent: 0.1558
+[ 13.9%] Iter 1362/9765 | SPS:  1097 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0270 | PLoss: -0.0061 | Ent: 0.1654
+[ 14.0%] Iter 1363/9765 | SPS:  1098 | Reward:  0.069 | Value:  0.690 | VLoss: 0.0374 | PLoss: -0.0063 | Ent: 0.1551
+[ 14.0%] Iter 1364/9765 | SPS:  1098 | Reward:  0.066 | Value:  0.650 | VLoss: 0.0298 | PLoss: -0.0006 | Ent: 0.1897
+[ 14.0%] Iter 1365/9765 | SPS:  1099 | Reward:  0.059 | Value:  0.634 | VLoss: 0.0407 | PLoss: 0.0002 | Ent: 0.1858
+[ 14.0%] Iter 1366/9765 | SPS:  1099 | Reward:  0.061 | Value:  0.653 | VLoss: 0.0236 | PLoss: -0.0081 | Ent: 0.1735
+[ 14.0%] Iter 1367/9765 | SPS:  1100 | Reward:  0.054 | Value:  0.616 | VLoss: 0.0292 | PLoss: -0.0063 | Ent: 0.1836
+[ 14.0%] Iter 1368/9765 | SPS:  1100 | Reward:  0.066 | Value:  0.659 | VLoss: 0.0430 | PLoss: -0.0104 | Ent: 0.1833
+[ 14.0%] Iter 1369/9765 | SPS:  1101 | Reward:  0.062 | Value:  0.700 | VLoss: 0.0304 | PLoss: -0.0137 | Ent: 0.1555
+[ 14.0%] Iter 1370/9765 | SPS:  1101 | Reward:  0.048 | Value:  0.654 | VLoss: 0.0287 | PLoss: -0.0051 | Ent: 0.1870
+[ 14.0%] Iter 1371/9765 | SPS:  1102 | Reward:  0.073 | Value:  0.680 | VLoss: 0.0402 | PLoss: -0.0053 | Ent: 0.1691
+[ 14.1%] Iter 1372/9765 | SPS:  1103 | Reward:  0.048 | Value:  0.650 | VLoss: 0.0283 | PLoss: -0.0103 | Ent: 0.1758
+[ 14.1%] Iter 1373/9765 | SPS:  1103 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0244 | PLoss: -0.0045 | Ent: 0.1863
+[ 14.1%] Iter 1374/9765 | SPS:  1104 | Reward:  0.066 | Value:  0.676 | VLoss: 0.0341 | PLoss: -0.0086 | Ent: 0.1637
+[ 14.1%] Iter 1375/9765 | SPS:  1104 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0245 | PLoss: -0.0051 | Ent: 0.1384
+[ 14.1%] Iter 1376/9765 | SPS:  1105 | Reward:  0.058 | Value:  0.705 | VLoss: 0.0301 | PLoss: -0.0090 | Ent: 0.1939
+[ 14.1%] Iter 1377/9765 | SPS:  1105 | Reward:  0.059 | Value:  0.612 | VLoss: 0.0383 | PLoss: -0.0082 | Ent: 0.1660
+[ 14.1%] Iter 1378/9765 | SPS:  1106 | Reward:  0.058 | Value:  0.627 | VLoss: 0.0320 | PLoss: -0.0068 | Ent: 0.1982
+[ 14.1%] Iter 1379/9765 | SPS:  1106 | Reward:  0.068 | Value:  0.659 | VLoss: 0.0328 | PLoss: -0.0066 | Ent: 0.2259
+[ 14.1%] Iter 1380/9765 | SPS:  1107 | Reward:  0.064 | Value:  0.632 | VLoss: 0.0294 | PLoss: -0.0137 | Ent: 0.1963
+[ 14.1%] Iter 1381/9765 | SPS:  1108 | Reward:  0.055 | Value:  0.673 | VLoss: 0.0294 | PLoss: -0.0106 | Ent: 0.1773
+[ 14.2%] Iter 1382/9765 | SPS:  1108 | Reward:  0.069 | Value:  0.625 | VLoss: 0.0315 | PLoss: -0.0048 | Ent: 0.1271
+[ 14.2%] Iter 1383/9765 | SPS:  1109 | Reward:  0.063 | Value:  0.681 | VLoss: 0.0263 | PLoss: -0.0146 | Ent: 0.2098
+[ 14.2%] Iter 1384/9765 | SPS:  1109 | Reward:  0.063 | Value:  0.741 | VLoss: 0.0358 | PLoss: -0.0083 | Ent: 0.1906
+[ 14.2%] Iter 1385/9765 | SPS:  1110 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0265 | PLoss: 0.0010 | Ent: 0.1882
+[ 14.2%] Iter 1386/9765 | SPS:  1110 | Reward:  0.084 | Value:  0.676 | VLoss: 0.0357 | PLoss: -0.0103 | Ent: 0.1916
+[ 14.2%] Iter 1387/9765 | SPS:  1111 | Reward:  0.076 | Value:  0.743 | VLoss: 0.0271 | PLoss: -0.0116 | Ent: 0.1990
+[ 14.2%] Iter 1388/9765 | SPS:  1112 | Reward:  0.065 | Value:  0.729 | VLoss: 0.0309 | PLoss: -0.0131 | Ent: 0.1454
+[ 14.2%] Iter 1389/9765 | SPS:  1112 | Reward:  0.063 | Value:  0.649 | VLoss: 0.0365 | PLoss: -0.0250 | Ent: 0.2132
+[ 14.2%] Iter 1390/9765 | SPS:  1113 | Reward:  0.053 | Value:  0.667 | VLoss: 0.0255 | PLoss: -0.0154 | Ent: 0.1684
+[ 14.2%] Iter 1391/9765 | SPS:  1113 | Reward:  0.057 | Value:  0.685 | VLoss: 0.0246 | PLoss: -0.0063 | Ent: 0.1728
+[ 14.3%] Iter 1392/9765 | SPS:  1114 | Reward:  0.056 | Value:  0.666 | VLoss: 0.0277 | PLoss: -0.0051 | Ent: 0.1632
+[ 14.3%] Iter 1393/9765 | SPS:  1114 | Reward:  0.060 | Value:  0.629 | VLoss: 0.0318 | PLoss: -0.0084 | Ent: 0.1196
+[ 14.3%] Iter 1394/9765 | SPS:  1115 | Reward:  0.061 | Value:  0.697 | VLoss: 0.0315 | PLoss: -0.0083 | Ent: 0.1136
+[ 14.3%] Iter 1395/9765 | SPS:  1115 | Reward:  0.050 | Value:  0.672 | VLoss: 0.0251 | PLoss: -0.0145 | Ent: 0.1890
+[ 14.3%] Iter 1396/9765 | SPS:  1116 | Reward:  0.058 | Value:  0.632 | VLoss: 0.0312 | PLoss: -0.0024 | Ent: 0.2026
+[ 14.3%] Iter 1397/9765 | SPS:  1117 | Reward:  0.076 | Value:  0.658 | VLoss: 0.0266 | PLoss: -0.0049 | Ent: 0.1761
+[ 14.3%] Iter 1398/9765 | SPS:  1117 | Reward:  0.060 | Value:  0.684 | VLoss: 0.0330 | PLoss: -0.0073 | Ent: 0.1704
+[ 14.3%] Iter 1399/9765 | SPS:  1118 | Reward:  0.071 | Value:  0.651 | VLoss: 0.0392 | PLoss: 0.0042 | Ent: 0.1572
+[ 14.3%] Iter 1400/9765 | SPS:  1118 | Reward:  0.066 | Value:  0.615 | VLoss: 0.0318 | PLoss: -0.0077 | Ent: 0.1916
+[ 14.3%] Iter 1401/9765 | SPS:  1119 | Reward:  0.066 | Value:  0.738 | VLoss: 0.0281 | PLoss: -0.0088 | Ent: 0.2062
+[ 14.4%] Iter 1402/9765 | SPS:  1119 | Reward:  0.073 | Value:  0.698 | VLoss: 0.0379 | PLoss: -0.0020 | Ent: 0.1454
+[ 14.4%] Iter 1403/9765 | SPS:  1120 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0247 | PLoss: -0.0099 | Ent: 0.2075
+[ 14.4%] Iter 1404/9765 | SPS:  1120 | Reward:  0.061 | Value:  0.670 | VLoss: 0.0277 | PLoss: -0.0066 | Ent: 0.1707
+[ 14.4%] Iter 1405/9765 | SPS:  1121 | Reward:  0.048 | Value:  0.713 | VLoss: 0.0267 | PLoss: -0.0168 | Ent: 0.2057
+[ 14.4%] Iter 1406/9765 | SPS:  1122 | Reward:  0.061 | Value:  0.728 | VLoss: 0.0220 | PLoss: -0.0051 | Ent: 0.1774
+[ 14.4%] Iter 1407/9765 | SPS:  1122 | Reward:  0.064 | Value:  0.682 | VLoss: 0.0291 | PLoss: -0.0014 | Ent: 0.1655
+[ 14.4%] Iter 1408/9765 | SPS:  1123 | Reward:  0.054 | Value:  0.670 | VLoss: 0.0290 | PLoss: -0.0038 | Ent: 0.1586
+[ 14.4%] Iter 1409/9765 | SPS:  1123 | Reward:  0.063 | Value:  0.665 | VLoss: 0.0325 | PLoss: -0.0089 | Ent: 0.1263
+[ 14.4%] Iter 1410/9765 | SPS:  1124 | Reward:  0.058 | Value:  0.667 | VLoss: 0.0362 | PLoss: -0.0049 | Ent: 0.1305
+[ 14.4%] Iter 1411/9765 | SPS:  1124 | Reward:  0.071 | Value:  0.645 | VLoss: 0.0293 | PLoss: -0.0158 | Ent: 0.1778
+[ 14.5%] Iter 1412/9765 | SPS:  1125 | Reward:  0.056 | Value:  0.669 | VLoss: 0.0328 | PLoss: -0.0131 | Ent: 0.2159
+[ 14.5%] Iter 1413/9765 | SPS:  1125 | Reward:  0.059 | Value:  0.676 | VLoss: 0.0377 | PLoss: -0.0107 | Ent: 0.1716
+[ 14.5%] Iter 1414/9765 | SPS:  1126 | Reward:  0.046 | Value:  0.709 | VLoss: 0.0171 | PLoss: -0.0068 | Ent: 0.1328
+[ 14.5%] Iter 1415/9765 | SPS:  1126 | Reward:  0.061 | Value:  0.690 | VLoss: 0.0255 | PLoss: -0.0088 | Ent: 0.1234
+[ 14.5%] Iter 1416/9765 | SPS:  1127 | Reward:  0.061 | Value:  0.658 | VLoss: 0.0199 | PLoss: -0.0051 | Ent: 0.1463
+[ 14.5%] Iter 1417/9765 | SPS:  1128 | Reward:  0.058 | Value:  0.627 | VLoss: 0.0254 | PLoss: -0.0080 | Ent: 0.1437
+[ 14.5%] Iter 1418/9765 | SPS:  1128 | Reward:  0.045 | Value:  0.595 | VLoss: 0.0246 | PLoss: -0.0075 | Ent: 0.1460
+[ 14.5%] Iter 1419/9765 | SPS:  1129 | Reward:  0.042 | Value:  0.622 | VLoss: 0.0235 | PLoss: -0.0106 | Ent: 0.1474
+[ 14.5%] Iter 1420/9765 | SPS:  1129 | Reward:  0.034 | Value:  0.529 | VLoss: 0.0193 | PLoss: 0.0019 | Ent: 0.1148
+[ 14.6%] Iter 1421/9765 | SPS:  1130 | Reward:  0.031 | Value:  0.535 | VLoss: 0.0227 | PLoss: -0.0040 | Ent: 0.1395
+[ 14.6%] Iter 1422/9765 | SPS:  1130 | Reward:  0.038 | Value:  0.545 | VLoss: 0.0165 | PLoss: -0.0078 | Ent: 0.0930
+[ 14.6%] Iter 1423/9765 | SPS:  1131 | Reward:  0.048 | Value:  0.605 | VLoss: 0.0178 | PLoss: -0.0070 | Ent: 0.1054
+[ 14.6%] Iter 1424/9765 | SPS:  1132 | Reward:  0.034 | Value:  0.602 | VLoss: 0.0130 | PLoss: -0.0078 | Ent: 0.1076
+[ 14.6%] Iter 1425/9765 | SPS:  1132 | Reward:  0.046 | Value:  0.624 | VLoss: 0.0212 | PLoss: -0.0177 | Ent: 0.1122
+[ 14.6%] Iter 1426/9765 | SPS:  1133 | Reward:  0.040 | Value:  0.598 | VLoss: 0.0239 | PLoss: -0.0060 | Ent: 0.1315
+[ 14.6%] Iter 1427/9765 | SPS:  1133 | Reward:  0.048 | Value:  0.532 | VLoss: 0.0260 | PLoss: -0.0023 | Ent: 0.1461
+[ 14.6%] Iter 1428/9765 | SPS:  1134 | Reward:  0.037 | Value:  0.541 | VLoss: 0.0174 | PLoss: -0.0086 | Ent: 0.1573
+[ 14.6%] Iter 1429/9765 | SPS:  1134 | Reward:  0.061 | Value:  0.607 | VLoss: 0.0318 | PLoss: -0.0036 | Ent: 0.1364
+[ 14.6%] Iter 1430/9765 | SPS:  1135 | Reward:  0.049 | Value:  0.565 | VLoss: 0.0212 | PLoss: -0.0035 | Ent: 0.1385
+[ 14.7%] Iter 1431/9765 | SPS:  1135 | Reward:  0.066 | Value:  0.599 | VLoss: 0.0299 | PLoss: -0.0050 | Ent: 0.1711
+[ 14.7%] Iter 1432/9765 | SPS:  1136 | Reward:  0.056 | Value:  0.587 | VLoss: 0.0332 | PLoss: -0.0123 | Ent: 0.1379
+[ 14.7%] Iter 1433/9765 | SPS:  1137 | Reward:  0.048 | Value:  0.589 | VLoss: 0.0225 | PLoss: -0.0017 | Ent: 0.1187
+[ 14.7%] Iter 1434/9765 | SPS:  1137 | Reward:  0.067 | Value:  0.581 | VLoss: 0.0467 | PLoss: -0.0059 | Ent: 0.1634
+[ 14.7%] Iter 1435/9765 | SPS:  1138 | Reward:  0.064 | Value:  0.666 | VLoss: 0.0339 | PLoss: -0.0087 | Ent: 0.1761
+[ 14.7%] Iter 1436/9765 | SPS:  1138 | Reward:  0.063 | Value:  0.709 | VLoss: 0.0264 | PLoss: -0.0103 | Ent: 0.1925
+[ 14.7%] Iter 1437/9765 | SPS:  1139 | Reward:  0.071 | Value:  0.661 | VLoss: 0.0352 | PLoss: -0.0071 | Ent: 0.1809
+[ 14.7%] Iter 1438/9765 | SPS:  1139 | Reward:  0.062 | Value:  0.651 | VLoss: 0.0338 | PLoss: -0.0058 | Ent: 0.1885
+[ 14.7%] Iter 1439/9765 | SPS:  1140 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0200 | PLoss: -0.0128 | Ent: 0.1772
+[ 14.7%] Iter 1440/9765 | SPS:  1140 | Reward:  0.070 | Value:  0.673 | VLoss: 0.0322 | PLoss: -0.0156 | Ent: 0.1863
+[ 14.8%] Iter 1441/9765 | SPS:  1141 | Reward:  0.070 | Value:  0.699 | VLoss: 0.0292 | PLoss: -0.0036 | Ent: 0.1893
+[ 14.8%] Iter 1442/9765 | SPS:  1141 | Reward:  0.058 | Value:  0.667 | VLoss: 0.0330 | PLoss: -0.0072 | Ent: 0.1718
+[ 14.8%] Iter 1443/9765 | SPS:  1142 | Reward:  0.063 | Value:  0.648 | VLoss: 0.0322 | PLoss: -0.0116 | Ent: 0.1435
+[ 14.8%] Iter 1444/9765 | SPS:  1143 | Reward:  0.061 | Value:  0.645 | VLoss: 0.0315 | PLoss: -0.0080 | Ent: 0.1545
+[ 14.8%] Iter 1445/9765 | SPS:  1143 | Reward:  0.064 | Value:  0.646 | VLoss: 0.0387 | PLoss: -0.0174 | Ent: 0.2094
+[ 14.8%] Iter 1446/9765 | SPS:  1144 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0313 | PLoss: -0.0111 | Ent: 0.1750
+[ 14.8%] Iter 1447/9765 | SPS:  1144 | Reward:  0.066 | Value:  0.624 | VLoss: 0.0329 | PLoss: -0.0086 | Ent: 0.1434
+[ 14.8%] Iter 1448/9765 | SPS:  1145 | Reward:  0.074 | Value:  0.639 | VLoss: 0.0296 | PLoss: -0.0058 | Ent: 0.2031
+[ 14.8%] Iter 1449/9765 | SPS:  1145 | Reward:  0.081 | Value:  0.679 | VLoss: 0.0324 | PLoss: -0.0112 | Ent: 0.1881
+[ 14.8%] Iter 1450/9765 | SPS:  1146 | Reward:  0.072 | Value:  0.683 | VLoss: 0.0322 | PLoss: -0.0154 | Ent: 0.1826
+[ 14.9%] Iter 1451/9765 | SPS:  1146 | Reward:  0.066 | Value:  0.680 | VLoss: 0.0299 | PLoss: -0.0097 | Ent: 0.1514
+[ 14.9%] Iter 1452/9765 | SPS:  1147 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0397 | PLoss: -0.0052 | Ent: 0.1541
+[ 14.9%] Iter 1453/9765 | SPS:  1147 | Reward:  0.059 | Value:  0.693 | VLoss: 0.0306 | PLoss: -0.0148 | Ent: 0.1619
+[ 14.9%] Iter 1454/9765 | SPS:  1148 | Reward:  0.060 | Value:  0.643 | VLoss: 0.0336 | PLoss: -0.0094 | Ent: 0.1521
+[ 14.9%] Iter 1455/9765 | SPS:  1148 | Reward:  0.071 | Value:  0.669 | VLoss: 0.0292 | PLoss: -0.0084 | Ent: 0.1487
+[ 14.9%] Iter 1456/9765 | SPS:  1149 | Reward:  0.067 | Value:  0.677 | VLoss: 0.0322 | PLoss: -0.0084 | Ent: 0.1960
+[ 14.9%] Iter 1457/9765 | SPS:  1150 | Reward:  0.066 | Value:  0.684 | VLoss: 0.0307 | PLoss: -0.0043 | Ent: 0.1038
+[ 14.9%] Iter 1458/9765 | SPS:  1150 | Reward:  0.072 | Value:  0.718 | VLoss: 0.0313 | PLoss: -0.0047 | Ent: 0.1491
+[ 14.9%] Iter 1459/9765 | SPS:  1151 | Reward:  0.061 | Value:  0.645 | VLoss: 0.0328 | PLoss: -0.0141 | Ent: 0.1652
+[ 15.0%] Iter 1460/9765 | SPS:  1151 | Reward:  0.052 | Value:  0.712 | VLoss: 0.0263 | PLoss: -0.0087 | Ent: 0.1592
+[ 15.0%] Iter 1461/9765 | SPS:  1152 | Reward:  0.054 | Value:  0.707 | VLoss: 0.0255 | PLoss: -0.0029 | Ent: 0.1345
+[ 15.0%] Iter 1462/9765 | SPS:  1152 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0394 | PLoss: -0.0063 | Ent: 0.1517
+[ 15.0%] Iter 1463/9765 | SPS:  1153 | Reward:  0.066 | Value:  0.678 | VLoss: 0.0300 | PLoss: -0.0076 | Ent: 0.1611
+[ 15.0%] Iter 1464/9765 | SPS:  1153 | Reward:  0.059 | Value:  0.606 | VLoss: 0.0398 | PLoss: -0.0020 | Ent: 0.1957
+[ 15.0%] Iter 1465/9765 | SPS:  1154 | Reward:  0.072 | Value:  0.700 | VLoss: 0.0317 | PLoss: -0.0201 | Ent: 0.1742
+[ 15.0%] Iter 1466/9765 | SPS:  1154 | Reward:  0.063 | Value:  0.679 | VLoss: 0.0260 | PLoss: -0.0118 | Ent: 0.1227
+[ 15.0%] Iter 1467/9765 | SPS:  1155 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0311 | PLoss: -0.0102 | Ent: 0.1733
+[ 15.0%] Iter 1468/9765 | SPS:  1155 | Reward:  0.062 | Value:  0.646 | VLoss: 0.0294 | PLoss: -0.0125 | Ent: 0.1850
+[ 15.0%] Iter 1469/9765 | SPS:  1156 | Reward:  0.069 | Value:  0.633 | VLoss: 0.0395 | PLoss: -0.0031 | Ent: 0.1405
+[ 15.1%] Iter 1470/9765 | SPS:  1157 | Reward:  0.077 | Value:  0.646 | VLoss: 0.0322 | PLoss: -0.0122 | Ent: 0.1486
+[ 15.1%] Iter 1471/9765 | SPS:  1157 | Reward:  0.071 | Value:  0.716 | VLoss: 0.0375 | PLoss: -0.0063 | Ent: 0.1837
+[ 15.1%] Iter 1472/9765 | SPS:  1158 | Reward:  0.072 | Value:  0.703 | VLoss: 0.0363 | PLoss: -0.0078 | Ent: 0.1876
+[ 15.1%] Iter 1473/9765 | SPS:  1158 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0377 | PLoss: -0.0102 | Ent: 0.1688
+[ 15.1%] Iter 1474/9765 | SPS:  1159 | Reward:  0.070 | Value:  0.612 | VLoss: 0.0395 | PLoss: -0.0062 | Ent: 0.1570
+[ 15.1%] Iter 1475/9765 | SPS:  1159 | Reward:  0.066 | Value:  0.642 | VLoss: 0.0343 | PLoss: -0.0106 | Ent: 0.2038
+[ 15.1%] Iter 1476/9765 | SPS:  1160 | Reward:  0.072 | Value:  0.705 | VLoss: 0.0295 | PLoss: -0.0040 | Ent: 0.1649
+[ 15.1%] Iter 1477/9765 | SPS:  1160 | Reward:  0.071 | Value:  0.692 | VLoss: 0.0349 | PLoss: -0.0088 | Ent: 0.1772
+[ 15.1%] Iter 1478/9765 | SPS:  1161 | Reward:  0.068 | Value:  0.698 | VLoss: 0.0304 | PLoss: -0.0060 | Ent: 0.1416
+[ 15.1%] Iter 1479/9765 | SPS:  1161 | Reward:  0.068 | Value:  0.685 | VLoss: 0.0285 | PLoss: -0.0091 | Ent: 0.1544
+[ 15.2%] Iter 1480/9765 | SPS:  1162 | Reward:  0.076 | Value:  0.680 | VLoss: 0.0289 | PLoss: -0.0114 | Ent: 0.1801
+[ 15.2%] Iter 1481/9765 | SPS:  1162 | Reward:  0.066 | Value:  0.689 | VLoss: 0.0237 | PLoss: -0.0001 | Ent: 0.1829
+[ 15.2%] Iter 1482/9765 | SPS:  1163 | Reward:  0.061 | Value:  0.703 | VLoss: 0.0311 | PLoss: -0.0084 | Ent: 0.1847
+[ 15.2%] Iter 1483/9765 | SPS:  1163 | Reward:  0.051 | Value:  0.614 | VLoss: 0.0267 | PLoss: -0.0111 | Ent: 0.1703
+[ 15.2%] Iter 1484/9765 | SPS:  1164 | Reward:  0.064 | Value:  0.663 | VLoss: 0.0321 | PLoss: -0.0109 | Ent: 0.1508
+[ 15.2%] Iter 1485/9765 | SPS:  1164 | Reward:  0.069 | Value:  0.664 | VLoss: 0.0382 | PLoss: -0.0118 | Ent: 0.1784
+[ 15.2%] Iter 1486/9765 | SPS:  1165 | Reward:  0.057 | Value:  0.634 | VLoss: 0.0331 | PLoss: -0.0133 | Ent: 0.1706
+[ 15.2%] Iter 1487/9765 | SPS:  1166 | Reward:  0.062 | Value:  0.640 | VLoss: 0.0367 | PLoss: -0.0006 | Ent: 0.2149
+[ 15.2%] Iter 1488/9765 | SPS:  1166 | Reward:  0.072 | Value:  0.659 | VLoss: 0.0321 | PLoss: -0.0105 | Ent: 0.1670
+[ 15.2%] Iter 1489/9765 | SPS:  1167 | Reward:  0.066 | Value:  0.631 | VLoss: 0.0342 | PLoss: -0.0078 | Ent: 0.2137
+[ 15.3%] Iter 1490/9765 | SPS:  1167 | Reward:  0.067 | Value:  0.621 | VLoss: 0.0262 | PLoss: -0.0081 | Ent: 0.1703
+[ 15.3%] Iter 1491/9765 | SPS:  1168 | Reward:  0.060 | Value:  0.661 | VLoss: 0.0279 | PLoss: -0.0095 | Ent: 0.1684
+[ 15.3%] Iter 1492/9765 | SPS:  1168 | Reward:  0.055 | Value:  0.655 | VLoss: 0.0260 | PLoss: -0.0107 | Ent: 0.1332
+[ 15.3%] Iter 1493/9765 | SPS:  1169 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0296 | PLoss: -0.0137 | Ent: 0.1818
+[ 15.3%] Iter 1494/9765 | SPS:  1169 | Reward:  0.053 | Value:  0.673 | VLoss: 0.0248 | PLoss: -0.0103 | Ent: 0.1669
+[ 15.3%] Iter 1495/9765 | SPS:  1170 | Reward:  0.059 | Value:  0.629 | VLoss: 0.0294 | PLoss: -0.0071 | Ent: 0.1610
+[ 15.3%] Iter 1496/9765 | SPS:  1170 | Reward:  0.053 | Value:  0.633 | VLoss: 0.0331 | PLoss: -0.0136 | Ent: 0.1997
+[ 15.3%] Iter 1497/9765 | SPS:  1171 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0208 | PLoss: -0.0087 | Ent: 0.1933
+[ 15.3%] Iter 1498/9765 | SPS:  1171 | Reward:  0.063 | Value:  0.603 | VLoss: 0.0298 | PLoss: -0.0033 | Ent: 0.2440
+[ 15.4%] Iter 1499/9765 | SPS:  1172 | Reward:  0.055 | Value:  0.610 | VLoss: 0.0261 | PLoss: -0.0106 | Ent: 0.2276
+[ 15.4%] Iter 1500/9765 | SPS:  1172 | Reward:  0.048 | Value:  0.675 | VLoss: 0.0245 | PLoss: -0.0101 | Ent: 0.2105
+[ 15.4%] Iter 1501/9765 | SPS:  1173 | Reward:  0.052 | Value:  0.542 | VLoss: 0.0324 | PLoss: -0.0087 | Ent: 0.1937
+[ 15.4%] Iter 1502/9765 | SPS:  1174 | Reward:  0.060 | Value:  0.579 | VLoss: 0.0242 | PLoss: -0.0094 | Ent: 0.1791
+[ 15.4%] Iter 1503/9765 | SPS:  1174 | Reward:  0.043 | Value:  0.648 | VLoss: 0.0244 | PLoss: -0.0055 | Ent: 0.1194
+[ 15.4%] Iter 1504/9765 | SPS:  1175 | Reward:  0.062 | Value:  0.575 | VLoss: 0.0289 | PLoss: -0.0072 | Ent: 0.1759
+[ 15.4%] Iter 1505/9765 | SPS:  1175 | Reward:  0.058 | Value:  0.581 | VLoss: 0.0309 | PLoss: -0.0085 | Ent: 0.2073
+[ 15.4%] Iter 1506/9765 | SPS:  1176 | Reward:  0.039 | Value:  0.610 | VLoss: 0.0280 | PLoss: -0.0124 | Ent: 0.1849
+[ 15.4%] Iter 1507/9765 | SPS:  1176 | Reward:  0.042 | Value:  0.625 | VLoss: 0.0202 | PLoss: -0.0094 | Ent: 0.1734
+[ 15.4%] Iter 1508/9765 | SPS:  1177 | Reward:  0.045 | Value:  0.631 | VLoss: 0.0190 | PLoss: -0.0082 | Ent: 0.1945
+[ 15.5%] Iter 1509/9765 | SPS:  1177 | Reward:  0.051 | Value:  0.649 | VLoss: 0.0287 | PLoss: -0.0087 | Ent: 0.1423
+[ 15.5%] Iter 1510/9765 | SPS:  1178 | Reward:  0.057 | Value:  0.638 | VLoss: 0.0248 | PLoss: -0.0020 | Ent: 0.1701
+[ 15.5%] Iter 1511/9765 | SPS:  1178 | Reward:  0.062 | Value:  0.660 | VLoss: 0.0268 | PLoss: -0.0084 | Ent: 0.1625
+[ 15.5%] Iter 1512/9765 | SPS:  1179 | Reward:  0.067 | Value:  0.703 | VLoss: 0.0252 | PLoss: -0.0109 | Ent: 0.2256
+[ 15.5%] Iter 1513/9765 | SPS:  1179 | Reward:  0.064 | Value:  0.662 | VLoss: 0.0295 | PLoss: -0.0137 | Ent: 0.1653
+[ 15.5%] Iter 1514/9765 | SPS:  1180 | Reward:  0.065 | Value:  0.727 | VLoss: 0.0174 | PLoss: -0.0065 | Ent: 0.1427
+[ 15.5%] Iter 1515/9765 | SPS:  1180 | Reward:  0.074 | Value:  0.738 | VLoss: 0.0279 | PLoss: -0.0031 | Ent: 0.1807
+[ 15.5%] Iter 1516/9765 | SPS:  1181 | Reward:  0.067 | Value:  0.747 | VLoss: 0.0290 | PLoss: -0.0053 | Ent: 0.1802
+[ 15.5%] Iter 1517/9765 | SPS:  1182 | Reward:  0.071 | Value:  0.706 | VLoss: 0.0306 | PLoss: -0.0115 | Ent: 0.1742
+[ 15.5%] Iter 1518/9765 | SPS:  1182 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0373 | PLoss: -0.0111 | Ent: 0.1610
+[ 15.6%] Iter 1519/9765 | SPS:  1183 | Reward:  0.061 | Value:  0.679 | VLoss: 0.0333 | PLoss: -0.0096 | Ent: 0.1418
+[ 15.6%] Iter 1520/9765 | SPS:  1183 | Reward:  0.069 | Value:  0.662 | VLoss: 0.0275 | PLoss: -0.0069 | Ent: 0.1416
+[ 15.6%] Iter 1521/9765 | SPS:  1184 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0317 | PLoss: -0.0099 | Ent: 0.1576
+[ 15.6%] Iter 1522/9765 | SPS:  1184 | Reward:  0.068 | Value:  0.669 | VLoss: 0.0292 | PLoss: -0.0074 | Ent: 0.1209
+[ 15.6%] Iter 1523/9765 | SPS:  1185 | Reward:  0.060 | Value:  0.652 | VLoss: 0.0321 | PLoss: -0.0132 | Ent: 0.1751
+[ 15.6%] Iter 1524/9765 | SPS:  1185 | Reward:  0.072 | Value:  0.622 | VLoss: 0.0398 | PLoss: -0.0074 | Ent: 0.1837
+[ 15.6%] Iter 1525/9765 | SPS:  1186 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0241 | PLoss: -0.0076 | Ent: 0.1187
+[ 15.6%] Iter 1526/9765 | SPS:  1186 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0307 | PLoss: -0.0077 | Ent: 0.1212
+[ 15.6%] Iter 1527/9765 | SPS:  1187 | Reward:  0.040 | Value:  0.649 | VLoss: 0.0238 | PLoss: -0.0137 | Ent: 0.1439
+[ 15.6%] Iter 1528/9765 | SPS:  1187 | Reward:  0.052 | Value:  0.618 | VLoss: 0.0291 | PLoss: -0.0039 | Ent: 0.1418
+[ 15.7%] Iter 1529/9765 | SPS:  1188 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0250 | PLoss: -0.0108 | Ent: 0.1820
+[ 15.7%] Iter 1530/9765 | SPS:  1188 | Reward:  0.062 | Value:  0.622 | VLoss: 0.0349 | PLoss: -0.0095 | Ent: 0.1706
+[ 15.7%] Iter 1531/9765 | SPS:  1189 | Reward:  0.059 | Value:  0.643 | VLoss: 0.0281 | PLoss: -0.0129 | Ent: 0.1664
+[ 15.7%] Iter 1532/9765 | SPS:  1189 | Reward:  0.067 | Value:  0.641 | VLoss: 0.0260 | PLoss: -0.0101 | Ent: 0.1480
+[ 15.7%] Iter 1533/9765 | SPS:  1190 | Reward:  0.065 | Value:  0.737 | VLoss: 0.0275 | PLoss: -0.0129 | Ent: 0.1770
+[ 15.7%] Iter 1534/9765 | SPS:  1190 | Reward:  0.074 | Value:  0.731 | VLoss: 0.0340 | PLoss: -0.0097 | Ent: 0.1936
+[ 15.7%] Iter 1535/9765 | SPS:  1191 | Reward:  0.063 | Value:  0.620 | VLoss: 0.0356 | PLoss: -0.0065 | Ent: 0.1471
+[ 15.7%] Iter 1536/9765 | SPS:  1191 | Reward:  0.069 | Value:  0.686 | VLoss: 0.0310 | PLoss: -0.0038 | Ent: 0.1700
+[ 15.7%] Iter 1537/9765 | SPS:  1192 | Reward:  0.075 | Value:  0.613 | VLoss: 0.0445 | PLoss: -0.0087 | Ent: 0.1761
+[ 15.8%] Iter 1538/9765 | SPS:  1193 | Reward:  0.052 | Value:  0.662 | VLoss: 0.0320 | PLoss: -0.0057 | Ent: 0.1528
+[ 15.8%] Iter 1539/9765 | SPS:  1193 | Reward:  0.058 | Value:  0.697 | VLoss: 0.0231 | PLoss: -0.0114 | Ent: 0.1831
+[ 15.8%] Iter 1540/9765 | SPS:  1194 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0333 | PLoss: -0.0012 | Ent: 0.1325
+[ 15.8%] Iter 1541/9765 | SPS:  1194 | Reward:  0.072 | Value:  0.660 | VLoss: 0.0367 | PLoss: -0.0046 | Ent: 0.1879
+[ 15.8%] Iter 1542/9765 | SPS:  1195 | Reward:  0.071 | Value:  0.713 | VLoss: 0.0337 | PLoss: -0.0051 | Ent: 0.1604
+[ 15.8%] Iter 1543/9765 | SPS:  1195 | Reward:  0.079 | Value:  0.684 | VLoss: 0.0343 | PLoss: -0.0109 | Ent: 0.1630
+[ 15.8%] Iter 1544/9765 | SPS:  1196 | Reward:  0.054 | Value:  0.694 | VLoss: 0.0189 | PLoss: -0.0069 | Ent: 0.1852
+[ 15.8%] Iter 1545/9765 | SPS:  1196 | Reward:  0.069 | Value:  0.673 | VLoss: 0.0414 | PLoss: -0.0055 | Ent: 0.1652
+[ 15.8%] Iter 1546/9765 | SPS:  1197 | Reward:  0.074 | Value:  0.669 | VLoss: 0.0308 | PLoss: -0.0114 | Ent: 0.1753
+[ 15.8%] Iter 1547/9765 | SPS:  1197 | Reward:  0.072 | Value:  0.635 | VLoss: 0.0357 | PLoss: -0.0028 | Ent: 0.1666
+[ 15.9%] Iter 1548/9765 | SPS:  1198 | Reward:  0.047 | Value:  0.617 | VLoss: 0.0345 | PLoss: -0.0097 | Ent: 0.1592
+[ 15.9%] Iter 1549/9765 | SPS:  1198 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0410 | PLoss: -0.0062 | Ent: 0.1815
+[ 15.9%] Iter 1550/9765 | SPS:  1199 | Reward:  0.067 | Value:  0.646 | VLoss: 0.0271 | PLoss: -0.0019 | Ent: 0.1444
+[ 15.9%] Iter 1551/9765 | SPS:  1199 | Reward:  0.058 | Value:  0.631 | VLoss: 0.0286 | PLoss: -0.0065 | Ent: 0.1337
+[ 15.9%] Iter 1552/9765 | SPS:  1200 | Reward:  0.047 | Value:  0.683 | VLoss: 0.0270 | PLoss: -0.0021 | Ent: 0.1510
+[ 15.9%] Iter 1553/9765 | SPS:  1200 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0401 | PLoss: -0.0080 | Ent: 0.1604
+[ 15.9%] Iter 1554/9765 | SPS:  1201 | Reward:  0.052 | Value:  0.664 | VLoss: 0.0277 | PLoss: -0.0036 | Ent: 0.1919
+[ 15.9%] Iter 1555/9765 | SPS:  1201 | Reward:  0.059 | Value:  0.685 | VLoss: 0.0302 | PLoss: -0.0027 | Ent: 0.1399
+[ 15.9%] Iter 1556/9765 | SPS:  1202 | Reward:  0.069 | Value:  0.707 | VLoss: 0.0315 | PLoss: -0.0141 | Ent: 0.1565
+[ 15.9%] Iter 1557/9765 | SPS:  1202 | Reward:  0.064 | Value:  0.646 | VLoss: 0.0286 | PLoss: -0.0047 | Ent: 0.1472
+[ 16.0%] Iter 1558/9765 | SPS:  1203 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0353 | PLoss: -0.0094 | Ent: 0.1496
+[ 16.0%] Iter 1559/9765 | SPS:  1203 | Reward:  0.065 | Value:  0.621 | VLoss: 0.0335 | PLoss: -0.0054 | Ent: 0.1548
+[ 16.0%] Iter 1560/9765 | SPS:  1204 | Reward:  0.063 | Value:  0.678 | VLoss: 0.0283 | PLoss: -0.0063 | Ent: 0.1382
+[ 16.0%] Iter 1561/9765 | SPS:  1204 | Reward:  0.053 | Value:  0.702 | VLoss: 0.0342 | PLoss: -0.0083 | Ent: 0.1482
+[ 16.0%] Iter 1562/9765 | SPS:  1205 | Reward:  0.054 | Value:  0.633 | VLoss: 0.0318 | PLoss: -0.0062 | Ent: 0.1954
+[ 16.0%] Iter 1563/9765 | SPS:  1205 | Reward:  0.069 | Value:  0.640 | VLoss: 0.0314 | PLoss: -0.0086 | Ent: 0.1305
+[ 16.0%] Iter 1564/9765 | SPS:  1206 | Reward:  0.078 | Value:  0.670 | VLoss: 0.0379 | PLoss: -0.0042 | Ent: 0.1409
+[ 16.0%] Iter 1565/9765 | SPS:  1206 | Reward:  0.073 | Value:  0.636 | VLoss: 0.0329 | PLoss: -0.0014 | Ent: 0.1700
+[ 16.0%] Iter 1566/9765 | SPS:  1207 | Reward:  0.065 | Value:  0.649 | VLoss: 0.0364 | PLoss: -0.0040 | Ent: 0.1595
+[ 16.0%] Iter 1567/9765 | SPS:  1208 | Reward:  0.055 | Value:  0.644 | VLoss: 0.0195 | PLoss: -0.0072 | Ent: 0.1340
+[ 16.1%] Iter 1568/9765 | SPS:  1208 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0290 | PLoss: -0.0119 | Ent: 0.1369
+[ 16.1%] Iter 1569/9765 | SPS:  1209 | Reward:  0.048 | Value:  0.666 | VLoss: 0.0264 | PLoss: -0.0027 | Ent: 0.1216
+[ 16.1%] Iter 1570/9765 | SPS:  1209 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0259 | PLoss: -0.0070 | Ent: 0.1296
+[ 16.1%] Iter 1571/9765 | SPS:  1210 | Reward:  0.044 | Value:  0.660 | VLoss: 0.0242 | PLoss: -0.0041 | Ent: 0.1343
+[ 16.1%] Iter 1572/9765 | SPS:  1210 | Reward:  0.071 | Value:  0.647 | VLoss: 0.0362 | PLoss: -0.0084 | Ent: 0.1707
+[ 16.1%] Iter 1573/9765 | SPS:  1211 | Reward:  0.057 | Value:  0.604 | VLoss: 0.0243 | PLoss: -0.0056 | Ent: 0.1299
+[ 16.1%] Iter 1574/9765 | SPS:  1211 | Reward:  0.067 | Value:  0.596 | VLoss: 0.0335 | PLoss: -0.0032 | Ent: 0.1381
+[ 16.1%] Iter 1575/9765 | SPS:  1212 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0372 | PLoss: -0.0077 | Ent: 0.2083
+[ 16.1%] Iter 1576/9765 | SPS:  1212 | Reward:  0.063 | Value:  0.680 | VLoss: 0.0372 | PLoss: -0.0106 | Ent: 0.1446
+[ 16.1%] Iter 1577/9765 | SPS:  1213 | Reward:  0.067 | Value:  0.667 | VLoss: 0.0365 | PLoss: -0.0102 | Ent: 0.1925
+[ 16.2%] Iter 1578/9765 | SPS:  1213 | Reward:  0.070 | Value:  0.627 | VLoss: 0.0365 | PLoss: -0.0110 | Ent: 0.1613
+[ 16.2%] Iter 1579/9765 | SPS:  1214 | Reward:  0.053 | Value:  0.609 | VLoss: 0.0274 | PLoss: -0.0114 | Ent: 0.1490
+[ 16.2%] Iter 1580/9765 | SPS:  1214 | Reward:  0.069 | Value:  0.695 | VLoss: 0.0293 | PLoss: -0.0096 | Ent: 0.1274
+[ 16.2%] Iter 1581/9765 | SPS:  1215 | Reward:  0.081 | Value:  0.655 | VLoss: 0.0373 | PLoss: -0.0045 | Ent: 0.1371
+[ 16.2%] Iter 1582/9765 | SPS:  1215 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0305 | PLoss: -0.0017 | Ent: 0.1780
+[ 16.2%] Iter 1583/9765 | SPS:  1216 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0309 | PLoss: -0.0073 | Ent: 0.1784
+[ 16.2%] Iter 1584/9765 | SPS:  1216 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0305 | PLoss: -0.0082 | Ent: 0.1455
+[ 16.2%] Iter 1585/9765 | SPS:  1217 | Reward:  0.075 | Value:  0.741 | VLoss: 0.0266 | PLoss: -0.0064 | Ent: 0.1318
+[ 16.2%] Iter 1586/9765 | SPS:  1217 | Reward:  0.057 | Value:  0.705 | VLoss: 0.0265 | PLoss: -0.0107 | Ent: 0.1877
+[ 16.3%] Iter 1587/9765 | SPS:  1218 | Reward:  0.065 | Value:  0.654 | VLoss: 0.0316 | PLoss: -0.0046 | Ent: 0.1765
+[ 16.3%] Iter 1588/9765 | SPS:  1218 | Reward:  0.068 | Value:  0.653 | VLoss: 0.0302 | PLoss: -0.0094 | Ent: 0.1861
+[ 16.3%] Iter 1589/9765 | SPS:  1219 | Reward:  0.075 | Value:  0.692 | VLoss: 0.0281 | PLoss: -0.0082 | Ent: 0.1221
+[ 16.3%] Iter 1590/9765 | SPS:  1219 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0294 | PLoss: -0.0160 | Ent: 0.1974
+[ 16.3%] Iter 1591/9765 | SPS:  1220 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0320 | PLoss: -0.0044 | Ent: 0.1519
+[ 16.3%] Iter 1592/9765 | SPS:  1220 | Reward:  0.071 | Value:  0.660 | VLoss: 0.0400 | PLoss: -0.0090 | Ent: 0.1995
+[ 16.3%] Iter 1593/9765 | SPS:  1221 | Reward:  0.066 | Value:  0.600 | VLoss: 0.0321 | PLoss: -0.0123 | Ent: 0.1577
+[ 16.3%] Iter 1594/9765 | SPS:  1221 | Reward:  0.062 | Value:  0.658 | VLoss: 0.0273 | PLoss: -0.0038 | Ent: 0.1680
+[ 16.3%] Iter 1595/9765 | SPS:  1222 | Reward:  0.074 | Value:  0.612 | VLoss: 0.0353 | PLoss: -0.0054 | Ent: 0.1384
+[ 16.3%] Iter 1596/9765 | SPS:  1222 | Reward:  0.057 | Value:  0.675 | VLoss: 0.0286 | PLoss: -0.0009 | Ent: 0.1656
+[ 16.4%] Iter 1597/9765 | SPS:  1223 | Reward:  0.064 | Value:  0.640 | VLoss: 0.0383 | PLoss: -0.0144 | Ent: 0.1858
+[ 16.4%] Iter 1598/9765 | SPS:  1223 | Reward:  0.058 | Value:  0.671 | VLoss: 0.0340 | PLoss: -0.0108 | Ent: 0.1548
+[ 16.4%] Iter 1599/9765 | SPS:  1224 | Reward:  0.053 | Value:  0.658 | VLoss: 0.0300 | PLoss: -0.0099 | Ent: 0.1566
+[ 16.4%] Iter 1600/9765 | SPS:  1224 | Reward:  0.051 | Value:  0.624 | VLoss: 0.0228 | PLoss: -0.0100 | Ent: 0.1324
+[ 16.4%] Iter 1601/9765 | SPS:  1225 | Reward:  0.064 | Value:  0.649 | VLoss: 0.0327 | PLoss: -0.0104 | Ent: 0.1433
+[ 16.4%] Iter 1602/9765 | SPS:  1225 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0304 | PLoss: -0.0082 | Ent: 0.1215
+[ 16.4%] Iter 1603/9765 | SPS:  1226 | Reward:  0.046 | Value:  0.596 | VLoss: 0.0326 | PLoss: -0.0085 | Ent: 0.1643
+[ 16.4%] Iter 1604/9765 | SPS:  1226 | Reward:  0.070 | Value:  0.584 | VLoss: 0.0354 | PLoss: -0.0093 | Ent: 0.2075
+[ 16.4%] Iter 1605/9765 | SPS:  1227 | Reward:  0.064 | Value:  0.623 | VLoss: 0.0314 | PLoss: -0.0096 | Ent: 0.1461
+[ 16.4%] Iter 1606/9765 | SPS:  1227 | Reward:  0.082 | Value:  0.687 | VLoss: 0.0367 | PLoss: -0.0062 | Ent: 0.1163
+[ 16.5%] Iter 1607/9765 | SPS:  1228 | Reward:  0.080 | Value:  0.648 | VLoss: 0.0305 | PLoss: -0.0085 | Ent: 0.1630
+[ 16.5%] Iter 1608/9765 | SPS:  1228 | Reward:  0.050 | Value:  0.651 | VLoss: 0.0275 | PLoss: -0.0074 | Ent: 0.1324
+[ 16.5%] Iter 1609/9765 | SPS:  1229 | Reward:  0.048 | Value:  0.701 | VLoss: 0.0291 | PLoss: -0.0087 | Ent: 0.1759
+[ 16.5%] Iter 1610/9765 | SPS:  1229 | Reward:  0.055 | Value:  0.674 | VLoss: 0.0262 | PLoss: -0.0065 | Ent: 0.1926
+[ 16.5%] Iter 1611/9765 | SPS:  1230 | Reward:  0.060 | Value:  0.665 | VLoss: 0.0314 | PLoss: -0.0070 | Ent: 0.1351
+[ 16.5%] Iter 1612/9765 | SPS:  1230 | Reward:  0.061 | Value:  0.661 | VLoss: 0.0225 | PLoss: -0.0100 | Ent: 0.1244
+[ 16.5%] Iter 1613/9765 | SPS:  1231 | Reward:  0.061 | Value:  0.627 | VLoss: 0.0380 | PLoss: -0.0142 | Ent: 0.1691
+[ 16.5%] Iter 1614/9765 | SPS:  1231 | Reward:  0.052 | Value:  0.579 | VLoss: 0.0339 | PLoss: -0.0066 | Ent: 0.1535
+[ 16.5%] Iter 1615/9765 | SPS:  1232 | Reward:  0.045 | Value:  0.506 | VLoss: 0.0223 | PLoss: -0.0025 | Ent: 0.1611
+[ 16.5%] Iter 1616/9765 | SPS:  1232 | Reward:  0.062 | Value:  0.604 | VLoss: 0.0284 | PLoss: -0.0058 | Ent: 0.1247
+[ 16.6%] Iter 1617/9765 | SPS:  1233 | Reward:  0.058 | Value:  0.635 | VLoss: 0.0303 | PLoss: -0.0039 | Ent: 0.1481
+[ 16.6%] Iter 1618/9765 | SPS:  1233 | Reward:  0.064 | Value:  0.620 | VLoss: 0.0247 | PLoss: -0.0092 | Ent: 0.1380
+[ 16.6%] Iter 1619/9765 | SPS:  1234 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0305 | PLoss: -0.0103 | Ent: 0.1666
+[ 16.6%] Iter 1620/9765 | SPS:  1234 | Reward:  0.069 | Value:  0.700 | VLoss: 0.0362 | PLoss: -0.0095 | Ent: 0.1511
+[ 16.6%] Iter 1621/9765 | SPS:  1235 | Reward:  0.050 | Value:  0.718 | VLoss: 0.0199 | PLoss: -0.0044 | Ent: 0.1465
+[ 16.6%] Iter 1622/9765 | SPS:  1236 | Reward:  0.059 | Value:  0.718 | VLoss: 0.0213 | PLoss: -0.0112 | Ent: 0.1599
+[ 16.6%] Iter 1623/9765 | SPS:  1236 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0335 | PLoss: -0.0084 | Ent: 0.1487
+[ 16.6%] Iter 1624/9765 | SPS:  1237 | Reward:  0.066 | Value:  0.684 | VLoss: 0.0310 | PLoss: -0.0059 | Ent: 0.1265
+[ 16.6%] Iter 1625/9765 | SPS:  1237 | Reward:  0.056 | Value:  0.641 | VLoss: 0.0278 | PLoss: -0.0141 | Ent: 0.1645
+[ 16.7%] Iter 1626/9765 | SPS:  1238 | Reward:  0.043 | Value:  0.663 | VLoss: 0.0253 | PLoss: -0.0074 | Ent: 0.1830
+[ 16.7%] Iter 1627/9765 | SPS:  1238 | Reward:  0.051 | Value:  0.620 | VLoss: 0.0249 | PLoss: -0.0077 | Ent: 0.1538
+[ 16.7%] Iter 1628/9765 | SPS:  1239 | Reward:  0.064 | Value:  0.602 | VLoss: 0.0338 | PLoss: -0.0131 | Ent: 0.1303
+[ 16.7%] Iter 1629/9765 | SPS:  1239 | Reward:  0.052 | Value:  0.634 | VLoss: 0.0394 | PLoss: -0.0016 | Ent: 0.1620
+[ 16.7%] Iter 1630/9765 | SPS:  1240 | Reward:  0.069 | Value:  0.660 | VLoss: 0.0318 | PLoss: -0.0042 | Ent: 0.1722
+[ 16.7%] Iter 1631/9765 | SPS:  1240 | Reward:  0.081 | Value:  0.677 | VLoss: 0.0330 | PLoss: -0.0062 | Ent: 0.1474
+[ 16.7%] Iter 1632/9765 | SPS:  1241 | Reward:  0.074 | Value:  0.671 | VLoss: 0.0330 | PLoss: -0.0076 | Ent: 0.1861
+[ 16.7%] Iter 1633/9765 | SPS:  1241 | Reward:  0.070 | Value:  0.729 | VLoss: 0.0209 | PLoss: -0.0115 | Ent: 0.1574
+[ 16.7%] Iter 1634/9765 | SPS:  1242 | Reward:  0.063 | Value:  0.730 | VLoss: 0.0246 | PLoss: -0.0047 | Ent: 0.1218
+[ 16.7%] Iter 1635/9765 | SPS:  1242 | Reward:  0.065 | Value:  0.698 | VLoss: 0.0318 | PLoss: -0.0116 | Ent: 0.2164
+[ 16.8%] Iter 1636/9765 | SPS:  1243 | Reward:  0.058 | Value:  0.637 | VLoss: 0.0357 | PLoss: -0.0066 | Ent: 0.1964
+[ 16.8%] Iter 1637/9765 | SPS:  1243 | Reward:  0.051 | Value:  0.631 | VLoss: 0.0302 | PLoss: -0.0096 | Ent: 0.1286
+[ 16.8%] Iter 1638/9765 | SPS:  1244 | Reward:  0.054 | Value:  0.619 | VLoss: 0.0282 | PLoss: -0.0116 | Ent: 0.1483
+[ 16.8%] Iter 1639/9765 | SPS:  1244 | Reward:  0.061 | Value:  0.626 | VLoss: 0.0261 | PLoss: -0.0054 | Ent: 0.1338
+[ 16.8%] Iter 1640/9765 | SPS:  1245 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0287 | PLoss: -0.0073 | Ent: 0.1996
+[ 16.8%] Iter 1641/9765 | SPS:  1245 | Reward:  0.055 | Value:  0.668 | VLoss: 0.0238 | PLoss: -0.0047 | Ent: 0.1418
+[ 16.8%] Iter 1642/9765 | SPS:  1246 | Reward:  0.056 | Value:  0.618 | VLoss: 0.0241 | PLoss: -0.0154 | Ent: 0.1180
+[ 16.8%] Iter 1643/9765 | SPS:  1246 | Reward:  0.058 | Value:  0.606 | VLoss: 0.0255 | PLoss: -0.0069 | Ent: 0.1433
+[ 16.8%] Iter 1644/9765 | SPS:  1247 | Reward:  0.060 | Value:  0.583 | VLoss: 0.0315 | PLoss: -0.0095 | Ent: 0.1505
+[ 16.8%] Iter 1645/9765 | SPS:  1247 | Reward:  0.061 | Value:  0.557 | VLoss: 0.0315 | PLoss: -0.0075 | Ent: 0.1379
+[ 16.9%] Iter 1646/9765 | SPS:  1248 | Reward:  0.053 | Value:  0.582 | VLoss: 0.0204 | PLoss: -0.0056 | Ent: 0.1002
+[ 16.9%] Iter 1647/9765 | SPS:  1248 | Reward:  0.053 | Value:  0.573 | VLoss: 0.0239 | PLoss: -0.0077 | Ent: 0.1358
+[ 16.9%] Iter 1648/9765 | SPS:  1249 | Reward:  0.057 | Value:  0.630 | VLoss: 0.0361 | PLoss: -0.0119 | Ent: 0.1654
+[ 16.9%] Iter 1649/9765 | SPS:  1249 | Reward:  0.043 | Value:  0.556 | VLoss: 0.0264 | PLoss: -0.0017 | Ent: 0.1617
+[ 16.9%] Iter 1650/9765 | SPS:  1250 | Reward:  0.036 | Value:  0.592 | VLoss: 0.0261 | PLoss: -0.0109 | Ent: 0.1808
+[ 16.9%] Iter 1651/9765 | SPS:  1250 | Reward:  0.055 | Value:  0.539 | VLoss: 0.0234 | PLoss: -0.0045 | Ent: 0.0985
+[ 16.9%] Iter 1652/9765 | SPS:  1251 | Reward:  0.052 | Value:  0.652 | VLoss: 0.0220 | PLoss: -0.0089 | Ent: 0.1414
+[ 16.9%] Iter 1653/9765 | SPS:  1251 | Reward:  0.049 | Value:  0.630 | VLoss: 0.0180 | PLoss: -0.0101 | Ent: 0.1777
+[ 16.9%] Iter 1654/9765 | SPS:  1252 | Reward:  0.059 | Value:  0.535 | VLoss: 0.0289 | PLoss: -0.0123 | Ent: 0.1663
+[ 16.9%] Iter 1655/9765 | SPS:  1252 | Reward:  0.062 | Value:  0.630 | VLoss: 0.0210 | PLoss: -0.0081 | Ent: 0.1904
+[ 17.0%] Iter 1656/9765 | SPS:  1253 | Reward:  0.056 | Value:  0.612 | VLoss: 0.0327 | PLoss: -0.0171 | Ent: 0.1323
+[ 17.0%] Iter 1657/9765 | SPS:  1253 | Reward:  0.046 | Value:  0.643 | VLoss: 0.0285 | PLoss: -0.0123 | Ent: 0.1295
+[ 17.0%] Iter 1658/9765 | SPS:  1254 | Reward:  0.063 | Value:  0.563 | VLoss: 0.0372 | PLoss: -0.0093 | Ent: 0.1269
+[ 17.0%] Iter 1659/9765 | SPS:  1254 | Reward:  0.052 | Value:  0.602 | VLoss: 0.0300 | PLoss: -0.0010 | Ent: 0.1240
+[ 17.0%] Iter 1660/9765 | SPS:  1255 | Reward:  0.058 | Value:  0.586 | VLoss: 0.0323 | PLoss: -0.0095 | Ent: 0.1083
+[ 17.0%] Iter 1661/9765 | SPS:  1255 | Reward:  0.053 | Value:  0.629 | VLoss: 0.0253 | PLoss: -0.0064 | Ent: 0.1497
+[ 17.0%] Iter 1662/9765 | SPS:  1256 | Reward:  0.056 | Value:  0.627 | VLoss: 0.0291 | PLoss: -0.0042 | Ent: 0.1725
+[ 17.0%] Iter 1663/9765 | SPS:  1256 | Reward:  0.065 | Value:  0.590 | VLoss: 0.0205 | PLoss: -0.0006 | Ent: 0.1520
+[ 17.0%] Iter 1664/9765 | SPS:  1257 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0233 | PLoss: -0.0039 | Ent: 0.1068
+[ 17.1%] Iter 1665/9765 | SPS:  1257 | Reward:  0.052 | Value:  0.654 | VLoss: 0.0312 | PLoss: 0.0018 | Ent: 0.1729
+[ 17.1%] Iter 1666/9765 | SPS:  1258 | Reward:  0.066 | Value:  0.680 | VLoss: 0.0254 | PLoss: -0.0027 | Ent: 0.1274
+[ 17.1%] Iter 1667/9765 | SPS:  1258 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0226 | PLoss: -0.0043 | Ent: 0.1370
+[ 17.1%] Iter 1668/9765 | SPS:  1259 | Reward:  0.053 | Value:  0.664 | VLoss: 0.0204 | PLoss: -0.0029 | Ent: 0.1330
+[ 17.1%] Iter 1669/9765 | SPS:  1259 | Reward:  0.048 | Value:  0.641 | VLoss: 0.0226 | PLoss: -0.0141 | Ent: 0.1408
+[ 17.1%] Iter 1670/9765 | SPS:  1260 | Reward:  0.058 | Value:  0.654 | VLoss: 0.0270 | PLoss: 0.0004 | Ent: 0.1236
+[ 17.1%] Iter 1671/9765 | SPS:  1260 | Reward:  0.047 | Value:  0.629 | VLoss: 0.0264 | PLoss: -0.0067 | Ent: 0.1726
+[ 17.1%] Iter 1672/9765 | SPS:  1261 | Reward:  0.040 | Value:  0.660 | VLoss: 0.0194 | PLoss: -0.0089 | Ent: 0.2080
+[ 17.1%] Iter 1673/9765 | SPS:  1261 | Reward:  0.056 | Value:  0.700 | VLoss: 0.0230 | PLoss: -0.0161 | Ent: 0.1685
+[ 17.1%] Iter 1674/9765 | SPS:  1262 | Reward:  0.057 | Value:  0.630 | VLoss: 0.0265 | PLoss: -0.0119 | Ent: 0.1780
+[ 17.2%] Iter 1675/9765 | SPS:  1262 | Reward:  0.049 | Value:  0.716 | VLoss: 0.0267 | PLoss: -0.0045 | Ent: 0.1564
+[ 17.2%] Iter 1676/9765 | SPS:  1263 | Reward:  0.071 | Value:  0.652 | VLoss: 0.0308 | PLoss: -0.0073 | Ent: 0.1936
+[ 17.2%] Iter 1677/9765 | SPS:  1263 | Reward:  0.057 | Value:  0.647 | VLoss: 0.0338 | PLoss: -0.0062 | Ent: 0.1628
+[ 17.2%] Iter 1678/9765 | SPS:  1264 | Reward:  0.066 | Value:  0.679 | VLoss: 0.0286 | PLoss: -0.0057 | Ent: 0.1654
+[ 17.2%] Iter 1679/9765 | SPS:  1264 | Reward:  0.071 | Value:  0.689 | VLoss: 0.0344 | PLoss: -0.0029 | Ent: 0.1717
+[ 17.2%] Iter 1680/9765 | SPS:  1265 | Reward:  0.078 | Value:  0.652 | VLoss: 0.0309 | PLoss: -0.0162 | Ent: 0.1741
+[ 17.2%] Iter 1681/9765 | SPS:  1265 | Reward:  0.073 | Value:  0.700 | VLoss: 0.0328 | PLoss: -0.0071 | Ent: 0.1717
+[ 17.2%] Iter 1682/9765 | SPS:  1266 | Reward:  0.064 | Value:  0.720 | VLoss: 0.0298 | PLoss: -0.0035 | Ent: 0.1883
+[ 17.2%] Iter 1683/9765 | SPS:  1266 | Reward:  0.069 | Value:  0.697 | VLoss: 0.0269 | PLoss: -0.0156 | Ent: 0.1732
+[ 17.2%] Iter 1684/9765 | SPS:  1267 | Reward:  0.062 | Value:  0.626 | VLoss: 0.0439 | PLoss: -0.0043 | Ent: 0.1786
+[ 17.3%] Iter 1685/9765 | SPS:  1267 | Reward:  0.070 | Value:  0.697 | VLoss: 0.0323 | PLoss: -0.0117 | Ent: 0.1398
+[ 17.3%] Iter 1686/9765 | SPS:  1268 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0281 | PLoss: -0.0084 | Ent: 0.1256
+[ 17.3%] Iter 1687/9765 | SPS:  1268 | Reward:  0.067 | Value:  0.656 | VLoss: 0.0339 | PLoss: -0.0034 | Ent: 0.1445
+[ 17.3%] Iter 1688/9765 | SPS:  1269 | Reward:  0.057 | Value:  0.683 | VLoss: 0.0296 | PLoss: -0.0070 | Ent: 0.1475
+[ 17.3%] Iter 1689/9765 | SPS:  1269 | Reward:  0.064 | Value:  0.659 | VLoss: 0.0341 | PLoss: -0.0078 | Ent: 0.1597
+[ 17.3%] Iter 1690/9765 | SPS:  1269 | Reward:  0.047 | Value:  0.660 | VLoss: 0.0256 | PLoss: -0.0074 | Ent: 0.1426
+[ 17.3%] Iter 1691/9765 | SPS:  1270 | Reward:  0.052 | Value:  0.612 | VLoss: 0.0327 | PLoss: -0.0112 | Ent: 0.1680
+[ 17.3%] Iter 1692/9765 | SPS:  1270 | Reward:  0.062 | Value:  0.581 | VLoss: 0.0302 | PLoss: -0.0003 | Ent: 0.1496
+[ 17.3%] Iter 1693/9765 | SPS:  1271 | Reward:  0.074 | Value:  0.694 | VLoss: 0.0236 | PLoss: -0.0050 | Ent: 0.1320
+[ 17.3%] Iter 1694/9765 | SPS:  1271 | Reward:  0.061 | Value:  0.769 | VLoss: 0.0194 | PLoss: -0.0150 | Ent: 0.1726
+[ 17.4%] Iter 1695/9765 | SPS:  1272 | Reward:  0.060 | Value:  0.648 | VLoss: 0.0364 | PLoss: -0.0059 | Ent: 0.1904
+[ 17.4%] Iter 1696/9765 | SPS:  1272 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0287 | PLoss: -0.0049 | Ent: 0.1885
+[ 17.4%] Iter 1697/9765 | SPS:  1273 | Reward:  0.075 | Value:  0.665 | VLoss: 0.0324 | PLoss: -0.0122 | Ent: 0.1941
+[ 17.4%] Iter 1698/9765 | SPS:  1273 | Reward:  0.050 | Value:  0.716 | VLoss: 0.0233 | PLoss: -0.0090 | Ent: 0.1355
+[ 17.4%] Iter 1699/9765 | SPS:  1274 | Reward:  0.048 | Value:  0.680 | VLoss: 0.0289 | PLoss: -0.0058 | Ent: 0.1902
+[ 17.4%] Iter 1700/9765 | SPS:  1274 | Reward:  0.068 | Value:  0.595 | VLoss: 0.0244 | PLoss: -0.0078 | Ent: 0.1957
+[ 17.4%] Iter 1701/9765 | SPS:  1275 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0252 | PLoss: -0.0054 | Ent: 0.1565
+[ 17.4%] Iter 1702/9765 | SPS:  1275 | Reward:  0.064 | Value:  0.654 | VLoss: 0.0297 | PLoss: -0.0097 | Ent: 0.2052
+[ 17.4%] Iter 1703/9765 | SPS:  1276 | Reward:  0.060 | Value:  0.606 | VLoss: 0.0358 | PLoss: -0.0113 | Ent: 0.2090
+[ 17.5%] Iter 1704/9765 | SPS:  1276 | Reward:  0.059 | Value:  0.656 | VLoss: 0.0386 | PLoss: -0.0091 | Ent: 0.1671
+[ 17.5%] Iter 1705/9765 | SPS:  1277 | Reward:  0.076 | Value:  0.650 | VLoss: 0.0409 | PLoss: -0.0110 | Ent: 0.1491
+[ 17.5%] Iter 1706/9765 | SPS:  1277 | Reward:  0.073 | Value:  0.621 | VLoss: 0.0375 | PLoss: -0.0010 | Ent: 0.1350
+[ 17.5%] Iter 1707/9765 | SPS:  1278 | Reward:  0.063 | Value:  0.691 | VLoss: 0.0314 | PLoss: -0.0103 | Ent: 0.1736
+[ 17.5%] Iter 1708/9765 | SPS:  1278 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0264 | PLoss: -0.0081 | Ent: 0.1518
+[ 17.5%] Iter 1709/9765 | SPS:  1279 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0325 | PLoss: -0.0071 | Ent: 0.2060
+[ 17.5%] Iter 1710/9765 | SPS:  1279 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0307 | PLoss: -0.0132 | Ent: 0.1976
+[ 17.5%] Iter 1711/9765 | SPS:  1280 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0282 | PLoss: -0.0108 | Ent: 0.1465
+[ 17.5%] Iter 1712/9765 | SPS:  1280 | Reward:  0.063 | Value:  0.639 | VLoss: 0.0349 | PLoss: -0.0133 | Ent: 0.1771
+[ 17.5%] Iter 1713/9765 | SPS:  1281 | Reward:  0.065 | Value:  0.633 | VLoss: 0.0284 | PLoss: -0.0108 | Ent: 0.1680
+[ 17.6%] Iter 1714/9765 | SPS:  1281 | Reward:  0.065 | Value:  0.655 | VLoss: 0.0256 | PLoss: -0.0044 | Ent: 0.1695
+[ 17.6%] Iter 1715/9765 | SPS:  1282 | Reward:  0.057 | Value:  0.713 | VLoss: 0.0240 | PLoss: -0.0086 | Ent: 0.1767
+[ 17.6%] Iter 1716/9765 | SPS:  1282 | Reward:  0.047 | Value:  0.724 | VLoss: 0.0252 | PLoss: -0.0040 | Ent: 0.1577
+[ 17.6%] Iter 1717/9765 | SPS:  1283 | Reward:  0.063 | Value:  0.772 | VLoss: 0.0255 | PLoss: -0.0051 | Ent: 0.1292
+[ 17.6%] Iter 1718/9765 | SPS:  1283 | Reward:  0.072 | Value:  0.647 | VLoss: 0.0267 | PLoss: -0.0157 | Ent: 0.2090
+[ 17.6%] Iter 1719/9765 | SPS:  1284 | Reward:  0.066 | Value:  0.707 | VLoss: 0.0271 | PLoss: -0.0050 | Ent: 0.1541
+[ 17.6%] Iter 1720/9765 | SPS:  1284 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0253 | PLoss: -0.0039 | Ent: 0.1754
+[ 17.6%] Iter 1721/9765 | SPS:  1285 | Reward:  0.059 | Value:  0.572 | VLoss: 0.0356 | PLoss: -0.0180 | Ent: 0.1792
+[ 17.6%] Iter 1722/9765 | SPS:  1285 | Reward:  0.061 | Value:  0.628 | VLoss: 0.0299 | PLoss: -0.0070 | Ent: 0.1582
+[ 17.6%] Iter 1723/9765 | SPS:  1286 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0318 | PLoss: -0.0050 | Ent: 0.1619
+[ 17.7%] Iter 1724/9765 | SPS:  1286 | Reward:  0.056 | Value:  0.718 | VLoss: 0.0265 | PLoss: -0.0092 | Ent: 0.1595
+[ 17.7%] Iter 1725/9765 | SPS:  1287 | Reward:  0.065 | Value:  0.690 | VLoss: 0.0317 | PLoss: -0.0089 | Ent: 0.1391
+[ 17.7%] Iter 1726/9765 | SPS:  1287 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0283 | PLoss: -0.0054 | Ent: 0.1374
+[ 17.7%] Iter 1727/9765 | SPS:  1288 | Reward:  0.068 | Value:  0.725 | VLoss: 0.0282 | PLoss: -0.0002 | Ent: 0.1777
+[ 17.7%] Iter 1728/9765 | SPS:  1288 | Reward:  0.064 | Value:  0.731 | VLoss: 0.0326 | PLoss: -0.0103 | Ent: 0.2001
+[ 17.7%] Iter 1729/9765 | SPS:  1289 | Reward:  0.043 | Value:  0.654 | VLoss: 0.0238 | PLoss: -0.0082 | Ent: 0.1853
+[ 17.7%] Iter 1730/9765 | SPS:  1289 | Reward:  0.059 | Value:  0.633 | VLoss: 0.0255 | PLoss: -0.0065 | Ent: 0.1826
+[ 17.7%] Iter 1731/9765 | SPS:  1289 | Reward:  0.065 | Value:  0.637 | VLoss: 0.0233 | PLoss: -0.0131 | Ent: 0.1560
+[ 17.7%] Iter 1732/9765 | SPS:  1290 | Reward:  0.054 | Value:  0.659 | VLoss: 0.0229 | PLoss: -0.0078 | Ent: 0.1427
+[ 17.7%] Iter 1733/9765 | SPS:  1290 | Reward:  0.060 | Value:  0.664 | VLoss: 0.0306 | PLoss: -0.0021 | Ent: 0.1356
+[ 17.8%] Iter 1734/9765 | SPS:  1291 | Reward:  0.047 | Value:  0.711 | VLoss: 0.0183 | PLoss: -0.0049 | Ent: 0.1160
+[ 17.8%] Iter 1735/9765 | SPS:  1291 | Reward:  0.055 | Value:  0.643 | VLoss: 0.0281 | PLoss: -0.0082 | Ent: 0.1235
+[ 17.8%] Iter 1736/9765 | SPS:  1292 | Reward:  0.051 | Value:  0.635 | VLoss: 0.0174 | PLoss: -0.0014 | Ent: 0.1490
+[ 17.8%] Iter 1737/9765 | SPS:  1292 | Reward:  0.067 | Value:  0.651 | VLoss: 0.0265 | PLoss: -0.0050 | Ent: 0.1353
+[ 17.8%] Iter 1738/9765 | SPS:  1293 | Reward:  0.057 | Value:  0.635 | VLoss: 0.0257 | PLoss: -0.0086 | Ent: 0.1530
+[ 17.8%] Iter 1739/9765 | SPS:  1293 | Reward:  0.055 | Value:  0.700 | VLoss: 0.0237 | PLoss: -0.0121 | Ent: 0.2262
+[ 17.8%] Iter 1740/9765 | SPS:  1294 | Reward:  0.059 | Value:  0.695 | VLoss: 0.0224 | PLoss: -0.0135 | Ent: 0.1768
+[ 17.8%] Iter 1741/9765 | SPS:  1294 | Reward:  0.075 | Value:  0.732 | VLoss: 0.0296 | PLoss: -0.0038 | Ent: 0.1594
+[ 17.8%] Iter 1742/9765 | SPS:  1295 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0334 | PLoss: -0.0072 | Ent: 0.1589
+[ 17.8%] Iter 1743/9765 | SPS:  1295 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0331 | PLoss: -0.0058 | Ent: 0.1603
+[ 17.9%] Iter 1744/9765 | SPS:  1296 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0275 | PLoss: -0.0049 | Ent: 0.1418
+[ 17.9%] Iter 1745/9765 | SPS:  1296 | Reward:  0.060 | Value:  0.714 | VLoss: 0.0237 | PLoss: -0.0095 | Ent: 0.1650
+[ 17.9%] Iter 1746/9765 | SPS:  1297 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0307 | PLoss: -0.0091 | Ent: 0.1522
+[ 17.9%] Iter 1747/9765 | SPS:  1297 | Reward:  0.041 | Value:  0.700 | VLoss: 0.0256 | PLoss: -0.0059 | Ent: 0.1777
+[ 17.9%] Iter 1748/9765 | SPS:  1298 | Reward:  0.036 | Value:  0.651 | VLoss: 0.0239 | PLoss: -0.0033 | Ent: 0.1902
+[ 17.9%] Iter 1749/9765 | SPS:  1298 | Reward:  0.052 | Value:  0.641 | VLoss: 0.0294 | PLoss: -0.0111 | Ent: 0.1502
+[ 17.9%] Iter 1750/9765 | SPS:  1299 | Reward:  0.054 | Value:  0.699 | VLoss: 0.0334 | PLoss: -0.0061 | Ent: 0.1314
+[ 17.9%] Iter 1751/9765 | SPS:  1299 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0195 | PLoss: -0.0022 | Ent: 0.1273
+[ 17.9%] Iter 1752/9765 | SPS:  1300 | Reward:  0.044 | Value:  0.582 | VLoss: 0.0285 | PLoss: -0.0040 | Ent: 0.1584
+[ 18.0%] Iter 1753/9765 | SPS:  1300 | Reward:  0.058 | Value:  0.671 | VLoss: 0.0238 | PLoss: -0.0170 | Ent: 0.1989
+[ 18.0%] Iter 1754/9765 | SPS:  1301 | Reward:  0.050 | Value:  0.645 | VLoss: 0.0207 | PLoss: -0.0036 | Ent: 0.1714
+[ 18.0%] Iter 1755/9765 | SPS:  1301 | Reward:  0.050 | Value:  0.691 | VLoss: 0.0265 | PLoss: -0.0063 | Ent: 0.1824
+[ 18.0%] Iter 1756/9765 | SPS:  1302 | Reward:  0.048 | Value:  0.660 | VLoss: 0.0345 | PLoss: -0.0126 | Ent: 0.1616
+[ 18.0%] Iter 1757/9765 | SPS:  1302 | Reward:  0.071 | Value:  0.613 | VLoss: 0.0268 | PLoss: -0.0077 | Ent: 0.1815
+[ 18.0%] Iter 1758/9765 | SPS:  1303 | Reward:  0.051 | Value:  0.572 | VLoss: 0.0348 | PLoss: 0.0002 | Ent: 0.1823
+[ 18.0%] Iter 1759/9765 | SPS:  1303 | Reward:  0.069 | Value:  0.666 | VLoss: 0.0307 | PLoss: -0.0096 | Ent: 0.1411
+[ 18.0%] Iter 1760/9765 | SPS:  1304 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0283 | PLoss: -0.0064 | Ent: 0.1551
+[ 18.0%] Iter 1761/9765 | SPS:  1304 | Reward:  0.045 | Value:  0.637 | VLoss: 0.0256 | PLoss: -0.0086 | Ent: 0.1392
+[ 18.0%] Iter 1762/9765 | SPS:  1304 | Reward:  0.045 | Value:  0.655 | VLoss: 0.0197 | PLoss: -0.0106 | Ent: 0.1517
+[ 18.1%] Iter 1763/9765 | SPS:  1305 | Reward:  0.050 | Value:  0.644 | VLoss: 0.0210 | PLoss: -0.0053 | Ent: 0.1458
+[ 18.1%] Iter 1764/9765 | SPS:  1305 | Reward:  0.058 | Value:  0.660 | VLoss: 0.0297 | PLoss: -0.0087 | Ent: 0.1943
+[ 18.1%] Iter 1765/9765 | SPS:  1306 | Reward:  0.068 | Value:  0.662 | VLoss: 0.0407 | PLoss: -0.0148 | Ent: 0.1944
+[ 18.1%] Iter 1766/9765 | SPS:  1306 | Reward:  0.062 | Value:  0.740 | VLoss: 0.0308 | PLoss: -0.0032 | Ent: 0.1402
+[ 18.1%] Iter 1767/9765 | SPS:  1307 | Reward:  0.075 | Value:  0.670 | VLoss: 0.0365 | PLoss: -0.0122 | Ent: 0.1461
+[ 18.1%] Iter 1768/9765 | SPS:  1307 | Reward:  0.064 | Value:  0.746 | VLoss: 0.0251 | PLoss: -0.0038 | Ent: 0.1319
+[ 18.1%] Iter 1769/9765 | SPS:  1308 | Reward:  0.055 | Value:  0.667 | VLoss: 0.0272 | PLoss: -0.0076 | Ent: 0.1450
+[ 18.1%] Iter 1770/9765 | SPS:  1308 | Reward:  0.070 | Value:  0.702 | VLoss: 0.0302 | PLoss: -0.0061 | Ent: 0.1563
+[ 18.1%] Iter 1771/9765 | SPS:  1309 | Reward:  0.056 | Value:  0.730 | VLoss: 0.0264 | PLoss: -0.0043 | Ent: 0.1144
+[ 18.1%] Iter 1772/9765 | SPS:  1309 | Reward:  0.066 | Value:  0.676 | VLoss: 0.0256 | PLoss: -0.0101 | Ent: 0.1570
+[ 18.2%] Iter 1773/9765 | SPS:  1310 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0269 | PLoss: -0.0175 | Ent: 0.1913
+[ 18.2%] Iter 1774/9765 | SPS:  1310 | Reward:  0.056 | Value:  0.702 | VLoss: 0.0255 | PLoss: -0.0141 | Ent: 0.2203
+[ 18.2%] Iter 1775/9765 | SPS:  1311 | Reward:  0.070 | Value:  0.676 | VLoss: 0.0235 | PLoss: -0.0114 | Ent: 0.1838
+[ 18.2%] Iter 1776/9765 | SPS:  1311 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0330 | PLoss: -0.0109 | Ent: 0.1647
+[ 18.2%] Iter 1777/9765 | SPS:  1312 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0273 | PLoss: -0.0125 | Ent: 0.1673
+[ 18.2%] Iter 1778/9765 | SPS:  1312 | Reward:  0.060 | Value:  0.721 | VLoss: 0.0260 | PLoss: 0.0015 | Ent: 0.1513
+[ 18.2%] Iter 1779/9765 | SPS:  1313 | Reward:  0.067 | Value:  0.658 | VLoss: 0.0309 | PLoss: -0.0042 | Ent: 0.1680
+[ 18.2%] Iter 1780/9765 | SPS:  1313 | Reward:  0.048 | Value:  0.708 | VLoss: 0.0237 | PLoss: -0.0073 | Ent: 0.1584
+[ 18.2%] Iter 1781/9765 | SPS:  1314 | Reward:  0.065 | Value:  0.680 | VLoss: 0.0289 | PLoss: -0.0041 | Ent: 0.1473
+[ 18.2%] Iter 1782/9765 | SPS:  1314 | Reward:  0.059 | Value:  0.620 | VLoss: 0.0387 | PLoss: -0.0095 | Ent: 0.1674
+[ 18.3%] Iter 1783/9765 | SPS:  1314 | Reward:  0.080 | Value:  0.647 | VLoss: 0.0289 | PLoss: -0.0080 | Ent: 0.1571
+[ 18.3%] Iter 1784/9765 | SPS:  1315 | Reward:  0.055 | Value:  0.685 | VLoss: 0.0278 | PLoss: -0.0019 | Ent: 0.1627
+[ 18.3%] Iter 1785/9765 | SPS:  1315 | Reward:  0.063 | Value:  0.659 | VLoss: 0.0274 | PLoss: -0.0051 | Ent: 0.1637
+[ 18.3%] Iter 1786/9765 | SPS:  1316 | Reward:  0.075 | Value:  0.677 | VLoss: 0.0298 | PLoss: 0.0013 | Ent: 0.0769
+[ 18.3%] Iter 1787/9765 | SPS:  1316 | Reward:  0.064 | Value:  0.715 | VLoss: 0.0358 | PLoss: -0.0081 | Ent: 0.1711
+[ 18.3%] Iter 1788/9765 | SPS:  1317 | Reward:  0.064 | Value:  0.641 | VLoss: 0.0309 | PLoss: -0.0049 | Ent: 0.1661
+[ 18.3%] Iter 1789/9765 | SPS:  1317 | Reward:  0.054 | Value:  0.678 | VLoss: 0.0259 | PLoss: -0.0053 | Ent: 0.1737
+[ 18.3%] Iter 1790/9765 | SPS:  1318 | Reward:  0.057 | Value:  0.637 | VLoss: 0.0242 | PLoss: -0.0026 | Ent: 0.1995
+[ 18.3%] Iter 1791/9765 | SPS:  1318 | Reward:  0.051 | Value:  0.650 | VLoss: 0.0269 | PLoss: -0.0110 | Ent: 0.1846
+[ 18.4%] Iter 1792/9765 | SPS:  1319 | Reward:  0.074 | Value:  0.707 | VLoss: 0.0238 | PLoss: -0.0173 | Ent: 0.1585
+[ 18.4%] Iter 1793/9765 | SPS:  1319 | Reward:  0.061 | Value:  0.684 | VLoss: 0.0238 | PLoss: -0.0031 | Ent: 0.1122
+[ 18.4%] Iter 1794/9765 | SPS:  1320 | Reward:  0.065 | Value:  0.660 | VLoss: 0.0359 | PLoss: -0.0059 | Ent: 0.1511
+[ 18.4%] Iter 1795/9765 | SPS:  1320 | Reward:  0.051 | Value:  0.608 | VLoss: 0.0260 | PLoss: -0.0082 | Ent: 0.1356
+[ 18.4%] Iter 1796/9765 | SPS:  1321 | Reward:  0.059 | Value:  0.616 | VLoss: 0.0279 | PLoss: -0.0051 | Ent: 0.1315
+[ 18.4%] Iter 1797/9765 | SPS:  1321 | Reward:  0.051 | Value:  0.709 | VLoss: 0.0230 | PLoss: -0.0127 | Ent: 0.1667
+[ 18.4%] Iter 1798/9765 | SPS:  1322 | Reward:  0.064 | Value:  0.665 | VLoss: 0.0303 | PLoss: -0.0104 | Ent: 0.1719
+[ 18.4%] Iter 1799/9765 | SPS:  1322 | Reward:  0.070 | Value:  0.657 | VLoss: 0.0294 | PLoss: -0.0066 | Ent: 0.1277
+[ 18.4%] Iter 1800/9765 | SPS:  1323 | Reward:  0.053 | Value:  0.721 | VLoss: 0.0276 | PLoss: -0.0106 | Ent: 0.1573
+[ 18.4%] Iter 1801/9765 | SPS:  1323 | Reward:  0.056 | Value:  0.632 | VLoss: 0.0298 | PLoss: -0.0055 | Ent: 0.1386
+[ 18.5%] Iter 1802/9765 | SPS:  1323 | Reward:  0.055 | Value:  0.692 | VLoss: 0.0251 | PLoss: -0.0070 | Ent: 0.1362
+[ 18.5%] Iter 1803/9765 | SPS:  1324 | Reward:  0.050 | Value:  0.641 | VLoss: 0.0202 | PLoss: -0.0123 | Ent: 0.1579
+[ 18.5%] Iter 1804/9765 | SPS:  1324 | Reward:  0.054 | Value:  0.687 | VLoss: 0.0267 | PLoss: -0.0126 | Ent: 0.1531
+[ 18.5%] Iter 1805/9765 | SPS:  1325 | Reward:  0.052 | Value:  0.630 | VLoss: 0.0304 | PLoss: -0.0146 | Ent: 0.1692
+[ 18.5%] Iter 1806/9765 | SPS:  1325 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0306 | PLoss: -0.0100 | Ent: 0.1424
+[ 18.5%] Iter 1807/9765 | SPS:  1326 | Reward:  0.067 | Value:  0.655 | VLoss: 0.0301 | PLoss: -0.0105 | Ent: 0.1499
+[ 18.5%] Iter 1808/9765 | SPS:  1326 | Reward:  0.060 | Value:  0.659 | VLoss: 0.0251 | PLoss: -0.0031 | Ent: 0.1727
+[ 18.5%] Iter 1809/9765 | SPS:  1327 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0289 | PLoss: -0.0063 | Ent: 0.1754
+[ 18.5%] Iter 1810/9765 | SPS:  1327 | Reward:  0.068 | Value:  0.658 | VLoss: 0.0339 | PLoss: -0.0059 | Ent: 0.1935
+[ 18.5%] Iter 1811/9765 | SPS:  1328 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0226 | PLoss: -0.0065 | Ent: 0.1708
+[ 18.6%] Iter 1812/9765 | SPS:  1328 | Reward:  0.063 | Value:  0.659 | VLoss: 0.0319 | PLoss: -0.0080 | Ent: 0.1772
+[ 18.6%] Iter 1813/9765 | SPS:  1329 | Reward:  0.069 | Value:  0.707 | VLoss: 0.0311 | PLoss: -0.0069 | Ent: 0.1712
+[ 18.6%] Iter 1814/9765 | SPS:  1329 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0296 | PLoss: -0.0119 | Ent: 0.1461
+[ 18.6%] Iter 1815/9765 | SPS:  1329 | Reward:  0.059 | Value:  0.698 | VLoss: 0.0239 | PLoss: -0.0097 | Ent: 0.1482
+[ 18.6%] Iter 1816/9765 | SPS:  1330 | Reward:  0.057 | Value:  0.679 | VLoss: 0.0169 | PLoss: -0.0123 | Ent: 0.1466
+[ 18.6%] Iter 1817/9765 | SPS:  1330 | Reward:  0.041 | Value:  0.656 | VLoss: 0.0245 | PLoss: -0.0010 | Ent: 0.1853
+[ 18.6%] Iter 1818/9765 | SPS:  1331 | Reward:  0.049 | Value:  0.661 | VLoss: 0.0197 | PLoss: -0.0029 | Ent: 0.1372
+[ 18.6%] Iter 1819/9765 | SPS:  1331 | Reward:  0.044 | Value:  0.655 | VLoss: 0.0234 | PLoss: -0.0070 | Ent: 0.1245
+[ 18.6%] Iter 1820/9765 | SPS:  1332 | Reward:  0.048 | Value:  0.639 | VLoss: 0.0186 | PLoss: -0.0138 | Ent: 0.1582
+[ 18.6%] Iter 1821/9765 | SPS:  1332 | Reward:  0.046 | Value:  0.643 | VLoss: 0.0216 | PLoss: -0.0081 | Ent: 0.1267
+[ 18.7%] Iter 1822/9765 | SPS:  1333 | Reward:  0.052 | Value:  0.630 | VLoss: 0.0268 | PLoss: -0.0098 | Ent: 0.1494
+[ 18.7%] Iter 1823/9765 | SPS:  1333 | Reward:  0.053 | Value:  0.615 | VLoss: 0.0259 | PLoss: -0.0075 | Ent: 0.1469
+[ 18.7%] Iter 1824/9765 | SPS:  1334 | Reward:  0.043 | Value:  0.642 | VLoss: 0.0268 | PLoss: -0.0061 | Ent: 0.1145
+[ 18.7%] Iter 1825/9765 | SPS:  1334 | Reward:  0.056 | Value:  0.623 | VLoss: 0.0225 | PLoss: -0.0057 | Ent: 0.1417
+[ 18.7%] Iter 1826/9765 | SPS:  1335 | Reward:  0.040 | Value:  0.591 | VLoss: 0.0166 | PLoss: -0.0113 | Ent: 0.1264
+[ 18.7%] Iter 1827/9765 | SPS:  1335 | Reward:  0.042 | Value:  0.573 | VLoss: 0.0217 | PLoss: -0.0040 | Ent: 0.1381
+[ 18.7%] Iter 1828/9765 | SPS:  1336 | Reward:  0.048 | Value:  0.571 | VLoss: 0.0212 | PLoss: -0.0138 | Ent: 0.1397
+[ 18.7%] Iter 1829/9765 | SPS:  1336 | Reward:  0.047 | Value:  0.661 | VLoss: 0.0219 | PLoss: -0.0068 | Ent: 0.1673
+[ 18.7%] Iter 1830/9765 | SPS:  1337 | Reward:  0.058 | Value:  0.652 | VLoss: 0.0267 | PLoss: -0.0051 | Ent: 0.1348
+[ 18.8%] Iter 1831/9765 | SPS:  1337 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0214 | PLoss: -0.0010 | Ent: 0.1944
+[ 18.8%] Iter 1832/9765 | SPS:  1338 | Reward:  0.073 | Value:  0.698 | VLoss: 0.0269 | PLoss: -0.0101 | Ent: 0.1631
+[ 18.8%] Iter 1833/9765 | SPS:  1338 | Reward:  0.059 | Value:  0.672 | VLoss: 0.0294 | PLoss: -0.0057 | Ent: 0.1813
+[ 18.8%] Iter 1834/9765 | SPS:  1339 | Reward:  0.068 | Value:  0.644 | VLoss: 0.0413 | PLoss: -0.0015 | Ent: 0.1586
+[ 18.8%] Iter 1835/9765 | SPS:  1339 | Reward:  0.042 | Value:  0.655 | VLoss: 0.0254 | PLoss: -0.0086 | Ent: 0.1302
+[ 18.8%] Iter 1836/9765 | SPS:  1340 | Reward:  0.057 | Value:  0.665 | VLoss: 0.0223 | PLoss: -0.0082 | Ent: 0.1331
+[ 18.8%] Iter 1837/9765 | SPS:  1340 | Reward:  0.067 | Value:  0.726 | VLoss: 0.0214 | PLoss: -0.0020 | Ent: 0.1386
+[ 18.8%] Iter 1838/9765 | SPS:  1340 | Reward:  0.055 | Value:  0.700 | VLoss: 0.0246 | PLoss: -0.0060 | Ent: 0.1661
+[ 18.8%] Iter 1839/9765 | SPS:  1341 | Reward:  0.048 | Value:  0.679 | VLoss: 0.0243 | PLoss: -0.0110 | Ent: 0.1870
+[ 18.8%] Iter 1840/9765 | SPS:  1341 | Reward:  0.080 | Value:  0.700 | VLoss: 0.0310 | PLoss: -0.0029 | Ent: 0.1444
+[ 18.9%] Iter 1841/9765 | SPS:  1342 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0211 | PLoss: -0.0095 | Ent: 0.1361
+[ 18.9%] Iter 1842/9765 | SPS:  1342 | Reward:  0.069 | Value:  0.688 | VLoss: 0.0287 | PLoss: -0.0094 | Ent: 0.1556
+[ 18.9%] Iter 1843/9765 | SPS:  1343 | Reward:  0.061 | Value:  0.669 | VLoss: 0.0298 | PLoss: -0.0073 | Ent: 0.1769
+[ 18.9%] Iter 1844/9765 | SPS:  1343 | Reward:  0.063 | Value:  0.673 | VLoss: 0.0398 | PLoss: -0.0080 | Ent: 0.1216
+[ 18.9%] Iter 1845/9765 | SPS:  1344 | Reward:  0.062 | Value:  0.630 | VLoss: 0.0299 | PLoss: -0.0139 | Ent: 0.1596
+[ 18.9%] Iter 1846/9765 | SPS:  1344 | Reward:  0.078 | Value:  0.673 | VLoss: 0.0318 | PLoss: -0.0086 | Ent: 0.1528
+[ 18.9%] Iter 1847/9765 | SPS:  1345 | Reward:  0.071 | Value:  0.634 | VLoss: 0.0435 | PLoss: -0.0087 | Ent: 0.1580
+[ 18.9%] Iter 1848/9765 | SPS:  1345 | Reward:  0.076 | Value:  0.684 | VLoss: 0.0342 | PLoss: -0.0087 | Ent: 0.1507
+[ 18.9%] Iter 1849/9765 | SPS:  1345 | Reward:  0.070 | Value:  0.702 | VLoss: 0.0331 | PLoss: -0.0093 | Ent: 0.1339
+[ 18.9%] Iter 1850/9765 | SPS:  1346 | Reward:  0.073 | Value:  0.737 | VLoss: 0.0262 | PLoss: -0.0153 | Ent: 0.1500
+[ 19.0%] Iter 1851/9765 | SPS:  1346 | Reward:  0.070 | Value:  0.740 | VLoss: 0.0297 | PLoss: -0.0084 | Ent: 0.1489
+[ 19.0%] Iter 1852/9765 | SPS:  1347 | Reward:  0.073 | Value:  0.651 | VLoss: 0.0272 | PLoss: -0.0124 | Ent: 0.1635
+[ 19.0%] Iter 1853/9765 | SPS:  1347 | Reward:  0.069 | Value:  0.737 | VLoss: 0.0394 | PLoss: -0.0057 | Ent: 0.1307
+[ 19.0%] Iter 1854/9765 | SPS:  1348 | Reward:  0.063 | Value:  0.734 | VLoss: 0.0288 | PLoss: -0.0084 | Ent: 0.1567
+[ 19.0%] Iter 1855/9765 | SPS:  1348 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0243 | PLoss: 0.0004 | Ent: 0.1457
+[ 19.0%] Iter 1856/9765 | SPS:  1349 | Reward:  0.051 | Value:  0.638 | VLoss: 0.0201 | PLoss: -0.0107 | Ent: 0.1322
+[ 19.0%] Iter 1857/9765 | SPS:  1349 | Reward:  0.042 | Value:  0.633 | VLoss: 0.0222 | PLoss: -0.0180 | Ent: 0.1676
+[ 19.0%] Iter 1858/9765 | SPS:  1350 | Reward:  0.054 | Value:  0.619 | VLoss: 0.0320 | PLoss: -0.0080 | Ent: 0.1578
+[ 19.0%] Iter 1859/9765 | SPS:  1350 | Reward:  0.053 | Value:  0.630 | VLoss: 0.0202 | PLoss: -0.0104 | Ent: 0.1591
+[ 19.0%] Iter 1860/9765 | SPS:  1351 | Reward:  0.045 | Value:  0.657 | VLoss: 0.0242 | PLoss: -0.0038 | Ent: 0.1894
+[ 19.1%] Iter 1861/9765 | SPS:  1351 | Reward:  0.052 | Value:  0.586 | VLoss: 0.0287 | PLoss: -0.0039 | Ent: 0.1787
+[ 19.1%] Iter 1862/9765 | SPS:  1351 | Reward:  0.057 | Value:  0.644 | VLoss: 0.0288 | PLoss: -0.0063 | Ent: 0.1405
+[ 19.1%] Iter 1863/9765 | SPS:  1352 | Reward:  0.050 | Value:  0.639 | VLoss: 0.0233 | PLoss: -0.0047 | Ent: 0.1810
+[ 19.1%] Iter 1864/9765 | SPS:  1352 | Reward:  0.050 | Value:  0.701 | VLoss: 0.0206 | PLoss: -0.0087 | Ent: 0.1070
+[ 19.1%] Iter 1865/9765 | SPS:  1353 | Reward:  0.048 | Value:  0.687 | VLoss: 0.0254 | PLoss: -0.0101 | Ent: 0.1534
+[ 19.1%] Iter 1866/9765 | SPS:  1353 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0281 | PLoss: -0.0107 | Ent: 0.1211
+[ 19.1%] Iter 1867/9765 | SPS:  1354 | Reward:  0.057 | Value:  0.598 | VLoss: 0.0263 | PLoss: 0.0008 | Ent: 0.1630
+[ 19.1%] Iter 1868/9765 | SPS:  1354 | Reward:  0.056 | Value:  0.604 | VLoss: 0.0267 | PLoss: -0.0118 | Ent: 0.1861
+[ 19.1%] Iter 1869/9765 | SPS:  1355 | Reward:  0.064 | Value:  0.686 | VLoss: 0.0296 | PLoss: -0.0097 | Ent: 0.1909
+[ 19.2%] Iter 1870/9765 | SPS:  1355 | Reward:  0.045 | Value:  0.647 | VLoss: 0.0308 | PLoss: -0.0108 | Ent: 0.2103
+[ 19.2%] Iter 1871/9765 | SPS:  1356 | Reward:  0.047 | Value:  0.667 | VLoss: 0.0323 | PLoss: -0.0097 | Ent: 0.1521
+[ 19.2%] Iter 1872/9765 | SPS:  1356 | Reward:  0.062 | Value:  0.588 | VLoss: 0.0267 | PLoss: -0.0092 | Ent: 0.1716
+[ 19.2%] Iter 1873/9765 | SPS:  1357 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0220 | PLoss: -0.0104 | Ent: 0.2120
+[ 19.2%] Iter 1874/9765 | SPS:  1357 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0340 | PLoss: -0.0108 | Ent: 0.1746
+[ 19.2%] Iter 1875/9765 | SPS:  1357 | Reward:  0.073 | Value:  0.725 | VLoss: 0.0295 | PLoss: -0.0051 | Ent: 0.1641
+[ 19.2%] Iter 1876/9765 | SPS:  1358 | Reward:  0.060 | Value:  0.679 | VLoss: 0.0297 | PLoss: -0.0051 | Ent: 0.1853
+[ 19.2%] Iter 1877/9765 | SPS:  1358 | Reward:  0.059 | Value:  0.642 | VLoss: 0.0358 | PLoss: -0.0099 | Ent: 0.1739
+[ 19.2%] Iter 1878/9765 | SPS:  1359 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0298 | PLoss: -0.0077 | Ent: 0.1430
+[ 19.2%] Iter 1879/9765 | SPS:  1359 | Reward:  0.055 | Value:  0.643 | VLoss: 0.0326 | PLoss: -0.0071 | Ent: 0.1768
+[ 19.3%] Iter 1880/9765 | SPS:  1360 | Reward:  0.075 | Value:  0.667 | VLoss: 0.0312 | PLoss: -0.0090 | Ent: 0.1800
+[ 19.3%] Iter 1881/9765 | SPS:  1360 | Reward:  0.050 | Value:  0.743 | VLoss: 0.0210 | PLoss: -0.0046 | Ent: 0.1167
+[ 19.3%] Iter 1882/9765 | SPS:  1361 | Reward:  0.064 | Value:  0.679 | VLoss: 0.0343 | PLoss: -0.0070 | Ent: 0.1653
+[ 19.3%] Iter 1883/9765 | SPS:  1361 | Reward:  0.063 | Value:  0.668 | VLoss: 0.0341 | PLoss: -0.0117 | Ent: 0.1420
+[ 19.3%] Iter 1884/9765 | SPS:  1361 | Reward:  0.059 | Value:  0.673 | VLoss: 0.0240 | PLoss: -0.0156 | Ent: 0.1412
+[ 19.3%] Iter 1885/9765 | SPS:  1362 | Reward:  0.072 | Value:  0.694 | VLoss: 0.0295 | PLoss: -0.0077 | Ent: 0.1586
+[ 19.3%] Iter 1886/9765 | SPS:  1362 | Reward:  0.050 | Value:  0.680 | VLoss: 0.0249 | PLoss: -0.0090 | Ent: 0.1689
+[ 19.3%] Iter 1887/9765 | SPS:  1363 | Reward:  0.049 | Value:  0.676 | VLoss: 0.0204 | PLoss: -0.0096 | Ent: 0.1351
+[ 19.3%] Iter 1888/9765 | SPS:  1363 | Reward:  0.044 | Value:  0.611 | VLoss: 0.0182 | PLoss: -0.0154 | Ent: 0.1610
+[ 19.3%] Iter 1889/9765 | SPS:  1364 | Reward:  0.048 | Value:  0.559 | VLoss: 0.0206 | PLoss: -0.0039 | Ent: 0.1670
+[ 19.4%] Iter 1890/9765 | SPS:  1364 | Reward:  0.040 | Value:  0.599 | VLoss: 0.0216 | PLoss: -0.0036 | Ent: 0.1544
+[ 19.4%] Iter 1891/9765 | SPS:  1365 | Reward:  0.038 | Value:  0.565 | VLoss: 0.0174 | PLoss: -0.0142 | Ent: 0.1964
+[ 19.4%] Iter 1892/9765 | SPS:  1365 | Reward:  0.047 | Value:  0.551 | VLoss: 0.0173 | PLoss: -0.0042 | Ent: 0.2545
+[ 19.4%] Iter 1893/9765 | SPS:  1366 | Reward:  0.058 | Value:  0.631 | VLoss: 0.0259 | PLoss: -0.0113 | Ent: 0.1648
+[ 19.4%] Iter 1894/9765 | SPS:  1366 | Reward:  0.052 | Value:  0.619 | VLoss: 0.0174 | PLoss: -0.0063 | Ent: 0.1646
+[ 19.4%] Iter 1895/9765 | SPS:  1367 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0261 | PLoss: -0.0092 | Ent: 0.1934
+[ 19.4%] Iter 1896/9765 | SPS:  1367 | Reward:  0.066 | Value:  0.658 | VLoss: 0.0281 | PLoss: -0.0024 | Ent: 0.1517
+[ 19.4%] Iter 1897/9765 | SPS:  1367 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0409 | PLoss: 0.0078 | Ent: 0.1725
+[ 19.4%] Iter 1898/9765 | SPS:  1368 | Reward:  0.072 | Value:  0.676 | VLoss: 0.0327 | PLoss: -0.0104 | Ent: 0.1341
+[ 19.4%] Iter 1899/9765 | SPS:  1368 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0371 | PLoss: -0.0135 | Ent: 0.1708
+[ 19.5%] Iter 1900/9765 | SPS:  1369 | Reward:  0.074 | Value:  0.673 | VLoss: 0.0310 | PLoss: -0.0098 | Ent: 0.1341
+[ 19.5%] Iter 1901/9765 | SPS:  1369 | Reward:  0.061 | Value:  0.672 | VLoss: 0.0266 | PLoss: -0.0105 | Ent: 0.1445
+[ 19.5%] Iter 1902/9765 | SPS:  1370 | Reward:  0.065 | Value:  0.699 | VLoss: 0.0250 | PLoss: -0.0054 | Ent: 0.1559
+[ 19.5%] Iter 1903/9765 | SPS:  1370 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0290 | PLoss: -0.0055 | Ent: 0.1593
+[ 19.5%] Iter 1904/9765 | SPS:  1371 | Reward:  0.065 | Value:  0.671 | VLoss: 0.0362 | PLoss: -0.0021 | Ent: 0.1515
+[ 19.5%] Iter 1905/9765 | SPS:  1371 | Reward:  0.069 | Value:  0.640 | VLoss: 0.0282 | PLoss: -0.0050 | Ent: 0.1456
+[ 19.5%] Iter 1906/9765 | SPS:  1371 | Reward:  0.067 | Value:  0.638 | VLoss: 0.0382 | PLoss: -0.0062 | Ent: 0.1871
+[ 19.5%] Iter 1907/9765 | SPS:  1372 | Reward:  0.064 | Value:  0.675 | VLoss: 0.0306 | PLoss: -0.0066 | Ent: 0.1411
+[ 19.5%] Iter 1908/9765 | SPS:  1372 | Reward:  0.049 | Value:  0.666 | VLoss: 0.0293 | PLoss: -0.0034 | Ent: 0.1056
+[ 19.5%] Iter 1909/9765 | SPS:  1373 | Reward:  0.070 | Value:  0.650 | VLoss: 0.0258 | PLoss: -0.0058 | Ent: 0.1326
+[ 19.6%] Iter 1910/9765 | SPS:  1373 | Reward:  0.061 | Value:  0.682 | VLoss: 0.0315 | PLoss: -0.0146 | Ent: 0.1628
+[ 19.6%] Iter 1911/9765 | SPS:  1374 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0228 | PLoss: -0.0067 | Ent: 0.1178
+[ 19.6%] Iter 1912/9765 | SPS:  1374 | Reward:  0.056 | Value:  0.624 | VLoss: 0.0333 | PLoss: -0.0017 | Ent: 0.1554
+[ 19.6%] Iter 1913/9765 | SPS:  1375 | Reward:  0.058 | Value:  0.614 | VLoss: 0.0278 | PLoss: -0.0069 | Ent: 0.1786
+[ 19.6%] Iter 1914/9765 | SPS:  1375 | Reward:  0.053 | Value:  0.684 | VLoss: 0.0308 | PLoss: -0.0087 | Ent: 0.1528
+[ 19.6%] Iter 1915/9765 | SPS:  1375 | Reward:  0.054 | Value:  0.681 | VLoss: 0.0298 | PLoss: -0.0056 | Ent: 0.1449
+[ 19.6%] Iter 1916/9765 | SPS:  1376 | Reward:  0.066 | Value:  0.633 | VLoss: 0.0309 | PLoss: -0.0092 | Ent: 0.1680
+[ 19.6%] Iter 1917/9765 | SPS:  1376 | Reward:  0.066 | Value:  0.655 | VLoss: 0.0290 | PLoss: -0.0083 | Ent: 0.1355
+[ 19.6%] Iter 1918/9765 | SPS:  1377 | Reward:  0.062 | Value:  0.629 | VLoss: 0.0245 | PLoss: -0.0033 | Ent: 0.1779
+[ 19.7%] Iter 1919/9765 | SPS:  1377 | Reward:  0.070 | Value:  0.668 | VLoss: 0.0288 | PLoss: -0.0077 | Ent: 0.1856
+[ 19.7%] Iter 1920/9765 | SPS:  1378 | Reward:  0.063 | Value:  0.680 | VLoss: 0.0258 | PLoss: -0.0080 | Ent: 0.1600
+[ 19.7%] Iter 1921/9765 | SPS:  1378 | Reward:  0.074 | Value:  0.688 | VLoss: 0.0283 | PLoss: -0.0058 | Ent: 0.1644
+[ 19.7%] Iter 1922/9765 | SPS:  1379 | Reward:  0.067 | Value:  0.663 | VLoss: 0.0342 | PLoss: -0.0100 | Ent: 0.1639
+[ 19.7%] Iter 1923/9765 | SPS:  1379 | Reward:  0.075 | Value:  0.670 | VLoss: 0.0250 | PLoss: -0.0027 | Ent: 0.1203
+[ 19.7%] Iter 1924/9765 | SPS:  1379 | Reward:  0.068 | Value:  0.693 | VLoss: 0.0316 | PLoss: -0.0124 | Ent: 0.1663
+[ 19.7%] Iter 1925/9765 | SPS:  1380 | Reward:  0.048 | Value:  0.668 | VLoss: 0.0268 | PLoss: -0.0012 | Ent: 0.1717
+[ 19.7%] Iter 1926/9765 | SPS:  1380 | Reward:  0.061 | Value:  0.631 | VLoss: 0.0297 | PLoss: 0.0002 | Ent: 0.1693
+[ 19.7%] Iter 1927/9765 | SPS:  1381 | Reward:  0.058 | Value:  0.663 | VLoss: 0.0249 | PLoss: -0.0024 | Ent: 0.1288
+[ 19.7%] Iter 1928/9765 | SPS:  1381 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0235 | PLoss: -0.0093 | Ent: 0.1613
+[ 19.8%] Iter 1929/9765 | SPS:  1382 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0304 | PLoss: -0.0086 | Ent: 0.1519
+[ 19.8%] Iter 1930/9765 | SPS:  1382 | Reward:  0.071 | Value:  0.726 | VLoss: 0.0258 | PLoss: -0.0062 | Ent: 0.1782
+[ 19.8%] Iter 1931/9765 | SPS:  1383 | Reward:  0.065 | Value:  0.710 | VLoss: 0.0255 | PLoss: -0.0085 | Ent: 0.1187
+[ 19.8%] Iter 1932/9765 | SPS:  1383 | Reward:  0.053 | Value:  0.645 | VLoss: 0.0380 | PLoss: -0.0057 | Ent: 0.1621
+[ 19.8%] Iter 1933/9765 | SPS:  1383 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0305 | PLoss: -0.0071 | Ent: 0.1620
+[ 19.8%] Iter 1934/9765 | SPS:  1384 | Reward:  0.064 | Value:  0.694 | VLoss: 0.0233 | PLoss: -0.0005 | Ent: 0.1600
+[ 19.8%] Iter 1935/9765 | SPS:  1384 | Reward:  0.062 | Value:  0.666 | VLoss: 0.0213 | PLoss: -0.0045 | Ent: 0.1635
+[ 19.8%] Iter 1936/9765 | SPS:  1385 | Reward:  0.067 | Value:  0.662 | VLoss: 0.0301 | PLoss: -0.0052 | Ent: 0.1698
+[ 19.8%] Iter 1937/9765 | SPS:  1385 | Reward:  0.031 | Value:  0.634 | VLoss: 0.0174 | PLoss: -0.0035 | Ent: 0.1342
+[ 19.8%] Iter 1938/9765 | SPS:  1386 | Reward:  0.063 | Value:  0.665 | VLoss: 0.0370 | PLoss: -0.0125 | Ent: 0.1456
+[ 19.9%] Iter 1939/9765 | SPS:  1386 | Reward:  0.051 | Value:  0.704 | VLoss: 0.0231 | PLoss: -0.0049 | Ent: 0.1314
+[ 19.9%] Iter 1940/9765 | SPS:  1387 | Reward:  0.044 | Value:  0.691 | VLoss: 0.0304 | PLoss: -0.0051 | Ent: 0.1212
+[ 19.9%] Iter 1941/9765 | SPS:  1387 | Reward:  0.048 | Value:  0.641 | VLoss: 0.0287 | PLoss: -0.0063 | Ent: 0.1123
+[ 19.9%] Iter 1942/9765 | SPS:  1388 | Reward:  0.056 | Value:  0.599 | VLoss: 0.0243 | PLoss: -0.0096 | Ent: 0.1889
+[ 19.9%] Iter 1943/9765 | SPS:  1388 | Reward:  0.054 | Value:  0.600 | VLoss: 0.0272 | PLoss: -0.0076 | Ent: 0.1298
+[ 19.9%] Iter 1944/9765 | SPS:  1388 | Reward:  0.045 | Value:  0.627 | VLoss: 0.0203 | PLoss: -0.0058 | Ent: 0.1403
+[ 19.9%] Iter 1945/9765 | SPS:  1389 | Reward:  0.060 | Value:  0.669 | VLoss: 0.0300 | PLoss: -0.0170 | Ent: 0.1434
+[ 19.9%] Iter 1946/9765 | SPS:  1389 | Reward:  0.068 | Value:  0.612 | VLoss: 0.0347 | PLoss: -0.0060 | Ent: 0.1304
+[ 19.9%] Iter 1947/9765 | SPS:  1390 | Reward:  0.041 | Value:  0.756 | VLoss: 0.0192 | PLoss: -0.0026 | Ent: 0.1255
+[ 19.9%] Iter 1948/9765 | SPS:  1390 | Reward:  0.053 | Value:  0.696 | VLoss: 0.0199 | PLoss: -0.0085 | Ent: 0.1228
+[ 20.0%] Iter 1949/9765 | SPS:  1391 | Reward:  0.061 | Value:  0.698 | VLoss: 0.0276 | PLoss: -0.0054 | Ent: 0.1376
+[ 20.0%] Iter 1950/9765 | SPS:  1391 | Reward:  0.067 | Value:  0.667 | VLoss: 0.0309 | PLoss: -0.0107 | Ent: 0.1899
+[ 20.0%] Iter 1951/9765 | SPS:  1392 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0297 | PLoss: -0.0174 | Ent: 0.1979
+[ 20.0%] Iter 1952/9765 | SPS:  1392 | Reward:  0.069 | Value:  0.619 | VLoss: 0.0298 | PLoss: -0.0065 | Ent: 0.1638
+[ 20.0%] Iter 1953/9765 | SPS:  1392 | Reward:  0.057 | Value:  0.738 | VLoss: 0.0332 | PLoss: -0.0094 | Ent: 0.1577
+[ 20.0%] Iter 1954/9765 | SPS:  1393 | Reward:  0.073 | Value:  0.704 | VLoss: 0.0340 | PLoss: -0.0099 | Ent: 0.1265
+[ 20.0%] Iter 1955/9765 | SPS:  1393 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0332 | PLoss: -0.0060 | Ent: 0.1408
+[ 20.0%] Iter 1956/9765 | SPS:  1394 | Reward:  0.057 | Value:  0.605 | VLoss: 0.0351 | PLoss: -0.0128 | Ent: 0.1617
+[ 20.0%] Iter 1957/9765 | SPS:  1394 | Reward:  0.072 | Value:  0.599 | VLoss: 0.0384 | PLoss: -0.0050 | Ent: 0.1104
+[ 20.1%] Iter 1958/9765 | SPS:  1395 | Reward:  0.066 | Value:  0.623 | VLoss: 0.0257 | PLoss: -0.0090 | Ent: 0.1546
+[ 20.1%] Iter 1959/9765 | SPS:  1395 | Reward:  0.065 | Value:  0.737 | VLoss: 0.0259 | PLoss: -0.0055 | Ent: 0.1303
+[ 20.1%] Iter 1960/9765 | SPS:  1395 | Reward:  0.061 | Value:  0.747 | VLoss: 0.0318 | PLoss: -0.0101 | Ent: 0.1536
+[ 20.1%] Iter 1961/9765 | SPS:  1396 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0273 | PLoss: -0.0016 | Ent: 0.1398
+[ 20.1%] Iter 1962/9765 | SPS:  1396 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0323 | PLoss: -0.0169 | Ent: 0.1975
+[ 20.1%] Iter 1963/9765 | SPS:  1397 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0244 | PLoss: -0.0140 | Ent: 0.2094
+[ 20.1%] Iter 1964/9765 | SPS:  1397 | Reward:  0.062 | Value:  0.663 | VLoss: 0.0280 | PLoss: -0.0070 | Ent: 0.1429
+[ 20.1%] Iter 1965/9765 | SPS:  1398 | Reward:  0.060 | Value:  0.613 | VLoss: 0.0272 | PLoss: -0.0091 | Ent: 0.1393
+[ 20.1%] Iter 1966/9765 | SPS:  1398 | Reward:  0.059 | Value:  0.652 | VLoss: 0.0320 | PLoss: -0.0082 | Ent: 0.1883
+[ 20.1%] Iter 1967/9765 | SPS:  1399 | Reward:  0.067 | Value:  0.605 | VLoss: 0.0332 | PLoss: -0.0068 | Ent: 0.1211
+[ 20.2%] Iter 1968/9765 | SPS:  1399 | Reward:  0.062 | Value:  0.620 | VLoss: 0.0294 | PLoss: -0.0184 | Ent: 0.2029
+[ 20.2%] Iter 1969/9765 | SPS:  1399 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0308 | PLoss: -0.0093 | Ent: 0.1585
+[ 20.2%] Iter 1970/9765 | SPS:  1400 | Reward:  0.062 | Value:  0.605 | VLoss: 0.0292 | PLoss: -0.0124 | Ent: 0.1203
+[ 20.2%] Iter 1971/9765 | SPS:  1400 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0248 | PLoss: -0.0051 | Ent: 0.1327
+[ 20.2%] Iter 1972/9765 | SPS:  1401 | Reward:  0.060 | Value:  0.707 | VLoss: 0.0202 | PLoss: -0.0066 | Ent: 0.1648
+[ 20.2%] Iter 1973/9765 | SPS:  1401 | Reward:  0.061 | Value:  0.646 | VLoss: 0.0291 | PLoss: -0.0048 | Ent: 0.1270
+[ 20.2%] Iter 1974/9765 | SPS:  1402 | Reward:  0.052 | Value:  0.686 | VLoss: 0.0253 | PLoss: -0.0070 | Ent: 0.1808
+[ 20.2%] Iter 1975/9765 | SPS:  1402 | Reward:  0.053 | Value:  0.643 | VLoss: 0.0251 | PLoss: -0.0044 | Ent: 0.2122
+[ 20.2%] Iter 1976/9765 | SPS:  1403 | Reward:  0.055 | Value:  0.649 | VLoss: 0.0216 | PLoss: -0.0116 | Ent: 0.1688
+[ 20.2%] Iter 1977/9765 | SPS:  1403 | Reward:  0.052 | Value:  0.608 | VLoss: 0.0237 | PLoss: -0.0052 | Ent: 0.1388
+[ 20.3%] Iter 1978/9765 | SPS:  1403 | Reward:  0.042 | Value:  0.632 | VLoss: 0.0201 | PLoss: -0.0084 | Ent: 0.1329
+[ 20.3%] Iter 1979/9765 | SPS:  1404 | Reward:  0.054 | Value:  0.696 | VLoss: 0.0191 | PLoss: -0.0066 | Ent: 0.1396
+[ 20.3%] Iter 1980/9765 | SPS:  1404 | Reward:  0.067 | Value:  0.697 | VLoss: 0.0201 | PLoss: -0.0035 | Ent: 0.1396
+[ 20.3%] Iter 1981/9765 | SPS:  1405 | Reward:  0.045 | Value:  0.676 | VLoss: 0.0178 | PLoss: -0.0102 | Ent: 0.1124
+[ 20.3%] Iter 1982/9765 | SPS:  1405 | Reward:  0.066 | Value:  0.675 | VLoss: 0.0223 | PLoss: -0.0064 | Ent: 0.1160
+[ 20.3%] Iter 1983/9765 | SPS:  1406 | Reward:  0.043 | Value:  0.667 | VLoss: 0.0192 | PLoss: -0.0066 | Ent: 0.1373
+[ 20.3%] Iter 1984/9765 | SPS:  1406 | Reward:  0.058 | Value:  0.666 | VLoss: 0.0284 | PLoss: -0.0090 | Ent: 0.1344
+[ 20.3%] Iter 1985/9765 | SPS:  1407 | Reward:  0.071 | Value:  0.640 | VLoss: 0.0362 | PLoss: -0.0047 | Ent: 0.1453
+[ 20.3%] Iter 1986/9765 | SPS:  1407 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0217 | PLoss: -0.0116 | Ent: 0.1443
+[ 20.3%] Iter 1987/9765 | SPS:  1407 | Reward:  0.069 | Value:  0.684 | VLoss: 0.0260 | PLoss: -0.0016 | Ent: 0.1543
+[ 20.4%] Iter 1988/9765 | SPS:  1408 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0329 | PLoss: -0.0087 | Ent: 0.1442
+[ 20.4%] Iter 1989/9765 | SPS:  1408 | Reward:  0.065 | Value:  0.698 | VLoss: 0.0306 | PLoss: -0.0066 | Ent: 0.1766
+[ 20.4%] Iter 1990/9765 | SPS:  1409 | Reward:  0.049 | Value:  0.731 | VLoss: 0.0199 | PLoss: -0.0111 | Ent: 0.1735
+[ 20.4%] Iter 1991/9765 | SPS:  1409 | Reward:  0.065 | Value:  0.646 | VLoss: 0.0197 | PLoss: -0.0053 | Ent: 0.1787
+[ 20.4%] Iter 1992/9765 | SPS:  1410 | Reward:  0.051 | Value:  0.639 | VLoss: 0.0241 | PLoss: -0.0107 | Ent: 0.1901
+[ 20.4%] Iter 1993/9765 | SPS:  1410 | Reward:  0.063 | Value:  0.626 | VLoss: 0.0205 | PLoss: -0.0026 | Ent: 0.1538
+[ 20.4%] Iter 1994/9765 | SPS:  1411 | Reward:  0.058 | Value:  0.664 | VLoss: 0.0213 | PLoss: -0.0084 | Ent: 0.1664
+[ 20.4%] Iter 1995/9765 | SPS:  1411 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0340 | PLoss: -0.0077 | Ent: 0.1771
+[ 20.4%] Iter 1996/9765 | SPS:  1411 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0235 | PLoss: -0.0177 | Ent: 0.1612
+[ 20.5%] Iter 1997/9765 | SPS:  1412 | Reward:  0.053 | Value:  0.659 | VLoss: 0.0255 | PLoss: -0.0049 | Ent: 0.1679
+[ 20.5%] Iter 1998/9765 | SPS:  1412 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0244 | PLoss: -0.0126 | Ent: 0.1586
+[ 20.5%] Iter 1999/9765 | SPS:  1413 | Reward:  0.072 | Value:  0.707 | VLoss: 0.0243 | PLoss: -0.0149 | Ent: 0.1511
+[ 20.5%] Iter 2000/9765 | SPS:  1413 | Reward:  0.064 | Value:  0.739 | VLoss: 0.0350 | PLoss: -0.0090 | Ent: 0.1600
+[ 20.5%] Iter 2001/9765 | SPS:  1414 | Reward:  0.073 | Value:  0.759 | VLoss: 0.0273 | PLoss: -0.0108 | Ent: 0.1534
+[ 20.5%] Iter 2002/9765 | SPS:  1414 | Reward:  0.066 | Value:  0.737 | VLoss: 0.0266 | PLoss: -0.0077 | Ent: 0.1472
+[ 20.5%] Iter 2003/9765 | SPS:  1414 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0291 | PLoss: -0.0113 | Ent: 0.1492
+[ 20.5%] Iter 2004/9765 | SPS:  1415 | Reward:  0.051 | Value:  0.683 | VLoss: 0.0270 | PLoss: -0.0148 | Ent: 0.1610
+[ 20.5%] Iter 2005/9765 | SPS:  1415 | Reward:  0.052 | Value:  0.647 | VLoss: 0.0206 | PLoss: -0.0148 | Ent: 0.1510
+[ 20.5%] Iter 2006/9765 | SPS:  1416 | Reward:  0.068 | Value:  0.673 | VLoss: 0.0360 | PLoss: -0.0068 | Ent: 0.1465
+[ 20.6%] Iter 2007/9765 | SPS:  1416 | Reward:  0.068 | Value:  0.682 | VLoss: 0.0326 | PLoss: -0.0064 | Ent: 0.1765
+[ 20.6%] Iter 2008/9765 | SPS:  1417 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0280 | PLoss: -0.0050 | Ent: 0.1572
+[ 20.6%] Iter 2009/9765 | SPS:  1417 | Reward:  0.066 | Value:  0.648 | VLoss: 0.0291 | PLoss: -0.0040 | Ent: 0.1568
+[ 20.6%] Iter 2010/9765 | SPS:  1418 | Reward:  0.073 | Value:  0.723 | VLoss: 0.0240 | PLoss: -0.0007 | Ent: 0.1400
+[ 20.6%] Iter 2011/9765 | SPS:  1418 | Reward:  0.066 | Value:  0.689 | VLoss: 0.0249 | PLoss: -0.0095 | Ent: 0.1610
+[ 20.6%] Iter 2012/9765 | SPS:  1418 | Reward:  0.073 | Value:  0.709 | VLoss: 0.0247 | PLoss: -0.0173 | Ent: 0.1593
+[ 20.6%] Iter 2013/9765 | SPS:  1419 | Reward:  0.076 | Value:  0.750 | VLoss: 0.0283 | PLoss: -0.0106 | Ent: 0.1649
+[ 20.6%] Iter 2014/9765 | SPS:  1419 | Reward:  0.072 | Value:  0.703 | VLoss: 0.0248 | PLoss: -0.0086 | Ent: 0.1566
+[ 20.6%] Iter 2015/9765 | SPS:  1420 | Reward:  0.056 | Value:  0.699 | VLoss: 0.0183 | PLoss: -0.0008 | Ent: 0.1706
+[ 20.6%] Iter 2016/9765 | SPS:  1420 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0249 | PLoss: -0.0039 | Ent: 0.1030
+[ 20.7%] Iter 2017/9765 | SPS:  1421 | Reward:  0.058 | Value:  0.698 | VLoss: 0.0289 | PLoss: -0.0054 | Ent: 0.1265
+[ 20.7%] Iter 2018/9765 | SPS:  1421 | Reward:  0.063 | Value:  0.657 | VLoss: 0.0333 | PLoss: -0.0072 | Ent: 0.1415
+[ 20.7%] Iter 2019/9765 | SPS:  1421 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0222 | PLoss: -0.0076 | Ent: 0.1694
+[ 20.7%] Iter 2020/9765 | SPS:  1422 | Reward:  0.061 | Value:  0.701 | VLoss: 0.0216 | PLoss: -0.0072 | Ent: 0.1372
+[ 20.7%] Iter 2021/9765 | SPS:  1422 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0370 | PLoss: -0.0096 | Ent: 0.1601
+[ 20.7%] Iter 2022/9765 | SPS:  1423 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0318 | PLoss: -0.0100 | Ent: 0.1358
+[ 20.7%] Iter 2023/9765 | SPS:  1423 | Reward:  0.050 | Value:  0.740 | VLoss: 0.0305 | PLoss: -0.0056 | Ent: 0.1472
+[ 20.7%] Iter 2024/9765 | SPS:  1424 | Reward:  0.057 | Value:  0.691 | VLoss: 0.0229 | PLoss: -0.0071 | Ent: 0.1650
+[ 20.7%] Iter 2025/9765 | SPS:  1424 | Reward:  0.060 | Value:  0.633 | VLoss: 0.0252 | PLoss: -0.0014 | Ent: 0.1224
+[ 20.7%] Iter 2026/9765 | SPS:  1424 | Reward:  0.062 | Value:  0.621 | VLoss: 0.0359 | PLoss: -0.0096 | Ent: 0.1205
+[ 20.8%] Iter 2027/9765 | SPS:  1425 | Reward:  0.064 | Value:  0.651 | VLoss: 0.0354 | PLoss: -0.0077 | Ent: 0.1743
+[ 20.8%] Iter 2028/9765 | SPS:  1425 | Reward:  0.069 | Value:  0.679 | VLoss: 0.0302 | PLoss: -0.0048 | Ent: 0.1235
+[ 20.8%] Iter 2029/9765 | SPS:  1426 | Reward:  0.073 | Value:  0.656 | VLoss: 0.0361 | PLoss: -0.0146 | Ent: 0.2032
+[ 20.8%] Iter 2030/9765 | SPS:  1426 | Reward:  0.059 | Value:  0.691 | VLoss: 0.0289 | PLoss: -0.0122 | Ent: 0.1548
+[ 20.8%] Iter 2031/9765 | SPS:  1427 | Reward:  0.070 | Value:  0.640 | VLoss: 0.0326 | PLoss: -0.0098 | Ent: 0.1549
+[ 20.8%] Iter 2032/9765 | SPS:  1427 | Reward:  0.063 | Value:  0.679 | VLoss: 0.0312 | PLoss: -0.0041 | Ent: 0.2051
+[ 20.8%] Iter 2033/9765 | SPS:  1427 | Reward:  0.067 | Value:  0.678 | VLoss: 0.0273 | PLoss: -0.0102 | Ent: 0.1728
+[ 20.8%] Iter 2034/9765 | SPS:  1428 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0334 | PLoss: -0.0082 | Ent: 0.1366
+[ 20.8%] Iter 2035/9765 | SPS:  1428 | Reward:  0.061 | Value:  0.715 | VLoss: 0.0186 | PLoss: -0.0164 | Ent: 0.1570
+[ 20.8%] Iter 2036/9765 | SPS:  1429 | Reward:  0.058 | Value:  0.687 | VLoss: 0.0243 | PLoss: -0.0040 | Ent: 0.1606
+[ 20.9%] Iter 2037/9765 | SPS:  1429 | Reward:  0.070 | Value:  0.625 | VLoss: 0.0382 | PLoss: -0.0047 | Ent: 0.1193
+[ 20.9%] Iter 2038/9765 | SPS:  1429 | Reward:  0.065 | Value:  0.675 | VLoss: 0.0331 | PLoss: -0.0046 | Ent: 0.1722
+[ 20.9%] Iter 2039/9765 | SPS:  1430 | Reward:  0.074 | Value:  0.689 | VLoss: 0.0332 | PLoss: -0.0045 | Ent: 0.1089
+[ 20.9%] Iter 2040/9765 | SPS:  1430 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0275 | PLoss: -0.0126 | Ent: 0.1571
+[ 20.9%] Iter 2041/9765 | SPS:  1431 | Reward:  0.058 | Value:  0.683 | VLoss: 0.0313 | PLoss: -0.0087 | Ent: 0.1368
+[ 20.9%] Iter 2042/9765 | SPS:  1431 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0357 | PLoss: 0.0003 | Ent: 0.1457
+[ 20.9%] Iter 2043/9765 | SPS:  1432 | Reward:  0.057 | Value:  0.637 | VLoss: 0.0300 | PLoss: -0.0006 | Ent: 0.1862
+[ 20.9%] Iter 2044/9765 | SPS:  1432 | Reward:  0.056 | Value:  0.646 | VLoss: 0.0211 | PLoss: -0.0042 | Ent: 0.1741
+[ 20.9%] Iter 2045/9765 | SPS:  1432 | Reward:  0.053 | Value:  0.647 | VLoss: 0.0291 | PLoss: -0.0130 | Ent: 0.1503
+[ 21.0%] Iter 2046/9765 | SPS:  1433 | Reward:  0.056 | Value:  0.676 | VLoss: 0.0223 | PLoss: -0.0072 | Ent: 0.1355
+[ 21.0%] Iter 2047/9765 | SPS:  1433 | Reward:  0.064 | Value:  0.663 | VLoss: 0.0285 | PLoss: -0.0100 | Ent: 0.1617
+[ 21.0%] Iter 2048/9765 | SPS:  1434 | Reward:  0.060 | Value:  0.605 | VLoss: 0.0247 | PLoss: -0.0005 | Ent: 0.1984
+[ 21.0%] Iter 2049/9765 | SPS:  1434 | Reward:  0.056 | Value:  0.672 | VLoss: 0.0255 | PLoss: -0.0081 | Ent: 0.1648
+[ 21.0%] Iter 2050/9765 | SPS:  1435 | Reward:  0.066 | Value:  0.645 | VLoss: 0.0269 | PLoss: -0.0086 | Ent: 0.1834
+[ 21.0%] Iter 2051/9765 | SPS:  1435 | Reward:  0.070 | Value:  0.654 | VLoss: 0.0298 | PLoss: -0.0062 | Ent: 0.1428
+[ 21.0%] Iter 2052/9765 | SPS:  1435 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0251 | PLoss: -0.0038 | Ent: 0.1662
+[ 21.0%] Iter 2053/9765 | SPS:  1436 | Reward:  0.072 | Value:  0.641 | VLoss: 0.0329 | PLoss: -0.0130 | Ent: 0.1513
+[ 21.0%] Iter 2054/9765 | SPS:  1436 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0324 | PLoss: -0.0129 | Ent: 0.1546
+[ 21.0%] Iter 2055/9765 | SPS:  1437 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0238 | PLoss: -0.0076 | Ent: 0.1576
+[ 21.1%] Iter 2056/9765 | SPS:  1437 | Reward:  0.066 | Value:  0.667 | VLoss: 0.0285 | PLoss: -0.0045 | Ent: 0.1821
+[ 21.1%] Iter 2057/9765 | SPS:  1437 | Reward:  0.070 | Value:  0.664 | VLoss: 0.0230 | PLoss: -0.0013 | Ent: 0.1427
+[ 21.1%] Iter 2058/9765 | SPS:  1438 | Reward:  0.063 | Value:  0.694 | VLoss: 0.0272 | PLoss: -0.0139 | Ent: 0.1683
+[ 21.1%] Iter 2059/9765 | SPS:  1438 | Reward:  0.068 | Value:  0.747 | VLoss: 0.0320 | PLoss: -0.0055 | Ent: 0.1895
+[ 21.1%] Iter 2060/9765 | SPS:  1439 | Reward:  0.057 | Value:  0.688 | VLoss: 0.0239 | PLoss: -0.0082 | Ent: 0.1752
+[ 21.1%] Iter 2061/9765 | SPS:  1439 | Reward:  0.065 | Value:  0.629 | VLoss: 0.0329 | PLoss: -0.0107 | Ent: 0.1636
+[ 21.1%] Iter 2062/9765 | SPS:  1440 | Reward:  0.066 | Value:  0.659 | VLoss: 0.0303 | PLoss: -0.0089 | Ent: 0.1401
+[ 21.1%] Iter 2063/9765 | SPS:  1440 | Reward:  0.062 | Value:  0.639 | VLoss: 0.0263 | PLoss: -0.0077 | Ent: 0.1288
+[ 21.1%] Iter 2064/9765 | SPS:  1440 | Reward:  0.046 | Value:  0.618 | VLoss: 0.0204 | PLoss: -0.0100 | Ent: 0.2028
+[ 21.1%] Iter 2065/9765 | SPS:  1441 | Reward:  0.069 | Value:  0.618 | VLoss: 0.0438 | PLoss: -0.0123 | Ent: 0.2037
+[ 21.2%] Iter 2066/9765 | SPS:  1441 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0278 | PLoss: -0.0043 | Ent: 0.1622
+[ 21.2%] Iter 2067/9765 | SPS:  1442 | Reward:  0.061 | Value:  0.642 | VLoss: 0.0305 | PLoss: -0.0094 | Ent: 0.1972
+[ 21.2%] Iter 2068/9765 | SPS:  1442 | Reward:  0.065 | Value:  0.635 | VLoss: 0.0278 | PLoss: -0.0016 | Ent: 0.1597
+[ 21.2%] Iter 2069/9765 | SPS:  1443 | Reward:  0.053 | Value:  0.597 | VLoss: 0.0252 | PLoss: -0.0083 | Ent: 0.1445
+[ 21.2%] Iter 2070/9765 | SPS:  1443 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0285 | PLoss: -0.0027 | Ent: 0.1017
+[ 21.2%] Iter 2071/9765 | SPS:  1443 | Reward:  0.049 | Value:  0.577 | VLoss: 0.0184 | PLoss: -0.0091 | Ent: 0.1145
+[ 21.2%] Iter 2072/9765 | SPS:  1444 | Reward:  0.055 | Value:  0.660 | VLoss: 0.0187 | PLoss: 0.0045 | Ent: 0.1695
+[ 21.2%] Iter 2073/9765 | SPS:  1444 | Reward:  0.059 | Value:  0.593 | VLoss: 0.0257 | PLoss: -0.0109 | Ent: 0.1552
+[ 21.2%] Iter 2074/9765 | SPS:  1445 | Reward:  0.049 | Value:  0.633 | VLoss: 0.0276 | PLoss: -0.0109 | Ent: 0.1313
+[ 21.2%] Iter 2075/9765 | SPS:  1445 | Reward:  0.051 | Value:  0.623 | VLoss: 0.0222 | PLoss: -0.0088 | Ent: 0.1293
+[ 21.3%] Iter 2076/9765 | SPS:  1446 | Reward:  0.042 | Value:  0.594 | VLoss: 0.0276 | PLoss: -0.0023 | Ent: 0.1668
+[ 21.3%] Iter 2077/9765 | SPS:  1446 | Reward:  0.056 | Value:  0.593 | VLoss: 0.0193 | PLoss: 0.0007 | Ent: 0.1573
+[ 21.3%] Iter 2078/9765 | SPS:  1446 | Reward:  0.055 | Value:  0.590 | VLoss: 0.0265 | PLoss: -0.0112 | Ent: 0.1612
+[ 21.3%] Iter 2079/9765 | SPS:  1447 | Reward:  0.046 | Value:  0.602 | VLoss: 0.0213 | PLoss: -0.0168 | Ent: 0.1593
+[ 21.3%] Iter 2080/9765 | SPS:  1447 | Reward:  0.032 | Value:  0.558 | VLoss: 0.0276 | PLoss: -0.0053 | Ent: 0.1459
+[ 21.3%] Iter 2081/9765 | SPS:  1448 | Reward:  0.058 | Value:  0.565 | VLoss: 0.0226 | PLoss: -0.0061 | Ent: 0.1303
+[ 21.3%] Iter 2082/9765 | SPS:  1448 | Reward:  0.079 | Value:  0.568 | VLoss: 0.0295 | PLoss: -0.0103 | Ent: 0.1612
+[ 21.3%] Iter 2083/9765 | SPS:  1449 | Reward:  0.056 | Value:  0.613 | VLoss: 0.0226 | PLoss: -0.0060 | Ent: 0.1606
+[ 21.3%] Iter 2084/9765 | SPS:  1449 | Reward:  0.061 | Value:  0.626 | VLoss: 0.0325 | PLoss: -0.0058 | Ent: 0.1384
+[ 21.4%] Iter 2085/9765 | SPS:  1449 | Reward:  0.059 | Value:  0.706 | VLoss: 0.0282 | PLoss: -0.0025 | Ent: 0.1093
+[ 21.4%] Iter 2086/9765 | SPS:  1450 | Reward:  0.051 | Value:  0.695 | VLoss: 0.0265 | PLoss: -0.0078 | Ent: 0.1566
+[ 21.4%] Iter 2087/9765 | SPS:  1450 | Reward:  0.053 | Value:  0.664 | VLoss: 0.0262 | PLoss: -0.0052 | Ent: 0.1927
+[ 21.4%] Iter 2088/9765 | SPS:  1451 | Reward:  0.067 | Value:  0.705 | VLoss: 0.0235 | PLoss: -0.0115 | Ent: 0.1707
+[ 21.4%] Iter 2089/9765 | SPS:  1451 | Reward:  0.072 | Value:  0.694 | VLoss: 0.0338 | PLoss: -0.0092 | Ent: 0.1813
+[ 21.4%] Iter 2090/9765 | SPS:  1451 | Reward:  0.066 | Value:  0.665 | VLoss: 0.0353 | PLoss: -0.0074 | Ent: 0.1699
+[ 21.4%] Iter 2091/9765 | SPS:  1452 | Reward:  0.070 | Value:  0.647 | VLoss: 0.0257 | PLoss: -0.0116 | Ent: 0.1681
+[ 21.4%] Iter 2092/9765 | SPS:  1452 | Reward:  0.072 | Value:  0.692 | VLoss: 0.0270 | PLoss: -0.0082 | Ent: 0.1710
+[ 21.4%] Iter 2093/9765 | SPS:  1453 | Reward:  0.056 | Value:  0.659 | VLoss: 0.0319 | PLoss: -0.0021 | Ent: 0.1470
+[ 21.4%] Iter 2094/9765 | SPS:  1453 | Reward:  0.059 | Value:  0.638 | VLoss: 0.0246 | PLoss: -0.0102 | Ent: 0.1729
+[ 21.5%] Iter 2095/9765 | SPS:  1454 | Reward:  0.077 | Value:  0.676 | VLoss: 0.0316 | PLoss: -0.0167 | Ent: 0.1694
+[ 21.5%] Iter 2096/9765 | SPS:  1454 | Reward:  0.067 | Value:  0.656 | VLoss: 0.0334 | PLoss: 0.0001 | Ent: 0.1394
+[ 21.5%] Iter 2097/9765 | SPS:  1454 | Reward:  0.068 | Value:  0.709 | VLoss: 0.0224 | PLoss: -0.0098 | Ent: 0.1329
+[ 21.5%] Iter 2098/9765 | SPS:  1455 | Reward:  0.057 | Value:  0.672 | VLoss: 0.0260 | PLoss: -0.0088 | Ent: 0.1767
+[ 21.5%] Iter 2099/9765 | SPS:  1455 | Reward:  0.056 | Value:  0.761 | VLoss: 0.0201 | PLoss: -0.0085 | Ent: 0.1588
+[ 21.5%] Iter 2100/9765 | SPS:  1456 | Reward:  0.062 | Value:  0.689 | VLoss: 0.0252 | PLoss: -0.0026 | Ent: 0.1828
+[ 21.5%] Iter 2101/9765 | SPS:  1456 | Reward:  0.064 | Value:  0.698 | VLoss: 0.0237 | PLoss: -0.0038 | Ent: 0.1513
+[ 21.5%] Iter 2102/9765 | SPS:  1456 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0329 | PLoss: -0.0070 | Ent: 0.1696
+[ 21.5%] Iter 2103/9765 | SPS:  1457 | Reward:  0.058 | Value:  0.730 | VLoss: 0.0277 | PLoss: -0.0047 | Ent: 0.1506
+[ 21.5%] Iter 2104/9765 | SPS:  1457 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0249 | PLoss: -0.0063 | Ent: 0.1450
+[ 21.6%] Iter 2105/9765 | SPS:  1458 | Reward:  0.055 | Value:  0.685 | VLoss: 0.0221 | PLoss: -0.0162 | Ent: 0.1797
+[ 21.6%] Iter 2106/9765 | SPS:  1458 | Reward:  0.072 | Value:  0.649 | VLoss: 0.0348 | PLoss: -0.0110 | Ent: 0.1657
+[ 21.6%] Iter 2107/9765 | SPS:  1459 | Reward:  0.055 | Value:  0.707 | VLoss: 0.0265 | PLoss: -0.0018 | Ent: 0.1201
+[ 21.6%] Iter 2108/9765 | SPS:  1459 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0265 | PLoss: -0.0062 | Ent: 0.1763
+[ 21.6%] Iter 2109/9765 | SPS:  1459 | Reward:  0.058 | Value:  0.770 | VLoss: 0.0215 | PLoss: -0.0087 | Ent: 0.1159
+[ 21.6%] Iter 2110/9765 | SPS:  1460 | Reward:  0.054 | Value:  0.655 | VLoss: 0.0305 | PLoss: -0.0152 | Ent: 0.1426
+[ 21.6%] Iter 2111/9765 | SPS:  1460 | Reward:  0.049 | Value:  0.681 | VLoss: 0.0205 | PLoss: -0.0113 | Ent: 0.1724
+[ 21.6%] Iter 2112/9765 | SPS:  1461 | Reward:  0.048 | Value:  0.675 | VLoss: 0.0218 | PLoss: -0.0038 | Ent: 0.1336
+[ 21.6%] Iter 2113/9765 | SPS:  1461 | Reward:  0.051 | Value:  0.625 | VLoss: 0.0186 | PLoss: -0.0034 | Ent: 0.1050
+[ 21.6%] Iter 2114/9765 | SPS:  1461 | Reward:  0.050 | Value:  0.662 | VLoss: 0.0342 | PLoss: -0.0102 | Ent: 0.1482
+[ 21.7%] Iter 2115/9765 | SPS:  1462 | Reward:  0.056 | Value:  0.610 | VLoss: 0.0329 | PLoss: -0.0112 | Ent: 0.1815
+[ 21.7%] Iter 2116/9765 | SPS:  1462 | Reward:  0.052 | Value:  0.583 | VLoss: 0.0280 | PLoss: -0.0080 | Ent: 0.1212
+[ 21.7%] Iter 2117/9765 | SPS:  1463 | Reward:  0.060 | Value:  0.590 | VLoss: 0.0207 | PLoss: -0.0091 | Ent: 0.1248
+[ 21.7%] Iter 2118/9765 | SPS:  1463 | Reward:  0.053 | Value:  0.649 | VLoss: 0.0230 | PLoss: -0.0064 | Ent: 0.1417
+[ 21.7%] Iter 2119/9765 | SPS:  1464 | Reward:  0.061 | Value:  0.626 | VLoss: 0.0321 | PLoss: -0.0120 | Ent: 0.1183
+[ 21.7%] Iter 2120/9765 | SPS:  1464 | Reward:  0.062 | Value:  0.685 | VLoss: 0.0252 | PLoss: -0.0139 | Ent: 0.1407
+[ 21.7%] Iter 2121/9765 | SPS:  1464 | Reward:  0.046 | Value:  0.625 | VLoss: 0.0144 | PLoss: -0.0033 | Ent: 0.1666
+[ 21.7%] Iter 2122/9765 | SPS:  1465 | Reward:  0.053 | Value:  0.559 | VLoss: 0.0266 | PLoss: -0.0043 | Ent: 0.1703
+[ 21.7%] Iter 2123/9765 | SPS:  1465 | Reward:  0.056 | Value:  0.591 | VLoss: 0.0274 | PLoss: -0.0058 | Ent: 0.1366
+[ 21.8%] Iter 2124/9765 | SPS:  1466 | Reward:  0.056 | Value:  0.656 | VLoss: 0.0256 | PLoss: -0.0148 | Ent: 0.1528
+[ 21.8%] Iter 2125/9765 | SPS:  1466 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0321 | PLoss: -0.0056 | Ent: 0.1451
+[ 21.8%] Iter 2126/9765 | SPS:  1467 | Reward:  0.052 | Value:  0.644 | VLoss: 0.0302 | PLoss: -0.0093 | Ent: 0.1317
+[ 21.8%] Iter 2127/9765 | SPS:  1467 | Reward:  0.038 | Value:  0.615 | VLoss: 0.0150 | PLoss: -0.0027 | Ent: 0.1168
+[ 21.8%] Iter 2128/9765 | SPS:  1467 | Reward:  0.046 | Value:  0.627 | VLoss: 0.0374 | PLoss: -0.0037 | Ent: 0.1899
+[ 21.8%] Iter 2129/9765 | SPS:  1468 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0283 | PLoss: -0.0107 | Ent: 0.1743
+[ 21.8%] Iter 2130/9765 | SPS:  1468 | Reward:  0.054 | Value:  0.688 | VLoss: 0.0288 | PLoss: -0.0100 | Ent: 0.1722
+[ 21.8%] Iter 2131/9765 | SPS:  1469 | Reward:  0.041 | Value:  0.692 | VLoss: 0.0178 | PLoss: -0.0052 | Ent: 0.1418
+[ 21.8%] Iter 2132/9765 | SPS:  1469 | Reward:  0.052 | Value:  0.668 | VLoss: 0.0217 | PLoss: -0.0035 | Ent: 0.1374
+[ 21.8%] Iter 2133/9765 | SPS:  1469 | Reward:  0.045 | Value:  0.651 | VLoss: 0.0152 | PLoss: -0.0118 | Ent: 0.1687
+[ 21.9%] Iter 2134/9765 | SPS:  1470 | Reward:  0.053 | Value:  0.624 | VLoss: 0.0262 | PLoss: -0.0073 | Ent: 0.1192
+[ 21.9%] Iter 2135/9765 | SPS:  1470 | Reward:  0.071 | Value:  0.670 | VLoss: 0.0320 | PLoss: -0.0071 | Ent: 0.1219
+[ 21.9%] Iter 2136/9765 | SPS:  1471 | Reward:  0.071 | Value:  0.654 | VLoss: 0.0340 | PLoss: -0.0086 | Ent: 0.1771
+[ 21.9%] Iter 2137/9765 | SPS:  1471 | Reward:  0.058 | Value:  0.671 | VLoss: 0.0247 | PLoss: -0.0055 | Ent: 0.1317
+[ 21.9%] Iter 2138/9765 | SPS:  1472 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0356 | PLoss: -0.0068 | Ent: 0.1290
+[ 21.9%] Iter 2139/9765 | SPS:  1472 | Reward:  0.048 | Value:  0.643 | VLoss: 0.0353 | PLoss: -0.0139 | Ent: 0.1586
+[ 21.9%] Iter 2140/9765 | SPS:  1472 | Reward:  0.058 | Value:  0.658 | VLoss: 0.0295 | PLoss: -0.0112 | Ent: 0.1528
+[ 21.9%] Iter 2141/9765 | SPS:  1473 | Reward:  0.053 | Value:  0.652 | VLoss: 0.0201 | PLoss: -0.0056 | Ent: 0.1368
+[ 21.9%] Iter 2142/9765 | SPS:  1473 | Reward:  0.043 | Value:  0.750 | VLoss: 0.0244 | PLoss: -0.0038 | Ent: 0.1280
+[ 21.9%] Iter 2143/9765 | SPS:  1474 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0256 | PLoss: -0.0067 | Ent: 0.1321
+[ 22.0%] Iter 2144/9765 | SPS:  1474 | Reward:  0.054 | Value:  0.635 | VLoss: 0.0240 | PLoss: -0.0109 | Ent: 0.1569
+[ 22.0%] Iter 2145/9765 | SPS:  1474 | Reward:  0.057 | Value:  0.694 | VLoss: 0.0209 | PLoss: -0.0139 | Ent: 0.1072
+[ 22.0%] Iter 2146/9765 | SPS:  1475 | Reward:  0.060 | Value:  0.684 | VLoss: 0.0246 | PLoss: -0.0026 | Ent: 0.1769
+[ 22.0%] Iter 2147/9765 | SPS:  1475 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0246 | PLoss: -0.0094 | Ent: 0.1508
+[ 22.0%] Iter 2148/9765 | SPS:  1476 | Reward:  0.059 | Value:  0.588 | VLoss: 0.0271 | PLoss: -0.0054 | Ent: 0.1195
+[ 22.0%] Iter 2149/9765 | SPS:  1476 | Reward:  0.055 | Value:  0.655 | VLoss: 0.0335 | PLoss: -0.0093 | Ent: 0.1630
+[ 22.0%] Iter 2150/9765 | SPS:  1476 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0258 | PLoss: -0.0032 | Ent: 0.1699
+[ 22.0%] Iter 2151/9765 | SPS:  1477 | Reward:  0.064 | Value:  0.681 | VLoss: 0.0304 | PLoss: -0.0073 | Ent: 0.1642
+[ 22.0%] Iter 2152/9765 | SPS:  1477 | Reward:  0.067 | Value:  0.661 | VLoss: 0.0305 | PLoss: -0.0027 | Ent: 0.1200
+[ 22.0%] Iter 2153/9765 | SPS:  1478 | Reward:  0.065 | Value:  0.706 | VLoss: 0.0227 | PLoss: -0.0074 | Ent: 0.1639
+[ 22.1%] Iter 2154/9765 | SPS:  1478 | Reward:  0.067 | Value:  0.663 | VLoss: 0.0268 | PLoss: -0.0056 | Ent: 0.1814
+[ 22.1%] Iter 2155/9765 | SPS:  1479 | Reward:  0.074 | Value:  0.666 | VLoss: 0.0337 | PLoss: -0.0119 | Ent: 0.1744
+[ 22.1%] Iter 2156/9765 | SPS:  1479 | Reward:  0.053 | Value:  0.672 | VLoss: 0.0270 | PLoss: -0.0063 | Ent: 0.1452
+[ 22.1%] Iter 2157/9765 | SPS:  1479 | Reward:  0.074 | Value:  0.662 | VLoss: 0.0275 | PLoss: -0.0053 | Ent: 0.1400
+[ 22.1%] Iter 2158/9765 | SPS:  1480 | Reward:  0.060 | Value:  0.684 | VLoss: 0.0387 | PLoss: -0.0086 | Ent: 0.1517
+[ 22.1%] Iter 2159/9765 | SPS:  1480 | Reward:  0.056 | Value:  0.716 | VLoss: 0.0186 | PLoss: -0.0145 | Ent: 0.1786
+[ 22.1%] Iter 2160/9765 | SPS:  1481 | Reward:  0.060 | Value:  0.668 | VLoss: 0.0222 | PLoss: -0.0060 | Ent: 0.1243
+[ 22.1%] Iter 2161/9765 | SPS:  1481 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0284 | PLoss: -0.0064 | Ent: 0.0982
+[ 22.1%] Iter 2162/9765 | SPS:  1481 | Reward:  0.058 | Value:  0.698 | VLoss: 0.0282 | PLoss: -0.0078 | Ent: 0.1639
+[ 22.2%] Iter 2163/9765 | SPS:  1482 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0366 | PLoss: -0.0026 | Ent: 0.1170
+[ 22.2%] Iter 2164/9765 | SPS:  1482 | Reward:  0.041 | Value:  0.667 | VLoss: 0.0269 | PLoss: -0.0008 | Ent: 0.1281
+[ 22.2%] Iter 2165/9765 | SPS:  1483 | Reward:  0.059 | Value:  0.639 | VLoss: 0.0236 | PLoss: -0.0138 | Ent: 0.1408
+[ 22.2%] Iter 2166/9765 | SPS:  1483 | Reward:  0.067 | Value:  0.681 | VLoss: 0.0330 | PLoss: -0.0075 | Ent: 0.1488
+[ 22.2%] Iter 2167/9765 | SPS:  1483 | Reward:  0.062 | Value:  0.658 | VLoss: 0.0315 | PLoss: -0.0084 | Ent: 0.1540
+[ 22.2%] Iter 2168/9765 | SPS:  1484 | Reward:  0.071 | Value:  0.673 | VLoss: 0.0309 | PLoss: -0.0113 | Ent: 0.1912
+[ 22.2%] Iter 2169/9765 | SPS:  1484 | Reward:  0.051 | Value:  0.661 | VLoss: 0.0343 | PLoss: -0.0052 | Ent: 0.1505
+[ 22.2%] Iter 2170/9765 | SPS:  1485 | Reward:  0.067 | Value:  0.677 | VLoss: 0.0305 | PLoss: -0.0135 | Ent: 0.1580
+[ 22.2%] Iter 2171/9765 | SPS:  1485 | Reward:  0.048 | Value:  0.695 | VLoss: 0.0262 | PLoss: -0.0041 | Ent: 0.1197
+[ 22.2%] Iter 2172/9765 | SPS:  1485 | Reward:  0.079 | Value:  0.655 | VLoss: 0.0330 | PLoss: -0.0089 | Ent: 0.1587
+[ 22.3%] Iter 2173/9765 | SPS:  1486 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0264 | PLoss: -0.0104 | Ent: 0.1466
+[ 22.3%] Iter 2174/9765 | SPS:  1486 | Reward:  0.068 | Value:  0.646 | VLoss: 0.0290 | PLoss: -0.0090 | Ent: 0.1721
+[ 22.3%] Iter 2175/9765 | SPS:  1487 | Reward:  0.060 | Value:  0.657 | VLoss: 0.0257 | PLoss: -0.0126 | Ent: 0.1330
+[ 22.3%] Iter 2176/9765 | SPS:  1487 | Reward:  0.068 | Value:  0.698 | VLoss: 0.0335 | PLoss: -0.0049 | Ent: 0.1396
+[ 22.3%] Iter 2177/9765 | SPS:  1487 | Reward:  0.053 | Value:  0.618 | VLoss: 0.0323 | PLoss: -0.0045 | Ent: 0.1451
+[ 22.3%] Iter 2178/9765 | SPS:  1488 | Reward:  0.045 | Value:  0.619 | VLoss: 0.0225 | PLoss: -0.0035 | Ent: 0.1426
+[ 22.3%] Iter 2179/9765 | SPS:  1488 | Reward:  0.061 | Value:  0.642 | VLoss: 0.0299 | PLoss: -0.0067 | Ent: 0.1353
+[ 22.3%] Iter 2180/9765 | SPS:  1489 | Reward:  0.057 | Value:  0.672 | VLoss: 0.0266 | PLoss: -0.0102 | Ent: 0.1553
+[ 22.3%] Iter 2181/9765 | SPS:  1489 | Reward:  0.069 | Value:  0.640 | VLoss: 0.0345 | PLoss: -0.0122 | Ent: 0.1549
+[ 22.3%] Iter 2182/9765 | SPS:  1489 | Reward:  0.060 | Value:  0.680 | VLoss: 0.0273 | PLoss: -0.0078 | Ent: 0.1211
+[ 22.4%] Iter 2183/9765 | SPS:  1490 | Reward:  0.053 | Value:  0.672 | VLoss: 0.0266 | PLoss: -0.0070 | Ent: 0.1685
+[ 22.4%] Iter 2184/9765 | SPS:  1490 | Reward:  0.065 | Value:  0.720 | VLoss: 0.0229 | PLoss: -0.0078 | Ent: 0.1525
+[ 22.4%] Iter 2185/9765 | SPS:  1491 | Reward:  0.075 | Value:  0.654 | VLoss: 0.0267 | PLoss: -0.0062 | Ent: 0.1161
+[ 22.4%] Iter 2186/9765 | SPS:  1491 | Reward:  0.068 | Value:  0.634 | VLoss: 0.0276 | PLoss: -0.0024 | Ent: 0.1495
+[ 22.4%] Iter 2187/9765 | SPS:  1491 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0355 | PLoss: -0.0036 | Ent: 0.1617
+[ 22.4%] Iter 2188/9765 | SPS:  1492 | Reward:  0.062 | Value:  0.613 | VLoss: 0.0288 | PLoss: -0.0064 | Ent: 0.1643
+[ 22.4%] Iter 2189/9765 | SPS:  1492 | Reward:  0.077 | Value:  0.663 | VLoss: 0.0318 | PLoss: -0.0044 | Ent: 0.1591
+[ 22.4%] Iter 2190/9765 | SPS:  1493 | Reward:  0.060 | Value:  0.643 | VLoss: 0.0316 | PLoss: -0.0035 | Ent: 0.1444
+[ 22.4%] Iter 2191/9765 | SPS:  1493 | Reward:  0.073 | Value:  0.697 | VLoss: 0.0285 | PLoss: -0.0001 | Ent: 0.1458
+[ 22.4%] Iter 2192/9765 | SPS:  1493 | Reward:  0.062 | Value:  0.610 | VLoss: 0.0416 | PLoss: -0.0107 | Ent: 0.1619
+[ 22.5%] Iter 2193/9765 | SPS:  1494 | Reward:  0.051 | Value:  0.687 | VLoss: 0.0268 | PLoss: -0.0005 | Ent: 0.1832
+[ 22.5%] Iter 2194/9765 | SPS:  1494 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0342 | PLoss: -0.0108 | Ent: 0.1209
+[ 22.5%] Iter 2195/9765 | SPS:  1495 | Reward:  0.059 | Value:  0.738 | VLoss: 0.0273 | PLoss: -0.0046 | Ent: 0.1536
+[ 22.5%] Iter 2196/9765 | SPS:  1495 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0346 | PLoss: -0.0115 | Ent: 0.1631
+[ 22.5%] Iter 2197/9765 | SPS:  1495 | Reward:  0.062 | Value:  0.645 | VLoss: 0.0385 | PLoss: -0.0094 | Ent: 0.1805
+[ 22.5%] Iter 2198/9765 | SPS:  1496 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0349 | PLoss: -0.0067 | Ent: 0.1497
+[ 22.5%] Iter 2199/9765 | SPS:  1496 | Reward:  0.049 | Value:  0.682 | VLoss: 0.0336 | PLoss: -0.0091 | Ent: 0.1317
+[ 22.5%] Iter 2200/9765 | SPS:  1497 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0249 | PLoss: -0.0059 | Ent: 0.1308
+[ 22.5%] Iter 2201/9765 | SPS:  1497 | Reward:  0.071 | Value:  0.655 | VLoss: 0.0368 | PLoss: -0.0095 | Ent: 0.1486
+[ 22.5%] Iter 2202/9765 | SPS:  1497 | Reward:  0.066 | Value:  0.641 | VLoss: 0.0327 | PLoss: -0.0054 | Ent: 0.1614
+[ 22.6%] Iter 2203/9765 | SPS:  1498 | Reward:  0.066 | Value:  0.707 | VLoss: 0.0229 | PLoss: -0.0058 | Ent: 0.1632
+[ 22.6%] Iter 2204/9765 | SPS:  1498 | Reward:  0.062 | Value:  0.614 | VLoss: 0.0335 | PLoss: -0.0097 | Ent: 0.1955
+[ 22.6%] Iter 2205/9765 | SPS:  1499 | Reward:  0.072 | Value:  0.600 | VLoss: 0.0310 | PLoss: -0.0083 | Ent: 0.1710
+[ 22.6%] Iter 2206/9765 | SPS:  1499 | Reward:  0.052 | Value:  0.648 | VLoss: 0.0236 | PLoss: -0.0072 | Ent: 0.1744
+[ 22.6%] Iter 2207/9765 | SPS:  1499 | Reward:  0.066 | Value:  0.664 | VLoss: 0.0229 | PLoss: -0.0136 | Ent: 0.1049
+[ 22.6%] Iter 2208/9765 | SPS:  1500 | Reward:  0.068 | Value:  0.683 | VLoss: 0.0251 | PLoss: -0.0066 | Ent: 0.1273
+[ 22.6%] Iter 2209/9765 | SPS:  1500 | Reward:  0.055 | Value:  0.634 | VLoss: 0.0334 | PLoss: -0.0093 | Ent: 0.1732
+[ 22.6%] Iter 2210/9765 | SPS:  1501 | Reward:  0.056 | Value:  0.660 | VLoss: 0.0287 | PLoss: -0.0097 | Ent: 0.1772
+[ 22.6%] Iter 2211/9765 | SPS:  1501 | Reward:  0.056 | Value:  0.620 | VLoss: 0.0271 | PLoss: -0.0072 | Ent: 0.1480
+[ 22.7%] Iter 2212/9765 | SPS:  1501 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0273 | PLoss: -0.0136 | Ent: 0.1482
+[ 22.7%] Iter 2213/9765 | SPS:  1502 | Reward:  0.070 | Value:  0.645 | VLoss: 0.0372 | PLoss: -0.0070 | Ent: 0.1808
+[ 22.7%] Iter 2214/9765 | SPS:  1502 | Reward:  0.077 | Value:  0.627 | VLoss: 0.0293 | PLoss: -0.0058 | Ent: 0.1835
+[ 22.7%] Iter 2215/9765 | SPS:  1503 | Reward:  0.055 | Value:  0.641 | VLoss: 0.0333 | PLoss: -0.0078 | Ent: 0.1759
+[ 22.7%] Iter 2216/9765 | SPS:  1503 | Reward:  0.070 | Value:  0.662 | VLoss: 0.0309 | PLoss: -0.0030 | Ent: 0.1683
+[ 22.7%] Iter 2217/9765 | SPS:  1503 | Reward:  0.072 | Value:  0.677 | VLoss: 0.0331 | PLoss: -0.0139 | Ent: 0.1771
+[ 22.7%] Iter 2218/9765 | SPS:  1504 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0282 | PLoss: -0.0096 | Ent: 0.1283
+[ 22.7%] Iter 2219/9765 | SPS:  1504 | Reward:  0.069 | Value:  0.665 | VLoss: 0.0249 | PLoss: -0.0043 | Ent: 0.1511
+[ 22.7%] Iter 2220/9765 | SPS:  1505 | Reward:  0.072 | Value:  0.711 | VLoss: 0.0307 | PLoss: -0.0079 | Ent: 0.1121
+[ 22.7%] Iter 2221/9765 | SPS:  1505 | Reward:  0.047 | Value:  0.671 | VLoss: 0.0158 | PLoss: -0.0012 | Ent: 0.1370
+[ 22.8%] Iter 2222/9765 | SPS:  1505 | Reward:  0.050 | Value:  0.690 | VLoss: 0.0248 | PLoss: -0.0107 | Ent: 0.1540
+[ 22.8%] Iter 2223/9765 | SPS:  1506 | Reward:  0.055 | Value:  0.672 | VLoss: 0.0175 | PLoss: -0.0062 | Ent: 0.1593
+[ 22.8%] Iter 2224/9765 | SPS:  1506 | Reward:  0.063 | Value:  0.636 | VLoss: 0.0321 | PLoss: -0.0007 | Ent: 0.1377
+[ 22.8%] Iter 2225/9765 | SPS:  1507 | Reward:  0.056 | Value:  0.653 | VLoss: 0.0413 | PLoss: -0.0127 | Ent: 0.1486
+[ 22.8%] Iter 2226/9765 | SPS:  1507 | Reward:  0.070 | Value:  0.627 | VLoss: 0.0310 | PLoss: -0.0122 | Ent: 0.1859
+[ 22.8%] Iter 2227/9765 | SPS:  1507 | Reward:  0.063 | Value:  0.666 | VLoss: 0.0253 | PLoss: -0.0102 | Ent: 0.1671
+[ 22.8%] Iter 2228/9765 | SPS:  1508 | Reward:  0.066 | Value:  0.635 | VLoss: 0.0342 | PLoss: -0.0036 | Ent: 0.1507
+[ 22.8%] Iter 2229/9765 | SPS:  1508 | Reward:  0.056 | Value:  0.629 | VLoss: 0.0311 | PLoss: -0.0035 | Ent: 0.1600
+[ 22.8%] Iter 2230/9765 | SPS:  1509 | Reward:  0.069 | Value:  0.638 | VLoss: 0.0253 | PLoss: -0.0108 | Ent: 0.1317
+[ 22.8%] Iter 2231/9765 | SPS:  1509 | Reward:  0.067 | Value:  0.663 | VLoss: 0.0308 | PLoss: -0.0025 | Ent: 0.1541
+[ 22.9%] Iter 2232/9765 | SPS:  1509 | Reward:  0.067 | Value:  0.683 | VLoss: 0.0273 | PLoss: -0.0141 | Ent: 0.1441
+[ 22.9%] Iter 2233/9765 | SPS:  1510 | Reward:  0.077 | Value:  0.659 | VLoss: 0.0307 | PLoss: -0.0101 | Ent: 0.1867
+[ 22.9%] Iter 2234/9765 | SPS:  1510 | Reward:  0.060 | Value:  0.697 | VLoss: 0.0236 | PLoss: -0.0137 | Ent: 0.1824
+[ 22.9%] Iter 2235/9765 | SPS:  1511 | Reward:  0.052 | Value:  0.662 | VLoss: 0.0265 | PLoss: -0.0032 | Ent: 0.1486
+[ 22.9%] Iter 2236/9765 | SPS:  1511 | Reward:  0.066 | Value:  0.644 | VLoss: 0.0393 | PLoss: -0.0093 | Ent: 0.1510
+[ 22.9%] Iter 2237/9765 | SPS:  1511 | Reward:  0.055 | Value:  0.640 | VLoss: 0.0281 | PLoss: -0.0082 | Ent: 0.1440
+[ 22.9%] Iter 2238/9765 | SPS:  1512 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0217 | PLoss: -0.0049 | Ent: 0.1553
+[ 22.9%] Iter 2239/9765 | SPS:  1512 | Reward:  0.072 | Value:  0.684 | VLoss: 0.0344 | PLoss: -0.0116 | Ent: 0.1769
+[ 22.9%] Iter 2240/9765 | SPS:  1513 | Reward:  0.063 | Value:  0.718 | VLoss: 0.0243 | PLoss: -0.0094 | Ent: 0.1465
+[ 22.9%] Iter 2241/9765 | SPS:  1513 | Reward:  0.063 | Value:  0.670 | VLoss: 0.0279 | PLoss: -0.0089 | Ent: 0.1594
+[ 23.0%] Iter 2242/9765 | SPS:  1513 | Reward:  0.071 | Value:  0.683 | VLoss: 0.0312 | PLoss: -0.0098 | Ent: 0.1533
+[ 23.0%] Iter 2243/9765 | SPS:  1514 | Reward:  0.063 | Value:  0.679 | VLoss: 0.0243 | PLoss: -0.0057 | Ent: 0.1298
+[ 23.0%] Iter 2244/9765 | SPS:  1514 | Reward:  0.057 | Value:  0.693 | VLoss: 0.0243 | PLoss: -0.0095 | Ent: 0.1815
+[ 23.0%] Iter 2245/9765 | SPS:  1514 | Reward:  0.069 | Value:  0.681 | VLoss: 0.0247 | PLoss: -0.0144 | Ent: 0.1654
+[ 23.0%] Iter 2246/9765 | SPS:  1515 | Reward:  0.066 | Value:  0.704 | VLoss: 0.0354 | PLoss: -0.0031 | Ent: 0.1567
+[ 23.0%] Iter 2247/9765 | SPS:  1515 | Reward:  0.077 | Value:  0.676 | VLoss: 0.0265 | PLoss: -0.0108 | Ent: 0.1476
+[ 23.0%] Iter 2248/9765 | SPS:  1516 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0331 | PLoss: -0.0057 | Ent: 0.1607
+[ 23.0%] Iter 2249/9765 | SPS:  1516 | Reward:  0.063 | Value:  0.685 | VLoss: 0.0346 | PLoss: -0.0083 | Ent: 0.1586
+[ 23.0%] Iter 2250/9765 | SPS:  1516 | Reward:  0.046 | Value:  0.671 | VLoss: 0.0223 | PLoss: -0.0115 | Ent: 0.1605
+[ 23.1%] Iter 2251/9765 | SPS:  1517 | Reward:  0.071 | Value:  0.600 | VLoss: 0.0285 | PLoss: -0.0049 | Ent: 0.1526
+[ 23.1%] Iter 2252/9765 | SPS:  1517 | Reward:  0.074 | Value:  0.685 | VLoss: 0.0279 | PLoss: -0.0019 | Ent: 0.1631
+[ 23.1%] Iter 2253/9765 | SPS:  1518 | Reward:  0.053 | Value:  0.715 | VLoss: 0.0278 | PLoss: -0.0094 | Ent: 0.1736
+[ 23.1%] Iter 2254/9765 | SPS:  1518 | Reward:  0.065 | Value:  0.699 | VLoss: 0.0247 | PLoss: -0.0050 | Ent: 0.1290
+[ 23.1%] Iter 2255/9765 | SPS:  1518 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0231 | PLoss: -0.0055 | Ent: 0.1200
+[ 23.1%] Iter 2256/9765 | SPS:  1519 | Reward:  0.050 | Value:  0.613 | VLoss: 0.0228 | PLoss: -0.0049 | Ent: 0.1348
+[ 23.1%] Iter 2257/9765 | SPS:  1519 | Reward:  0.044 | Value:  0.642 | VLoss: 0.0225 | PLoss: -0.0077 | Ent: 0.1327
+[ 23.1%] Iter 2258/9765 | SPS:  1520 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0246 | PLoss: -0.0117 | Ent: 0.1120
+[ 23.1%] Iter 2259/9765 | SPS:  1520 | Reward:  0.060 | Value:  0.637 | VLoss: 0.0286 | PLoss: -0.0056 | Ent: 0.1438
+[ 23.1%] Iter 2260/9765 | SPS:  1520 | Reward:  0.056 | Value:  0.645 | VLoss: 0.0304 | PLoss: -0.0130 | Ent: 0.1710
+[ 23.2%] Iter 2261/9765 | SPS:  1521 | Reward:  0.052 | Value:  0.624 | VLoss: 0.0261 | PLoss: -0.0063 | Ent: 0.1918
+[ 23.2%] Iter 2262/9765 | SPS:  1521 | Reward:  0.064 | Value:  0.727 | VLoss: 0.0348 | PLoss: -0.0085 | Ent: 0.1682
+[ 23.2%] Iter 2263/9765 | SPS:  1522 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0258 | PLoss: -0.0152 | Ent: 0.1839
+[ 23.2%] Iter 2264/9765 | SPS:  1522 | Reward:  0.049 | Value:  0.670 | VLoss: 0.0222 | PLoss: -0.0061 | Ent: 0.1514
+[ 23.2%] Iter 2265/9765 | SPS:  1522 | Reward:  0.060 | Value:  0.639 | VLoss: 0.0255 | PLoss: -0.0111 | Ent: 0.1506
+[ 23.2%] Iter 2266/9765 | SPS:  1523 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0247 | PLoss: -0.0101 | Ent: 0.1573
+[ 23.2%] Iter 2267/9765 | SPS:  1523 | Reward:  0.066 | Value:  0.763 | VLoss: 0.0319 | PLoss: -0.0088 | Ent: 0.1648
+[ 23.2%] Iter 2268/9765 | SPS:  1524 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0284 | PLoss: -0.0108 | Ent: 0.1978
+[ 23.2%] Iter 2269/9765 | SPS:  1524 | Reward:  0.073 | Value:  0.673 | VLoss: 0.0247 | PLoss: -0.0093 | Ent: 0.1504
+[ 23.2%] Iter 2270/9765 | SPS:  1524 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0388 | PLoss: -0.0037 | Ent: 0.1430
+[ 23.3%] Iter 2271/9765 | SPS:  1525 | Reward:  0.048 | Value:  0.635 | VLoss: 0.0271 | PLoss: -0.0039 | Ent: 0.1030
+[ 23.3%] Iter 2272/9765 | SPS:  1525 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0215 | PLoss: -0.0050 | Ent: 0.1389
+[ 23.3%] Iter 2273/9765 | SPS:  1525 | Reward:  0.062 | Value:  0.660 | VLoss: 0.0187 | PLoss: -0.0066 | Ent: 0.1523
+[ 23.3%] Iter 2274/9765 | SPS:  1526 | Reward:  0.052 | Value:  0.674 | VLoss: 0.0203 | PLoss: -0.0090 | Ent: 0.1613
+[ 23.3%] Iter 2275/9765 | SPS:  1526 | Reward:  0.057 | Value:  0.621 | VLoss: 0.0253 | PLoss: -0.0079 | Ent: 0.1585
+[ 23.3%] Iter 2276/9765 | SPS:  1527 | Reward:  0.070 | Value:  0.721 | VLoss: 0.0262 | PLoss: -0.0022 | Ent: 0.1419
+[ 23.3%] Iter 2277/9765 | SPS:  1527 | Reward:  0.057 | Value:  0.660 | VLoss: 0.0279 | PLoss: -0.0096 | Ent: 0.1643
+[ 23.3%] Iter 2278/9765 | SPS:  1527 | Reward:  0.070 | Value:  0.637 | VLoss: 0.0359 | PLoss: -0.0035 | Ent: 0.1658
+[ 23.3%] Iter 2279/9765 | SPS:  1528 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0186 | PLoss: -0.0089 | Ent: 0.1267
+[ 23.3%] Iter 2280/9765 | SPS:  1528 | Reward:  0.051 | Value:  0.668 | VLoss: 0.0229 | PLoss: -0.0079 | Ent: 0.1091
+[ 23.4%] Iter 2281/9765 | SPS:  1529 | Reward:  0.054 | Value:  0.634 | VLoss: 0.0301 | PLoss: -0.0107 | Ent: 0.1683
+[ 23.4%] Iter 2282/9765 | SPS:  1529 | Reward:  0.069 | Value:  0.698 | VLoss: 0.0241 | PLoss: -0.0095 | Ent: 0.1221
+[ 23.4%] Iter 2283/9765 | SPS:  1529 | Reward:  0.068 | Value:  0.702 | VLoss: 0.0279 | PLoss: -0.0090 | Ent: 0.1314
+[ 23.4%] Iter 2284/9765 | SPS:  1530 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0238 | PLoss: -0.0055 | Ent: 0.1737
+[ 23.4%] Iter 2285/9765 | SPS:  1530 | Reward:  0.059 | Value:  0.597 | VLoss: 0.0360 | PLoss: -0.0067 | Ent: 0.1569
+[ 23.4%] Iter 2286/9765 | SPS:  1531 | Reward:  0.057 | Value:  0.667 | VLoss: 0.0292 | PLoss: -0.0146 | Ent: 0.1484
+[ 23.4%] Iter 2287/9765 | SPS:  1531 | Reward:  0.065 | Value:  0.678 | VLoss: 0.0329 | PLoss: -0.0114 | Ent: 0.1421
+[ 23.4%] Iter 2288/9765 | SPS:  1531 | Reward:  0.073 | Value:  0.716 | VLoss: 0.0285 | PLoss: -0.0016 | Ent: 0.1142
+[ 23.4%] Iter 2289/9765 | SPS:  1532 | Reward:  0.072 | Value:  0.659 | VLoss: 0.0307 | PLoss: -0.0088 | Ent: 0.1885
+[ 23.5%] Iter 2290/9765 | SPS:  1532 | Reward:  0.057 | Value:  0.672 | VLoss: 0.0288 | PLoss: 0.0020 | Ent: 0.1603
+[ 23.5%] Iter 2291/9765 | SPS:  1533 | Reward:  0.059 | Value:  0.637 | VLoss: 0.0220 | PLoss: -0.0038 | Ent: 0.1504
+[ 23.5%] Iter 2292/9765 | SPS:  1533 | Reward:  0.055 | Value:  0.592 | VLoss: 0.0255 | PLoss: -0.0105 | Ent: 0.1801
+[ 23.5%] Iter 2293/9765 | SPS:  1533 | Reward:  0.058 | Value:  0.606 | VLoss: 0.0240 | PLoss: -0.0068 | Ent: 0.1149
+[ 23.5%] Iter 2294/9765 | SPS:  1534 | Reward:  0.049 | Value:  0.656 | VLoss: 0.0275 | PLoss: -0.0056 | Ent: 0.1467
+[ 23.5%] Iter 2295/9765 | SPS:  1534 | Reward:  0.067 | Value:  0.675 | VLoss: 0.0326 | PLoss: -0.0058 | Ent: 0.1465
+[ 23.5%] Iter 2296/9765 | SPS:  1534 | Reward:  0.067 | Value:  0.705 | VLoss: 0.0214 | PLoss: -0.0026 | Ent: 0.1377
+[ 23.5%] Iter 2297/9765 | SPS:  1535 | Reward:  0.060 | Value:  0.658 | VLoss: 0.0303 | PLoss: -0.0097 | Ent: 0.1674
+[ 23.5%] Iter 2298/9765 | SPS:  1535 | Reward:  0.055 | Value:  0.692 | VLoss: 0.0225 | PLoss: -0.0071 | Ent: 0.1302
+[ 23.5%] Iter 2299/9765 | SPS:  1536 | Reward:  0.049 | Value:  0.619 | VLoss: 0.0326 | PLoss: -0.0067 | Ent: 0.1477
+[ 23.6%] Iter 2300/9765 | SPS:  1536 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0331 | PLoss: -0.0090 | Ent: 0.1922
+[ 23.6%] Iter 2301/9765 | SPS:  1536 | Reward:  0.080 | Value:  0.700 | VLoss: 0.0202 | PLoss: -0.0083 | Ent: 0.1444
+[ 23.6%] Iter 2302/9765 | SPS:  1537 | Reward:  0.080 | Value:  0.753 | VLoss: 0.0274 | PLoss: -0.0066 | Ent: 0.1420
+[ 23.6%] Iter 2303/9765 | SPS:  1537 | Reward:  0.052 | Value:  0.703 | VLoss: 0.0265 | PLoss: -0.0030 | Ent: 0.1734
+[ 23.6%] Iter 2304/9765 | SPS:  1538 | Reward:  0.066 | Value:  0.675 | VLoss: 0.0383 | PLoss: -0.0072 | Ent: 0.1829
+[ 23.6%] Iter 2305/9765 | SPS:  1538 | Reward:  0.064 | Value:  0.689 | VLoss: 0.0213 | PLoss: -0.0022 | Ent: 0.1319
+[ 23.6%] Iter 2306/9765 | SPS:  1538 | Reward:  0.069 | Value:  0.680 | VLoss: 0.0305 | PLoss: -0.0081 | Ent: 0.1650
+[ 23.6%] Iter 2307/9765 | SPS:  1539 | Reward:  0.061 | Value:  0.649 | VLoss: 0.0322 | PLoss: -0.0012 | Ent: 0.1224
+[ 23.6%] Iter 2308/9765 | SPS:  1539 | Reward:  0.073 | Value:  0.613 | VLoss: 0.0340 | PLoss: -0.0009 | Ent: 0.1244
+[ 23.6%] Iter 2309/9765 | SPS:  1539 | Reward:  0.077 | Value:  0.703 | VLoss: 0.0286 | PLoss: -0.0034 | Ent: 0.1386
+[ 23.7%] Iter 2310/9765 | SPS:  1540 | Reward:  0.060 | Value:  0.694 | VLoss: 0.0306 | PLoss: -0.0093 | Ent: 0.1763
+[ 23.7%] Iter 2311/9765 | SPS:  1540 | Reward:  0.063 | Value:  0.634 | VLoss: 0.0255 | PLoss: -0.0113 | Ent: 0.1614
+[ 23.7%] Iter 2312/9765 | SPS:  1541 | Reward:  0.060 | Value:  0.683 | VLoss: 0.0231 | PLoss: -0.0073 | Ent: 0.1221
+[ 23.7%] Iter 2313/9765 | SPS:  1541 | Reward:  0.072 | Value:  0.661 | VLoss: 0.0325 | PLoss: -0.0109 | Ent: 0.1678
+[ 23.7%] Iter 2314/9765 | SPS:  1541 | Reward:  0.069 | Value:  0.663 | VLoss: 0.0382 | PLoss: -0.0090 | Ent: 0.1330
+[ 23.7%] Iter 2315/9765 | SPS:  1542 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0264 | PLoss: 0.0018 | Ent: 0.1166
+[ 23.7%] Iter 2316/9765 | SPS:  1542 | Reward:  0.070 | Value:  0.713 | VLoss: 0.0241 | PLoss: -0.0060 | Ent: 0.1419
+[ 23.7%] Iter 2317/9765 | SPS:  1543 | Reward:  0.053 | Value:  0.654 | VLoss: 0.0287 | PLoss: -0.0031 | Ent: 0.1557
+[ 23.7%] Iter 2318/9765 | SPS:  1543 | Reward:  0.052 | Value:  0.641 | VLoss: 0.0248 | PLoss: -0.0098 | Ent: 0.1273
+[ 23.7%] Iter 2319/9765 | SPS:  1543 | Reward:  0.069 | Value:  0.661 | VLoss: 0.0315 | PLoss: -0.0034 | Ent: 0.1650
+[ 23.8%] Iter 2320/9765 | SPS:  1544 | Reward:  0.065 | Value:  0.717 | VLoss: 0.0255 | PLoss: -0.0061 | Ent: 0.1548
+[ 23.8%] Iter 2321/9765 | SPS:  1544 | Reward:  0.076 | Value:  0.684 | VLoss: 0.0368 | PLoss: -0.0068 | Ent: 0.1477
+[ 23.8%] Iter 2322/9765 | SPS:  1544 | Reward:  0.075 | Value:  0.688 | VLoss: 0.0259 | PLoss: -0.0099 | Ent: 0.1512
+[ 23.8%] Iter 2323/9765 | SPS:  1545 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0234 | PLoss: -0.0078 | Ent: 0.1648
+[ 23.8%] Iter 2324/9765 | SPS:  1545 | Reward:  0.085 | Value:  0.694 | VLoss: 0.0289 | PLoss: -0.0067 | Ent: 0.1620
+[ 23.8%] Iter 2325/9765 | SPS:  1546 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0264 | PLoss: -0.0052 | Ent: 0.1423
+[ 23.8%] Iter 2326/9765 | SPS:  1546 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0217 | PLoss: -0.0100 | Ent: 0.1756
+[ 23.8%] Iter 2327/9765 | SPS:  1546 | Reward:  0.055 | Value:  0.724 | VLoss: 0.0239 | PLoss: -0.0085 | Ent: 0.1581
+[ 23.8%] Iter 2328/9765 | SPS:  1547 | Reward:  0.066 | Value:  0.696 | VLoss: 0.0191 | PLoss: -0.0068 | Ent: 0.1333
+[ 23.9%] Iter 2329/9765 | SPS:  1547 | Reward:  0.070 | Value:  0.731 | VLoss: 0.0208 | PLoss: -0.0202 | Ent: 0.2026
+[ 23.9%] Iter 2330/9765 | SPS:  1547 | Reward:  0.063 | Value:  0.740 | VLoss: 0.0252 | PLoss: -0.0105 | Ent: 0.1518
+[ 23.9%] Iter 2331/9765 | SPS:  1548 | Reward:  0.052 | Value:  0.810 | VLoss: 0.0197 | PLoss: -0.0074 | Ent: 0.1382
+[ 23.9%] Iter 2332/9765 | SPS:  1548 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0271 | PLoss: 0.0032 | Ent: 0.1564
+[ 23.9%] Iter 2333/9765 | SPS:  1549 | Reward:  0.066 | Value:  0.624 | VLoss: 0.0374 | PLoss: -0.0062 | Ent: 0.1416
+[ 23.9%] Iter 2334/9765 | SPS:  1549 | Reward:  0.060 | Value:  0.663 | VLoss: 0.0328 | PLoss: -0.0022 | Ent: 0.1350
+[ 23.9%] Iter 2335/9765 | SPS:  1549 | Reward:  0.062 | Value:  0.629 | VLoss: 0.0293 | PLoss: -0.0065 | Ent: 0.1757
+[ 23.9%] Iter 2336/9765 | SPS:  1550 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0347 | PLoss: 0.0078 | Ent: 0.1446
+[ 23.9%] Iter 2337/9765 | SPS:  1550 | Reward:  0.066 | Value:  0.644 | VLoss: 0.0343 | PLoss: -0.0047 | Ent: 0.1531
+[ 23.9%] Iter 2338/9765 | SPS:  1551 | Reward:  0.059 | Value:  0.588 | VLoss: 0.0319 | PLoss: -0.0114 | Ent: 0.1870
+[ 24.0%] Iter 2339/9765 | SPS:  1551 | Reward:  0.054 | Value:  0.596 | VLoss: 0.0311 | PLoss: -0.0106 | Ent: 0.1420
+[ 24.0%] Iter 2340/9765 | SPS:  1551 | Reward:  0.068 | Value:  0.648 | VLoss: 0.0339 | PLoss: -0.0040 | Ent: 0.1493
+[ 24.0%] Iter 2341/9765 | SPS:  1552 | Reward:  0.081 | Value:  0.686 | VLoss: 0.0264 | PLoss: 0.0008 | Ent: 0.1257
+[ 24.0%] Iter 2342/9765 | SPS:  1552 | Reward:  0.069 | Value:  0.694 | VLoss: 0.0355 | PLoss: -0.0103 | Ent: 0.1564
+[ 24.0%] Iter 2343/9765 | SPS:  1552 | Reward:  0.064 | Value:  0.687 | VLoss: 0.0354 | PLoss: -0.0055 | Ent: 0.2032
+[ 24.0%] Iter 2344/9765 | SPS:  1553 | Reward:  0.070 | Value:  0.720 | VLoss: 0.0332 | PLoss: 0.0025 | Ent: 0.1417
+[ 24.0%] Iter 2345/9765 | SPS:  1553 | Reward:  0.063 | Value:  0.716 | VLoss: 0.0198 | PLoss: -0.0018 | Ent: 0.1304
+[ 24.0%] Iter 2346/9765 | SPS:  1554 | Reward:  0.056 | Value:  0.705 | VLoss: 0.0314 | PLoss: -0.0009 | Ent: 0.1628
+[ 24.0%] Iter 2347/9765 | SPS:  1554 | Reward:  0.054 | Value:  0.676 | VLoss: 0.0240 | PLoss: 0.0053 | Ent: 0.1658
+[ 24.0%] Iter 2348/9765 | SPS:  1554 | Reward:  0.061 | Value:  0.678 | VLoss: 0.0247 | PLoss: -0.0050 | Ent: 0.1226
+[ 24.1%] Iter 2349/9765 | SPS:  1555 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0307 | PLoss: -0.0041 | Ent: 0.1690
+[ 24.1%] Iter 2350/9765 | SPS:  1555 | Reward:  0.062 | Value:  0.639 | VLoss: 0.0195 | PLoss: -0.0032 | Ent: 0.1307
+[ 24.1%] Iter 2351/9765 | SPS:  1555 | Reward:  0.053 | Value:  0.569 | VLoss: 0.0376 | PLoss: -0.0133 | Ent: 0.1258
+[ 24.1%] Iter 2352/9765 | SPS:  1556 | Reward:  0.059 | Value:  0.634 | VLoss: 0.0338 | PLoss: -0.0077 | Ent: 0.1573
+[ 24.1%] Iter 2353/9765 | SPS:  1556 | Reward:  0.065 | Value:  0.647 | VLoss: 0.0399 | PLoss: -0.0075 | Ent: 0.1580
+[ 24.1%] Iter 2354/9765 | SPS:  1557 | Reward:  0.075 | Value:  0.665 | VLoss: 0.0387 | PLoss: -0.0036 | Ent: 0.1554
+[ 24.1%] Iter 2355/9765 | SPS:  1557 | Reward:  0.081 | Value:  0.702 | VLoss: 0.0300 | PLoss: -0.0081 | Ent: 0.1747
+[ 24.1%] Iter 2356/9765 | SPS:  1557 | Reward:  0.063 | Value:  0.753 | VLoss: 0.0315 | PLoss: -0.0125 | Ent: 0.1633
+[ 24.1%] Iter 2357/9765 | SPS:  1558 | Reward:  0.067 | Value:  0.736 | VLoss: 0.0274 | PLoss: -0.0025 | Ent: 0.1413
+[ 24.1%] Iter 2358/9765 | SPS:  1558 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0222 | PLoss: -0.0078 | Ent: 0.1762
+[ 24.2%] Iter 2359/9765 | SPS:  1558 | Reward:  0.073 | Value:  0.705 | VLoss: 0.0251 | PLoss: -0.0089 | Ent: 0.1501
+[ 24.2%] Iter 2360/9765 | SPS:  1559 | Reward:  0.070 | Value:  0.694 | VLoss: 0.0308 | PLoss: -0.0085 | Ent: 0.1469
+[ 24.2%] Iter 2361/9765 | SPS:  1559 | Reward:  0.062 | Value:  0.671 | VLoss: 0.0308 | PLoss: -0.0045 | Ent: 0.1802
+[ 24.2%] Iter 2362/9765 | SPS:  1560 | Reward:  0.065 | Value:  0.704 | VLoss: 0.0272 | PLoss: -0.0081 | Ent: 0.1809
+[ 24.2%] Iter 2363/9765 | SPS:  1560 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0343 | PLoss: -0.0107 | Ent: 0.1952
+[ 24.2%] Iter 2364/9765 | SPS:  1560 | Reward:  0.064 | Value:  0.658 | VLoss: 0.0389 | PLoss: -0.0087 | Ent: 0.1368
+[ 24.2%] Iter 2365/9765 | SPS:  1561 | Reward:  0.059 | Value:  0.622 | VLoss: 0.0249 | PLoss: -0.0032 | Ent: 0.1416
+[ 24.2%] Iter 2366/9765 | SPS:  1561 | Reward:  0.050 | Value:  0.683 | VLoss: 0.0235 | PLoss: -0.0061 | Ent: 0.1594
+[ 24.2%] Iter 2367/9765 | SPS:  1562 | Reward:  0.065 | Value:  0.672 | VLoss: 0.0254 | PLoss: -0.0086 | Ent: 0.1554
+[ 24.2%] Iter 2368/9765 | SPS:  1562 | Reward:  0.069 | Value:  0.682 | VLoss: 0.0275 | PLoss: -0.0073 | Ent: 0.1391
+[ 24.3%] Iter 2369/9765 | SPS:  1562 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0284 | PLoss: -0.0076 | Ent: 0.1410
+[ 24.3%] Iter 2370/9765 | SPS:  1563 | Reward:  0.060 | Value:  0.598 | VLoss: 0.0242 | PLoss: -0.0049 | Ent: 0.1164
+[ 24.3%] Iter 2371/9765 | SPS:  1563 | Reward:  0.064 | Value:  0.694 | VLoss: 0.0191 | PLoss: -0.0118 | Ent: 0.1241
+[ 24.3%] Iter 2372/9765 | SPS:  1563 | Reward:  0.071 | Value:  0.684 | VLoss: 0.0308 | PLoss: -0.0093 | Ent: 0.1623
+[ 24.3%] Iter 2373/9765 | SPS:  1564 | Reward:  0.063 | Value:  0.659 | VLoss: 0.0328 | PLoss: -0.0067 | Ent: 0.1338
+[ 24.3%] Iter 2374/9765 | SPS:  1564 | Reward:  0.069 | Value:  0.620 | VLoss: 0.0255 | PLoss: -0.0138 | Ent: 0.1657
+[ 24.3%] Iter 2375/9765 | SPS:  1565 | Reward:  0.065 | Value:  0.663 | VLoss: 0.0293 | PLoss: -0.0111 | Ent: 0.1901
+[ 24.3%] Iter 2376/9765 | SPS:  1565 | Reward:  0.066 | Value:  0.710 | VLoss: 0.0364 | PLoss: 0.0037 | Ent: 0.1539
+[ 24.3%] Iter 2377/9765 | SPS:  1565 | Reward:  0.076 | Value:  0.684 | VLoss: 0.0302 | PLoss: -0.0117 | Ent: 0.1299
+[ 24.4%] Iter 2378/9765 | SPS:  1566 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0275 | PLoss: -0.0062 | Ent: 0.1332
+[ 24.4%] Iter 2379/9765 | SPS:  1566 | Reward:  0.062 | Value:  0.616 | VLoss: 0.0328 | PLoss: -0.0145 | Ent: 0.1856
+[ 24.4%] Iter 2380/9765 | SPS:  1566 | Reward:  0.062 | Value:  0.633 | VLoss: 0.0226 | PLoss: -0.0075 | Ent: 0.1556
+[ 24.4%] Iter 2381/9765 | SPS:  1567 | Reward:  0.057 | Value:  0.576 | VLoss: 0.0369 | PLoss: -0.0098 | Ent: 0.1625
+[ 24.4%] Iter 2382/9765 | SPS:  1567 | Reward:  0.065 | Value:  0.563 | VLoss: 0.0315 | PLoss: -0.0030 | Ent: 0.1580
+[ 24.4%] Iter 2383/9765 | SPS:  1567 | Reward:  0.053 | Value:  0.707 | VLoss: 0.0322 | PLoss: -0.0115 | Ent: 0.2191
+[ 24.4%] Iter 2384/9765 | SPS:  1568 | Reward:  0.065 | Value:  0.662 | VLoss: 0.0219 | PLoss: -0.0101 | Ent: 0.1498
+[ 24.4%] Iter 2385/9765 | SPS:  1568 | Reward:  0.069 | Value:  0.640 | VLoss: 0.0312 | PLoss: -0.0125 | Ent: 0.1752
+[ 24.4%] Iter 2386/9765 | SPS:  1569 | Reward:  0.055 | Value:  0.694 | VLoss: 0.0281 | PLoss: -0.0078 | Ent: 0.1741
+[ 24.4%] Iter 2387/9765 | SPS:  1569 | Reward:  0.054 | Value:  0.636 | VLoss: 0.0230 | PLoss: -0.0048 | Ent: 0.1620
+[ 24.5%] Iter 2388/9765 | SPS:  1569 | Reward:  0.071 | Value:  0.601 | VLoss: 0.0351 | PLoss: -0.0112 | Ent: 0.1781
+[ 24.5%] Iter 2389/9765 | SPS:  1570 | Reward:  0.069 | Value:  0.634 | VLoss: 0.0311 | PLoss: -0.0005 | Ent: 0.0995
+[ 24.5%] Iter 2390/9765 | SPS:  1570 | Reward:  0.072 | Value:  0.667 | VLoss: 0.0261 | PLoss: -0.0119 | Ent: 0.1473
+[ 24.5%] Iter 2391/9765 | SPS:  1570 | Reward:  0.074 | Value:  0.682 | VLoss: 0.0305 | PLoss: -0.0067 | Ent: 0.1912
+[ 24.5%] Iter 2392/9765 | SPS:  1571 | Reward:  0.056 | Value:  0.726 | VLoss: 0.0227 | PLoss: -0.0112 | Ent: 0.1715
+[ 24.5%] Iter 2393/9765 | SPS:  1571 | Reward:  0.061 | Value:  0.665 | VLoss: 0.0279 | PLoss: -0.0080 | Ent: 0.1677
+[ 24.5%] Iter 2394/9765 | SPS:  1572 | Reward:  0.060 | Value:  0.676 | VLoss: 0.0304 | PLoss: -0.0064 | Ent: 0.1474
+[ 24.5%] Iter 2395/9765 | SPS:  1572 | Reward:  0.053 | Value:  0.608 | VLoss: 0.0328 | PLoss: -0.0062 | Ent: 0.1941
+[ 24.5%] Iter 2396/9765 | SPS:  1572 | Reward:  0.070 | Value:  0.632 | VLoss: 0.0298 | PLoss: -0.0044 | Ent: 0.1520
+[ 24.5%] Iter 2397/9765 | SPS:  1573 | Reward:  0.057 | Value:  0.721 | VLoss: 0.0284 | PLoss: -0.0015 | Ent: 0.1595
+[ 24.6%] Iter 2398/9765 | SPS:  1573 | Reward:  0.056 | Value:  0.630 | VLoss: 0.0301 | PLoss: -0.0027 | Ent: 0.1629
+[ 24.6%] Iter 2399/9765 | SPS:  1573 | Reward:  0.057 | Value:  0.671 | VLoss: 0.0268 | PLoss: -0.0054 | Ent: 0.1558
+[ 24.6%] Iter 2400/9765 | SPS:  1574 | Reward:  0.058 | Value:  0.675 | VLoss: 0.0231 | PLoss: -0.0025 | Ent: 0.1583
+[ 24.6%] Iter 2401/9765 | SPS:  1574 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0300 | PLoss: -0.0097 | Ent: 0.1464
+[ 24.6%] Iter 2402/9765 | SPS:  1575 | Reward:  0.066 | Value:  0.673 | VLoss: 0.0330 | PLoss: -0.0030 | Ent: 0.1231
+[ 24.6%] Iter 2403/9765 | SPS:  1575 | Reward:  0.048 | Value:  0.671 | VLoss: 0.0230 | PLoss: -0.0101 | Ent: 0.1268
+[ 24.6%] Iter 2404/9765 | SPS:  1575 | Reward:  0.055 | Value:  0.679 | VLoss: 0.0250 | PLoss: -0.0136 | Ent: 0.1302
+[ 24.6%] Iter 2405/9765 | SPS:  1576 | Reward:  0.073 | Value:  0.687 | VLoss: 0.0329 | PLoss: -0.0064 | Ent: 0.1382
+[ 24.6%] Iter 2406/9765 | SPS:  1576 | Reward:  0.062 | Value:  0.631 | VLoss: 0.0290 | PLoss: -0.0048 | Ent: 0.1794
+[ 24.6%] Iter 2407/9765 | SPS:  1576 | Reward:  0.054 | Value:  0.576 | VLoss: 0.0382 | PLoss: -0.0192 | Ent: 0.1460
+[ 24.7%] Iter 2408/9765 | SPS:  1577 | Reward:  0.046 | Value:  0.586 | VLoss: 0.0312 | PLoss: -0.0069 | Ent: 0.1448
+[ 24.7%] Iter 2409/9765 | SPS:  1577 | Reward:  0.059 | Value:  0.598 | VLoss: 0.0285 | PLoss: -0.0108 | Ent: 0.1807
+[ 24.7%] Iter 2410/9765 | SPS:  1578 | Reward:  0.053 | Value:  0.695 | VLoss: 0.0281 | PLoss: -0.0046 | Ent: 0.1226
+[ 24.7%] Iter 2411/9765 | SPS:  1578 | Reward:  0.054 | Value:  0.714 | VLoss: 0.0178 | PLoss: -0.0108 | Ent: 0.1753
+[ 24.7%] Iter 2412/9765 | SPS:  1578 | Reward:  0.065 | Value:  0.733 | VLoss: 0.0224 | PLoss: -0.0101 | Ent: 0.1151
+[ 24.7%] Iter 2413/9765 | SPS:  1579 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0236 | PLoss: -0.0137 | Ent: 0.1369
+[ 24.7%] Iter 2414/9765 | SPS:  1579 | Reward:  0.047 | Value:  0.678 | VLoss: 0.0236 | PLoss: -0.0060 | Ent: 0.1303
+[ 24.7%] Iter 2415/9765 | SPS:  1580 | Reward:  0.064 | Value:  0.622 | VLoss: 0.0248 | PLoss: -0.0088 | Ent: 0.1589
+[ 24.7%] Iter 2416/9765 | SPS:  1580 | Reward:  0.049 | Value:  0.623 | VLoss: 0.0206 | PLoss: -0.0065 | Ent: 0.1395
+[ 24.8%] Iter 2417/9765 | SPS:  1580 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0248 | PLoss: -0.0044 | Ent: 0.1275
+[ 24.8%] Iter 2418/9765 | SPS:  1581 | Reward:  0.048 | Value:  0.631 | VLoss: 0.0273 | PLoss: -0.0073 | Ent: 0.1158
+[ 24.8%] Iter 2419/9765 | SPS:  1581 | Reward:  0.065 | Value:  0.687 | VLoss: 0.0331 | PLoss: -0.0045 | Ent: 0.1171
+[ 24.8%] Iter 2420/9765 | SPS:  1581 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0260 | PLoss: -0.0041 | Ent: 0.1528
+[ 24.8%] Iter 2421/9765 | SPS:  1582 | Reward:  0.076 | Value:  0.681 | VLoss: 0.0214 | PLoss: -0.0079 | Ent: 0.2029
+[ 24.8%] Iter 2422/9765 | SPS:  1582 | Reward:  0.061 | Value:  0.723 | VLoss: 0.0276 | PLoss: -0.0042 | Ent: 0.1807
+[ 24.8%] Iter 2423/9765 | SPS:  1583 | Reward:  0.063 | Value:  0.660 | VLoss: 0.0339 | PLoss: -0.0102 | Ent: 0.1381
+[ 24.8%] Iter 2424/9765 | SPS:  1583 | Reward:  0.069 | Value:  0.621 | VLoss: 0.0271 | PLoss: -0.0048 | Ent: 0.1609
+[ 24.8%] Iter 2425/9765 | SPS:  1583 | Reward:  0.057 | Value:  0.660 | VLoss: 0.0280 | PLoss: -0.0110 | Ent: 0.1257
+[ 24.8%] Iter 2426/9765 | SPS:  1584 | Reward:  0.062 | Value:  0.610 | VLoss: 0.0304 | PLoss: -0.0094 | Ent: 0.1870
+[ 24.9%] Iter 2427/9765 | SPS:  1584 | Reward:  0.046 | Value:  0.672 | VLoss: 0.0367 | PLoss: -0.0080 | Ent: 0.1566
+[ 24.9%] Iter 2428/9765 | SPS:  1584 | Reward:  0.050 | Value:  0.693 | VLoss: 0.0242 | PLoss: -0.0030 | Ent: 0.1608
+[ 24.9%] Iter 2429/9765 | SPS:  1585 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0270 | PLoss: -0.0135 | Ent: 0.1621
+[ 24.9%] Iter 2430/9765 | SPS:  1585 | Reward:  0.079 | Value:  0.732 | VLoss: 0.0238 | PLoss: -0.0101 | Ent: 0.1631
+[ 24.9%] Iter 2431/9765 | SPS:  1585 | Reward:  0.059 | Value:  0.676 | VLoss: 0.0310 | PLoss: -0.0180 | Ent: 0.1757
+[ 24.9%] Iter 2432/9765 | SPS:  1586 | Reward:  0.062 | Value:  0.644 | VLoss: 0.0324 | PLoss: -0.0115 | Ent: 0.1514
+[ 24.9%] Iter 2433/9765 | SPS:  1586 | Reward:  0.062 | Value:  0.611 | VLoss: 0.0266 | PLoss: -0.0038 | Ent: 0.1597
+[ 24.9%] Iter 2434/9765 | SPS:  1587 | Reward:  0.052 | Value:  0.659 | VLoss: 0.0288 | PLoss: -0.0120 | Ent: 0.1490
+[ 24.9%] Iter 2435/9765 | SPS:  1587 | Reward:  0.051 | Value:  0.640 | VLoss: 0.0222 | PLoss: -0.0020 | Ent: 0.1632
+[ 24.9%] Iter 2436/9765 | SPS:  1587 | Reward:  0.065 | Value:  0.709 | VLoss: 0.0143 | PLoss: -0.0136 | Ent: 0.1336
+[ 25.0%] Iter 2437/9765 | SPS:  1588 | Reward:  0.066 | Value:  0.697 | VLoss: 0.0242 | PLoss: -0.0137 | Ent: 0.1448
+[ 25.0%] Iter 2438/9765 | SPS:  1588 | Reward:  0.059 | Value:  0.723 | VLoss: 0.0202 | PLoss: -0.0091 | Ent: 0.1744
+[ 25.0%] Iter 2439/9765 | SPS:  1588 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0208 | PLoss: -0.0066 | Ent: 0.1739
+[ 25.0%] Iter 2440/9765 | SPS:  1589 | Reward:  0.053 | Value:  0.693 | VLoss: 0.0225 | PLoss: -0.0049 | Ent: 0.2138
+[ 25.0%] Iter 2441/9765 | SPS:  1589 | Reward:  0.074 | Value:  0.597 | VLoss: 0.0313 | PLoss: -0.0145 | Ent: 0.1469
+[ 25.0%] Iter 2442/9765 | SPS:  1590 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0270 | PLoss: -0.0021 | Ent: 0.1845
+[ 25.0%] Iter 2443/9765 | SPS:  1590 | Reward:  0.078 | Value:  0.672 | VLoss: 0.0350 | PLoss: -0.0074 | Ent: 0.1772
+[ 25.0%] Iter 2444/9765 | SPS:  1590 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0311 | PLoss: -0.0077 | Ent: 0.1716
+[ 25.0%] Iter 2445/9765 | SPS:  1591 | Reward:  0.059 | Value:  0.713 | VLoss: 0.0253 | PLoss: -0.0027 | Ent: 0.1676
+[ 25.0%] Iter 2446/9765 | SPS:  1591 | Reward:  0.059 | Value:  0.807 | VLoss: 0.0263 | PLoss: -0.0096 | Ent: 0.1090
+[ 25.1%] Iter 2447/9765 | SPS:  1591 | Reward:  0.069 | Value:  0.723 | VLoss: 0.0286 | PLoss: -0.0114 | Ent: 0.1703
+[ 25.1%] Iter 2448/9765 | SPS:  1592 | Reward:  0.061 | Value:  0.689 | VLoss: 0.0299 | PLoss: -0.0077 | Ent: 0.2108
+[ 25.1%] Iter 2449/9765 | SPS:  1592 | Reward:  0.060 | Value:  0.645 | VLoss: 0.0276 | PLoss: -0.0069 | Ent: 0.1226
+[ 25.1%] Iter 2450/9765 | SPS:  1592 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0251 | PLoss: -0.0087 | Ent: 0.1412
+[ 25.1%] Iter 2451/9765 | SPS:  1593 | Reward:  0.061 | Value:  0.612 | VLoss: 0.0421 | PLoss: -0.0096 | Ent: 0.1870
+[ 25.1%] Iter 2452/9765 | SPS:  1593 | Reward:  0.060 | Value:  0.677 | VLoss: 0.0318 | PLoss: -0.0119 | Ent: 0.1749
+[ 25.1%] Iter 2453/9765 | SPS:  1594 | Reward:  0.055 | Value:  0.625 | VLoss: 0.0292 | PLoss: -0.0048 | Ent: 0.1764
+[ 25.1%] Iter 2454/9765 | SPS:  1594 | Reward:  0.083 | Value:  0.654 | VLoss: 0.0320 | PLoss: -0.0064 | Ent: 0.1869
+[ 25.1%] Iter 2455/9765 | SPS:  1594 | Reward:  0.064 | Value:  0.750 | VLoss: 0.0291 | PLoss: -0.0123 | Ent: 0.1373
+[ 25.2%] Iter 2456/9765 | SPS:  1595 | Reward:  0.063 | Value:  0.650 | VLoss: 0.0228 | PLoss: -0.0098 | Ent: 0.1543
+[ 25.2%] Iter 2457/9765 | SPS:  1595 | Reward:  0.057 | Value:  0.686 | VLoss: 0.0282 | PLoss: -0.0014 | Ent: 0.1552
+[ 25.2%] Iter 2458/9765 | SPS:  1595 | Reward:  0.054 | Value:  0.658 | VLoss: 0.0192 | PLoss: -0.0068 | Ent: 0.1568
+[ 25.2%] Iter 2459/9765 | SPS:  1596 | Reward:  0.050 | Value:  0.631 | VLoss: 0.0222 | PLoss: -0.0071 | Ent: 0.1702
+[ 25.2%] Iter 2460/9765 | SPS:  1596 | Reward:  0.058 | Value:  0.644 | VLoss: 0.0350 | PLoss: -0.0077 | Ent: 0.1265
+[ 25.2%] Iter 2461/9765 | SPS:  1596 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0271 | PLoss: -0.0100 | Ent: 0.1435
+[ 25.2%] Iter 2462/9765 | SPS:  1597 | Reward:  0.065 | Value:  0.626 | VLoss: 0.0326 | PLoss: -0.0111 | Ent: 0.1835
+[ 25.2%] Iter 2463/9765 | SPS:  1597 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0262 | PLoss: -0.0045 | Ent: 0.1808
+[ 25.2%] Iter 2464/9765 | SPS:  1598 | Reward:  0.071 | Value:  0.725 | VLoss: 0.0388 | PLoss: -0.0105 | Ent: 0.1630
+[ 25.2%] Iter 2465/9765 | SPS:  1598 | Reward:  0.051 | Value:  0.762 | VLoss: 0.0219 | PLoss: -0.0085 | Ent: 0.1639
+[ 25.3%] Iter 2466/9765 | SPS:  1598 | Reward:  0.048 | Value:  0.666 | VLoss: 0.0349 | PLoss: -0.0096 | Ent: 0.1423
+[ 25.3%] Iter 2467/9765 | SPS:  1599 | Reward:  0.069 | Value:  0.618 | VLoss: 0.0300 | PLoss: -0.0082 | Ent: 0.1633
+[ 25.3%] Iter 2468/9765 | SPS:  1599 | Reward:  0.066 | Value:  0.669 | VLoss: 0.0253 | PLoss: -0.0070 | Ent: 0.1594
+[ 25.3%] Iter 2469/9765 | SPS:  1599 | Reward:  0.057 | Value:  0.699 | VLoss: 0.0267 | PLoss: -0.0040 | Ent: 0.1738
+[ 25.3%] Iter 2470/9765 | SPS:  1600 | Reward:  0.077 | Value:  0.712 | VLoss: 0.0278 | PLoss: -0.0014 | Ent: 0.1773
+[ 25.3%] Iter 2471/9765 | SPS:  1600 | Reward:  0.050 | Value:  0.705 | VLoss: 0.0260 | PLoss: -0.0061 | Ent: 0.2071
+[ 25.3%] Iter 2472/9765 | SPS:  1601 | Reward:  0.057 | Value:  0.660 | VLoss: 0.0244 | PLoss: -0.0123 | Ent: 0.2087
+[ 25.3%] Iter 2473/9765 | SPS:  1601 | Reward:  0.078 | Value:  0.664 | VLoss: 0.0260 | PLoss: -0.0125 | Ent: 0.1414
+[ 25.3%] Iter 2474/9765 | SPS:  1601 | Reward:  0.058 | Value:  0.760 | VLoss: 0.0320 | PLoss: 0.0006 | Ent: 0.2031
+[ 25.3%] Iter 2475/9765 | SPS:  1602 | Reward:  0.070 | Value:  0.699 | VLoss: 0.0268 | PLoss: -0.0062 | Ent: 0.2039
+[ 25.4%] Iter 2476/9765 | SPS:  1602 | Reward:  0.055 | Value:  0.677 | VLoss: 0.0259 | PLoss: -0.0082 | Ent: 0.1569
+[ 25.4%] Iter 2477/9765 | SPS:  1602 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0431 | PLoss: -0.0047 | Ent: 0.1476
+[ 25.4%] Iter 2478/9765 | SPS:  1603 | Reward:  0.063 | Value:  0.649 | VLoss: 0.0236 | PLoss: -0.0040 | Ent: 0.1752
+[ 25.4%] Iter 2479/9765 | SPS:  1603 | Reward:  0.057 | Value:  0.625 | VLoss: 0.0246 | PLoss: -0.0109 | Ent: 0.1767
+[ 25.4%] Iter 2480/9765 | SPS:  1603 | Reward:  0.065 | Value:  0.622 | VLoss: 0.0334 | PLoss: -0.0061 | Ent: 0.1588
+[ 25.4%] Iter 2481/9765 | SPS:  1604 | Reward:  0.067 | Value:  0.653 | VLoss: 0.0275 | PLoss: -0.0119 | Ent: 0.1370
+[ 25.4%] Iter 2482/9765 | SPS:  1604 | Reward:  0.063 | Value:  0.600 | VLoss: 0.0270 | PLoss: -0.0070 | Ent: 0.1177
+[ 25.4%] Iter 2483/9765 | SPS:  1604 | Reward:  0.078 | Value:  0.661 | VLoss: 0.0435 | PLoss: -0.0123 | Ent: 0.1850
+[ 25.4%] Iter 2484/9765 | SPS:  1605 | Reward:  0.054 | Value:  0.671 | VLoss: 0.0208 | PLoss: -0.0092 | Ent: 0.1828
+[ 25.4%] Iter 2485/9765 | SPS:  1605 | Reward:  0.060 | Value:  0.685 | VLoss: 0.0285 | PLoss: -0.0019 | Ent: 0.1447
+[ 25.5%] Iter 2486/9765 | SPS:  1606 | Reward:  0.065 | Value:  0.677 | VLoss: 0.0311 | PLoss: -0.0202 | Ent: 0.1421
+[ 25.5%] Iter 2487/9765 | SPS:  1606 | Reward:  0.068 | Value:  0.687 | VLoss: 0.0305 | PLoss: -0.0135 | Ent: 0.1584
+[ 25.5%] Iter 2488/9765 | SPS:  1606 | Reward:  0.069 | Value:  0.650 | VLoss: 0.0296 | PLoss: -0.0076 | Ent: 0.2162
+[ 25.5%] Iter 2489/9765 | SPS:  1607 | Reward:  0.071 | Value:  0.647 | VLoss: 0.0258 | PLoss: -0.0054 | Ent: 0.1864
+[ 25.5%] Iter 2490/9765 | SPS:  1607 | Reward:  0.055 | Value:  0.641 | VLoss: 0.0320 | PLoss: -0.0111 | Ent: 0.1763
+[ 25.5%] Iter 2491/9765 | SPS:  1607 | Reward:  0.075 | Value:  0.656 | VLoss: 0.0330 | PLoss: -0.0046 | Ent: 0.2023
+[ 25.5%] Iter 2492/9765 | SPS:  1608 | Reward:  0.055 | Value:  0.673 | VLoss: 0.0257 | PLoss: -0.0039 | Ent: 0.1645
+[ 25.5%] Iter 2493/9765 | SPS:  1608 | Reward:  0.063 | Value:  0.657 | VLoss: 0.0300 | PLoss: -0.0092 | Ent: 0.1690
+[ 25.5%] Iter 2494/9765 | SPS:  1608 | Reward:  0.053 | Value:  0.689 | VLoss: 0.0303 | PLoss: -0.0078 | Ent: 0.1726
+[ 25.6%] Iter 2495/9765 | SPS:  1609 | Reward:  0.069 | Value:  0.651 | VLoss: 0.0345 | PLoss: -0.0133 | Ent: 0.1970
+[ 25.6%] Iter 2496/9765 | SPS:  1609 | Reward:  0.059 | Value:  0.620 | VLoss: 0.0293 | PLoss: -0.0137 | Ent: 0.1846
+[ 25.6%] Iter 2497/9765 | SPS:  1609 | Reward:  0.060 | Value:  0.703 | VLoss: 0.0173 | PLoss: -0.0170 | Ent: 0.1599
+[ 25.6%] Iter 2498/9765 | SPS:  1610 | Reward:  0.057 | Value:  0.663 | VLoss: 0.0269 | PLoss: -0.0055 | Ent: 0.1985
+[ 25.6%] Iter 2499/9765 | SPS:  1610 | Reward:  0.061 | Value:  0.626 | VLoss: 0.0323 | PLoss: -0.0031 | Ent: 0.1453
+[ 25.6%] Iter 2500/9765 | SPS:  1610 | Reward:  0.055 | Value:  0.669 | VLoss: 0.0256 | PLoss: -0.0086 | Ent: 0.1780
+[ 25.6%] Iter 2501/9765 | SPS:  1611 | Reward:  0.071 | Value:  0.668 | VLoss: 0.0304 | PLoss: -0.0056 | Ent: 0.1835
+[ 25.6%] Iter 2502/9765 | SPS:  1611 | Reward:  0.071 | Value:  0.671 | VLoss: 0.0269 | PLoss: -0.0116 | Ent: 0.1552
+[ 25.6%] Iter 2503/9765 | SPS:  1612 | Reward:  0.050 | Value:  0.611 | VLoss: 0.0257 | PLoss: -0.0123 | Ent: 0.2249
+[ 25.6%] Iter 2504/9765 | SPS:  1612 | Reward:  0.068 | Value:  0.654 | VLoss: 0.0321 | PLoss: -0.0004 | Ent: 0.2172
+[ 25.7%] Iter 2505/9765 | SPS:  1612 | Reward:  0.060 | Value:  0.696 | VLoss: 0.0240 | PLoss: -0.0145 | Ent: 0.1764
+[ 25.7%] Iter 2506/9765 | SPS:  1613 | Reward:  0.068 | Value:  0.668 | VLoss: 0.0294 | PLoss: -0.0057 | Ent: 0.1447
+[ 25.7%] Iter 2507/9765 | SPS:  1613 | Reward:  0.052 | Value:  0.678 | VLoss: 0.0276 | PLoss: -0.0081 | Ent: 0.1391
+[ 25.7%] Iter 2508/9765 | SPS:  1613 | Reward:  0.067 | Value:  0.652 | VLoss: 0.0262 | PLoss: -0.0087 | Ent: 0.1658
+[ 25.7%] Iter 2509/9765 | SPS:  1614 | Reward:  0.059 | Value:  0.667 | VLoss: 0.0349 | PLoss: -0.0033 | Ent: 0.1564
+[ 25.7%] Iter 2510/9765 | SPS:  1614 | Reward:  0.081 | Value:  0.649 | VLoss: 0.0307 | PLoss: -0.0009 | Ent: 0.1602
+[ 25.7%] Iter 2511/9765 | SPS:  1614 | Reward:  0.071 | Value:  0.661 | VLoss: 0.0339 | PLoss: -0.0035 | Ent: 0.1885
+[ 25.7%] Iter 2512/9765 | SPS:  1615 | Reward:  0.056 | Value:  0.649 | VLoss: 0.0356 | PLoss: -0.0096 | Ent: 0.1702
+[ 25.7%] Iter 2513/9765 | SPS:  1615 | Reward:  0.058 | Value:  0.719 | VLoss: 0.0322 | PLoss: -0.0020 | Ent: 0.1378
+[ 25.7%] Iter 2514/9765 | SPS:  1615 | Reward:  0.060 | Value:  0.704 | VLoss: 0.0205 | PLoss: -0.0101 | Ent: 0.1463
+[ 25.8%] Iter 2515/9765 | SPS:  1616 | Reward:  0.055 | Value:  0.665 | VLoss: 0.0228 | PLoss: -0.0133 | Ent: 0.1573
+[ 25.8%] Iter 2516/9765 | SPS:  1616 | Reward:  0.065 | Value:  0.692 | VLoss: 0.0328 | PLoss: -0.0032 | Ent: 0.1612
+[ 25.8%] Iter 2517/9765 | SPS:  1616 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0291 | PLoss: -0.0121 | Ent: 0.2189
+[ 25.8%] Iter 2518/9765 | SPS:  1617 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0274 | PLoss: -0.0038 | Ent: 0.1655
+[ 25.8%] Iter 2519/9765 | SPS:  1617 | Reward:  0.055 | Value:  0.702 | VLoss: 0.0261 | PLoss: -0.0063 | Ent: 0.1363
+[ 25.8%] Iter 2520/9765 | SPS:  1617 | Reward:  0.057 | Value:  0.673 | VLoss: 0.0286 | PLoss: 0.0012 | Ent: 0.1105
+[ 25.8%] Iter 2521/9765 | SPS:  1618 | Reward:  0.073 | Value:  0.751 | VLoss: 0.0233 | PLoss: -0.0056 | Ent: 0.1655
+[ 25.8%] Iter 2522/9765 | SPS:  1618 | Reward:  0.050 | Value:  0.707 | VLoss: 0.0254 | PLoss: -0.0014 | Ent: 0.1092
+[ 25.8%] Iter 2523/9765 | SPS:  1619 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0251 | PLoss: -0.0084 | Ent: 0.1651
+[ 25.8%] Iter 2524/9765 | SPS:  1619 | Reward:  0.067 | Value:  0.627 | VLoss: 0.0251 | PLoss: -0.0111 | Ent: 0.1655
+[ 25.9%] Iter 2525/9765 | SPS:  1619 | Reward:  0.045 | Value:  0.615 | VLoss: 0.0260 | PLoss: -0.0020 | Ent: 0.1685
+[ 25.9%] Iter 2526/9765 | SPS:  1620 | Reward:  0.059 | Value:  0.614 | VLoss: 0.0275 | PLoss: -0.0032 | Ent: 0.1581
+[ 25.9%] Iter 2527/9765 | SPS:  1620 | Reward:  0.065 | Value:  0.693 | VLoss: 0.0216 | PLoss: -0.0075 | Ent: 0.1377
+[ 25.9%] Iter 2528/9765 | SPS:  1620 | Reward:  0.066 | Value:  0.734 | VLoss: 0.0172 | PLoss: -0.0049 | Ent: 0.1519
+[ 25.9%] Iter 2529/9765 | SPS:  1621 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0314 | PLoss: -0.0140 | Ent: 0.2044
+[ 25.9%] Iter 2530/9765 | SPS:  1621 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0282 | PLoss: -0.0057 | Ent: 0.1531
+[ 25.9%] Iter 2531/9765 | SPS:  1621 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0216 | PLoss: -0.0042 | Ent: 0.1198
+[ 25.9%] Iter 2532/9765 | SPS:  1622 | Reward:  0.068 | Value:  0.677 | VLoss: 0.0359 | PLoss: -0.0092 | Ent: 0.1755
+[ 25.9%] Iter 2533/9765 | SPS:  1622 | Reward:  0.059 | Value:  0.679 | VLoss: 0.0313 | PLoss: -0.0083 | Ent: 0.1744
+[ 25.9%] Iter 2534/9765 | SPS:  1622 | Reward:  0.072 | Value:  0.684 | VLoss: 0.0186 | PLoss: -0.0025 | Ent: 0.1484
+[ 26.0%] Iter 2535/9765 | SPS:  1623 | Reward:  0.075 | Value:  0.722 | VLoss: 0.0378 | PLoss: -0.0049 | Ent: 0.1412
+[ 26.0%] Iter 2536/9765 | SPS:  1623 | Reward:  0.076 | Value:  0.700 | VLoss: 0.0304 | PLoss: -0.0050 | Ent: 0.1706
+[ 26.0%] Iter 2537/9765 | SPS:  1623 | Reward:  0.072 | Value:  0.677 | VLoss: 0.0276 | PLoss: -0.0090 | Ent: 0.1209
+[ 26.0%] Iter 2538/9765 | SPS:  1624 | Reward:  0.062 | Value:  0.618 | VLoss: 0.0254 | PLoss: -0.0101 | Ent: 0.2107
+[ 26.0%] Iter 2539/9765 | SPS:  1624 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0247 | PLoss: -0.0073 | Ent: 0.1530
+[ 26.0%] Iter 2540/9765 | SPS:  1624 | Reward:  0.062 | Value:  0.627 | VLoss: 0.0221 | PLoss: -0.0093 | Ent: 0.1543
+[ 26.0%] Iter 2541/9765 | SPS:  1625 | Reward:  0.066 | Value:  0.658 | VLoss: 0.0378 | PLoss: -0.0080 | Ent: 0.1609
+[ 26.0%] Iter 2542/9765 | SPS:  1625 | Reward:  0.054 | Value:  0.638 | VLoss: 0.0257 | PLoss: -0.0110 | Ent: 0.1839
+[ 26.0%] Iter 2543/9765 | SPS:  1626 | Reward:  0.063 | Value:  0.594 | VLoss: 0.0340 | PLoss: 0.0012 | Ent: 0.1184
+[ 26.1%] Iter 2544/9765 | SPS:  1626 | Reward:  0.053 | Value:  0.691 | VLoss: 0.0203 | PLoss: -0.0066 | Ent: 0.1617
+[ 26.1%] Iter 2545/9765 | SPS:  1626 | Reward:  0.056 | Value:  0.597 | VLoss: 0.0360 | PLoss: -0.0044 | Ent: 0.1599
+[ 26.1%] Iter 2546/9765 | SPS:  1627 | Reward:  0.063 | Value:  0.765 | VLoss: 0.0213 | PLoss: -0.0080 | Ent: 0.1263
+[ 26.1%] Iter 2547/9765 | SPS:  1627 | Reward:  0.066 | Value:  0.737 | VLoss: 0.0255 | PLoss: -0.0142 | Ent: 0.1398
+[ 26.1%] Iter 2548/9765 | SPS:  1627 | Reward:  0.055 | Value:  0.729 | VLoss: 0.0230 | PLoss: -0.0027 | Ent: 0.1356
+[ 26.1%] Iter 2549/9765 | SPS:  1628 | Reward:  0.070 | Value:  0.659 | VLoss: 0.0270 | PLoss: -0.0054 | Ent: 0.1638
+[ 26.1%] Iter 2550/9765 | SPS:  1628 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0249 | PLoss: -0.0077 | Ent: 0.1619
+[ 26.1%] Iter 2551/9765 | SPS:  1628 | Reward:  0.071 | Value:  0.666 | VLoss: 0.0251 | PLoss: -0.0077 | Ent: 0.1943
+[ 26.1%] Iter 2552/9765 | SPS:  1629 | Reward:  0.053 | Value:  0.724 | VLoss: 0.0178 | PLoss: -0.0076 | Ent: 0.1324
+[ 26.1%] Iter 2553/9765 | SPS:  1629 | Reward:  0.055 | Value:  0.634 | VLoss: 0.0370 | PLoss: -0.0131 | Ent: 0.1598
+[ 26.2%] Iter 2554/9765 | SPS:  1629 | Reward:  0.065 | Value:  0.636 | VLoss: 0.0221 | PLoss: -0.0083 | Ent: 0.1370
+[ 26.2%] Iter 2555/9765 | SPS:  1630 | Reward:  0.070 | Value:  0.635 | VLoss: 0.0366 | PLoss: -0.0063 | Ent: 0.1481
+[ 26.2%] Iter 2556/9765 | SPS:  1630 | Reward:  0.062 | Value:  0.685 | VLoss: 0.0293 | PLoss: -0.0049 | Ent: 0.2068
+[ 26.2%] Iter 2557/9765 | SPS:  1631 | Reward:  0.069 | Value:  0.666 | VLoss: 0.0207 | PLoss: -0.0042 | Ent: 0.1471
+[ 26.2%] Iter 2558/9765 | SPS:  1631 | Reward:  0.060 | Value:  0.724 | VLoss: 0.0220 | PLoss: -0.0090 | Ent: 0.1524
+[ 26.2%] Iter 2559/9765 | SPS:  1631 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0308 | PLoss: -0.0100 | Ent: 0.1689
+[ 26.2%] Iter 2560/9765 | SPS:  1632 | Reward:  0.070 | Value:  0.642 | VLoss: 0.0399 | PLoss: -0.0027 | Ent: 0.1966
+[ 26.2%] Iter 2561/9765 | SPS:  1632 | Reward:  0.060 | Value:  0.713 | VLoss: 0.0261 | PLoss: -0.0043 | Ent: 0.1555
+[ 26.2%] Iter 2562/9765 | SPS:  1632 | Reward:  0.059 | Value:  0.773 | VLoss: 0.0209 | PLoss: -0.0007 | Ent: 0.1588
+[ 26.2%] Iter 2563/9765 | SPS:  1633 | Reward:  0.072 | Value:  0.719 | VLoss: 0.0277 | PLoss: -0.0027 | Ent: 0.1369
+[ 26.3%] Iter 2564/9765 | SPS:  1633 | Reward:  0.066 | Value:  0.702 | VLoss: 0.0326 | PLoss: -0.0111 | Ent: 0.1653
+[ 26.3%] Iter 2565/9765 | SPS:  1633 | Reward:  0.058 | Value:  0.733 | VLoss: 0.0202 | PLoss: -0.0077 | Ent: 0.1500
+[ 26.3%] Iter 2566/9765 | SPS:  1634 | Reward:  0.067 | Value:  0.693 | VLoss: 0.0250 | PLoss: -0.0176 | Ent: 0.1497
+[ 26.3%] Iter 2567/9765 | SPS:  1634 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0273 | PLoss: -0.0048 | Ent: 0.2026
+[ 26.3%] Iter 2568/9765 | SPS:  1634 | Reward:  0.076 | Value:  0.627 | VLoss: 0.0316 | PLoss: -0.0024 | Ent: 0.1460
+[ 26.3%] Iter 2569/9765 | SPS:  1635 | Reward:  0.066 | Value:  0.715 | VLoss: 0.0288 | PLoss: -0.0073 | Ent: 0.1924
+[ 26.3%] Iter 2570/9765 | SPS:  1635 | Reward:  0.066 | Value:  0.743 | VLoss: 0.0268 | PLoss: -0.0044 | Ent: 0.1760
+[ 26.3%] Iter 2571/9765 | SPS:  1635 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0332 | PLoss: -0.0034 | Ent: 0.1496
+[ 26.3%] Iter 2572/9765 | SPS:  1636 | Reward:  0.064 | Value:  0.643 | VLoss: 0.0254 | PLoss: -0.0118 | Ent: 0.1772
+[ 26.3%] Iter 2573/9765 | SPS:  1636 | Reward:  0.055 | Value:  0.733 | VLoss: 0.0271 | PLoss: -0.0071 | Ent: 0.1469
+[ 26.4%] Iter 2574/9765 | SPS:  1636 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0317 | PLoss: -0.0110 | Ent: 0.1656
+[ 26.4%] Iter 2575/9765 | SPS:  1637 | Reward:  0.086 | Value:  0.675 | VLoss: 0.0310 | PLoss: -0.0031 | Ent: 0.1358
+[ 26.4%] Iter 2576/9765 | SPS:  1637 | Reward:  0.058 | Value:  0.722 | VLoss: 0.0318 | PLoss: -0.0104 | Ent: 0.1777
+[ 26.4%] Iter 2577/9765 | SPS:  1637 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0264 | PLoss: -0.0011 | Ent: 0.1419
+[ 26.4%] Iter 2578/9765 | SPS:  1638 | Reward:  0.067 | Value:  0.648 | VLoss: 0.0246 | PLoss: -0.0022 | Ent: 0.1374
+[ 26.4%] Iter 2579/9765 | SPS:  1638 | Reward:  0.066 | Value:  0.656 | VLoss: 0.0258 | PLoss: -0.0088 | Ent: 0.1920
+[ 26.4%] Iter 2580/9765 | SPS:  1639 | Reward:  0.056 | Value:  0.722 | VLoss: 0.0173 | PLoss: 0.0008 | Ent: 0.2045
+[ 26.4%] Iter 2581/9765 | SPS:  1639 | Reward:  0.056 | Value:  0.670 | VLoss: 0.0228 | PLoss: -0.0045 | Ent: 0.1568
+[ 26.4%] Iter 2582/9765 | SPS:  1639 | Reward:  0.052 | Value:  0.677 | VLoss: 0.0238 | PLoss: -0.0027 | Ent: 0.1754
+[ 26.5%] Iter 2583/9765 | SPS:  1640 | Reward:  0.058 | Value:  0.724 | VLoss: 0.0270 | PLoss: -0.0098 | Ent: 0.1440
+[ 26.5%] Iter 2584/9765 | SPS:  1640 | Reward:  0.067 | Value:  0.659 | VLoss: 0.0281 | PLoss: -0.0065 | Ent: 0.1743
+[ 26.5%] Iter 2585/9765 | SPS:  1640 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0269 | PLoss: -0.0030 | Ent: 0.1635
+[ 26.5%] Iter 2586/9765 | SPS:  1641 | Reward:  0.062 | Value:  0.692 | VLoss: 0.0236 | PLoss: -0.0079 | Ent: 0.1284
+[ 26.5%] Iter 2587/9765 | SPS:  1641 | Reward:  0.075 | Value:  0.732 | VLoss: 0.0357 | PLoss: -0.0059 | Ent: 0.1452
+[ 26.5%] Iter 2588/9765 | SPS:  1641 | Reward:  0.059 | Value:  0.646 | VLoss: 0.0234 | PLoss: -0.0108 | Ent: 0.1727
+[ 26.5%] Iter 2589/9765 | SPS:  1642 | Reward:  0.075 | Value:  0.702 | VLoss: 0.0311 | PLoss: -0.0091 | Ent: 0.1181
+[ 26.5%] Iter 2590/9765 | SPS:  1642 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0251 | PLoss: -0.0085 | Ent: 0.1415
+[ 26.5%] Iter 2591/9765 | SPS:  1642 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0242 | PLoss: -0.0075 | Ent: 0.1541
+[ 26.5%] Iter 2592/9765 | SPS:  1643 | Reward:  0.071 | Value:  0.684 | VLoss: 0.0297 | PLoss: -0.0090 | Ent: 0.1396
+[ 26.6%] Iter 2593/9765 | SPS:  1643 | Reward:  0.071 | Value:  0.688 | VLoss: 0.0282 | PLoss: -0.0085 | Ent: 0.1402
+[ 26.6%] Iter 2594/9765 | SPS:  1643 | Reward:  0.053 | Value:  0.609 | VLoss: 0.0327 | PLoss: -0.0048 | Ent: 0.1460
+[ 26.6%] Iter 2595/9765 | SPS:  1644 | Reward:  0.063 | Value:  0.619 | VLoss: 0.0296 | PLoss: -0.0090 | Ent: 0.1508
+[ 26.6%] Iter 2596/9765 | SPS:  1644 | Reward:  0.060 | Value:  0.561 | VLoss: 0.0300 | PLoss: -0.0141 | Ent: 0.1248
+[ 26.6%] Iter 2597/9765 | SPS:  1644 | Reward:  0.066 | Value:  0.690 | VLoss: 0.0238 | PLoss: -0.0099 | Ent: 0.1512
+[ 26.6%] Iter 2598/9765 | SPS:  1645 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0314 | PLoss: -0.0076 | Ent: 0.1864
+[ 26.6%] Iter 2599/9765 | SPS:  1645 | Reward:  0.078 | Value:  0.714 | VLoss: 0.0325 | PLoss: -0.0045 | Ent: 0.1469
+[ 26.6%] Iter 2600/9765 | SPS:  1645 | Reward:  0.070 | Value:  0.719 | VLoss: 0.0315 | PLoss: -0.0140 | Ent: 0.1220
+[ 26.6%] Iter 2601/9765 | SPS:  1646 | Reward:  0.060 | Value:  0.763 | VLoss: 0.0299 | PLoss: -0.0048 | Ent: 0.1283
+[ 26.6%] Iter 2602/9765 | SPS:  1646 | Reward:  0.057 | Value:  0.679 | VLoss: 0.0213 | PLoss: -0.0059 | Ent: 0.1354
+[ 26.7%] Iter 2603/9765 | SPS:  1646 | Reward:  0.065 | Value:  0.659 | VLoss: 0.0255 | PLoss: -0.0070 | Ent: 0.1369
+[ 26.7%] Iter 2604/9765 | SPS:  1647 | Reward:  0.067 | Value:  0.619 | VLoss: 0.0291 | PLoss: -0.0193 | Ent: 0.1483
+[ 26.7%] Iter 2605/9765 | SPS:  1647 | Reward:  0.080 | Value:  0.669 | VLoss: 0.0225 | PLoss: -0.0081 | Ent: 0.1047
+[ 26.7%] Iter 2606/9765 | SPS:  1648 | Reward:  0.058 | Value:  0.667 | VLoss: 0.0348 | PLoss: -0.0090 | Ent: 0.1338
+[ 26.7%] Iter 2607/9765 | SPS:  1648 | Reward:  0.067 | Value:  0.663 | VLoss: 0.0370 | PLoss: -0.0039 | Ent: 0.1236
+[ 26.7%] Iter 2608/9765 | SPS:  1648 | Reward:  0.063 | Value:  0.626 | VLoss: 0.0314 | PLoss: -0.0062 | Ent: 0.1456
+[ 26.7%] Iter 2609/9765 | SPS:  1649 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0260 | PLoss: -0.0040 | Ent: 0.1838
+[ 26.7%] Iter 2610/9765 | SPS:  1649 | Reward:  0.067 | Value:  0.631 | VLoss: 0.0298 | PLoss: -0.0140 | Ent: 0.1725
+[ 26.7%] Iter 2611/9765 | SPS:  1649 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0328 | PLoss: -0.0042 | Ent: 0.1337
+[ 26.7%] Iter 2612/9765 | SPS:  1650 | Reward:  0.067 | Value:  0.653 | VLoss: 0.0280 | PLoss: -0.0094 | Ent: 0.1519
+[ 26.8%] Iter 2613/9765 | SPS:  1650 | Reward:  0.045 | Value:  0.671 | VLoss: 0.0213 | PLoss: -0.0072 | Ent: 0.1126
+[ 26.8%] Iter 2614/9765 | SPS:  1650 | Reward:  0.060 | Value:  0.648 | VLoss: 0.0275 | PLoss: -0.0014 | Ent: 0.1304
+[ 26.8%] Iter 2615/9765 | SPS:  1651 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0205 | PLoss: -0.0010 | Ent: 0.1021
+[ 26.8%] Iter 2616/9765 | SPS:  1651 | Reward:  0.054 | Value:  0.677 | VLoss: 0.0270 | PLoss: -0.0042 | Ent: 0.1372
+[ 26.8%] Iter 2617/9765 | SPS:  1651 | Reward:  0.060 | Value:  0.620 | VLoss: 0.0305 | PLoss: -0.0106 | Ent: 0.1050
+[ 26.8%] Iter 2618/9765 | SPS:  1652 | Reward:  0.056 | Value:  0.651 | VLoss: 0.0242 | PLoss: -0.0003 | Ent: 0.1718
+[ 26.8%] Iter 2619/9765 | SPS:  1652 | Reward:  0.065 | Value:  0.604 | VLoss: 0.0269 | PLoss: -0.0087 | Ent: 0.1339
+[ 26.8%] Iter 2620/9765 | SPS:  1652 | Reward:  0.054 | Value:  0.657 | VLoss: 0.0292 | PLoss: -0.0102 | Ent: 0.1238
+[ 26.8%] Iter 2621/9765 | SPS:  1653 | Reward:  0.049 | Value:  0.597 | VLoss: 0.0216 | PLoss: -0.0051 | Ent: 0.1012
+[ 26.9%] Iter 2622/9765 | SPS:  1653 | Reward:  0.045 | Value:  0.602 | VLoss: 0.0175 | PLoss: -0.0083 | Ent: 0.1201
+[ 26.9%] Iter 2623/9765 | SPS:  1653 | Reward:  0.053 | Value:  0.611 | VLoss: 0.0225 | PLoss: -0.0035 | Ent: 0.1166
+[ 26.9%] Iter 2624/9765 | SPS:  1654 | Reward:  0.052 | Value:  0.602 | VLoss: 0.0283 | PLoss: -0.0104 | Ent: 0.1393
+[ 26.9%] Iter 2625/9765 | SPS:  1654 | Reward:  0.040 | Value:  0.537 | VLoss: 0.0192 | PLoss: -0.0108 | Ent: 0.1051
+[ 26.9%] Iter 2626/9765 | SPS:  1655 | Reward:  0.054 | Value:  0.631 | VLoss: 0.0358 | PLoss: -0.0116 | Ent: 0.1205
+[ 26.9%] Iter 2627/9765 | SPS:  1655 | Reward:  0.068 | Value:  0.626 | VLoss: 0.0253 | PLoss: -0.0013 | Ent: 0.0876
+[ 26.9%] Iter 2628/9765 | SPS:  1655 | Reward:  0.056 | Value:  0.730 | VLoss: 0.0212 | PLoss: -0.0044 | Ent: 0.1502
+[ 26.9%] Iter 2629/9765 | SPS:  1656 | Reward:  0.068 | Value:  0.681 | VLoss: 0.0252 | PLoss: -0.0100 | Ent: 0.1604
+[ 26.9%] Iter 2630/9765 | SPS:  1656 | Reward:  0.066 | Value:  0.625 | VLoss: 0.0356 | PLoss: -0.0061 | Ent: 0.1560
+[ 26.9%] Iter 2631/9765 | SPS:  1656 | Reward:  0.053 | Value:  0.651 | VLoss: 0.0230 | PLoss: -0.0056 | Ent: 0.1204
+[ 27.0%] Iter 2632/9765 | SPS:  1657 | Reward:  0.076 | Value:  0.627 | VLoss: 0.0301 | PLoss: -0.0040 | Ent: 0.1281
+[ 27.0%] Iter 2633/9765 | SPS:  1657 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0222 | PLoss: -0.0045 | Ent: 0.1493
+[ 27.0%] Iter 2634/9765 | SPS:  1657 | Reward:  0.058 | Value:  0.682 | VLoss: 0.0215 | PLoss: -0.0143 | Ent: 0.1797
+[ 27.0%] Iter 2635/9765 | SPS:  1658 | Reward:  0.043 | Value:  0.675 | VLoss: 0.0249 | PLoss: -0.0045 | Ent: 0.1778
+[ 27.0%] Iter 2636/9765 | SPS:  1658 | Reward:  0.051 | Value:  0.629 | VLoss: 0.0255 | PLoss: -0.0025 | Ent: 0.1475
+[ 27.0%] Iter 2637/9765 | SPS:  1658 | Reward:  0.052 | Value:  0.575 | VLoss: 0.0336 | PLoss: -0.0032 | Ent: 0.1444
+[ 27.0%] Iter 2638/9765 | SPS:  1659 | Reward:  0.066 | Value:  0.660 | VLoss: 0.0339 | PLoss: -0.0031 | Ent: 0.1510
+[ 27.0%] Iter 2639/9765 | SPS:  1659 | Reward:  0.056 | Value:  0.697 | VLoss: 0.0223 | PLoss: -0.0051 | Ent: 0.1558
+[ 27.0%] Iter 2640/9765 | SPS:  1659 | Reward:  0.072 | Value:  0.657 | VLoss: 0.0363 | PLoss: -0.0090 | Ent: 0.1501
+[ 27.0%] Iter 2641/9765 | SPS:  1660 | Reward:  0.068 | Value:  0.623 | VLoss: 0.0320 | PLoss: -0.0049 | Ent: 0.1329
+[ 27.1%] Iter 2642/9765 | SPS:  1660 | Reward:  0.070 | Value:  0.670 | VLoss: 0.0324 | PLoss: -0.0139 | Ent: 0.1644
+[ 27.1%] Iter 2643/9765 | SPS:  1660 | Reward:  0.066 | Value:  0.650 | VLoss: 0.0304 | PLoss: -0.0114 | Ent: 0.1803
+[ 27.1%] Iter 2644/9765 | SPS:  1661 | Reward:  0.063 | Value:  0.653 | VLoss: 0.0247 | PLoss: -0.0051 | Ent: 0.1559
+[ 27.1%] Iter 2645/9765 | SPS:  1661 | Reward:  0.068 | Value:  0.664 | VLoss: 0.0372 | PLoss: -0.0034 | Ent: 0.1087
+[ 27.1%] Iter 2646/9765 | SPS:  1661 | Reward:  0.061 | Value:  0.680 | VLoss: 0.0343 | PLoss: -0.0090 | Ent: 0.1538
+[ 27.1%] Iter 2647/9765 | SPS:  1662 | Reward:  0.051 | Value:  0.716 | VLoss: 0.0252 | PLoss: -0.0071 | Ent: 0.1872
+[ 27.1%] Iter 2648/9765 | SPS:  1662 | Reward:  0.063 | Value:  0.607 | VLoss: 0.0241 | PLoss: -0.0109 | Ent: 0.1639
+[ 27.1%] Iter 2649/9765 | SPS:  1662 | Reward:  0.070 | Value:  0.633 | VLoss: 0.0277 | PLoss: -0.0116 | Ent: 0.1799
+[ 27.1%] Iter 2650/9765 | SPS:  1663 | Reward:  0.073 | Value:  0.689 | VLoss: 0.0302 | PLoss: -0.0093 | Ent: 0.1621
+[ 27.1%] Iter 2651/9765 | SPS:  1663 | Reward:  0.051 | Value:  0.692 | VLoss: 0.0268 | PLoss: -0.0044 | Ent: 0.1420
+[ 27.2%] Iter 2652/9765 | SPS:  1663 | Reward:  0.060 | Value:  0.640 | VLoss: 0.0344 | PLoss: -0.0149 | Ent: 0.1875
+[ 27.2%] Iter 2653/9765 | SPS:  1664 | Reward:  0.060 | Value:  0.661 | VLoss: 0.0339 | PLoss: -0.0109 | Ent: 0.1415
+[ 27.2%] Iter 2654/9765 | SPS:  1664 | Reward:  0.065 | Value:  0.737 | VLoss: 0.0288 | PLoss: -0.0166 | Ent: 0.1758
+[ 27.2%] Iter 2655/9765 | SPS:  1664 | Reward:  0.067 | Value:  0.647 | VLoss: 0.0192 | PLoss: -0.0055 | Ent: 0.1615
+[ 27.2%] Iter 2656/9765 | SPS:  1665 | Reward:  0.068 | Value:  0.720 | VLoss: 0.0215 | PLoss: -0.0097 | Ent: 0.1551
+[ 27.2%] Iter 2657/9765 | SPS:  1665 | Reward:  0.054 | Value:  0.692 | VLoss: 0.0232 | PLoss: -0.0117 | Ent: 0.1512
+[ 27.2%] Iter 2658/9765 | SPS:  1665 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0347 | PLoss: -0.0114 | Ent: 0.1309
+[ 27.2%] Iter 2659/9765 | SPS:  1666 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0320 | PLoss: -0.0053 | Ent: 0.1382
+[ 27.2%] Iter 2660/9765 | SPS:  1666 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0307 | PLoss: -0.0047 | Ent: 0.1775
+[ 27.3%] Iter 2661/9765 | SPS:  1666 | Reward:  0.065 | Value:  0.685 | VLoss: 0.0276 | PLoss: -0.0078 | Ent: 0.1735
+[ 27.3%] Iter 2662/9765 | SPS:  1667 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0278 | PLoss: -0.0138 | Ent: 0.1644
+[ 27.3%] Iter 2663/9765 | SPS:  1667 | Reward:  0.068 | Value:  0.702 | VLoss: 0.0208 | PLoss: -0.0069 | Ent: 0.1547
+[ 27.3%] Iter 2664/9765 | SPS:  1667 | Reward:  0.056 | Value:  0.780 | VLoss: 0.0197 | PLoss: -0.0093 | Ent: 0.1623
+[ 27.3%] Iter 2665/9765 | SPS:  1668 | Reward:  0.067 | Value:  0.729 | VLoss: 0.0235 | PLoss: -0.0140 | Ent: 0.1769
+[ 27.3%] Iter 2666/9765 | SPS:  1668 | Reward:  0.074 | Value:  0.709 | VLoss: 0.0378 | PLoss: -0.0037 | Ent: 0.1873
+[ 27.3%] Iter 2667/9765 | SPS:  1668 | Reward:  0.072 | Value:  0.655 | VLoss: 0.0253 | PLoss: -0.0043 | Ent: 0.1229
+[ 27.3%] Iter 2668/9765 | SPS:  1669 | Reward:  0.067 | Value:  0.700 | VLoss: 0.0229 | PLoss: -0.0074 | Ent: 0.1229
+[ 27.3%] Iter 2669/9765 | SPS:  1669 | Reward:  0.063 | Value:  0.744 | VLoss: 0.0228 | PLoss: -0.0126 | Ent: 0.1494
+[ 27.3%] Iter 2670/9765 | SPS:  1669 | Reward:  0.069 | Value:  0.674 | VLoss: 0.0301 | PLoss: -0.0138 | Ent: 0.1795
+[ 27.4%] Iter 2671/9765 | SPS:  1670 | Reward:  0.062 | Value:  0.685 | VLoss: 0.0281 | PLoss: -0.0063 | Ent: 0.1274
+[ 27.4%] Iter 2672/9765 | SPS:  1670 | Reward:  0.064 | Value:  0.683 | VLoss: 0.0218 | PLoss: -0.0048 | Ent: 0.1236
+[ 27.4%] Iter 2673/9765 | SPS:  1670 | Reward:  0.061 | Value:  0.696 | VLoss: 0.0288 | PLoss: -0.0147 | Ent: 0.1745
+[ 27.4%] Iter 2674/9765 | SPS:  1671 | Reward:  0.052 | Value:  0.740 | VLoss: 0.0191 | PLoss: -0.0106 | Ent: 0.1743
+[ 27.4%] Iter 2675/9765 | SPS:  1671 | Reward:  0.057 | Value:  0.703 | VLoss: 0.0296 | PLoss: -0.0113 | Ent: 0.1536
+[ 27.4%] Iter 2676/9765 | SPS:  1671 | Reward:  0.053 | Value:  0.636 | VLoss: 0.0348 | PLoss: -0.0098 | Ent: 0.1699
+[ 27.4%] Iter 2677/9765 | SPS:  1672 | Reward:  0.065 | Value:  0.647 | VLoss: 0.0291 | PLoss: -0.0110 | Ent: 0.1577
+[ 27.4%] Iter 2678/9765 | SPS:  1672 | Reward:  0.070 | Value:  0.638 | VLoss: 0.0327 | PLoss: -0.0041 | Ent: 0.1252
+[ 27.4%] Iter 2679/9765 | SPS:  1672 | Reward:  0.064 | Value:  0.631 | VLoss: 0.0308 | PLoss: -0.0044 | Ent: 0.1564
+[ 27.4%] Iter 2680/9765 | SPS:  1673 | Reward:  0.069 | Value:  0.657 | VLoss: 0.0311 | PLoss: -0.0052 | Ent: 0.1628
+[ 27.5%] Iter 2681/9765 | SPS:  1673 | Reward:  0.067 | Value:  0.751 | VLoss: 0.0323 | PLoss: -0.0046 | Ent: 0.1643
+[ 27.5%] Iter 2682/9765 | SPS:  1673 | Reward:  0.051 | Value:  0.708 | VLoss: 0.0197 | PLoss: -0.0057 | Ent: 0.1400
+[ 27.5%] Iter 2683/9765 | SPS:  1674 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0161 | PLoss: -0.0094 | Ent: 0.1388
+[ 27.5%] Iter 2684/9765 | SPS:  1674 | Reward:  0.050 | Value:  0.667 | VLoss: 0.0275 | PLoss: -0.0082 | Ent: 0.1579
+[ 27.5%] Iter 2685/9765 | SPS:  1675 | Reward:  0.057 | Value:  0.649 | VLoss: 0.0214 | PLoss: -0.0050 | Ent: 0.1399
+[ 27.5%] Iter 2686/9765 | SPS:  1675 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0298 | PLoss: -0.0043 | Ent: 0.1374
+[ 27.5%] Iter 2687/9765 | SPS:  1675 | Reward:  0.064 | Value:  0.687 | VLoss: 0.0275 | PLoss: -0.0119 | Ent: 0.1573
+[ 27.5%] Iter 2688/9765 | SPS:  1676 | Reward:  0.062 | Value:  0.695 | VLoss: 0.0197 | PLoss: -0.0061 | Ent: 0.1494
+[ 27.5%] Iter 2689/9765 | SPS:  1676 | Reward:  0.068 | Value:  0.723 | VLoss: 0.0255 | PLoss: -0.0096 | Ent: 0.1347
+[ 27.5%] Iter 2690/9765 | SPS:  1676 | Reward:  0.067 | Value:  0.708 | VLoss: 0.0436 | PLoss: -0.0103 | Ent: 0.1556
+[ 27.6%] Iter 2691/9765 | SPS:  1677 | Reward:  0.062 | Value:  0.690 | VLoss: 0.0222 | PLoss: -0.0045 | Ent: 0.1508
+[ 27.6%] Iter 2692/9765 | SPS:  1677 | Reward:  0.071 | Value:  0.693 | VLoss: 0.0350 | PLoss: -0.0061 | Ent: 0.1407
+[ 27.6%] Iter 2693/9765 | SPS:  1677 | Reward:  0.052 | Value:  0.686 | VLoss: 0.0243 | PLoss: -0.0020 | Ent: 0.1795
+[ 27.6%] Iter 2694/9765 | SPS:  1678 | Reward:  0.052 | Value:  0.665 | VLoss: 0.0215 | PLoss: -0.0118 | Ent: 0.1364
+[ 27.6%] Iter 2695/9765 | SPS:  1678 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0262 | PLoss: -0.0029 | Ent: 0.1245
+[ 27.6%] Iter 2696/9765 | SPS:  1678 | Reward:  0.066 | Value:  0.682 | VLoss: 0.0283 | PLoss: -0.0025 | Ent: 0.1451
+[ 27.6%] Iter 2697/9765 | SPS:  1679 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0244 | PLoss: -0.0040 | Ent: 0.1585
+[ 27.6%] Iter 2698/9765 | SPS:  1679 | Reward:  0.066 | Value:  0.751 | VLoss: 0.0203 | PLoss: -0.0088 | Ent: 0.1354
+[ 27.6%] Iter 2699/9765 | SPS:  1679 | Reward:  0.063 | Value:  0.679 | VLoss: 0.0291 | PLoss: -0.0069 | Ent: 0.1437
+[ 27.6%] Iter 2700/9765 | SPS:  1680 | Reward:  0.058 | Value:  0.696 | VLoss: 0.0249 | PLoss: -0.0070 | Ent: 0.1256
+[ 27.7%] Iter 2701/9765 | SPS:  1680 | Reward:  0.067 | Value:  0.605 | VLoss: 0.0351 | PLoss: -0.0195 | Ent: 0.1839
+[ 27.7%] Iter 2702/9765 | SPS:  1680 | Reward:  0.073 | Value:  0.720 | VLoss: 0.0212 | PLoss: -0.0036 | Ent: 0.1558
+[ 27.7%] Iter 2703/9765 | SPS:  1681 | Reward:  0.066 | Value:  0.760 | VLoss: 0.0281 | PLoss: -0.0077 | Ent: 0.1292
+[ 27.7%] Iter 2704/9765 | SPS:  1681 | Reward:  0.067 | Value:  0.649 | VLoss: 0.0389 | PLoss: -0.0069 | Ent: 0.1645
+[ 27.7%] Iter 2705/9765 | SPS:  1681 | Reward:  0.069 | Value:  0.694 | VLoss: 0.0317 | PLoss: -0.0053 | Ent: 0.1350
+[ 27.7%] Iter 2706/9765 | SPS:  1682 | Reward:  0.066 | Value:  0.720 | VLoss: 0.0263 | PLoss: -0.0145 | Ent: 0.1327
+[ 27.7%] Iter 2707/9765 | SPS:  1682 | Reward:  0.064 | Value:  0.737 | VLoss: 0.0246 | PLoss: -0.0097 | Ent: 0.1295
+[ 27.7%] Iter 2708/9765 | SPS:  1682 | Reward:  0.074 | Value:  0.720 | VLoss: 0.0309 | PLoss: -0.0130 | Ent: 0.1415
+[ 27.7%] Iter 2709/9765 | SPS:  1683 | Reward:  0.066 | Value:  0.696 | VLoss: 0.0228 | PLoss: -0.0031 | Ent: 0.1127
+[ 27.8%] Iter 2710/9765 | SPS:  1683 | Reward:  0.080 | Value:  0.695 | VLoss: 0.0289 | PLoss: -0.0121 | Ent: 0.1585
+[ 27.8%] Iter 2711/9765 | SPS:  1683 | Reward:  0.076 | Value:  0.703 | VLoss: 0.0274 | PLoss: -0.0042 | Ent: 0.1375
+[ 27.8%] Iter 2712/9765 | SPS:  1684 | Reward:  0.071 | Value:  0.734 | VLoss: 0.0320 | PLoss: -0.0031 | Ent: 0.1431
+[ 27.8%] Iter 2713/9765 | SPS:  1684 | Reward:  0.064 | Value:  0.669 | VLoss: 0.0290 | PLoss: -0.0086 | Ent: 0.1444
+[ 27.8%] Iter 2714/9765 | SPS:  1684 | Reward:  0.065 | Value:  0.734 | VLoss: 0.0200 | PLoss: -0.0069 | Ent: 0.1282
+[ 27.8%] Iter 2715/9765 | SPS:  1684 | Reward:  0.076 | Value:  0.736 | VLoss: 0.0328 | PLoss: -0.0078 | Ent: 0.1281
+[ 27.8%] Iter 2716/9765 | SPS:  1685 | Reward:  0.072 | Value:  0.696 | VLoss: 0.0264 | PLoss: -0.0062 | Ent: 0.1601
+[ 27.8%] Iter 2717/9765 | SPS:  1685 | Reward:  0.057 | Value:  0.731 | VLoss: 0.0213 | PLoss: -0.0117 | Ent: 0.1803
+[ 27.8%] Iter 2718/9765 | SPS:  1685 | Reward:  0.061 | Value:  0.727 | VLoss: 0.0243 | PLoss: 0.0013 | Ent: 0.1665
+[ 27.8%] Iter 2719/9765 | SPS:  1686 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0233 | PLoss: -0.0090 | Ent: 0.1681
+[ 27.9%] Iter 2720/9765 | SPS:  1686 | Reward:  0.055 | Value:  0.643 | VLoss: 0.0277 | PLoss: -0.0080 | Ent: 0.1476
+[ 27.9%] Iter 2721/9765 | SPS:  1687 | Reward:  0.063 | Value:  0.656 | VLoss: 0.0252 | PLoss: -0.0090 | Ent: 0.1558
+[ 27.9%] Iter 2722/9765 | SPS:  1687 | Reward:  0.065 | Value:  0.642 | VLoss: 0.0242 | PLoss: -0.0139 | Ent: 0.1772
+[ 27.9%] Iter 2723/9765 | SPS:  1687 | Reward:  0.063 | Value:  0.698 | VLoss: 0.0273 | PLoss: -0.0069 | Ent: 0.1730
+[ 27.9%] Iter 2724/9765 | SPS:  1687 | Reward:  0.057 | Value:  0.675 | VLoss: 0.0262 | PLoss: -0.0073 | Ent: 0.1806
+[ 27.9%] Iter 2725/9765 | SPS:  1688 | Reward:  0.073 | Value:  0.694 | VLoss: 0.0302 | PLoss: -0.0036 | Ent: 0.1392
+[ 27.9%] Iter 2726/9765 | SPS:  1688 | Reward:  0.067 | Value:  0.682 | VLoss: 0.0302 | PLoss: -0.0072 | Ent: 0.2035
+[ 27.9%] Iter 2727/9765 | SPS:  1688 | Reward:  0.061 | Value:  0.678 | VLoss: 0.0277 | PLoss: -0.0099 | Ent: 0.1531
+[ 27.9%] Iter 2728/9765 | SPS:  1689 | Reward:  0.074 | Value:  0.726 | VLoss: 0.0406 | PLoss: -0.0043 | Ent: 0.1545
+[ 27.9%] Iter 2729/9765 | SPS:  1689 | Reward:  0.054 | Value:  0.631 | VLoss: 0.0288 | PLoss: -0.0162 | Ent: 0.1389
+[ 28.0%] Iter 2730/9765 | SPS:  1689 | Reward:  0.066 | Value:  0.630 | VLoss: 0.0286 | PLoss: -0.0110 | Ent: 0.1672
+[ 28.0%] Iter 2731/9765 | SPS:  1690 | Reward:  0.066 | Value:  0.687 | VLoss: 0.0268 | PLoss: -0.0023 | Ent: 0.1738
+[ 28.0%] Iter 2732/9765 | SPS:  1690 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0248 | PLoss: -0.0184 | Ent: 0.2308
+[ 28.0%] Iter 2733/9765 | SPS:  1690 | Reward:  0.070 | Value:  0.665 | VLoss: 0.0330 | PLoss: -0.0135 | Ent: 0.1603
+[ 28.0%] Iter 2734/9765 | SPS:  1691 | Reward:  0.062 | Value:  0.652 | VLoss: 0.0208 | PLoss: -0.0115 | Ent: 0.1863
+[ 28.0%] Iter 2735/9765 | SPS:  1691 | Reward:  0.068 | Value:  0.637 | VLoss: 0.0320 | PLoss: -0.0072 | Ent: 0.1860
+[ 28.0%] Iter 2736/9765 | SPS:  1691 | Reward:  0.069 | Value:  0.665 | VLoss: 0.0314 | PLoss: -0.0075 | Ent: 0.1774
+[ 28.0%] Iter 2737/9765 | SPS:  1692 | Reward:  0.065 | Value:  0.713 | VLoss: 0.0282 | PLoss: -0.0052 | Ent: 0.1379
+[ 28.0%] Iter 2738/9765 | SPS:  1692 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0288 | PLoss: -0.0107 | Ent: 0.1809
+[ 28.0%] Iter 2739/9765 | SPS:  1692 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0263 | PLoss: -0.0069 | Ent: 0.1775
+[ 28.1%] Iter 2740/9765 | SPS:  1693 | Reward:  0.070 | Value:  0.678 | VLoss: 0.0280 | PLoss: -0.0011 | Ent: 0.1467
+[ 28.1%] Iter 2741/9765 | SPS:  1693 | Reward:  0.067 | Value:  0.692 | VLoss: 0.0352 | PLoss: -0.0070 | Ent: 0.1789
+[ 28.1%] Iter 2742/9765 | SPS:  1693 | Reward:  0.047 | Value:  0.719 | VLoss: 0.0233 | PLoss: -0.0035 | Ent: 0.2207
+[ 28.1%] Iter 2743/9765 | SPS:  1694 | Reward:  0.073 | Value:  0.713 | VLoss: 0.0264 | PLoss: -0.0061 | Ent: 0.1473
+[ 28.1%] Iter 2744/9765 | SPS:  1694 | Reward:  0.063 | Value:  0.735 | VLoss: 0.0236 | PLoss: -0.0137 | Ent: 0.1674
+[ 28.1%] Iter 2745/9765 | SPS:  1694 | Reward:  0.065 | Value:  0.694 | VLoss: 0.0242 | PLoss: -0.0055 | Ent: 0.1410
+[ 28.1%] Iter 2746/9765 | SPS:  1695 | Reward:  0.069 | Value:  0.706 | VLoss: 0.0262 | PLoss: -0.0085 | Ent: 0.1315
+[ 28.1%] Iter 2747/9765 | SPS:  1695 | Reward:  0.070 | Value:  0.641 | VLoss: 0.0386 | PLoss: -0.0120 | Ent: 0.1760
+[ 28.1%] Iter 2748/9765 | SPS:  1695 | Reward:  0.077 | Value:  0.696 | VLoss: 0.0257 | PLoss: -0.0056 | Ent: 0.1254
+[ 28.2%] Iter 2749/9765 | SPS:  1696 | Reward:  0.059 | Value:  0.696 | VLoss: 0.0276 | PLoss: -0.0171 | Ent: 0.2498
+[ 28.2%] Iter 2750/9765 | SPS:  1696 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0262 | PLoss: -0.0115 | Ent: 0.1577
+[ 28.2%] Iter 2751/9765 | SPS:  1696 | Reward:  0.064 | Value:  0.736 | VLoss: 0.0216 | PLoss: -0.0105 | Ent: 0.1531
+[ 28.2%] Iter 2752/9765 | SPS:  1697 | Reward:  0.067 | Value:  0.757 | VLoss: 0.0226 | PLoss: -0.0055 | Ent: 0.1610
+[ 28.2%] Iter 2753/9765 | SPS:  1697 | Reward:  0.068 | Value:  0.701 | VLoss: 0.0277 | PLoss: -0.0012 | Ent: 0.1387
+[ 28.2%] Iter 2754/9765 | SPS:  1697 | Reward:  0.081 | Value:  0.742 | VLoss: 0.0238 | PLoss: -0.0067 | Ent: 0.1513
+[ 28.2%] Iter 2755/9765 | SPS:  1698 | Reward:  0.051 | Value:  0.701 | VLoss: 0.0217 | PLoss: -0.0094 | Ent: 0.1705
+[ 28.2%] Iter 2756/9765 | SPS:  1698 | Reward:  0.054 | Value:  0.656 | VLoss: 0.0275 | PLoss: -0.0037 | Ent: 0.1493
+[ 28.2%] Iter 2757/9765 | SPS:  1698 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0220 | PLoss: -0.0084 | Ent: 0.1777
+[ 28.2%] Iter 2758/9765 | SPS:  1699 | Reward:  0.070 | Value:  0.636 | VLoss: 0.0310 | PLoss: -0.0093 | Ent: 0.2044
+[ 28.3%] Iter 2759/9765 | SPS:  1699 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0307 | PLoss: -0.0030 | Ent: 0.1347
+[ 28.3%] Iter 2760/9765 | SPS:  1699 | Reward:  0.069 | Value:  0.707 | VLoss: 0.0290 | PLoss: -0.0015 | Ent: 0.1375
+[ 28.3%] Iter 2761/9765 | SPS:  1700 | Reward:  0.064 | Value:  0.700 | VLoss: 0.0256 | PLoss: -0.0136 | Ent: 0.1723
+[ 28.3%] Iter 2762/9765 | SPS:  1700 | Reward:  0.064 | Value:  0.654 | VLoss: 0.0251 | PLoss: -0.0064 | Ent: 0.1612
+[ 28.3%] Iter 2763/9765 | SPS:  1700 | Reward:  0.064 | Value:  0.628 | VLoss: 0.0273 | PLoss: -0.0092 | Ent: 0.1605
+[ 28.3%] Iter 2764/9765 | SPS:  1701 | Reward:  0.060 | Value:  0.644 | VLoss: 0.0277 | PLoss: -0.0097 | Ent: 0.1481
+[ 28.3%] Iter 2765/9765 | SPS:  1701 | Reward:  0.065 | Value:  0.686 | VLoss: 0.0307 | PLoss: -0.0159 | Ent: 0.1456
+[ 28.3%] Iter 2766/9765 | SPS:  1701 | Reward:  0.074 | Value:  0.703 | VLoss: 0.0306 | PLoss: -0.0084 | Ent: 0.2122
+[ 28.3%] Iter 2767/9765 | SPS:  1702 | Reward:  0.076 | Value:  0.680 | VLoss: 0.0273 | PLoss: -0.0070 | Ent: 0.1525
+[ 28.3%] Iter 2768/9765 | SPS:  1702 | Reward:  0.063 | Value:  0.726 | VLoss: 0.0257 | PLoss: -0.0086 | Ent: 0.1698
+[ 28.4%] Iter 2769/9765 | SPS:  1702 | Reward:  0.058 | Value:  0.751 | VLoss: 0.0223 | PLoss: -0.0102 | Ent: 0.1834
+[ 28.4%] Iter 2770/9765 | SPS:  1703 | Reward:  0.067 | Value:  0.744 | VLoss: 0.0248 | PLoss: -0.0038 | Ent: 0.1321
+[ 28.4%] Iter 2771/9765 | SPS:  1703 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0273 | PLoss: -0.0067 | Ent: 0.1888
+[ 28.4%] Iter 2772/9765 | SPS:  1703 | Reward:  0.061 | Value:  0.642 | VLoss: 0.0263 | PLoss: -0.0080 | Ent: 0.1369
+[ 28.4%] Iter 2773/9765 | SPS:  1704 | Reward:  0.076 | Value:  0.622 | VLoss: 0.0339 | PLoss: -0.0101 | Ent: 0.1671
+[ 28.4%] Iter 2774/9765 | SPS:  1704 | Reward:  0.058 | Value:  0.688 | VLoss: 0.0244 | PLoss: 0.0013 | Ent: 0.1424
+[ 28.4%] Iter 2775/9765 | SPS:  1704 | Reward:  0.055 | Value:  0.612 | VLoss: 0.0243 | PLoss: -0.0050 | Ent: 0.1856
+[ 28.4%] Iter 2776/9765 | SPS:  1705 | Reward:  0.050 | Value:  0.593 | VLoss: 0.0191 | PLoss: -0.0048 | Ent: 0.1842
+[ 28.4%] Iter 2777/9765 | SPS:  1705 | Reward:  0.069 | Value:  0.718 | VLoss: 0.0288 | PLoss: -0.0059 | Ent: 0.1506
+[ 28.4%] Iter 2778/9765 | SPS:  1705 | Reward:  0.069 | Value:  0.707 | VLoss: 0.0265 | PLoss: -0.0024 | Ent: 0.1638
+[ 28.5%] Iter 2779/9765 | SPS:  1706 | Reward:  0.060 | Value:  0.670 | VLoss: 0.0336 | PLoss: -0.0081 | Ent: 0.1563
+[ 28.5%] Iter 2780/9765 | SPS:  1706 | Reward:  0.073 | Value:  0.647 | VLoss: 0.0333 | PLoss: -0.0004 | Ent: 0.1634
+[ 28.5%] Iter 2781/9765 | SPS:  1706 | Reward:  0.065 | Value:  0.754 | VLoss: 0.0304 | PLoss: -0.0087 | Ent: 0.1382
+[ 28.5%] Iter 2782/9765 | SPS:  1707 | Reward:  0.075 | Value:  0.731 | VLoss: 0.0326 | PLoss: -0.0065 | Ent: 0.1723
+[ 28.5%] Iter 2783/9765 | SPS:  1707 | Reward:  0.063 | Value:  0.687 | VLoss: 0.0352 | PLoss: -0.0057 | Ent: 0.2166
+[ 28.5%] Iter 2784/9765 | SPS:  1707 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0258 | PLoss: -0.0095 | Ent: 0.1446
+[ 28.5%] Iter 2785/9765 | SPS:  1708 | Reward:  0.054 | Value:  0.705 | VLoss: 0.0240 | PLoss: -0.0053 | Ent: 0.1373
+[ 28.5%] Iter 2786/9765 | SPS:  1708 | Reward:  0.054 | Value:  0.662 | VLoss: 0.0243 | PLoss: -0.0102 | Ent: 0.1699
+[ 28.5%] Iter 2787/9765 | SPS:  1708 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0304 | PLoss: -0.0054 | Ent: 0.1686
+[ 28.6%] Iter 2788/9765 | SPS:  1709 | Reward:  0.057 | Value:  0.630 | VLoss: 0.0286 | PLoss: -0.0096 | Ent: 0.1362
+[ 28.6%] Iter 2789/9765 | SPS:  1709 | Reward:  0.069 | Value:  0.674 | VLoss: 0.0213 | PLoss: -0.0052 | Ent: 0.1822
+[ 28.6%] Iter 2790/9765 | SPS:  1709 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0322 | PLoss: -0.0079 | Ent: 0.1642
+[ 28.6%] Iter 2791/9765 | SPS:  1710 | Reward:  0.066 | Value:  0.686 | VLoss: 0.0262 | PLoss: -0.0062 | Ent: 0.1287
+[ 28.6%] Iter 2792/9765 | SPS:  1710 | Reward:  0.056 | Value:  0.653 | VLoss: 0.0292 | PLoss: -0.0117 | Ent: 0.1706
+[ 28.6%] Iter 2793/9765 | SPS:  1710 | Reward:  0.069 | Value:  0.582 | VLoss: 0.0369 | PLoss: -0.0040 | Ent: 0.1308
+[ 28.6%] Iter 2794/9765 | SPS:  1711 | Reward:  0.071 | Value:  0.630 | VLoss: 0.0298 | PLoss: -0.0045 | Ent: 0.1648
+[ 28.6%] Iter 2795/9765 | SPS:  1711 | Reward:  0.048 | Value:  0.643 | VLoss: 0.0228 | PLoss: -0.0027 | Ent: 0.1570
+[ 28.6%] Iter 2796/9765 | SPS:  1711 | Reward:  0.071 | Value:  0.672 | VLoss: 0.0383 | PLoss: -0.0044 | Ent: 0.1373
+[ 28.6%] Iter 2797/9765 | SPS:  1712 | Reward:  0.058 | Value:  0.688 | VLoss: 0.0249 | PLoss: -0.0023 | Ent: 0.1430
+[ 28.7%] Iter 2798/9765 | SPS:  1712 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0273 | PLoss: -0.0032 | Ent: 0.1765
+[ 28.7%] Iter 2799/9765 | SPS:  1712 | Reward:  0.073 | Value:  0.685 | VLoss: 0.0287 | PLoss: -0.0069 | Ent: 0.1686
+[ 28.7%] Iter 2800/9765 | SPS:  1713 | Reward:  0.075 | Value:  0.679 | VLoss: 0.0253 | PLoss: -0.0091 | Ent: 0.1632
+[ 28.7%] Iter 2801/9765 | SPS:  1713 | Reward:  0.059 | Value:  0.698 | VLoss: 0.0211 | PLoss: -0.0110 | Ent: 0.1953
+[ 28.7%] Iter 2802/9765 | SPS:  1713 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0326 | PLoss: -0.0152 | Ent: 0.1633
+[ 28.7%] Iter 2803/9765 | SPS:  1713 | Reward:  0.074 | Value:  0.622 | VLoss: 0.0357 | PLoss: -0.0092 | Ent: 0.1650
+[ 28.7%] Iter 2804/9765 | SPS:  1714 | Reward:  0.059 | Value:  0.759 | VLoss: 0.0299 | PLoss: -0.0139 | Ent: 0.2224
+[ 28.7%] Iter 2805/9765 | SPS:  1714 | Reward:  0.073 | Value:  0.695 | VLoss: 0.0285 | PLoss: -0.0118 | Ent: 0.1635
+[ 28.7%] Iter 2806/9765 | SPS:  1714 | Reward:  0.071 | Value:  0.684 | VLoss: 0.0296 | PLoss: -0.0077 | Ent: 0.1342
+[ 28.7%] Iter 2807/9765 | SPS:  1715 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0338 | PLoss: -0.0048 | Ent: 0.1532
+[ 28.8%] Iter 2808/9765 | SPS:  1715 | Reward:  0.069 | Value:  0.635 | VLoss: 0.0293 | PLoss: -0.0003 | Ent: 0.1712
+[ 28.8%] Iter 2809/9765 | SPS:  1715 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0267 | PLoss: -0.0052 | Ent: 0.1319
+[ 28.8%] Iter 2810/9765 | SPS:  1716 | Reward:  0.067 | Value:  0.662 | VLoss: 0.0276 | PLoss: -0.0037 | Ent: 0.1641
+[ 28.8%] Iter 2811/9765 | SPS:  1716 | Reward:  0.072 | Value:  0.703 | VLoss: 0.0329 | PLoss: -0.0072 | Ent: 0.1591
+[ 28.8%] Iter 2812/9765 | SPS:  1716 | Reward:  0.077 | Value:  0.725 | VLoss: 0.0326 | PLoss: -0.0081 | Ent: 0.1385
+[ 28.8%] Iter 2813/9765 | SPS:  1717 | Reward:  0.079 | Value:  0.729 | VLoss: 0.0297 | PLoss: -0.0119 | Ent: 0.1374
+[ 28.8%] Iter 2814/9765 | SPS:  1717 | Reward:  0.073 | Value:  0.709 | VLoss: 0.0241 | PLoss: -0.0044 | Ent: 0.1637
+[ 28.8%] Iter 2815/9765 | SPS:  1717 | Reward:  0.059 | Value:  0.728 | VLoss: 0.0337 | PLoss: -0.0045 | Ent: 0.1594
+[ 28.8%] Iter 2816/9765 | SPS:  1718 | Reward:  0.069 | Value:  0.659 | VLoss: 0.0284 | PLoss: -0.0097 | Ent: 0.1678
+[ 28.8%] Iter 2817/9765 | SPS:  1718 | Reward:  0.065 | Value:  0.689 | VLoss: 0.0218 | PLoss: -0.0017 | Ent: 0.1187
+[ 28.9%] Iter 2818/9765 | SPS:  1718 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0346 | PLoss: -0.0120 | Ent: 0.1738
+[ 28.9%] Iter 2819/9765 | SPS:  1719 | Reward:  0.057 | Value:  0.688 | VLoss: 0.0218 | PLoss: -0.0060 | Ent: 0.1892
+[ 28.9%] Iter 2820/9765 | SPS:  1719 | Reward:  0.065 | Value:  0.655 | VLoss: 0.0323 | PLoss: -0.0061 | Ent: 0.1904
+[ 28.9%] Iter 2821/9765 | SPS:  1719 | Reward:  0.050 | Value:  0.687 | VLoss: 0.0242 | PLoss: -0.0085 | Ent: 0.1325
+[ 28.9%] Iter 2822/9765 | SPS:  1720 | Reward:  0.065 | Value:  0.700 | VLoss: 0.0228 | PLoss: -0.0027 | Ent: 0.1518
+[ 28.9%] Iter 2823/9765 | SPS:  1720 | Reward:  0.051 | Value:  0.662 | VLoss: 0.0284 | PLoss: -0.0064 | Ent: 0.2028
+[ 28.9%] Iter 2824/9765 | SPS:  1720 | Reward:  0.057 | Value:  0.700 | VLoss: 0.0342 | PLoss: -0.0093 | Ent: 0.1792
+[ 28.9%] Iter 2825/9765 | SPS:  1721 | Reward:  0.065 | Value:  0.637 | VLoss: 0.0284 | PLoss: -0.0069 | Ent: 0.1434
+[ 28.9%] Iter 2826/9765 | SPS:  1721 | Reward:  0.067 | Value:  0.687 | VLoss: 0.0312 | PLoss: -0.0018 | Ent: 0.1483
+[ 29.0%] Iter 2827/9765 | SPS:  1721 | Reward:  0.075 | Value:  0.705 | VLoss: 0.0244 | PLoss: -0.0041 | Ent: 0.1377
+[ 29.0%] Iter 2828/9765 | SPS:  1722 | Reward:  0.074 | Value:  0.737 | VLoss: 0.0269 | PLoss: -0.0145 | Ent: 0.1707
+[ 29.0%] Iter 2829/9765 | SPS:  1722 | Reward:  0.074 | Value:  0.761 | VLoss: 0.0284 | PLoss: -0.0060 | Ent: 0.1131
+[ 29.0%] Iter 2830/9765 | SPS:  1722 | Reward:  0.062 | Value:  0.825 | VLoss: 0.0225 | PLoss: -0.0056 | Ent: 0.1542
+[ 29.0%] Iter 2831/9765 | SPS:  1723 | Reward:  0.072 | Value:  0.685 | VLoss: 0.0267 | PLoss: -0.0060 | Ent: 0.2031
+[ 29.0%] Iter 2832/9765 | SPS:  1723 | Reward:  0.071 | Value:  0.672 | VLoss: 0.0309 | PLoss: -0.0004 | Ent: 0.0918
+[ 29.0%] Iter 2833/9765 | SPS:  1723 | Reward:  0.083 | Value:  0.684 | VLoss: 0.0293 | PLoss: -0.0065 | Ent: 0.1718
+[ 29.0%] Iter 2834/9765 | SPS:  1723 | Reward:  0.071 | Value:  0.702 | VLoss: 0.0321 | PLoss: -0.0087 | Ent: 0.1802
+[ 29.0%] Iter 2835/9765 | SPS:  1724 | Reward:  0.061 | Value:  0.765 | VLoss: 0.0307 | PLoss: -0.0024 | Ent: 0.2203
+[ 29.0%] Iter 2836/9765 | SPS:  1724 | Reward:  0.055 | Value:  0.703 | VLoss: 0.0282 | PLoss: -0.0141 | Ent: 0.1459
+[ 29.1%] Iter 2837/9765 | SPS:  1724 | Reward:  0.067 | Value:  0.672 | VLoss: 0.0231 | PLoss: -0.0095 | Ent: 0.1566
+[ 29.1%] Iter 2838/9765 | SPS:  1725 | Reward:  0.051 | Value:  0.712 | VLoss: 0.0268 | PLoss: -0.0103 | Ent: 0.1612
+[ 29.1%] Iter 2839/9765 | SPS:  1725 | Reward:  0.048 | Value:  0.616 | VLoss: 0.0228 | PLoss: -0.0062 | Ent: 0.1156
+[ 29.1%] Iter 2840/9765 | SPS:  1725 | Reward:  0.056 | Value:  0.602 | VLoss: 0.0214 | PLoss: -0.0036 | Ent: 0.1332
+[ 29.1%] Iter 2841/9765 | SPS:  1726 | Reward:  0.060 | Value:  0.649 | VLoss: 0.0348 | PLoss: -0.0027 | Ent: 0.1516
+[ 29.1%] Iter 2842/9765 | SPS:  1726 | Reward:  0.050 | Value:  0.592 | VLoss: 0.0299 | PLoss: -0.0105 | Ent: 0.1333
+[ 29.1%] Iter 2843/9765 | SPS:  1726 | Reward:  0.061 | Value:  0.613 | VLoss: 0.0282 | PLoss: -0.0084 | Ent: 0.1587
+[ 29.1%] Iter 2844/9765 | SPS:  1727 | Reward:  0.051 | Value:  0.603 | VLoss: 0.0195 | PLoss: -0.0036 | Ent: 0.1402
+[ 29.1%] Iter 2845/9765 | SPS:  1727 | Reward:  0.059 | Value:  0.616 | VLoss: 0.0240 | PLoss: -0.0161 | Ent: 0.1549
+[ 29.1%] Iter 2846/9765 | SPS:  1727 | Reward:  0.060 | Value:  0.544 | VLoss: 0.0285 | PLoss: -0.0117 | Ent: 0.1726
+[ 29.2%] Iter 2847/9765 | SPS:  1728 | Reward:  0.055 | Value:  0.556 | VLoss: 0.0282 | PLoss: -0.0024 | Ent: 0.1519
+[ 29.2%] Iter 2848/9765 | SPS:  1728 | Reward:  0.058 | Value:  0.599 | VLoss: 0.0239 | PLoss: -0.0103 | Ent: 0.1564
+[ 29.2%] Iter 2849/9765 | SPS:  1728 | Reward:  0.075 | Value:  0.676 | VLoss: 0.0299 | PLoss: -0.0098 | Ent: 0.1722
+[ 29.2%] Iter 2850/9765 | SPS:  1729 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0336 | PLoss: -0.0078 | Ent: 0.1594
+[ 29.2%] Iter 2851/9765 | SPS:  1729 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0298 | PLoss: -0.0087 | Ent: 0.1365
+[ 29.2%] Iter 2852/9765 | SPS:  1729 | Reward:  0.070 | Value:  0.676 | VLoss: 0.0453 | PLoss: -0.0089 | Ent: 0.1651
+[ 29.2%] Iter 2853/9765 | SPS:  1730 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0215 | PLoss: -0.0052 | Ent: 0.1318
+[ 29.2%] Iter 2854/9765 | SPS:  1730 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0286 | PLoss: -0.0168 | Ent: 0.1639
+[ 29.2%] Iter 2855/9765 | SPS:  1730 | Reward:  0.067 | Value:  0.664 | VLoss: 0.0314 | PLoss: -0.0057 | Ent: 0.1646
+[ 29.2%] Iter 2856/9765 | SPS:  1730 | Reward:  0.077 | Value:  0.707 | VLoss: 0.0329 | PLoss: -0.0013 | Ent: 0.1429
+[ 29.3%] Iter 2857/9765 | SPS:  1731 | Reward:  0.064 | Value:  0.695 | VLoss: 0.0270 | PLoss: -0.0024 | Ent: 0.1756
+[ 29.3%] Iter 2858/9765 | SPS:  1731 | Reward:  0.054 | Value:  0.667 | VLoss: 0.0275 | PLoss: -0.0169 | Ent: 0.1603
+[ 29.3%] Iter 2859/9765 | SPS:  1731 | Reward:  0.071 | Value:  0.659 | VLoss: 0.0304 | PLoss: 0.0066 | Ent: 0.1507
+[ 29.3%] Iter 2860/9765 | SPS:  1732 | Reward:  0.064 | Value:  0.649 | VLoss: 0.0325 | PLoss: -0.0057 | Ent: 0.1550
+[ 29.3%] Iter 2861/9765 | SPS:  1732 | Reward:  0.072 | Value:  0.678 | VLoss: 0.0347 | PLoss: -0.0055 | Ent: 0.1615
+[ 29.3%] Iter 2862/9765 | SPS:  1732 | Reward:  0.055 | Value:  0.747 | VLoss: 0.0270 | PLoss: -0.0073 | Ent: 0.1739
+[ 29.3%] Iter 2863/9765 | SPS:  1733 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0248 | PLoss: -0.0059 | Ent: 0.1606
+[ 29.3%] Iter 2864/9765 | SPS:  1733 | Reward:  0.055 | Value:  0.639 | VLoss: 0.0294 | PLoss: -0.0024 | Ent: 0.2316
+[ 29.3%] Iter 2865/9765 | SPS:  1733 | Reward:  0.064 | Value:  0.666 | VLoss: 0.0278 | PLoss: -0.0061 | Ent: 0.1727
+[ 29.3%] Iter 2866/9765 | SPS:  1734 | Reward:  0.059 | Value:  0.664 | VLoss: 0.0261 | PLoss: -0.0120 | Ent: 0.1644
+[ 29.4%] Iter 2867/9765 | SPS:  1734 | Reward:  0.055 | Value:  0.650 | VLoss: 0.0245 | PLoss: -0.0129 | Ent: 0.1748
+[ 29.4%] Iter 2868/9765 | SPS:  1734 | Reward:  0.058 | Value:  0.689 | VLoss: 0.0258 | PLoss: -0.0049 | Ent: 0.1260
+[ 29.4%] Iter 2869/9765 | SPS:  1735 | Reward:  0.053 | Value:  0.678 | VLoss: 0.0194 | PLoss: -0.0055 | Ent: 0.1549
+[ 29.4%] Iter 2870/9765 | SPS:  1735 | Reward:  0.062 | Value:  0.664 | VLoss: 0.0263 | PLoss: -0.0093 | Ent: 0.1525
+[ 29.4%] Iter 2871/9765 | SPS:  1735 | Reward:  0.059 | Value:  0.601 | VLoss: 0.0256 | PLoss: -0.0076 | Ent: 0.1369
+[ 29.4%] Iter 2872/9765 | SPS:  1736 | Reward:  0.052 | Value:  0.599 | VLoss: 0.0324 | PLoss: -0.0121 | Ent: 0.1498
+[ 29.4%] Iter 2873/9765 | SPS:  1736 | Reward:  0.044 | Value:  0.529 | VLoss: 0.0336 | PLoss: -0.0065 | Ent: 0.1774
+[ 29.4%] Iter 2874/9765 | SPS:  1736 | Reward:  0.056 | Value:  0.589 | VLoss: 0.0191 | PLoss: -0.0177 | Ent: 0.1566
+[ 29.4%] Iter 2875/9765 | SPS:  1737 | Reward:  0.046 | Value:  0.593 | VLoss: 0.0246 | PLoss: -0.0096 | Ent: 0.1379
+[ 29.5%] Iter 2876/9765 | SPS:  1737 | Reward:  0.041 | Value:  0.518 | VLoss: 0.0185 | PLoss: -0.0052 | Ent: 0.1352
+[ 29.5%] Iter 2877/9765 | SPS:  1737 | Reward:  0.050 | Value:  0.605 | VLoss: 0.0235 | PLoss: -0.0134 | Ent: 0.1506
+[ 29.5%] Iter 2878/9765 | SPS:  1738 | Reward:  0.052 | Value:  0.604 | VLoss: 0.0171 | PLoss: -0.0110 | Ent: 0.1692
+[ 29.5%] Iter 2879/9765 | SPS:  1738 | Reward:  0.057 | Value:  0.608 | VLoss: 0.0306 | PLoss: -0.0054 | Ent: 0.1226
+[ 29.5%] Iter 2880/9765 | SPS:  1738 | Reward:  0.057 | Value:  0.651 | VLoss: 0.0272 | PLoss: -0.0013 | Ent: 0.1536
+[ 29.5%] Iter 2881/9765 | SPS:  1739 | Reward:  0.051 | Value:  0.650 | VLoss: 0.0304 | PLoss: -0.0000 | Ent: 0.1802
+[ 29.5%] Iter 2882/9765 | SPS:  1739 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0284 | PLoss: -0.0070 | Ent: 0.1810
+[ 29.5%] Iter 2883/9765 | SPS:  1739 | Reward:  0.061 | Value:  0.665 | VLoss: 0.0241 | PLoss: -0.0057 | Ent: 0.1432
+[ 29.5%] Iter 2884/9765 | SPS:  1740 | Reward:  0.049 | Value:  0.592 | VLoss: 0.0330 | PLoss: -0.0058 | Ent: 0.1313
+[ 29.5%] Iter 2885/9765 | SPS:  1740 | Reward:  0.056 | Value:  0.658 | VLoss: 0.0215 | PLoss: -0.0056 | Ent: 0.1587
+[ 29.6%] Iter 2886/9765 | SPS:  1740 | Reward:  0.062 | Value:  0.617 | VLoss: 0.0346 | PLoss: -0.0127 | Ent: 0.1692
+[ 29.6%] Iter 2887/9765 | SPS:  1740 | Reward:  0.063 | Value:  0.678 | VLoss: 0.0236 | PLoss: -0.0035 | Ent: 0.1281
+[ 29.6%] Iter 2888/9765 | SPS:  1741 | Reward:  0.052 | Value:  0.745 | VLoss: 0.0192 | PLoss: -0.0109 | Ent: 0.1728
+[ 29.6%] Iter 2889/9765 | SPS:  1741 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0281 | PLoss: -0.0060 | Ent: 0.1676
+[ 29.6%] Iter 2890/9765 | SPS:  1741 | Reward:  0.057 | Value:  0.668 | VLoss: 0.0229 | PLoss: -0.0041 | Ent: 0.2055
+[ 29.6%] Iter 2891/9765 | SPS:  1742 | Reward:  0.049 | Value:  0.640 | VLoss: 0.0237 | PLoss: -0.0100 | Ent: 0.1697
+[ 29.6%] Iter 2892/9765 | SPS:  1742 | Reward:  0.079 | Value:  0.670 | VLoss: 0.0329 | PLoss: -0.0054 | Ent: 0.1718
+[ 29.6%] Iter 2893/9765 | SPS:  1742 | Reward:  0.060 | Value:  0.690 | VLoss: 0.0273 | PLoss: -0.0170 | Ent: 0.1933
+[ 29.6%] Iter 2894/9765 | SPS:  1743 | Reward:  0.059 | Value:  0.675 | VLoss: 0.0202 | PLoss: -0.0116 | Ent: 0.1989
+[ 29.6%] Iter 2895/9765 | SPS:  1743 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0297 | PLoss: -0.0030 | Ent: 0.1517
+[ 29.7%] Iter 2896/9765 | SPS:  1743 | Reward:  0.077 | Value:  0.706 | VLoss: 0.0299 | PLoss: -0.0128 | Ent: 0.1672
+[ 29.7%] Iter 2897/9765 | SPS:  1744 | Reward:  0.059 | Value:  0.745 | VLoss: 0.0211 | PLoss: -0.0100 | Ent: 0.1777
+[ 29.7%] Iter 2898/9765 | SPS:  1744 | Reward:  0.062 | Value:  0.655 | VLoss: 0.0358 | PLoss: -0.0097 | Ent: 0.1828
+[ 29.7%] Iter 2899/9765 | SPS:  1744 | Reward:  0.070 | Value:  0.714 | VLoss: 0.0385 | PLoss: -0.0082 | Ent: 0.1234
+[ 29.7%] Iter 2900/9765 | SPS:  1745 | Reward:  0.068 | Value:  0.760 | VLoss: 0.0227 | PLoss: -0.0048 | Ent: 0.1387
+[ 29.7%] Iter 2901/9765 | SPS:  1745 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0270 | PLoss: -0.0032 | Ent: 0.1469
+[ 29.7%] Iter 2902/9765 | SPS:  1745 | Reward:  0.063 | Value:  0.606 | VLoss: 0.0365 | PLoss: -0.0092 | Ent: 0.1830
+[ 29.7%] Iter 2903/9765 | SPS:  1746 | Reward:  0.056 | Value:  0.669 | VLoss: 0.0250 | PLoss: -0.0078 | Ent: 0.1677
+[ 29.7%] Iter 2904/9765 | SPS:  1746 | Reward:  0.057 | Value:  0.711 | VLoss: 0.0318 | PLoss: -0.0062 | Ent: 0.1738
+[ 29.7%] Iter 2905/9765 | SPS:  1746 | Reward:  0.062 | Value:  0.601 | VLoss: 0.0271 | PLoss: -0.0112 | Ent: 0.1686
+[ 29.8%] Iter 2906/9765 | SPS:  1747 | Reward:  0.042 | Value:  0.654 | VLoss: 0.0278 | PLoss: -0.0036 | Ent: 0.1386
+[ 29.8%] Iter 2907/9765 | SPS:  1747 | Reward:  0.035 | Value:  0.656 | VLoss: 0.0163 | PLoss: -0.0026 | Ent: 0.1359
+[ 29.8%] Iter 2908/9765 | SPS:  1747 | Reward:  0.044 | Value:  0.698 | VLoss: 0.0210 | PLoss: -0.0095 | Ent: 0.1370
+[ 29.8%] Iter 2909/9765 | SPS:  1748 | Reward:  0.062 | Value:  0.620 | VLoss: 0.0226 | PLoss: -0.0057 | Ent: 0.1493
+[ 29.8%] Iter 2910/9765 | SPS:  1748 | Reward:  0.050 | Value:  0.648 | VLoss: 0.0195 | PLoss: -0.0042 | Ent: 0.1417
+[ 29.8%] Iter 2911/9765 | SPS:  1748 | Reward:  0.040 | Value:  0.677 | VLoss: 0.0157 | PLoss: -0.0128 | Ent: 0.1730
+[ 29.8%] Iter 2912/9765 | SPS:  1749 | Reward:  0.062 | Value:  0.610 | VLoss: 0.0285 | PLoss: -0.0046 | Ent: 0.1885
+[ 29.8%] Iter 2913/9765 | SPS:  1749 | Reward:  0.047 | Value:  0.626 | VLoss: 0.0168 | PLoss: -0.0100 | Ent: 0.1631
+[ 29.8%] Iter 2914/9765 | SPS:  1749 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0245 | PLoss: -0.0130 | Ent: 0.1980
+[ 29.9%] Iter 2915/9765 | SPS:  1750 | Reward:  0.052 | Value:  0.663 | VLoss: 0.0177 | PLoss: -0.0056 | Ent: 0.1960
+[ 29.9%] Iter 2916/9765 | SPS:  1750 | Reward:  0.074 | Value:  0.710 | VLoss: 0.0277 | PLoss: -0.0081 | Ent: 0.1638
+[ 29.9%] Iter 2917/9765 | SPS:  1750 | Reward:  0.046 | Value:  0.747 | VLoss: 0.0289 | PLoss: -0.0131 | Ent: 0.1335
+[ 29.9%] Iter 2918/9765 | SPS:  1750 | Reward:  0.058 | Value:  0.656 | VLoss: 0.0273 | PLoss: -0.0097 | Ent: 0.1736
+[ 29.9%] Iter 2919/9765 | SPS:  1751 | Reward:  0.056 | Value:  0.671 | VLoss: 0.0187 | PLoss: -0.0047 | Ent: 0.1530
+[ 29.9%] Iter 2920/9765 | SPS:  1751 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0221 | PLoss: -0.0071 | Ent: 0.1668
+[ 29.9%] Iter 2921/9765 | SPS:  1751 | Reward:  0.082 | Value:  0.688 | VLoss: 0.0315 | PLoss: -0.0087 | Ent: 0.2080
+[ 29.9%] Iter 2922/9765 | SPS:  1752 | Reward:  0.065 | Value:  0.654 | VLoss: 0.0280 | PLoss: -0.0067 | Ent: 0.1326
+[ 29.9%] Iter 2923/9765 | SPS:  1752 | Reward:  0.050 | Value:  0.676 | VLoss: 0.0254 | PLoss: -0.0096 | Ent: 0.1558
+[ 29.9%] Iter 2924/9765 | SPS:  1752 | Reward:  0.057 | Value:  0.616 | VLoss: 0.0299 | PLoss: -0.0114 | Ent: 0.1766
+[ 30.0%] Iter 2925/9765 | SPS:  1753 | Reward:  0.055 | Value:  0.661 | VLoss: 0.0222 | PLoss: -0.0057 | Ent: 0.1626
+[ 30.0%] Iter 2926/9765 | SPS:  1753 | Reward:  0.065 | Value:  0.648 | VLoss: 0.0375 | PLoss: -0.0051 | Ent: 0.1780
+[ 30.0%] Iter 2927/9765 | SPS:  1753 | Reward:  0.057 | Value:  0.711 | VLoss: 0.0283 | PLoss: -0.0053 | Ent: 0.1681
+[ 30.0%] Iter 2928/9765 | SPS:  1754 | Reward:  0.064 | Value:  0.682 | VLoss: 0.0244 | PLoss: -0.0109 | Ent: 0.1934
+[ 30.0%] Iter 2929/9765 | SPS:  1754 | Reward:  0.077 | Value:  0.677 | VLoss: 0.0354 | PLoss: -0.0067 | Ent: 0.1778
+[ 30.0%] Iter 2930/9765 | SPS:  1754 | Reward:  0.049 | Value:  0.683 | VLoss: 0.0220 | PLoss: -0.0042 | Ent: 0.2126
+[ 30.0%] Iter 2931/9765 | SPS:  1755 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0256 | PLoss: -0.0081 | Ent: 0.1698
+[ 30.0%] Iter 2932/9765 | SPS:  1755 | Reward:  0.061 | Value:  0.665 | VLoss: 0.0292 | PLoss: -0.0098 | Ent: 0.1660
+[ 30.0%] Iter 2933/9765 | SPS:  1755 | Reward:  0.061 | Value:  0.628 | VLoss: 0.0296 | PLoss: -0.0120 | Ent: 0.1798
+[ 30.0%] Iter 2934/9765 | SPS:  1755 | Reward:  0.059 | Value:  0.690 | VLoss: 0.0267 | PLoss: -0.0099 | Ent: 0.1075
+[ 30.1%] Iter 2935/9765 | SPS:  1756 | Reward:  0.051 | Value:  0.635 | VLoss: 0.0257 | PLoss: -0.0060 | Ent: 0.1944
+[ 30.1%] Iter 2936/9765 | SPS:  1756 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0249 | PLoss: -0.0035 | Ent: 0.1809
+[ 30.1%] Iter 2937/9765 | SPS:  1756 | Reward:  0.060 | Value:  0.651 | VLoss: 0.0280 | PLoss: -0.0058 | Ent: 0.1504
+[ 30.1%] Iter 2938/9765 | SPS:  1757 | Reward:  0.072 | Value:  0.627 | VLoss: 0.0382 | PLoss: -0.0085 | Ent: 0.1851
+[ 30.1%] Iter 2939/9765 | SPS:  1757 | Reward:  0.056 | Value:  0.647 | VLoss: 0.0298 | PLoss: -0.0031 | Ent: 0.1716
+[ 30.1%] Iter 2940/9765 | SPS:  1757 | Reward:  0.065 | Value:  0.632 | VLoss: 0.0319 | PLoss: -0.0120 | Ent: 0.1353
+[ 30.1%] Iter 2941/9765 | SPS:  1758 | Reward:  0.069 | Value:  0.670 | VLoss: 0.0301 | PLoss: -0.0077 | Ent: 0.1474
+[ 30.1%] Iter 2942/9765 | SPS:  1758 | Reward:  0.058 | Value:  0.683 | VLoss: 0.0259 | PLoss: -0.0091 | Ent: 0.1640
+[ 30.1%] Iter 2943/9765 | SPS:  1758 | Reward:  0.053 | Value:  0.692 | VLoss: 0.0256 | PLoss: -0.0062 | Ent: 0.1768
+[ 30.1%] Iter 2944/9765 | SPS:  1759 | Reward:  0.061 | Value:  0.670 | VLoss: 0.0314 | PLoss: -0.0075 | Ent: 0.1707
+[ 30.2%] Iter 2945/9765 | SPS:  1759 | Reward:  0.065 | Value:  0.618 | VLoss: 0.0334 | PLoss: -0.0149 | Ent: 0.1777
+[ 30.2%] Iter 2946/9765 | SPS:  1759 | Reward:  0.063 | Value:  0.699 | VLoss: 0.0216 | PLoss: -0.0047 | Ent: 0.1867
+[ 30.2%] Iter 2947/9765 | SPS:  1760 | Reward:  0.062 | Value:  0.699 | VLoss: 0.0347 | PLoss: -0.0118 | Ent: 0.1631
+[ 30.2%] Iter 2948/9765 | SPS:  1760 | Reward:  0.061 | Value:  0.756 | VLoss: 0.0353 | PLoss: -0.0087 | Ent: 0.1650
+[ 30.2%] Iter 2949/9765 | SPS:  1760 | Reward:  0.057 | Value:  0.686 | VLoss: 0.0343 | PLoss: -0.0069 | Ent: 0.1597
+[ 30.2%] Iter 2950/9765 | SPS:  1760 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0203 | PLoss: -0.0038 | Ent: 0.1416
+[ 30.2%] Iter 2951/9765 | SPS:  1761 | Reward:  0.057 | Value:  0.733 | VLoss: 0.0238 | PLoss: -0.0073 | Ent: 0.1219
+[ 30.2%] Iter 2952/9765 | SPS:  1761 | Reward:  0.068 | Value:  0.657 | VLoss: 0.0299 | PLoss: -0.0089 | Ent: 0.1690
+[ 30.2%] Iter 2953/9765 | SPS:  1761 | Reward:  0.060 | Value:  0.657 | VLoss: 0.0266 | PLoss: -0.0119 | Ent: 0.1649
+[ 30.3%] Iter 2954/9765 | SPS:  1762 | Reward:  0.058 | Value:  0.702 | VLoss: 0.0274 | PLoss: -0.0081 | Ent: 0.1655
+[ 30.3%] Iter 2955/9765 | SPS:  1762 | Reward:  0.077 | Value:  0.707 | VLoss: 0.0302 | PLoss: -0.0115 | Ent: 0.1626
+[ 30.3%] Iter 2956/9765 | SPS:  1762 | Reward:  0.055 | Value:  0.623 | VLoss: 0.0384 | PLoss: -0.0130 | Ent: 0.1808
+[ 30.3%] Iter 2957/9765 | SPS:  1763 | Reward:  0.073 | Value:  0.685 | VLoss: 0.0264 | PLoss: -0.0059 | Ent: 0.1485
+[ 30.3%] Iter 2958/9765 | SPS:  1763 | Reward:  0.072 | Value:  0.708 | VLoss: 0.0323 | PLoss: -0.0123 | Ent: 0.1684
+[ 30.3%] Iter 2959/9765 | SPS:  1763 | Reward:  0.071 | Value:  0.743 | VLoss: 0.0203 | PLoss: -0.0060 | Ent: 0.1368
+[ 30.3%] Iter 2960/9765 | SPS:  1764 | Reward:  0.068 | Value:  0.713 | VLoss: 0.0158 | PLoss: -0.0063 | Ent: 0.1549
+[ 30.3%] Iter 2961/9765 | SPS:  1764 | Reward:  0.067 | Value:  0.759 | VLoss: 0.0279 | PLoss: -0.0129 | Ent: 0.1645
+[ 30.3%] Iter 2962/9765 | SPS:  1764 | Reward:  0.052 | Value:  0.672 | VLoss: 0.0192 | PLoss: -0.0126 | Ent: 0.1552
+[ 30.3%] Iter 2963/9765 | SPS:  1764 | Reward:  0.047 | Value:  0.603 | VLoss: 0.0317 | PLoss: -0.0116 | Ent: 0.1409
+[ 30.4%] Iter 2964/9765 | SPS:  1765 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0258 | PLoss: -0.0045 | Ent: 0.1638
+[ 30.4%] Iter 2965/9765 | SPS:  1765 | Reward:  0.055 | Value:  0.710 | VLoss: 0.0159 | PLoss: -0.0082 | Ent: 0.1624
+[ 30.4%] Iter 2966/9765 | SPS:  1765 | Reward:  0.051 | Value:  0.682 | VLoss: 0.0210 | PLoss: -0.0047 | Ent: 0.1658
+[ 30.4%] Iter 2967/9765 | SPS:  1766 | Reward:  0.056 | Value:  0.669 | VLoss: 0.0257 | PLoss: -0.0046 | Ent: 0.2003
+[ 30.4%] Iter 2968/9765 | SPS:  1766 | Reward:  0.065 | Value:  0.670 | VLoss: 0.0348 | PLoss: -0.0093 | Ent: 0.2185
+[ 30.4%] Iter 2969/9765 | SPS:  1766 | Reward:  0.066 | Value:  0.634 | VLoss: 0.0279 | PLoss: -0.0051 | Ent: 0.2171
+[ 30.4%] Iter 2970/9765 | SPS:  1767 | Reward:  0.069 | Value:  0.612 | VLoss: 0.0314 | PLoss: -0.0006 | Ent: 0.1715
+[ 30.4%] Iter 2971/9765 | SPS:  1767 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0247 | PLoss: -0.0090 | Ent: 0.1383
+[ 30.4%] Iter 2972/9765 | SPS:  1767 | Reward:  0.069 | Value:  0.646 | VLoss: 0.0358 | PLoss: -0.0036 | Ent: 0.1786
+[ 30.4%] Iter 2973/9765 | SPS:  1768 | Reward:  0.052 | Value:  0.638 | VLoss: 0.0237 | PLoss: -0.0066 | Ent: 0.2011
+[ 30.5%] Iter 2974/9765 | SPS:  1768 | Reward:  0.052 | Value:  0.641 | VLoss: 0.0307 | PLoss: -0.0127 | Ent: 0.1773
+[ 30.5%] Iter 2975/9765 | SPS:  1768 | Reward:  0.053 | Value:  0.569 | VLoss: 0.0287 | PLoss: -0.0056 | Ent: 0.1910
+[ 30.5%] Iter 2976/9765 | SPS:  1769 | Reward:  0.047 | Value:  0.604 | VLoss: 0.0214 | PLoss: -0.0131 | Ent: 0.1792
+[ 30.5%] Iter 2977/9765 | SPS:  1769 | Reward:  0.052 | Value:  0.579 | VLoss: 0.0226 | PLoss: -0.0099 | Ent: 0.1680
+[ 30.5%] Iter 2978/9765 | SPS:  1769 | Reward:  0.066 | Value:  0.677 | VLoss: 0.0295 | PLoss: -0.0030 | Ent: 0.1629
+[ 30.5%] Iter 2979/9765 | SPS:  1770 | Reward:  0.053 | Value:  0.613 | VLoss: 0.0291 | PLoss: -0.0101 | Ent: 0.1856
+[ 30.5%] Iter 2980/9765 | SPS:  1770 | Reward:  0.077 | Value:  0.732 | VLoss: 0.0332 | PLoss: -0.0145 | Ent: 0.1571
+[ 30.5%] Iter 2981/9765 | SPS:  1770 | Reward:  0.077 | Value:  0.710 | VLoss: 0.0330 | PLoss: -0.0158 | Ent: 0.1984
+[ 30.5%] Iter 2982/9765 | SPS:  1770 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0281 | PLoss: -0.0085 | Ent: 0.1756
+[ 30.5%] Iter 2983/9765 | SPS:  1771 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0263 | PLoss: -0.0089 | Ent: 0.1639
+[ 30.6%] Iter 2984/9765 | SPS:  1771 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0248 | PLoss: -0.0054 | Ent: 0.2185
+[ 30.6%] Iter 2985/9765 | SPS:  1771 | Reward:  0.053 | Value:  0.691 | VLoss: 0.0272 | PLoss: -0.0068 | Ent: 0.2317
+[ 30.6%] Iter 2986/9765 | SPS:  1772 | Reward:  0.067 | Value:  0.644 | VLoss: 0.0297 | PLoss: -0.0103 | Ent: 0.1508
+[ 30.6%] Iter 2987/9765 | SPS:  1772 | Reward:  0.064 | Value:  0.701 | VLoss: 0.0282 | PLoss: -0.0076 | Ent: 0.1473
+[ 30.6%] Iter 2988/9765 | SPS:  1772 | Reward:  0.065 | Value:  0.720 | VLoss: 0.0228 | PLoss: 0.0003 | Ent: 0.1541
+[ 30.6%] Iter 2989/9765 | SPS:  1773 | Reward:  0.048 | Value:  0.678 | VLoss: 0.0186 | PLoss: -0.0110 | Ent: 0.1788
+[ 30.6%] Iter 2990/9765 | SPS:  1773 | Reward:  0.063 | Value:  0.653 | VLoss: 0.0233 | PLoss: -0.0020 | Ent: 0.1878
+[ 30.6%] Iter 2991/9765 | SPS:  1773 | Reward:  0.059 | Value:  0.627 | VLoss: 0.0249 | PLoss: -0.0104 | Ent: 0.1835
+[ 30.6%] Iter 2992/9765 | SPS:  1773 | Reward:  0.076 | Value:  0.709 | VLoss: 0.0343 | PLoss: -0.0078 | Ent: 0.1786
+[ 30.7%] Iter 2993/9765 | SPS:  1774 | Reward:  0.067 | Value:  0.699 | VLoss: 0.0321 | PLoss: -0.0018 | Ent: 0.1781
+[ 30.7%] Iter 2994/9765 | SPS:  1774 | Reward:  0.055 | Value:  0.682 | VLoss: 0.0247 | PLoss: -0.0027 | Ent: 0.1807
+[ 30.7%] Iter 2995/9765 | SPS:  1774 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0259 | PLoss: -0.0050 | Ent: 0.1619
+[ 30.7%] Iter 2996/9765 | SPS:  1775 | Reward:  0.055 | Value:  0.712 | VLoss: 0.0223 | PLoss: -0.0046 | Ent: 0.2100
+[ 30.7%] Iter 2997/9765 | SPS:  1775 | Reward:  0.053 | Value:  0.689 | VLoss: 0.0309 | PLoss: -0.0094 | Ent: 0.1690
+[ 30.7%] Iter 2998/9765 | SPS:  1775 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0252 | PLoss: -0.0057 | Ent: 0.1643
+[ 30.7%] Iter 2999/9765 | SPS:  1776 | Reward:  0.058 | Value:  0.648 | VLoss: 0.0234 | PLoss: -0.0124 | Ent: 0.1571
+[ 30.7%] Iter 3000/9765 | SPS:  1776 | Reward:  0.054 | Value:  0.664 | VLoss: 0.0231 | PLoss: -0.0029 | Ent: 0.2004
+[ 30.7%] Iter 3001/9765 | SPS:  1776 | Reward:  0.055 | Value:  0.593 | VLoss: 0.0276 | PLoss: -0.0044 | Ent: 0.1592
+[ 30.7%] Iter 3002/9765 | SPS:  1777 | Reward:  0.061 | Value:  0.644 | VLoss: 0.0246 | PLoss: -0.0052 | Ent: 0.1370
+[ 30.8%] Iter 3003/9765 | SPS:  1777 | Reward:  0.064 | Value:  0.644 | VLoss: 0.0255 | PLoss: -0.0061 | Ent: 0.1791
+[ 30.8%] Iter 3004/9765 | SPS:  1777 | Reward:  0.055 | Value:  0.684 | VLoss: 0.0275 | PLoss: -0.0027 | Ent: 0.1453
+[ 30.8%] Iter 3005/9765 | SPS:  1777 | Reward:  0.071 | Value:  0.646 | VLoss: 0.0262 | PLoss: -0.0045 | Ent: 0.1550
+[ 30.8%] Iter 3006/9765 | SPS:  1778 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0200 | PLoss: -0.0067 | Ent: 0.1873
+[ 30.8%] Iter 3007/9765 | SPS:  1778 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0234 | PLoss: -0.0081 | Ent: 0.2243
+[ 30.8%] Iter 3008/9765 | SPS:  1778 | Reward:  0.057 | Value:  0.652 | VLoss: 0.0244 | PLoss: -0.0045 | Ent: 0.1994
+[ 30.8%] Iter 3009/9765 | SPS:  1779 | Reward:  0.053 | Value:  0.642 | VLoss: 0.0256 | PLoss: -0.0116 | Ent: 0.2605
+[ 30.8%] Iter 3010/9765 | SPS:  1779 | Reward:  0.053 | Value:  0.651 | VLoss: 0.0324 | PLoss: -0.0143 | Ent: 0.2156
+[ 30.8%] Iter 3011/9765 | SPS:  1779 | Reward:  0.061 | Value:  0.626 | VLoss: 0.0259 | PLoss: -0.0089 | Ent: 0.1926
+[ 30.8%] Iter 3012/9765 | SPS:  1780 | Reward:  0.042 | Value:  0.562 | VLoss: 0.0257 | PLoss: -0.0097 | Ent: 0.2072
+[ 30.9%] Iter 3013/9765 | SPS:  1780 | Reward:  0.051 | Value:  0.615 | VLoss: 0.0229 | PLoss: -0.0143 | Ent: 0.1872
+[ 30.9%] Iter 3014/9765 | SPS:  1780 | Reward:  0.068 | Value:  0.661 | VLoss: 0.0248 | PLoss: -0.0110 | Ent: 0.1570
+[ 30.9%] Iter 3015/9765 | SPS:  1781 | Reward:  0.064 | Value:  0.600 | VLoss: 0.0346 | PLoss: -0.0059 | Ent: 0.1943
+[ 30.9%] Iter 3016/9765 | SPS:  1781 | Reward:  0.041 | Value:  0.643 | VLoss: 0.0237 | PLoss: -0.0098 | Ent: 0.2059
+[ 30.9%] Iter 3017/9765 | SPS:  1781 | Reward:  0.046 | Value:  0.664 | VLoss: 0.0277 | PLoss: -0.0076 | Ent: 0.1680
+[ 30.9%] Iter 3018/9765 | SPS:  1782 | Reward:  0.063 | Value:  0.615 | VLoss: 0.0299 | PLoss: -0.0061 | Ent: 0.1725
+[ 30.9%] Iter 3019/9765 | SPS:  1782 | Reward:  0.050 | Value:  0.550 | VLoss: 0.0316 | PLoss: -0.0005 | Ent: 0.1644
+[ 30.9%] Iter 3020/9765 | SPS:  1782 | Reward:  0.065 | Value:  0.637 | VLoss: 0.0255 | PLoss: -0.0045 | Ent: 0.1346
+[ 30.9%] Iter 3021/9765 | SPS:  1782 | Reward:  0.060 | Value:  0.567 | VLoss: 0.0281 | PLoss: -0.0052 | Ent: 0.2202
+[ 30.9%] Iter 3022/9765 | SPS:  1783 | Reward:  0.050 | Value:  0.635 | VLoss: 0.0213 | PLoss: -0.0021 | Ent: 0.1826
+[ 31.0%] Iter 3023/9765 | SPS:  1783 | Reward:  0.064 | Value:  0.610 | VLoss: 0.0302 | PLoss: -0.0089 | Ent: 0.1983
+[ 31.0%] Iter 3024/9765 | SPS:  1783 | Reward:  0.077 | Value:  0.682 | VLoss: 0.0227 | PLoss: -0.0070 | Ent: 0.1690
+[ 31.0%] Iter 3025/9765 | SPS:  1784 | Reward:  0.053 | Value:  0.744 | VLoss: 0.0209 | PLoss: -0.0103 | Ent: 0.1893
+[ 31.0%] Iter 3026/9765 | SPS:  1784 | Reward:  0.063 | Value:  0.673 | VLoss: 0.0265 | PLoss: -0.0127 | Ent: 0.2025
+[ 31.0%] Iter 3027/9765 | SPS:  1784 | Reward:  0.078 | Value:  0.766 | VLoss: 0.0251 | PLoss: -0.0062 | Ent: 0.1899
+[ 31.0%] Iter 3028/9765 | SPS:  1785 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0266 | PLoss: -0.0032 | Ent: 0.1660
+[ 31.0%] Iter 3029/9765 | SPS:  1785 | Reward:  0.057 | Value:  0.737 | VLoss: 0.0236 | PLoss: -0.0163 | Ent: 0.1977
+[ 31.0%] Iter 3030/9765 | SPS:  1785 | Reward:  0.055 | Value:  0.642 | VLoss: 0.0344 | PLoss: -0.0129 | Ent: 0.1583
+[ 31.0%] Iter 3031/9765 | SPS:  1785 | Reward:  0.059 | Value:  0.646 | VLoss: 0.0281 | PLoss: -0.0055 | Ent: 0.1544
+[ 31.0%] Iter 3032/9765 | SPS:  1786 | Reward:  0.068 | Value:  0.687 | VLoss: 0.0253 | PLoss: -0.0067 | Ent: 0.1837
+[ 31.1%] Iter 3033/9765 | SPS:  1786 | Reward:  0.059 | Value:  0.741 | VLoss: 0.0179 | PLoss: -0.0007 | Ent: 0.1219
+[ 31.1%] Iter 3034/9765 | SPS:  1786 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0290 | PLoss: -0.0075 | Ent: 0.1639
+[ 31.1%] Iter 3035/9765 | SPS:  1787 | Reward:  0.073 | Value:  0.713 | VLoss: 0.0278 | PLoss: -0.0094 | Ent: 0.1503
+[ 31.1%] Iter 3036/9765 | SPS:  1787 | Reward:  0.051 | Value:  0.665 | VLoss: 0.0273 | PLoss: -0.0088 | Ent: 0.1598
+[ 31.1%] Iter 3037/9765 | SPS:  1787 | Reward:  0.065 | Value:  0.698 | VLoss: 0.0209 | PLoss: -0.0074 | Ent: 0.1643
+[ 31.1%] Iter 3038/9765 | SPS:  1788 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0266 | PLoss: -0.0111 | Ent: 0.2030
+[ 31.1%] Iter 3039/9765 | SPS:  1788 | Reward:  0.064 | Value:  0.749 | VLoss: 0.0208 | PLoss: -0.0114 | Ent: 0.1930
+[ 31.1%] Iter 3040/9765 | SPS:  1788 | Reward:  0.062 | Value:  0.618 | VLoss: 0.0319 | PLoss: -0.0061 | Ent: 0.1509
+[ 31.1%] Iter 3041/9765 | SPS:  1788 | Reward:  0.078 | Value:  0.688 | VLoss: 0.0282 | PLoss: -0.0042 | Ent: 0.1305
+[ 31.2%] Iter 3042/9765 | SPS:  1789 | Reward:  0.059 | Value:  0.719 | VLoss: 0.0255 | PLoss: -0.0156 | Ent: 0.1934
+[ 31.2%] Iter 3043/9765 | SPS:  1789 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0286 | PLoss: -0.0018 | Ent: 0.1524
+[ 31.2%] Iter 3044/9765 | SPS:  1789 | Reward:  0.061 | Value:  0.743 | VLoss: 0.0179 | PLoss: -0.0102 | Ent: 0.1516
+[ 31.2%] Iter 3045/9765 | SPS:  1790 | Reward:  0.057 | Value:  0.666 | VLoss: 0.0282 | PLoss: -0.0058 | Ent: 0.1526
+[ 31.2%] Iter 3046/9765 | SPS:  1790 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0233 | PLoss: -0.0037 | Ent: 0.1773
+[ 31.2%] Iter 3047/9765 | SPS:  1790 | Reward:  0.071 | Value:  0.687 | VLoss: 0.0396 | PLoss: -0.0068 | Ent: 0.1802
+[ 31.2%] Iter 3048/9765 | SPS:  1791 | Reward:  0.063 | Value:  0.693 | VLoss: 0.0240 | PLoss: -0.0039 | Ent: 0.1560
+[ 31.2%] Iter 3049/9765 | SPS:  1791 | Reward:  0.065 | Value:  0.668 | VLoss: 0.0255 | PLoss: -0.0134 | Ent: 0.1769
+[ 31.2%] Iter 3050/9765 | SPS:  1791 | Reward:  0.054 | Value:  0.664 | VLoss: 0.0281 | PLoss: -0.0093 | Ent: 0.1575
+[ 31.2%] Iter 3051/9765 | SPS:  1791 | Reward:  0.069 | Value:  0.685 | VLoss: 0.0266 | PLoss: -0.0047 | Ent: 0.1610
+[ 31.3%] Iter 3052/9765 | SPS:  1792 | Reward:  0.048 | Value:  0.678 | VLoss: 0.0208 | PLoss: -0.0035 | Ent: 0.1204
+[ 31.3%] Iter 3053/9765 | SPS:  1792 | Reward:  0.043 | Value:  0.671 | VLoss: 0.0234 | PLoss: 0.0016 | Ent: 0.1491
+[ 31.3%] Iter 3054/9765 | SPS:  1792 | Reward:  0.049 | Value:  0.598 | VLoss: 0.0204 | PLoss: -0.0066 | Ent: 0.1103
+[ 31.3%] Iter 3055/9765 | SPS:  1793 | Reward:  0.053 | Value:  0.633 | VLoss: 0.0210 | PLoss: -0.0057 | Ent: 0.1461
+[ 31.3%] Iter 3056/9765 | SPS:  1793 | Reward:  0.051 | Value:  0.745 | VLoss: 0.0203 | PLoss: 0.0011 | Ent: 0.1359
+[ 31.3%] Iter 3057/9765 | SPS:  1793 | Reward:  0.054 | Value:  0.646 | VLoss: 0.0283 | PLoss: -0.0083 | Ent: 0.1585
+[ 31.3%] Iter 3058/9765 | SPS:  1794 | Reward:  0.065 | Value:  0.650 | VLoss: 0.0276 | PLoss: -0.0130 | Ent: 0.1129
+[ 31.3%] Iter 3059/9765 | SPS:  1794 | Reward:  0.044 | Value:  0.668 | VLoss: 0.0224 | PLoss: 0.0014 | Ent: 0.1368
+[ 31.3%] Iter 3060/9765 | SPS:  1794 | Reward:  0.048 | Value:  0.617 | VLoss: 0.0166 | PLoss: -0.0082 | Ent: 0.1898
+[ 31.3%] Iter 3061/9765 | SPS:  1795 | Reward:  0.051 | Value:  0.608 | VLoss: 0.0228 | PLoss: -0.0105 | Ent: 0.1674
+[ 31.4%] Iter 3062/9765 | SPS:  1795 | Reward:  0.060 | Value:  0.612 | VLoss: 0.0150 | PLoss: -0.0067 | Ent: 0.1599
+[ 31.4%] Iter 3063/9765 | SPS:  1795 | Reward:  0.058 | Value:  0.672 | VLoss: 0.0279 | PLoss: -0.0021 | Ent: 0.1482
+[ 31.4%] Iter 3064/9765 | SPS:  1795 | Reward:  0.052 | Value:  0.660 | VLoss: 0.0264 | PLoss: -0.0129 | Ent: 0.1747
+[ 31.4%] Iter 3065/9765 | SPS:  1796 | Reward:  0.065 | Value:  0.597 | VLoss: 0.0299 | PLoss: -0.0115 | Ent: 0.1399
+[ 31.4%] Iter 3066/9765 | SPS:  1796 | Reward:  0.078 | Value:  0.714 | VLoss: 0.0359 | PLoss: -0.0131 | Ent: 0.1653
+[ 31.4%] Iter 3067/9765 | SPS:  1796 | Reward:  0.059 | Value:  0.718 | VLoss: 0.0313 | PLoss: -0.0009 | Ent: 0.1912
+[ 31.4%] Iter 3068/9765 | SPS:  1797 | Reward:  0.064 | Value:  0.703 | VLoss: 0.0235 | PLoss: -0.0166 | Ent: 0.1904
+[ 31.4%] Iter 3069/9765 | SPS:  1797 | Reward:  0.066 | Value:  0.724 | VLoss: 0.0298 | PLoss: -0.0067 | Ent: 0.1934
+[ 31.4%] Iter 3070/9765 | SPS:  1797 | Reward:  0.070 | Value:  0.687 | VLoss: 0.0296 | PLoss: -0.0035 | Ent: 0.1434
+[ 31.4%] Iter 3071/9765 | SPS:  1797 | Reward:  0.060 | Value:  0.691 | VLoss: 0.0238 | PLoss: -0.0162 | Ent: 0.1805
+[ 31.5%] Iter 3072/9765 | SPS:  1798 | Reward:  0.069 | Value:  0.664 | VLoss: 0.0241 | PLoss: -0.0031 | Ent: 0.1614
+[ 31.5%] Iter 3073/9765 | SPS:  1798 | Reward:  0.057 | Value:  0.732 | VLoss: 0.0303 | PLoss: -0.0127 | Ent: 0.1541
+[ 31.5%] Iter 3074/9765 | SPS:  1798 | Reward:  0.067 | Value:  0.679 | VLoss: 0.0223 | PLoss: -0.0129 | Ent: 0.1792
+[ 31.5%] Iter 3075/9765 | SPS:  1799 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0295 | PLoss: -0.0056 | Ent: 0.1601
+[ 31.5%] Iter 3076/9765 | SPS:  1799 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0332 | PLoss: -0.0074 | Ent: 0.1558
+[ 31.5%] Iter 3077/9765 | SPS:  1799 | Reward:  0.072 | Value:  0.722 | VLoss: 0.0328 | PLoss: -0.0050 | Ent: 0.1758
+[ 31.5%] Iter 3078/9765 | SPS:  1800 | Reward:  0.063 | Value:  0.709 | VLoss: 0.0237 | PLoss: -0.0091 | Ent: 0.1559
+[ 31.5%] Iter 3079/9765 | SPS:  1800 | Reward:  0.046 | Value:  0.613 | VLoss: 0.0302 | PLoss: -0.0106 | Ent: 0.2163
+[ 31.5%] Iter 3080/9765 | SPS:  1800 | Reward:  0.055 | Value:  0.676 | VLoss: 0.0259 | PLoss: -0.0008 | Ent: 0.1607
+[ 31.6%] Iter 3081/9765 | SPS:  1800 | Reward:  0.055 | Value:  0.717 | VLoss: 0.0208 | PLoss: -0.0059 | Ent: 0.1891
+[ 31.6%] Iter 3082/9765 | SPS:  1801 | Reward:  0.070 | Value:  0.717 | VLoss: 0.0212 | PLoss: -0.0124 | Ent: 0.1731
+[ 31.6%] Iter 3083/9765 | SPS:  1801 | Reward:  0.066 | Value:  0.745 | VLoss: 0.0213 | PLoss: -0.0079 | Ent: 0.1635
+[ 31.6%] Iter 3084/9765 | SPS:  1801 | Reward:  0.071 | Value:  0.749 | VLoss: 0.0266 | PLoss: -0.0166 | Ent: 0.1974
+[ 31.6%] Iter 3085/9765 | SPS:  1802 | Reward:  0.073 | Value:  0.727 | VLoss: 0.0276 | PLoss: -0.0113 | Ent: 0.2022
+[ 31.6%] Iter 3086/9765 | SPS:  1802 | Reward:  0.052 | Value:  0.717 | VLoss: 0.0271 | PLoss: -0.0019 | Ent: 0.1656
+[ 31.6%] Iter 3087/9765 | SPS:  1802 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0281 | PLoss: -0.0032 | Ent: 0.1900
+[ 31.6%] Iter 3088/9765 | SPS:  1803 | Reward:  0.060 | Value:  0.762 | VLoss: 0.0177 | PLoss: -0.0031 | Ent: 0.1748
+[ 31.6%] Iter 3089/9765 | SPS:  1803 | Reward:  0.062 | Value:  0.734 | VLoss: 0.0264 | PLoss: -0.0093 | Ent: 0.1559
+[ 31.6%] Iter 3090/9765 | SPS:  1803 | Reward:  0.072 | Value:  0.678 | VLoss: 0.0249 | PLoss: -0.0101 | Ent: 0.1897
+[ 31.7%] Iter 3091/9765 | SPS:  1803 | Reward:  0.056 | Value:  0.624 | VLoss: 0.0269 | PLoss: -0.0072 | Ent: 0.1730
+[ 31.7%] Iter 3092/9765 | SPS:  1804 | Reward:  0.046 | Value:  0.615 | VLoss: 0.0126 | PLoss: -0.0082 | Ent: 0.1392
+[ 31.7%] Iter 3093/9765 | SPS:  1804 | Reward:  0.047 | Value:  0.581 | VLoss: 0.0238 | PLoss: -0.0143 | Ent: 0.1688
+[ 31.7%] Iter 3094/9765 | SPS:  1804 | Reward:  0.054 | Value:  0.678 | VLoss: 0.0321 | PLoss: -0.0095 | Ent: 0.2079
+[ 31.7%] Iter 3095/9765 | SPS:  1805 | Reward:  0.079 | Value:  0.663 | VLoss: 0.0286 | PLoss: -0.0091 | Ent: 0.1625
+[ 31.7%] Iter 3096/9765 | SPS:  1805 | Reward:  0.067 | Value:  0.742 | VLoss: 0.0269 | PLoss: -0.0031 | Ent: 0.1718
+[ 31.7%] Iter 3097/9765 | SPS:  1805 | Reward:  0.068 | Value:  0.708 | VLoss: 0.0304 | PLoss: 0.0004 | Ent: 0.1627
+[ 31.7%] Iter 3098/9765 | SPS:  1806 | Reward:  0.052 | Value:  0.677 | VLoss: 0.0209 | PLoss: -0.0089 | Ent: 0.1637
+[ 31.7%] Iter 3099/9765 | SPS:  1806 | Reward:  0.067 | Value:  0.661 | VLoss: 0.0290 | PLoss: -0.0072 | Ent: 0.1762
+[ 31.7%] Iter 3100/9765 | SPS:  1806 | Reward:  0.073 | Value:  0.702 | VLoss: 0.0215 | PLoss: -0.0064 | Ent: 0.1762
+[ 31.8%] Iter 3101/9765 | SPS:  1806 | Reward:  0.060 | Value:  0.688 | VLoss: 0.0307 | PLoss: -0.0083 | Ent: 0.1918
+[ 31.8%] Iter 3102/9765 | SPS:  1807 | Reward:  0.070 | Value:  0.683 | VLoss: 0.0245 | PLoss: -0.0078 | Ent: 0.1676
+[ 31.8%] Iter 3103/9765 | SPS:  1807 | Reward:  0.066 | Value:  0.671 | VLoss: 0.0271 | PLoss: -0.0070 | Ent: 0.1966
+[ 31.8%] Iter 3104/9765 | SPS:  1807 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0194 | PLoss: -0.0118 | Ent: 0.1864
+[ 31.8%] Iter 3105/9765 | SPS:  1808 | Reward:  0.056 | Value:  0.702 | VLoss: 0.0237 | PLoss: -0.0135 | Ent: 0.1574
+[ 31.8%] Iter 3106/9765 | SPS:  1808 | Reward:  0.079 | Value:  0.672 | VLoss: 0.0310 | PLoss: -0.0144 | Ent: 0.1938
+[ 31.8%] Iter 3107/9765 | SPS:  1808 | Reward:  0.048 | Value:  0.580 | VLoss: 0.0253 | PLoss: -0.0051 | Ent: 0.2164
+[ 31.8%] Iter 3108/9765 | SPS:  1808 | Reward:  0.045 | Value:  0.637 | VLoss: 0.0172 | PLoss: -0.0123 | Ent: 0.1859
+[ 31.8%] Iter 3109/9765 | SPS:  1809 | Reward:  0.049 | Value:  0.656 | VLoss: 0.0238 | PLoss: -0.0052 | Ent: 0.1827
+[ 31.8%] Iter 3110/9765 | SPS:  1809 | Reward:  0.051 | Value:  0.733 | VLoss: 0.0219 | PLoss: -0.0053 | Ent: 0.2037
+[ 31.9%] Iter 3111/9765 | SPS:  1809 | Reward:  0.066 | Value:  0.640 | VLoss: 0.0316 | PLoss: -0.0082 | Ent: 0.1440
+[ 31.9%] Iter 3112/9765 | SPS:  1810 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0251 | PLoss: -0.0068 | Ent: 0.2074
+[ 31.9%] Iter 3113/9765 | SPS:  1810 | Reward:  0.076 | Value:  0.641 | VLoss: 0.0275 | PLoss: -0.0087 | Ent: 0.2115
+[ 31.9%] Iter 3114/9765 | SPS:  1810 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0283 | PLoss: -0.0106 | Ent: 0.1829
+[ 31.9%] Iter 3115/9765 | SPS:  1811 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0288 | PLoss: -0.0161 | Ent: 0.2055
+[ 31.9%] Iter 3116/9765 | SPS:  1811 | Reward:  0.062 | Value:  0.655 | VLoss: 0.0320 | PLoss: -0.0059 | Ent: 0.1828
+[ 31.9%] Iter 3117/9765 | SPS:  1811 | Reward:  0.049 | Value:  0.653 | VLoss: 0.0251 | PLoss: -0.0094 | Ent: 0.2041
+[ 31.9%] Iter 3118/9765 | SPS:  1811 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0245 | PLoss: -0.0104 | Ent: 0.1803
+[ 31.9%] Iter 3119/9765 | SPS:  1812 | Reward:  0.070 | Value:  0.684 | VLoss: 0.0316 | PLoss: -0.0085 | Ent: 0.1469
+[ 32.0%] Iter 3120/9765 | SPS:  1812 | Reward:  0.056 | Value:  0.671 | VLoss: 0.0209 | PLoss: -0.0105 | Ent: 0.1879
+[ 32.0%] Iter 3121/9765 | SPS:  1812 | Reward:  0.045 | Value:  0.596 | VLoss: 0.0214 | PLoss: -0.0036 | Ent: 0.1809
+[ 32.0%] Iter 3122/9765 | SPS:  1813 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0270 | PLoss: -0.0071 | Ent: 0.1697
+[ 32.0%] Iter 3123/9765 | SPS:  1813 | Reward:  0.053 | Value:  0.678 | VLoss: 0.0214 | PLoss: -0.0091 | Ent: 0.1726
+[ 32.0%] Iter 3124/9765 | SPS:  1813 | Reward:  0.061 | Value:  0.665 | VLoss: 0.0234 | PLoss: -0.0046 | Ent: 0.2007
+[ 32.0%] Iter 3125/9765 | SPS:  1814 | Reward:  0.049 | Value:  0.666 | VLoss: 0.0232 | PLoss: -0.0129 | Ent: 0.2055
+[ 32.0%] Iter 3126/9765 | SPS:  1814 | Reward:  0.062 | Value:  0.647 | VLoss: 0.0292 | PLoss: -0.0103 | Ent: 0.2155
+[ 32.0%] Iter 3127/9765 | SPS:  1814 | Reward:  0.058 | Value:  0.712 | VLoss: 0.0317 | PLoss: -0.0064 | Ent: 0.1728
+[ 32.0%] Iter 3128/9765 | SPS:  1814 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0207 | PLoss: -0.0083 | Ent: 0.1848
+[ 32.0%] Iter 3129/9765 | SPS:  1815 | Reward:  0.053 | Value:  0.692 | VLoss: 0.0228 | PLoss: -0.0050 | Ent: 0.1728
+[ 32.1%] Iter 3130/9765 | SPS:  1815 | Reward:  0.070 | Value:  0.726 | VLoss: 0.0249 | PLoss: -0.0086 | Ent: 0.1802
+[ 32.1%] Iter 3131/9765 | SPS:  1815 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0343 | PLoss: -0.0050 | Ent: 0.1637
+[ 32.1%] Iter 3132/9765 | SPS:  1816 | Reward:  0.065 | Value:  0.695 | VLoss: 0.0227 | PLoss: -0.0121 | Ent: 0.1650
+[ 32.1%] Iter 3133/9765 | SPS:  1816 | Reward:  0.068 | Value:  0.752 | VLoss: 0.0294 | PLoss: -0.0082 | Ent: 0.1338
+[ 32.1%] Iter 3134/9765 | SPS:  1816 | Reward:  0.060 | Value:  0.752 | VLoss: 0.0225 | PLoss: -0.0111 | Ent: 0.1622
+[ 32.1%] Iter 3135/9765 | SPS:  1816 | Reward:  0.065 | Value:  0.755 | VLoss: 0.0314 | PLoss: -0.0098 | Ent: 0.1946
+[ 32.1%] Iter 3136/9765 | SPS:  1817 | Reward:  0.068 | Value:  0.696 | VLoss: 0.0257 | PLoss: -0.0085 | Ent: 0.1694
+[ 32.1%] Iter 3137/9765 | SPS:  1817 | Reward:  0.059 | Value:  0.707 | VLoss: 0.0224 | PLoss: -0.0081 | Ent: 0.2370
+[ 32.1%] Iter 3138/9765 | SPS:  1817 | Reward:  0.050 | Value:  0.664 | VLoss: 0.0274 | PLoss: -0.0085 | Ent: 0.1594
+[ 32.1%] Iter 3139/9765 | SPS:  1818 | Reward:  0.057 | Value:  0.629 | VLoss: 0.0278 | PLoss: -0.0079 | Ent: 0.2033
+[ 32.2%] Iter 3140/9765 | SPS:  1818 | Reward:  0.054 | Value:  0.660 | VLoss: 0.0261 | PLoss: -0.0025 | Ent: 0.1879
+[ 32.2%] Iter 3141/9765 | SPS:  1818 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0249 | PLoss: -0.0072 | Ent: 0.1580
+[ 32.2%] Iter 3142/9765 | SPS:  1818 | Reward:  0.059 | Value:  0.656 | VLoss: 0.0245 | PLoss: -0.0100 | Ent: 0.2018
+[ 32.2%] Iter 3143/9765 | SPS:  1819 | Reward:  0.085 | Value:  0.642 | VLoss: 0.0303 | PLoss: -0.0125 | Ent: 0.1528
+[ 32.2%] Iter 3144/9765 | SPS:  1819 | Reward:  0.048 | Value:  0.667 | VLoss: 0.0241 | PLoss: -0.0025 | Ent: 0.2396
+[ 32.2%] Iter 3145/9765 | SPS:  1819 | Reward:  0.059 | Value:  0.678 | VLoss: 0.0263 | PLoss: -0.0083 | Ent: 0.1587
+[ 32.2%] Iter 3146/9765 | SPS:  1820 | Reward:  0.044 | Value:  0.686 | VLoss: 0.0243 | PLoss: -0.0097 | Ent: 0.1363
+[ 32.2%] Iter 3147/9765 | SPS:  1820 | Reward:  0.053 | Value:  0.576 | VLoss: 0.0290 | PLoss: -0.0113 | Ent: 0.1672
+[ 32.2%] Iter 3148/9765 | SPS:  1820 | Reward:  0.068 | Value:  0.673 | VLoss: 0.0249 | PLoss: -0.0124 | Ent: 0.1224
+[ 32.2%] Iter 3149/9765 | SPS:  1820 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0300 | PLoss: -0.0102 | Ent: 0.1906
+[ 32.3%] Iter 3150/9765 | SPS:  1821 | Reward:  0.072 | Value:  0.729 | VLoss: 0.0235 | PLoss: -0.0091 | Ent: 0.1687
+[ 32.3%] Iter 3151/9765 | SPS:  1821 | Reward:  0.051 | Value:  0.692 | VLoss: 0.0246 | PLoss: -0.0040 | Ent: 0.1659
+[ 32.3%] Iter 3152/9765 | SPS:  1821 | Reward:  0.061 | Value:  0.660 | VLoss: 0.0240 | PLoss: -0.0036 | Ent: 0.1687
+[ 32.3%] Iter 3153/9765 | SPS:  1822 | Reward:  0.050 | Value:  0.604 | VLoss: 0.0284 | PLoss: -0.0163 | Ent: 0.1886
+[ 32.3%] Iter 3154/9765 | SPS:  1822 | Reward:  0.058 | Value:  0.655 | VLoss: 0.0226 | PLoss: -0.0057 | Ent: 0.1736
+[ 32.3%] Iter 3155/9765 | SPS:  1822 | Reward:  0.065 | Value:  0.687 | VLoss: 0.0245 | PLoss: -0.0057 | Ent: 0.1982
+[ 32.3%] Iter 3156/9765 | SPS:  1822 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0299 | PLoss: -0.0064 | Ent: 0.1498
+[ 32.3%] Iter 3157/9765 | SPS:  1823 | Reward:  0.072 | Value:  0.646 | VLoss: 0.0336 | PLoss: -0.0066 | Ent: 0.1974
+[ 32.3%] Iter 3158/9765 | SPS:  1823 | Reward:  0.046 | Value:  0.671 | VLoss: 0.0256 | PLoss: -0.0140 | Ent: 0.1680
+[ 32.4%] Iter 3159/9765 | SPS:  1823 | Reward:  0.042 | Value:  0.666 | VLoss: 0.0221 | PLoss: -0.0060 | Ent: 0.1776
+[ 32.4%] Iter 3160/9765 | SPS:  1824 | Reward:  0.050 | Value:  0.662 | VLoss: 0.0270 | PLoss: -0.0081 | Ent: 0.1426
+[ 32.4%] Iter 3161/9765 | SPS:  1824 | Reward:  0.070 | Value:  0.658 | VLoss: 0.0330 | PLoss: -0.0135 | Ent: 0.1915
+[ 32.4%] Iter 3162/9765 | SPS:  1824 | Reward:  0.061 | Value:  0.653 | VLoss: 0.0294 | PLoss: -0.0012 | Ent: 0.2029
+[ 32.4%] Iter 3163/9765 | SPS:  1825 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0221 | PLoss: -0.0096 | Ent: 0.2052
+[ 32.4%] Iter 3164/9765 | SPS:  1825 | Reward:  0.059 | Value:  0.685 | VLoss: 0.0268 | PLoss: -0.0091 | Ent: 0.2290
+[ 32.4%] Iter 3165/9765 | SPS:  1825 | Reward:  0.039 | Value:  0.617 | VLoss: 0.0259 | PLoss: -0.0017 | Ent: 0.2426
+[ 32.4%] Iter 3166/9765 | SPS:  1825 | Reward:  0.052 | Value:  0.653 | VLoss: 0.0183 | PLoss: -0.0014 | Ent: 0.1862
+[ 32.4%] Iter 3167/9765 | SPS:  1826 | Reward:  0.065 | Value:  0.612 | VLoss: 0.0246 | PLoss: -0.0046 | Ent: 0.1916
+[ 32.4%] Iter 3168/9765 | SPS:  1826 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0256 | PLoss: -0.0046 | Ent: 0.1622
+[ 32.5%] Iter 3169/9765 | SPS:  1826 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0228 | PLoss: -0.0022 | Ent: 0.2126
+[ 32.5%] Iter 3170/9765 | SPS:  1827 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0272 | PLoss: -0.0077 | Ent: 0.1662
+[ 32.5%] Iter 3171/9765 | SPS:  1827 | Reward:  0.072 | Value:  0.703 | VLoss: 0.0311 | PLoss: -0.0080 | Ent: 0.1953
+[ 32.5%] Iter 3172/9765 | SPS:  1827 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0340 | PLoss: -0.0098 | Ent: 0.1789
+[ 32.5%] Iter 3173/9765 | SPS:  1827 | Reward:  0.063 | Value:  0.690 | VLoss: 0.0280 | PLoss: -0.0052 | Ent: 0.1400
+[ 32.5%] Iter 3174/9765 | SPS:  1828 | Reward:  0.055 | Value:  0.689 | VLoss: 0.0193 | PLoss: -0.0050 | Ent: 0.1615
+[ 32.5%] Iter 3175/9765 | SPS:  1828 | Reward:  0.058 | Value:  0.680 | VLoss: 0.0331 | PLoss: -0.0145 | Ent: 0.1978
+[ 32.5%] Iter 3176/9765 | SPS:  1828 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0249 | PLoss: -0.0039 | Ent: 0.1216
+[ 32.5%] Iter 3177/9765 | SPS:  1829 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0280 | PLoss: -0.0130 | Ent: 0.1961
+[ 32.5%] Iter 3178/9765 | SPS:  1829 | Reward:  0.058 | Value:  0.727 | VLoss: 0.0285 | PLoss: -0.0008 | Ent: 0.1811
+[ 32.6%] Iter 3179/9765 | SPS:  1829 | Reward:  0.074 | Value:  0.706 | VLoss: 0.0300 | PLoss: -0.0140 | Ent: 0.1628
+[ 32.6%] Iter 3180/9765 | SPS:  1829 | Reward:  0.054 | Value:  0.658 | VLoss: 0.0257 | PLoss: -0.0086 | Ent: 0.1608
+[ 32.6%] Iter 3181/9765 | SPS:  1830 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0155 | PLoss: -0.0158 | Ent: 0.1602
+[ 32.6%] Iter 3182/9765 | SPS:  1830 | Reward:  0.058 | Value:  0.632 | VLoss: 0.0290 | PLoss: -0.0088 | Ent: 0.2026
+[ 32.6%] Iter 3183/9765 | SPS:  1830 | Reward:  0.048 | Value:  0.731 | VLoss: 0.0169 | PLoss: -0.0112 | Ent: 0.2326
+[ 32.6%] Iter 3184/9765 | SPS:  1831 | Reward:  0.060 | Value:  0.696 | VLoss: 0.0322 | PLoss: -0.0071 | Ent: 0.1682
+[ 32.6%] Iter 3185/9765 | SPS:  1831 | Reward:  0.063 | Value:  0.715 | VLoss: 0.0198 | PLoss: -0.0114 | Ent: 0.1437
+[ 32.6%] Iter 3186/9765 | SPS:  1831 | Reward:  0.076 | Value:  0.682 | VLoss: 0.0301 | PLoss: -0.0120 | Ent: 0.1947
+[ 32.6%] Iter 3187/9765 | SPS:  1831 | Reward:  0.061 | Value:  0.671 | VLoss: 0.0332 | PLoss: -0.0090 | Ent: 0.1549
+[ 32.6%] Iter 3188/9765 | SPS:  1832 | Reward:  0.058 | Value:  0.683 | VLoss: 0.0206 | PLoss: -0.0046 | Ent: 0.1474
+[ 32.7%] Iter 3189/9765 | SPS:  1832 | Reward:  0.059 | Value:  0.639 | VLoss: 0.0303 | PLoss: -0.0029 | Ent: 0.1220
+[ 32.7%] Iter 3190/9765 | SPS:  1832 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0278 | PLoss: -0.0074 | Ent: 0.2134
+[ 32.7%] Iter 3191/9765 | SPS:  1833 | Reward:  0.057 | Value:  0.771 | VLoss: 0.0271 | PLoss: -0.0158 | Ent: 0.1832
+[ 32.7%] Iter 3192/9765 | SPS:  1833 | Reward:  0.068 | Value:  0.621 | VLoss: 0.0292 | PLoss: -0.0109 | Ent: 0.1827
+[ 32.7%] Iter 3193/9765 | SPS:  1833 | Reward:  0.055 | Value:  0.623 | VLoss: 0.0224 | PLoss: -0.0086 | Ent: 0.1856
+[ 32.7%] Iter 3194/9765 | SPS:  1833 | Reward:  0.053 | Value:  0.673 | VLoss: 0.0266 | PLoss: -0.0101 | Ent: 0.1812
+[ 32.7%] Iter 3195/9765 | SPS:  1834 | Reward:  0.071 | Value:  0.690 | VLoss: 0.0298 | PLoss: -0.0068 | Ent: 0.1715
+[ 32.7%] Iter 3196/9765 | SPS:  1834 | Reward:  0.055 | Value:  0.734 | VLoss: 0.0215 | PLoss: -0.0072 | Ent: 0.2026
+[ 32.7%] Iter 3197/9765 | SPS:  1834 | Reward:  0.066 | Value:  0.702 | VLoss: 0.0289 | PLoss: -0.0062 | Ent: 0.1961
+[ 32.7%] Iter 3198/9765 | SPS:  1835 | Reward:  0.056 | Value:  0.628 | VLoss: 0.0294 | PLoss: -0.0054 | Ent: 0.2033
+[ 32.8%] Iter 3199/9765 | SPS:  1835 | Reward:  0.051 | Value:  0.705 | VLoss: 0.0234 | PLoss: -0.0049 | Ent: 0.1851
+[ 32.8%] Iter 3200/9765 | SPS:  1835 | Reward:  0.056 | Value:  0.656 | VLoss: 0.0295 | PLoss: -0.0042 | Ent: 0.1628
+[ 32.8%] Iter 3201/9765 | SPS:  1835 | Reward:  0.063 | Value:  0.687 | VLoss: 0.0237 | PLoss: -0.0130 | Ent: 0.1968
+[ 32.8%] Iter 3202/9765 | SPS:  1836 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0213 | PLoss: -0.0070 | Ent: 0.1780
+[ 32.8%] Iter 3203/9765 | SPS:  1836 | Reward:  0.065 | Value:  0.720 | VLoss: 0.0280 | PLoss: -0.0022 | Ent: 0.1624
+[ 32.8%] Iter 3204/9765 | SPS:  1836 | Reward:  0.065 | Value:  0.730 | VLoss: 0.0217 | PLoss: -0.0128 | Ent: 0.1687
+[ 32.8%] Iter 3205/9765 | SPS:  1837 | Reward:  0.067 | Value:  0.678 | VLoss: 0.0247 | PLoss: -0.0043 | Ent: 0.1824
+[ 32.8%] Iter 3206/9765 | SPS:  1837 | Reward:  0.061 | Value:  0.617 | VLoss: 0.0326 | PLoss: -0.0024 | Ent: 0.1996
+[ 32.8%] Iter 3207/9765 | SPS:  1837 | Reward:  0.067 | Value:  0.751 | VLoss: 0.0262 | PLoss: -0.0042 | Ent: 0.1763
+[ 32.9%] Iter 3208/9765 | SPS:  1837 | Reward:  0.066 | Value:  0.681 | VLoss: 0.0299 | PLoss: -0.0101 | Ent: 0.1855
+[ 32.9%] Iter 3209/9765 | SPS:  1838 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0269 | PLoss: -0.0113 | Ent: 0.2050
+[ 32.9%] Iter 3210/9765 | SPS:  1838 | Reward:  0.061 | Value:  0.668 | VLoss: 0.0247 | PLoss: -0.0120 | Ent: 0.2052
+[ 32.9%] Iter 3211/9765 | SPS:  1838 | Reward:  0.050 | Value:  0.723 | VLoss: 0.0169 | PLoss: 0.0000 | Ent: 0.1505
+[ 32.9%] Iter 3212/9765 | SPS:  1839 | Reward:  0.071 | Value:  0.718 | VLoss: 0.0294 | PLoss: -0.0115 | Ent: 0.1695
+[ 32.9%] Iter 3213/9765 | SPS:  1839 | Reward:  0.051 | Value:  0.718 | VLoss: 0.0277 | PLoss: -0.0095 | Ent: 0.1818
+[ 32.9%] Iter 3214/9765 | SPS:  1839 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0279 | PLoss: -0.0064 | Ent: 0.1889
+[ 32.9%] Iter 3215/9765 | SPS:  1839 | Reward:  0.072 | Value:  0.720 | VLoss: 0.0195 | PLoss: -0.0080 | Ent: 0.1529
+[ 32.9%] Iter 3216/9765 | SPS:  1840 | Reward:  0.068 | Value:  0.663 | VLoss: 0.0340 | PLoss: -0.0072 | Ent: 0.1411
+[ 32.9%] Iter 3217/9765 | SPS:  1840 | Reward:  0.062 | Value:  0.667 | VLoss: 0.0302 | PLoss: -0.0040 | Ent: 0.1498
+[ 33.0%] Iter 3218/9765 | SPS:  1840 | Reward:  0.038 | Value:  0.619 | VLoss: 0.0252 | PLoss: -0.0147 | Ent: 0.1419
+[ 33.0%] Iter 3219/9765 | SPS:  1841 | Reward:  0.054 | Value:  0.673 | VLoss: 0.0267 | PLoss: -0.0119 | Ent: 0.1536
+[ 33.0%] Iter 3220/9765 | SPS:  1841 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0248 | PLoss: -0.0059 | Ent: 0.1682
+[ 33.0%] Iter 3221/9765 | SPS:  1841 | Reward:  0.058 | Value:  0.700 | VLoss: 0.0261 | PLoss: -0.0034 | Ent: 0.1606
+[ 33.0%] Iter 3222/9765 | SPS:  1841 | Reward:  0.056 | Value:  0.643 | VLoss: 0.0294 | PLoss: -0.0114 | Ent: 0.1771
+[ 33.0%] Iter 3223/9765 | SPS:  1842 | Reward:  0.052 | Value:  0.672 | VLoss: 0.0218 | PLoss: -0.0083 | Ent: 0.1739
+[ 33.0%] Iter 3224/9765 | SPS:  1842 | Reward:  0.057 | Value:  0.653 | VLoss: 0.0229 | PLoss: -0.0090 | Ent: 0.1531
+[ 33.0%] Iter 3225/9765 | SPS:  1842 | Reward:  0.067 | Value:  0.625 | VLoss: 0.0194 | PLoss: 0.0026 | Ent: 0.1371
+[ 33.0%] Iter 3226/9765 | SPS:  1843 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0204 | PLoss: -0.0087 | Ent: 0.1556
+[ 33.0%] Iter 3227/9765 | SPS:  1843 | Reward:  0.050 | Value:  0.649 | VLoss: 0.0175 | PLoss: -0.0160 | Ent: 0.1774
+[ 33.1%] Iter 3228/9765 | SPS:  1843 | Reward:  0.050 | Value:  0.622 | VLoss: 0.0207 | PLoss: -0.0039 | Ent: 0.1582
+[ 33.1%] Iter 3229/9765 | SPS:  1843 | Reward:  0.053 | Value:  0.643 | VLoss: 0.0196 | PLoss: -0.0067 | Ent: 0.1770
+[ 33.1%] Iter 3230/9765 | SPS:  1844 | Reward:  0.050 | Value:  0.622 | VLoss: 0.0192 | PLoss: -0.0040 | Ent: 0.1156
+[ 33.1%] Iter 3231/9765 | SPS:  1844 | Reward:  0.035 | Value:  0.587 | VLoss: 0.0116 | PLoss: -0.0081 | Ent: 0.1271
+[ 33.1%] Iter 3232/9765 | SPS:  1844 | Reward:  0.038 | Value:  0.585 | VLoss: 0.0144 | PLoss: -0.0052 | Ent: 0.1302
+[ 33.1%] Iter 3233/9765 | SPS:  1845 | Reward:  0.045 | Value:  0.587 | VLoss: 0.0155 | PLoss: -0.0087 | Ent: 0.1557
+[ 33.1%] Iter 3234/9765 | SPS:  1845 | Reward:  0.034 | Value:  0.571 | VLoss: 0.0140 | PLoss: -0.0002 | Ent: 0.1100
+[ 33.1%] Iter 3235/9765 | SPS:  1845 | Reward:  0.059 | Value:  0.549 | VLoss: 0.0195 | PLoss: -0.0018 | Ent: 0.1522
+[ 33.1%] Iter 3236/9765 | SPS:  1846 | Reward:  0.043 | Value:  0.549 | VLoss: 0.0220 | PLoss: -0.0028 | Ent: 0.1543
+[ 33.1%] Iter 3237/9765 | SPS:  1846 | Reward:  0.048 | Value:  0.549 | VLoss: 0.0196 | PLoss: -0.0005 | Ent: 0.1410
+[ 33.2%] Iter 3238/9765 | SPS:  1846 | Reward:  0.044 | Value:  0.513 | VLoss: 0.0186 | PLoss: -0.0060 | Ent: 0.1593
+[ 33.2%] Iter 3239/9765 | SPS:  1847 | Reward:  0.043 | Value:  0.546 | VLoss: 0.0279 | PLoss: -0.0108 | Ent: 0.1240
+[ 33.2%] Iter 3240/9765 | SPS:  1847 | Reward:  0.053 | Value:  0.560 | VLoss: 0.0190 | PLoss: -0.0125 | Ent: 0.1401
+[ 33.2%] Iter 3241/9765 | SPS:  1847 | Reward:  0.057 | Value:  0.563 | VLoss: 0.0229 | PLoss: -0.0117 | Ent: 0.1414
+[ 33.2%] Iter 3242/9765 | SPS:  1847 | Reward:  0.047 | Value:  0.554 | VLoss: 0.0210 | PLoss: -0.0009 | Ent: 0.1420
+[ 33.2%] Iter 3243/9765 | SPS:  1848 | Reward:  0.056 | Value:  0.584 | VLoss: 0.0170 | PLoss: -0.0051 | Ent: 0.1179
+[ 33.2%] Iter 3244/9765 | SPS:  1848 | Reward:  0.056 | Value:  0.627 | VLoss: 0.0150 | PLoss: -0.0105 | Ent: 0.1592
+[ 33.2%] Iter 3245/9765 | SPS:  1848 | Reward:  0.058 | Value:  0.631 | VLoss: 0.0335 | PLoss: -0.0074 | Ent: 0.1556
+[ 33.2%] Iter 3246/9765 | SPS:  1849 | Reward:  0.042 | Value:  0.600 | VLoss: 0.0253 | PLoss: -0.0095 | Ent: 0.1642
+[ 33.3%] Iter 3247/9765 | SPS:  1849 | Reward:  0.050 | Value:  0.594 | VLoss: 0.0196 | PLoss: -0.0035 | Ent: 0.1063
+[ 33.3%] Iter 3248/9765 | SPS:  1849 | Reward:  0.050 | Value:  0.552 | VLoss: 0.0179 | PLoss: -0.0061 | Ent: 0.1384
+[ 33.3%] Iter 3249/9765 | SPS:  1849 | Reward:  0.040 | Value:  0.635 | VLoss: 0.0261 | PLoss: -0.0094 | Ent: 0.1326
+[ 33.3%] Iter 3250/9765 | SPS:  1850 | Reward:  0.048 | Value:  0.614 | VLoss: 0.0196 | PLoss: -0.0052 | Ent: 0.1528
+[ 33.3%] Iter 3251/9765 | SPS:  1850 | Reward:  0.053 | Value:  0.595 | VLoss: 0.0144 | PLoss: -0.0100 | Ent: 0.1327
+[ 33.3%] Iter 3252/9765 | SPS:  1850 | Reward:  0.054 | Value:  0.584 | VLoss: 0.0227 | PLoss: -0.0078 | Ent: 0.1341
+[ 33.3%] Iter 3253/9765 | SPS:  1851 | Reward:  0.042 | Value:  0.580 | VLoss: 0.0165 | PLoss: -0.0071 | Ent: 0.1005
+[ 33.3%] Iter 3254/9765 | SPS:  1851 | Reward:  0.038 | Value:  0.625 | VLoss: 0.0168 | PLoss: -0.0095 | Ent: 0.1320
+[ 33.3%] Iter 3255/9765 | SPS:  1851 | Reward:  0.046 | Value:  0.576 | VLoss: 0.0164 | PLoss: -0.0011 | Ent: 0.1696
+[ 33.3%] Iter 3256/9765 | SPS:  1852 | Reward:  0.054 | Value:  0.634 | VLoss: 0.0225 | PLoss: -0.0112 | Ent: 0.2003
+[ 33.4%] Iter 3257/9765 | SPS:  1852 | Reward:  0.057 | Value:  0.729 | VLoss: 0.0227 | PLoss: -0.0078 | Ent: 0.2191
+[ 33.4%] Iter 3258/9765 | SPS:  1852 | Reward:  0.059 | Value:  0.633 | VLoss: 0.0240 | PLoss: -0.0100 | Ent: 0.1997
+[ 33.4%] Iter 3259/9765 | SPS:  1852 | Reward:  0.054 | Value:  0.693 | VLoss: 0.0203 | PLoss: -0.0143 | Ent: 0.1537
+[ 33.4%] Iter 3260/9765 | SPS:  1853 | Reward:  0.075 | Value:  0.695 | VLoss: 0.0289 | PLoss: -0.0082 | Ent: 0.1713
+[ 33.4%] Iter 3261/9765 | SPS:  1853 | Reward:  0.067 | Value:  0.706 | VLoss: 0.0259 | PLoss: -0.0073 | Ent: 0.1628
+[ 33.4%] Iter 3262/9765 | SPS:  1853 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0256 | PLoss: 0.0060 | Ent: 0.1834
+[ 33.4%] Iter 3263/9765 | SPS:  1853 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0207 | PLoss: -0.0104 | Ent: 0.2015
+[ 33.4%] Iter 3264/9765 | SPS:  1854 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0258 | PLoss: -0.0110 | Ent: 0.1883
+[ 33.4%] Iter 3265/9765 | SPS:  1854 | Reward:  0.059 | Value:  0.726 | VLoss: 0.0292 | PLoss: -0.0104 | Ent: 0.1984
+[ 33.4%] Iter 3266/9765 | SPS:  1854 | Reward:  0.062 | Value:  0.633 | VLoss: 0.0306 | PLoss: -0.0130 | Ent: 0.2023
+[ 33.5%] Iter 3267/9765 | SPS:  1855 | Reward:  0.054 | Value:  0.670 | VLoss: 0.0247 | PLoss: -0.0077 | Ent: 0.2029
+[ 33.5%] Iter 3268/9765 | SPS:  1855 | Reward:  0.073 | Value:  0.668 | VLoss: 0.0304 | PLoss: -0.0070 | Ent: 0.1658
+[ 33.5%] Iter 3269/9765 | SPS:  1855 | Reward:  0.061 | Value:  0.636 | VLoss: 0.0288 | PLoss: -0.0102 | Ent: 0.1697
+[ 33.5%] Iter 3270/9765 | SPS:  1855 | Reward:  0.058 | Value:  0.607 | VLoss: 0.0329 | PLoss: -0.0105 | Ent: 0.2138
+[ 33.5%] Iter 3271/9765 | SPS:  1856 | Reward:  0.064 | Value:  0.712 | VLoss: 0.0250 | PLoss: -0.0082 | Ent: 0.1787
+[ 33.5%] Iter 3272/9765 | SPS:  1856 | Reward:  0.049 | Value:  0.704 | VLoss: 0.0218 | PLoss: -0.0091 | Ent: 0.1663
+[ 33.5%] Iter 3273/9765 | SPS:  1856 | Reward:  0.044 | Value:  0.708 | VLoss: 0.0185 | PLoss: 0.0008 | Ent: 0.1596
+[ 33.5%] Iter 3274/9765 | SPS:  1857 | Reward:  0.051 | Value:  0.599 | VLoss: 0.0179 | PLoss: -0.0092 | Ent: 0.1875
+[ 33.5%] Iter 3275/9765 | SPS:  1857 | Reward:  0.050 | Value:  0.628 | VLoss: 0.0260 | PLoss: 0.0012 | Ent: 0.1780
+[ 33.5%] Iter 3276/9765 | SPS:  1857 | Reward:  0.046 | Value:  0.640 | VLoss: 0.0211 | PLoss: 0.0009 | Ent: 0.1860
+[ 33.6%] Iter 3277/9765 | SPS:  1857 | Reward:  0.062 | Value:  0.603 | VLoss: 0.0307 | PLoss: -0.0135 | Ent: 0.1721
+[ 33.6%] Iter 3278/9765 | SPS:  1858 | Reward:  0.069 | Value:  0.655 | VLoss: 0.0282 | PLoss: -0.0082 | Ent: 0.1333
+[ 33.6%] Iter 3279/9765 | SPS:  1858 | Reward:  0.058 | Value:  0.634 | VLoss: 0.0260 | PLoss: -0.0057 | Ent: 0.1733
+[ 33.6%] Iter 3280/9765 | SPS:  1858 | Reward:  0.055 | Value:  0.611 | VLoss: 0.0275 | PLoss: -0.0147 | Ent: 0.1996
+[ 33.6%] Iter 3281/9765 | SPS:  1859 | Reward:  0.070 | Value:  0.605 | VLoss: 0.0230 | PLoss: -0.0025 | Ent: 0.1655
+[ 33.6%] Iter 3282/9765 | SPS:  1859 | Reward:  0.049 | Value:  0.630 | VLoss: 0.0183 | PLoss: -0.0034 | Ent: 0.1728
+[ 33.6%] Iter 3283/9765 | SPS:  1859 | Reward:  0.061 | Value:  0.642 | VLoss: 0.0179 | PLoss: -0.0027 | Ent: 0.1676
+[ 33.6%] Iter 3284/9765 | SPS:  1859 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0133 | PLoss: 0.0001 | Ent: 0.1377
+[ 33.6%] Iter 3285/9765 | SPS:  1860 | Reward:  0.042 | Value:  0.639 | VLoss: 0.0220 | PLoss: -0.0172 | Ent: 0.1586
+[ 33.7%] Iter 3286/9765 | SPS:  1860 | Reward:  0.044 | Value:  0.591 | VLoss: 0.0165 | PLoss: -0.0029 | Ent: 0.1591
+[ 33.7%] Iter 3287/9765 | SPS:  1860 | Reward:  0.046 | Value:  0.519 | VLoss: 0.0239 | PLoss: -0.0103 | Ent: 0.2048
+[ 33.7%] Iter 3288/9765 | SPS:  1861 | Reward:  0.062 | Value:  0.652 | VLoss: 0.0224 | PLoss: -0.0045 | Ent: 0.1228
+[ 33.7%] Iter 3289/9765 | SPS:  1861 | Reward:  0.066 | Value:  0.679 | VLoss: 0.0284 | PLoss: -0.0025 | Ent: 0.1568
+[ 33.7%] Iter 3290/9765 | SPS:  1861 | Reward:  0.085 | Value:  0.685 | VLoss: 0.0363 | PLoss: -0.0053 | Ent: 0.1158
+[ 33.7%] Iter 3291/9765 | SPS:  1861 | Reward:  0.071 | Value:  0.687 | VLoss: 0.0234 | PLoss: -0.0029 | Ent: 0.1591
+[ 33.7%] Iter 3292/9765 | SPS:  1862 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0271 | PLoss: -0.0082 | Ent: 0.1466
+[ 33.7%] Iter 3293/9765 | SPS:  1862 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0272 | PLoss: -0.0107 | Ent: 0.1676
+[ 33.7%] Iter 3294/9765 | SPS:  1862 | Reward:  0.062 | Value:  0.668 | VLoss: 0.0218 | PLoss: -0.0117 | Ent: 0.1603
+[ 33.7%] Iter 3295/9765 | SPS:  1862 | Reward:  0.061 | Value:  0.749 | VLoss: 0.0319 | PLoss: -0.0005 | Ent: 0.1504
+[ 33.8%] Iter 3296/9765 | SPS:  1863 | Reward:  0.054 | Value:  0.683 | VLoss: 0.0189 | PLoss: 0.0035 | Ent: 0.1529
+[ 33.8%] Iter 3297/9765 | SPS:  1863 | Reward:  0.049 | Value:  0.685 | VLoss: 0.0197 | PLoss: -0.0056 | Ent: 0.1824
+[ 33.8%] Iter 3298/9765 | SPS:  1863 | Reward:  0.059 | Value:  0.711 | VLoss: 0.0318 | PLoss: -0.0094 | Ent: 0.1539
+[ 33.8%] Iter 3299/9765 | SPS:  1864 | Reward:  0.054 | Value:  0.689 | VLoss: 0.0239 | PLoss: -0.0133 | Ent: 0.1563
+[ 33.8%] Iter 3300/9765 | SPS:  1864 | Reward:  0.055 | Value:  0.620 | VLoss: 0.0276 | PLoss: -0.0055 | Ent: 0.1927
+[ 33.8%] Iter 3301/9765 | SPS:  1864 | Reward:  0.054 | Value:  0.693 | VLoss: 0.0211 | PLoss: -0.0029 | Ent: 0.1771
+[ 33.8%] Iter 3302/9765 | SPS:  1864 | Reward:  0.066 | Value:  0.746 | VLoss: 0.0247 | PLoss: -0.0115 | Ent: 0.1700
+[ 33.8%] Iter 3303/9765 | SPS:  1865 | Reward:  0.062 | Value:  0.700 | VLoss: 0.0265 | PLoss: -0.0088 | Ent: 0.1814
+[ 33.8%] Iter 3304/9765 | SPS:  1865 | Reward:  0.061 | Value:  0.668 | VLoss: 0.0263 | PLoss: -0.0090 | Ent: 0.1191
+[ 33.8%] Iter 3305/9765 | SPS:  1865 | Reward:  0.073 | Value:  0.726 | VLoss: 0.0323 | PLoss: -0.0121 | Ent: 0.1656
+[ 33.9%] Iter 3306/9765 | SPS:  1866 | Reward:  0.064 | Value:  0.703 | VLoss: 0.0219 | PLoss: -0.0038 | Ent: 0.2066
+[ 33.9%] Iter 3307/9765 | SPS:  1866 | Reward:  0.056 | Value:  0.665 | VLoss: 0.0161 | PLoss: -0.0072 | Ent: 0.1889
+[ 33.9%] Iter 3308/9765 | SPS:  1866 | Reward:  0.074 | Value:  0.690 | VLoss: 0.0285 | PLoss: -0.0070 | Ent: 0.1566
+[ 33.9%] Iter 3309/9765 | SPS:  1866 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0269 | PLoss: -0.0095 | Ent: 0.2078
+[ 33.9%] Iter 3310/9765 | SPS:  1867 | Reward:  0.048 | Value:  0.678 | VLoss: 0.0210 | PLoss: -0.0100 | Ent: 0.2322
+[ 33.9%] Iter 3311/9765 | SPS:  1867 | Reward:  0.056 | Value:  0.679 | VLoss: 0.0241 | PLoss: -0.0153 | Ent: 0.1973
+[ 33.9%] Iter 3312/9765 | SPS:  1867 | Reward:  0.049 | Value:  0.672 | VLoss: 0.0245 | PLoss: -0.0139 | Ent: 0.2098
+[ 33.9%] Iter 3313/9765 | SPS:  1868 | Reward:  0.064 | Value:  0.687 | VLoss: 0.0282 | PLoss: -0.0152 | Ent: 0.1629
+[ 33.9%] Iter 3314/9765 | SPS:  1868 | Reward:  0.072 | Value:  0.648 | VLoss: 0.0305 | PLoss: -0.0035 | Ent: 0.1664
+[ 33.9%] Iter 3315/9765 | SPS:  1868 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0245 | PLoss: -0.0061 | Ent: 0.1750
+[ 34.0%] Iter 3316/9765 | SPS:  1868 | Reward:  0.075 | Value:  0.663 | VLoss: 0.0291 | PLoss: -0.0081 | Ent: 0.2031
+[ 34.0%] Iter 3317/9765 | SPS:  1869 | Reward:  0.077 | Value:  0.693 | VLoss: 0.0283 | PLoss: -0.0106 | Ent: 0.2083
+[ 34.0%] Iter 3318/9765 | SPS:  1869 | Reward:  0.078 | Value:  0.697 | VLoss: 0.0313 | PLoss: -0.0101 | Ent: 0.2239
+[ 34.0%] Iter 3319/9765 | SPS:  1869 | Reward:  0.065 | Value:  0.672 | VLoss: 0.0314 | PLoss: -0.0053 | Ent: 0.1674
+[ 34.0%] Iter 3320/9765 | SPS:  1869 | Reward:  0.073 | Value:  0.699 | VLoss: 0.0166 | PLoss: -0.0013 | Ent: 0.1681
+[ 34.0%] Iter 3321/9765 | SPS:  1870 | Reward:  0.058 | Value:  0.677 | VLoss: 0.0280 | PLoss: -0.0037 | Ent: 0.2320
+[ 34.0%] Iter 3322/9765 | SPS:  1870 | Reward:  0.057 | Value:  0.666 | VLoss: 0.0261 | PLoss: -0.0025 | Ent: 0.1781
+[ 34.0%] Iter 3323/9765 | SPS:  1870 | Reward:  0.045 | Value:  0.600 | VLoss: 0.0154 | PLoss: -0.0043 | Ent: 0.1996
+[ 34.0%] Iter 3324/9765 | SPS:  1871 | Reward:  0.061 | Value:  0.627 | VLoss: 0.0275 | PLoss: -0.0071 | Ent: 0.1426
+[ 34.1%] Iter 3325/9765 | SPS:  1871 | Reward:  0.055 | Value:  0.679 | VLoss: 0.0236 | PLoss: -0.0120 | Ent: 0.1948
+[ 34.1%] Iter 3326/9765 | SPS:  1871 | Reward:  0.083 | Value:  0.624 | VLoss: 0.0380 | PLoss: -0.0099 | Ent: 0.1471
+[ 34.1%] Iter 3327/9765 | SPS:  1871 | Reward:  0.071 | Value:  0.731 | VLoss: 0.0306 | PLoss: -0.0139 | Ent: 0.1907
+[ 34.1%] Iter 3328/9765 | SPS:  1872 | Reward:  0.053 | Value:  0.635 | VLoss: 0.0298 | PLoss: -0.0057 | Ent: 0.1470
+[ 34.1%] Iter 3329/9765 | SPS:  1872 | Reward:  0.061 | Value:  0.650 | VLoss: 0.0270 | PLoss: -0.0092 | Ent: 0.1477
+[ 34.1%] Iter 3330/9765 | SPS:  1872 | Reward:  0.061 | Value:  0.647 | VLoss: 0.0286 | PLoss: -0.0013 | Ent: 0.1626
+[ 34.1%] Iter 3331/9765 | SPS:  1872 | Reward:  0.063 | Value:  0.714 | VLoss: 0.0204 | PLoss: -0.0098 | Ent: 0.1973
+[ 34.1%] Iter 3332/9765 | SPS:  1873 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0222 | PLoss: -0.0064 | Ent: 0.1846
+[ 34.1%] Iter 3333/9765 | SPS:  1873 | Reward:  0.059 | Value:  0.716 | VLoss: 0.0164 | PLoss: -0.0064 | Ent: 0.1992
+[ 34.1%] Iter 3334/9765 | SPS:  1873 | Reward:  0.056 | Value:  0.639 | VLoss: 0.0266 | PLoss: -0.0160 | Ent: 0.1804
+[ 34.2%] Iter 3335/9765 | SPS:  1874 | Reward:  0.065 | Value:  0.682 | VLoss: 0.0305 | PLoss: -0.0059 | Ent: 0.1439
+[ 34.2%] Iter 3336/9765 | SPS:  1874 | Reward:  0.068 | Value:  0.756 | VLoss: 0.0300 | PLoss: -0.0100 | Ent: 0.1809
+[ 34.2%] Iter 3337/9765 | SPS:  1874 | Reward:  0.060 | Value:  0.679 | VLoss: 0.0182 | PLoss: -0.0095 | Ent: 0.2248
+[ 34.2%] Iter 3338/9765 | SPS:  1874 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0322 | PLoss: -0.0057 | Ent: 0.1347
+[ 34.2%] Iter 3339/9765 | SPS:  1875 | Reward:  0.071 | Value:  0.667 | VLoss: 0.0218 | PLoss: -0.0009 | Ent: 0.1932
+[ 34.2%] Iter 3340/9765 | SPS:  1875 | Reward:  0.068 | Value:  0.705 | VLoss: 0.0280 | PLoss: -0.0095 | Ent: 0.1965
+[ 34.2%] Iter 3341/9765 | SPS:  1875 | Reward:  0.067 | Value:  0.749 | VLoss: 0.0291 | PLoss: -0.0111 | Ent: 0.2122
+[ 34.2%] Iter 3342/9765 | SPS:  1876 | Reward:  0.054 | Value:  0.734 | VLoss: 0.0263 | PLoss: -0.0106 | Ent: 0.1999
+[ 34.2%] Iter 3343/9765 | SPS:  1876 | Reward:  0.053 | Value:  0.659 | VLoss: 0.0298 | PLoss: -0.0148 | Ent: 0.2079
+[ 34.2%] Iter 3344/9765 | SPS:  1876 | Reward:  0.059 | Value:  0.655 | VLoss: 0.0202 | PLoss: -0.0076 | Ent: 0.1754
+[ 34.3%] Iter 3345/9765 | SPS:  1876 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0239 | PLoss: -0.0128 | Ent: 0.1554
+[ 34.3%] Iter 3346/9765 | SPS:  1877 | Reward:  0.055 | Value:  0.648 | VLoss: 0.0198 | PLoss: -0.0110 | Ent: 0.2214
+[ 34.3%] Iter 3347/9765 | SPS:  1877 | Reward:  0.060 | Value:  0.709 | VLoss: 0.0197 | PLoss: -0.0114 | Ent: 0.1729
+[ 34.3%] Iter 3348/9765 | SPS:  1877 | Reward:  0.061 | Value:  0.686 | VLoss: 0.0254 | PLoss: -0.0097 | Ent: 0.1694
+[ 34.3%] Iter 3349/9765 | SPS:  1877 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0239 | PLoss: -0.0082 | Ent: 0.2082
+[ 34.3%] Iter 3350/9765 | SPS:  1878 | Reward:  0.061 | Value:  0.687 | VLoss: 0.0205 | PLoss: -0.0146 | Ent: 0.2172
+[ 34.3%] Iter 3351/9765 | SPS:  1878 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0166 | PLoss: -0.0058 | Ent: 0.1925
+[ 34.3%] Iter 3352/9765 | SPS:  1878 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0254 | PLoss: -0.0055 | Ent: 0.1821
+[ 34.3%] Iter 3353/9765 | SPS:  1879 | Reward:  0.047 | Value:  0.734 | VLoss: 0.0142 | PLoss: -0.0054 | Ent: 0.2286
+[ 34.3%] Iter 3354/9765 | SPS:  1879 | Reward:  0.049 | Value:  0.654 | VLoss: 0.0207 | PLoss: -0.0055 | Ent: 0.1827
+[ 34.4%] Iter 3355/9765 | SPS:  1879 | Reward:  0.059 | Value:  0.601 | VLoss: 0.0311 | PLoss: -0.0116 | Ent: 0.1464
+[ 34.4%] Iter 3356/9765 | SPS:  1879 | Reward:  0.056 | Value:  0.677 | VLoss: 0.0238 | PLoss: -0.0063 | Ent: 0.1483
+[ 34.4%] Iter 3357/9765 | SPS:  1880 | Reward:  0.064 | Value:  0.665 | VLoss: 0.0279 | PLoss: -0.0013 | Ent: 0.1634
+[ 34.4%] Iter 3358/9765 | SPS:  1880 | Reward:  0.066 | Value:  0.771 | VLoss: 0.0216 | PLoss: -0.0081 | Ent: 0.1566
+[ 34.4%] Iter 3359/9765 | SPS:  1880 | Reward:  0.057 | Value:  0.715 | VLoss: 0.0281 | PLoss: -0.0169 | Ent: 0.1248
+[ 34.4%] Iter 3360/9765 | SPS:  1881 | Reward:  0.051 | Value:  0.694 | VLoss: 0.0212 | PLoss: -0.0042 | Ent: 0.2130
+[ 34.4%] Iter 3361/9765 | SPS:  1881 | Reward:  0.066 | Value:  0.652 | VLoss: 0.0307 | PLoss: -0.0115 | Ent: 0.1870
+[ 34.4%] Iter 3362/9765 | SPS:  1881 | Reward:  0.061 | Value:  0.718 | VLoss: 0.0200 | PLoss: -0.0131 | Ent: 0.2125
+[ 34.4%] Iter 3363/9765 | SPS:  1881 | Reward:  0.055 | Value:  0.700 | VLoss: 0.0207 | PLoss: -0.0068 | Ent: 0.1852
+[ 34.4%] Iter 3364/9765 | SPS:  1882 | Reward:  0.060 | Value:  0.734 | VLoss: 0.0223 | PLoss: -0.0057 | Ent: 0.1926
+[ 34.5%] Iter 3365/9765 | SPS:  1882 | Reward:  0.073 | Value:  0.701 | VLoss: 0.0363 | PLoss: -0.0135 | Ent: 0.1729
+[ 34.5%] Iter 3366/9765 | SPS:  1882 | Reward:  0.060 | Value:  0.690 | VLoss: 0.0262 | PLoss: -0.0057 | Ent: 0.1711
+[ 34.5%] Iter 3367/9765 | SPS:  1882 | Reward:  0.071 | Value:  0.729 | VLoss: 0.0235 | PLoss: -0.0098 | Ent: 0.1709
+[ 34.5%] Iter 3368/9765 | SPS:  1883 | Reward:  0.077 | Value:  0.767 | VLoss: 0.0233 | PLoss: -0.0026 | Ent: 0.1676
+[ 34.5%] Iter 3369/9765 | SPS:  1883 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0362 | PLoss: -0.0133 | Ent: 0.2423
+[ 34.5%] Iter 3370/9765 | SPS:  1883 | Reward:  0.065 | Value:  0.678 | VLoss: 0.0343 | PLoss: -0.0096 | Ent: 0.1669
+[ 34.5%] Iter 3371/9765 | SPS:  1884 | Reward:  0.070 | Value:  0.713 | VLoss: 0.0287 | PLoss: -0.0031 | Ent: 0.1444
+[ 34.5%] Iter 3372/9765 | SPS:  1884 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0328 | PLoss: -0.0079 | Ent: 0.1965
+[ 34.5%] Iter 3373/9765 | SPS:  1884 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0334 | PLoss: -0.0107 | Ent: 0.1573
+[ 34.6%] Iter 3374/9765 | SPS:  1884 | Reward:  0.066 | Value:  0.681 | VLoss: 0.0298 | PLoss: -0.0093 | Ent: 0.1680
+[ 34.6%] Iter 3375/9765 | SPS:  1885 | Reward:  0.056 | Value:  0.696 | VLoss: 0.0287 | PLoss: -0.0073 | Ent: 0.1889
+[ 34.6%] Iter 3376/9765 | SPS:  1885 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0240 | PLoss: -0.0076 | Ent: 0.1491
+[ 34.6%] Iter 3377/9765 | SPS:  1885 | Reward:  0.057 | Value:  0.666 | VLoss: 0.0326 | PLoss: -0.0182 | Ent: 0.1911
+[ 34.6%] Iter 3378/9765 | SPS:  1885 | Reward:  0.049 | Value:  0.676 | VLoss: 0.0233 | PLoss: -0.0037 | Ent: 0.1807
+[ 34.6%] Iter 3379/9765 | SPS:  1886 | Reward:  0.061 | Value:  0.680 | VLoss: 0.0255 | PLoss: -0.0090 | Ent: 0.1657
+[ 34.6%] Iter 3380/9765 | SPS:  1886 | Reward:  0.071 | Value:  0.697 | VLoss: 0.0282 | PLoss: -0.0048 | Ent: 0.2007
+[ 34.6%] Iter 3381/9765 | SPS:  1886 | Reward:  0.056 | Value:  0.702 | VLoss: 0.0290 | PLoss: -0.0050 | Ent: 0.1972
+[ 34.6%] Iter 3382/9765 | SPS:  1887 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0256 | PLoss: -0.0149 | Ent: 0.1729
+[ 34.6%] Iter 3383/9765 | SPS:  1887 | Reward:  0.051 | Value:  0.724 | VLoss: 0.0261 | PLoss: -0.0104 | Ent: 0.1919
+[ 34.7%] Iter 3384/9765 | SPS:  1887 | Reward:  0.061 | Value:  0.660 | VLoss: 0.0243 | PLoss: -0.0055 | Ent: 0.2013
+[ 34.7%] Iter 3385/9765 | SPS:  1887 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0215 | PLoss: -0.0129 | Ent: 0.2238
+[ 34.7%] Iter 3386/9765 | SPS:  1888 | Reward:  0.068 | Value:  0.735 | VLoss: 0.0282 | PLoss: -0.0075 | Ent: 0.1331
+[ 34.7%] Iter 3387/9765 | SPS:  1888 | Reward:  0.071 | Value:  0.719 | VLoss: 0.0317 | PLoss: -0.0088 | Ent: 0.1626
+[ 34.7%] Iter 3388/9765 | SPS:  1888 | Reward:  0.061 | Value:  0.705 | VLoss: 0.0300 | PLoss: -0.0097 | Ent: 0.1673
+[ 34.7%] Iter 3389/9765 | SPS:  1888 | Reward:  0.063 | Value:  0.735 | VLoss: 0.0242 | PLoss: -0.0093 | Ent: 0.1959
+[ 34.7%] Iter 3390/9765 | SPS:  1889 | Reward:  0.052 | Value:  0.694 | VLoss: 0.0230 | PLoss: -0.0047 | Ent: 0.1728
+[ 34.7%] Iter 3391/9765 | SPS:  1889 | Reward:  0.066 | Value:  0.671 | VLoss: 0.0264 | PLoss: -0.0085 | Ent: 0.1758
+[ 34.7%] Iter 3392/9765 | SPS:  1889 | Reward:  0.063 | Value:  0.670 | VLoss: 0.0320 | PLoss: -0.0112 | Ent: 0.2295
+[ 34.7%] Iter 3393/9765 | SPS:  1890 | Reward:  0.083 | Value:  0.664 | VLoss: 0.0234 | PLoss: -0.0148 | Ent: 0.1859
+[ 34.8%] Iter 3394/9765 | SPS:  1890 | Reward:  0.068 | Value:  0.759 | VLoss: 0.0210 | PLoss: -0.0035 | Ent: 0.1871
+[ 34.8%] Iter 3395/9765 | SPS:  1890 | Reward:  0.051 | Value:  0.759 | VLoss: 0.0233 | PLoss: -0.0072 | Ent: 0.1606
+[ 34.8%] Iter 3396/9765 | SPS:  1890 | Reward:  0.065 | Value:  0.663 | VLoss: 0.0355 | PLoss: -0.0100 | Ent: 0.2104
+[ 34.8%] Iter 3397/9765 | SPS:  1891 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0228 | PLoss: -0.0046 | Ent: 0.1883
+[ 34.8%] Iter 3398/9765 | SPS:  1891 | Reward:  0.073 | Value:  0.668 | VLoss: 0.0253 | PLoss: -0.0108 | Ent: 0.2320
+[ 34.8%] Iter 3399/9765 | SPS:  1891 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0235 | PLoss: -0.0090 | Ent: 0.1957
+[ 34.8%] Iter 3400/9765 | SPS:  1891 | Reward:  0.064 | Value:  0.800 | VLoss: 0.0187 | PLoss: -0.0062 | Ent: 0.1706
+[ 34.8%] Iter 3401/9765 | SPS:  1892 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0348 | PLoss: -0.0039 | Ent: 0.2008
+[ 34.8%] Iter 3402/9765 | SPS:  1892 | Reward:  0.063 | Value:  0.636 | VLoss: 0.0339 | PLoss: -0.0003 | Ent: 0.1892
+[ 34.8%] Iter 3403/9765 | SPS:  1892 | Reward:  0.051 | Value:  0.649 | VLoss: 0.0294 | PLoss: -0.0047 | Ent: 0.1807
+[ 34.9%] Iter 3404/9765 | SPS:  1893 | Reward:  0.068 | Value:  0.625 | VLoss: 0.0251 | PLoss: -0.0100 | Ent: 0.1834
+[ 34.9%] Iter 3405/9765 | SPS:  1893 | Reward:  0.079 | Value:  0.660 | VLoss: 0.0254 | PLoss: -0.0074 | Ent: 0.1883
+[ 34.9%] Iter 3406/9765 | SPS:  1893 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0301 | PLoss: 0.0006 | Ent: 0.2306
+[ 34.9%] Iter 3407/9765 | SPS:  1893 | Reward:  0.059 | Value:  0.663 | VLoss: 0.0227 | PLoss: -0.0095 | Ent: 0.1473
+[ 34.9%] Iter 3408/9765 | SPS:  1894 | Reward:  0.057 | Value:  0.650 | VLoss: 0.0286 | PLoss: -0.0025 | Ent: 0.2092
+[ 34.9%] Iter 3409/9765 | SPS:  1894 | Reward:  0.052 | Value:  0.624 | VLoss: 0.0296 | PLoss: -0.0108 | Ent: 0.1897
+[ 34.9%] Iter 3410/9765 | SPS:  1894 | Reward:  0.053 | Value:  0.669 | VLoss: 0.0218 | PLoss: -0.0091 | Ent: 0.2078
+[ 34.9%] Iter 3411/9765 | SPS:  1894 | Reward:  0.058 | Value:  0.664 | VLoss: 0.0232 | PLoss: -0.0097 | Ent: 0.2015
+[ 34.9%] Iter 3412/9765 | SPS:  1895 | Reward:  0.068 | Value:  0.758 | VLoss: 0.0223 | PLoss: -0.0113 | Ent: 0.1920
+[ 35.0%] Iter 3413/9765 | SPS:  1895 | Reward:  0.058 | Value:  0.658 | VLoss: 0.0303 | PLoss: -0.0110 | Ent: 0.1797
+[ 35.0%] Iter 3414/9765 | SPS:  1895 | Reward:  0.057 | Value:  0.701 | VLoss: 0.0174 | PLoss: -0.0102 | Ent: 0.2142
+[ 35.0%] Iter 3415/9765 | SPS:  1895 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0308 | PLoss: -0.0050 | Ent: 0.1795
+[ 35.0%] Iter 3416/9765 | SPS:  1896 | Reward:  0.053 | Value:  0.648 | VLoss: 0.0253 | PLoss: -0.0006 | Ent: 0.2513
+[ 35.0%] Iter 3417/9765 | SPS:  1896 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0258 | PLoss: -0.0084 | Ent: 0.1620
+[ 35.0%] Iter 3418/9765 | SPS:  1896 | Reward:  0.058 | Value:  0.738 | VLoss: 0.0247 | PLoss: -0.0142 | Ent: 0.2270
+[ 35.0%] Iter 3419/9765 | SPS:  1897 | Reward:  0.063 | Value:  0.717 | VLoss: 0.0232 | PLoss: -0.0060 | Ent: 0.1616
+[ 35.0%] Iter 3420/9765 | SPS:  1897 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0294 | PLoss: -0.0082 | Ent: 0.2049
+[ 35.0%] Iter 3421/9765 | SPS:  1897 | Reward:  0.087 | Value:  0.745 | VLoss: 0.0312 | PLoss: -0.0073 | Ent: 0.1648
+[ 35.0%] Iter 3422/9765 | SPS:  1897 | Reward:  0.061 | Value:  0.736 | VLoss: 0.0261 | PLoss: -0.0121 | Ent: 0.1752
+[ 35.1%] Iter 3423/9765 | SPS:  1898 | Reward:  0.074 | Value:  0.737 | VLoss: 0.0241 | PLoss: -0.0162 | Ent: 0.2057
+[ 35.1%] Iter 3424/9765 | SPS:  1898 | Reward:  0.081 | Value:  0.693 | VLoss: 0.0403 | PLoss: -0.0067 | Ent: 0.1691
+[ 35.1%] Iter 3425/9765 | SPS:  1898 | Reward:  0.072 | Value:  0.715 | VLoss: 0.0285 | PLoss: -0.0114 | Ent: 0.2025
+[ 35.1%] Iter 3426/9765 | SPS:  1898 | Reward:  0.062 | Value:  0.738 | VLoss: 0.0297 | PLoss: -0.0042 | Ent: 0.1685
+[ 35.1%] Iter 3427/9765 | SPS:  1899 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0213 | PLoss: -0.0091 | Ent: 0.1752
+[ 35.1%] Iter 3428/9765 | SPS:  1899 | Reward:  0.056 | Value:  0.682 | VLoss: 0.0273 | PLoss: -0.0110 | Ent: 0.1634
+[ 35.1%] Iter 3429/9765 | SPS:  1899 | Reward:  0.070 | Value:  0.655 | VLoss: 0.0226 | PLoss: -0.0179 | Ent: 0.1796
+[ 35.1%] Iter 3430/9765 | SPS:  1899 | Reward:  0.068 | Value:  0.724 | VLoss: 0.0212 | PLoss: -0.0049 | Ent: 0.1612
+[ 35.1%] Iter 3431/9765 | SPS:  1900 | Reward:  0.073 | Value:  0.725 | VLoss: 0.0255 | PLoss: -0.0042 | Ent: 0.1771
+[ 35.1%] Iter 3432/9765 | SPS:  1900 | Reward:  0.061 | Value:  0.701 | VLoss: 0.0233 | PLoss: -0.0130 | Ent: 0.2248
+[ 35.2%] Iter 3433/9765 | SPS:  1900 | Reward:  0.054 | Value:  0.635 | VLoss: 0.0250 | PLoss: -0.0132 | Ent: 0.2070
+[ 35.2%] Iter 3434/9765 | SPS:  1901 | Reward:  0.055 | Value:  0.612 | VLoss: 0.0311 | PLoss: -0.0129 | Ent: 0.2255
+[ 35.2%] Iter 3435/9765 | SPS:  1901 | Reward:  0.070 | Value:  0.612 | VLoss: 0.0311 | PLoss: -0.0081 | Ent: 0.1593
+[ 35.2%] Iter 3436/9765 | SPS:  1901 | Reward:  0.073 | Value:  0.654 | VLoss: 0.0294 | PLoss: -0.0121 | Ent: 0.1707
+[ 35.2%] Iter 3437/9765 | SPS:  1901 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0220 | PLoss: -0.0186 | Ent: 0.2056
+[ 35.2%] Iter 3438/9765 | SPS:  1902 | Reward:  0.067 | Value:  0.748 | VLoss: 0.0230 | PLoss: -0.0087 | Ent: 0.1908
+[ 35.2%] Iter 3439/9765 | SPS:  1902 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0221 | PLoss: -0.0069 | Ent: 0.2292
+[ 35.2%] Iter 3440/9765 | SPS:  1902 | Reward:  0.059 | Value:  0.686 | VLoss: 0.0289 | PLoss: -0.0023 | Ent: 0.1548
+[ 35.2%] Iter 3441/9765 | SPS:  1902 | Reward:  0.069 | Value:  0.676 | VLoss: 0.0262 | PLoss: -0.0116 | Ent: 0.1972
+[ 35.2%] Iter 3442/9765 | SPS:  1903 | Reward:  0.074 | Value:  0.677 | VLoss: 0.0335 | PLoss: -0.0144 | Ent: 0.1822
+[ 35.3%] Iter 3443/9765 | SPS:  1903 | Reward:  0.078 | Value:  0.717 | VLoss: 0.0294 | PLoss: -0.0141 | Ent: 0.1935
+[ 35.3%] Iter 3444/9765 | SPS:  1903 | Reward:  0.058 | Value:  0.753 | VLoss: 0.0217 | PLoss: -0.0002 | Ent: 0.2319
+[ 35.3%] Iter 3445/9765 | SPS:  1903 | Reward:  0.061 | Value:  0.736 | VLoss: 0.0293 | PLoss: -0.0060 | Ent: 0.1776
+[ 35.3%] Iter 3446/9765 | SPS:  1904 | Reward:  0.068 | Value:  0.683 | VLoss: 0.0306 | PLoss: -0.0031 | Ent: 0.1805
+[ 35.3%] Iter 3447/9765 | SPS:  1904 | Reward:  0.072 | Value:  0.704 | VLoss: 0.0210 | PLoss: -0.0144 | Ent: 0.2456
+[ 35.3%] Iter 3448/9765 | SPS:  1904 | Reward:  0.065 | Value:  0.749 | VLoss: 0.0252 | PLoss: -0.0094 | Ent: 0.2060
+[ 35.3%] Iter 3449/9765 | SPS:  1904 | Reward:  0.064 | Value:  0.695 | VLoss: 0.0321 | PLoss: -0.0106 | Ent: 0.1626
+[ 35.3%] Iter 3450/9765 | SPS:  1905 | Reward:  0.056 | Value:  0.722 | VLoss: 0.0227 | PLoss: -0.0037 | Ent: 0.1420
+[ 35.3%] Iter 3451/9765 | SPS:  1905 | Reward:  0.057 | Value:  0.711 | VLoss: 0.0338 | PLoss: -0.0173 | Ent: 0.2037
+[ 35.4%] Iter 3452/9765 | SPS:  1905 | Reward:  0.063 | Value:  0.631 | VLoss: 0.0340 | PLoss: -0.0036 | Ent: 0.1810
+[ 35.4%] Iter 3453/9765 | SPS:  1906 | Reward:  0.064 | Value:  0.702 | VLoss: 0.0278 | PLoss: -0.0091 | Ent: 0.1842
+[ 35.4%] Iter 3454/9765 | SPS:  1906 | Reward:  0.070 | Value:  0.703 | VLoss: 0.0283 | PLoss: -0.0101 | Ent: 0.1547
+[ 35.4%] Iter 3455/9765 | SPS:  1906 | Reward:  0.064 | Value:  0.679 | VLoss: 0.0344 | PLoss: -0.0068 | Ent: 0.2093
+[ 35.4%] Iter 3456/9765 | SPS:  1906 | Reward:  0.061 | Value:  0.730 | VLoss: 0.0231 | PLoss: -0.0079 | Ent: 0.2047
+[ 35.4%] Iter 3457/9765 | SPS:  1907 | Reward:  0.066 | Value:  0.710 | VLoss: 0.0190 | PLoss: -0.0029 | Ent: 0.1538
+[ 35.4%] Iter 3458/9765 | SPS:  1907 | Reward:  0.060 | Value:  0.667 | VLoss: 0.0246 | PLoss: -0.0088 | Ent: 0.1768
+[ 35.4%] Iter 3459/9765 | SPS:  1907 | Reward:  0.056 | Value:  0.724 | VLoss: 0.0202 | PLoss: -0.0079 | Ent: 0.1673
+[ 35.4%] Iter 3460/9765 | SPS:  1907 | Reward:  0.053 | Value:  0.734 | VLoss: 0.0236 | PLoss: -0.0106 | Ent: 0.1577
+[ 35.4%] Iter 3461/9765 | SPS:  1908 | Reward:  0.066 | Value:  0.672 | VLoss: 0.0210 | PLoss: -0.0031 | Ent: 0.1860
+[ 35.5%] Iter 3462/9765 | SPS:  1908 | Reward:  0.068 | Value:  0.706 | VLoss: 0.0326 | PLoss: -0.0094 | Ent: 0.1767
+[ 35.5%] Iter 3463/9765 | SPS:  1908 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0257 | PLoss: -0.0067 | Ent: 0.1893
+[ 35.5%] Iter 3464/9765 | SPS:  1908 | Reward:  0.061 | Value:  0.702 | VLoss: 0.0278 | PLoss: -0.0064 | Ent: 0.1815
+[ 35.5%] Iter 3465/9765 | SPS:  1909 | Reward:  0.060 | Value:  0.651 | VLoss: 0.0338 | PLoss: -0.0113 | Ent: 0.2189
+[ 35.5%] Iter 3466/9765 | SPS:  1909 | Reward:  0.067 | Value:  0.727 | VLoss: 0.0186 | PLoss: -0.0002 | Ent: 0.1693
+[ 35.5%] Iter 3467/9765 | SPS:  1909 | Reward:  0.060 | Value:  0.627 | VLoss: 0.0361 | PLoss: -0.0094 | Ent: 0.2065
+[ 35.5%] Iter 3468/9765 | SPS:  1909 | Reward:  0.065 | Value:  0.664 | VLoss: 0.0306 | PLoss: -0.0101 | Ent: 0.2045
+[ 35.5%] Iter 3469/9765 | SPS:  1910 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0349 | PLoss: -0.0031 | Ent: 0.2097
+[ 35.5%] Iter 3470/9765 | SPS:  1910 | Reward:  0.071 | Value:  0.659 | VLoss: 0.0319 | PLoss: -0.0065 | Ent: 0.2331
+[ 35.5%] Iter 3471/9765 | SPS:  1910 | Reward:  0.052 | Value:  0.620 | VLoss: 0.0252 | PLoss: -0.0033 | Ent: 0.1575
+[ 35.6%] Iter 3472/9765 | SPS:  1911 | Reward:  0.053 | Value:  0.629 | VLoss: 0.0265 | PLoss: -0.0055 | Ent: 0.1762
+[ 35.6%] Iter 3473/9765 | SPS:  1911 | Reward:  0.043 | Value:  0.621 | VLoss: 0.0174 | PLoss: -0.0043 | Ent: 0.2429
+[ 35.6%] Iter 3474/9765 | SPS:  1911 | Reward:  0.048 | Value:  0.641 | VLoss: 0.0187 | PLoss: -0.0013 | Ent: 0.2049
+[ 35.6%] Iter 3475/9765 | SPS:  1911 | Reward:  0.058 | Value:  0.648 | VLoss: 0.0220 | PLoss: -0.0130 | Ent: 0.2949
+[ 35.6%] Iter 3476/9765 | SPS:  1912 | Reward:  0.057 | Value:  0.672 | VLoss: 0.0205 | PLoss: -0.0083 | Ent: 0.1724
+[ 35.6%] Iter 3477/9765 | SPS:  1912 | Reward:  0.067 | Value:  0.637 | VLoss: 0.0305 | PLoss: -0.0025 | Ent: 0.2019
+[ 35.6%] Iter 3478/9765 | SPS:  1912 | Reward:  0.054 | Value:  0.669 | VLoss: 0.0228 | PLoss: -0.0098 | Ent: 0.2044
+[ 35.6%] Iter 3479/9765 | SPS:  1912 | Reward:  0.049 | Value:  0.637 | VLoss: 0.0287 | PLoss: -0.0087 | Ent: 0.2138
+[ 35.6%] Iter 3480/9765 | SPS:  1913 | Reward:  0.058 | Value:  0.604 | VLoss: 0.0346 | PLoss: -0.0080 | Ent: 0.1997
+[ 35.6%] Iter 3481/9765 | SPS:  1913 | Reward:  0.050 | Value:  0.557 | VLoss: 0.0173 | PLoss: -0.0124 | Ent: 0.2029
+[ 35.7%] Iter 3482/9765 | SPS:  1913 | Reward:  0.055 | Value:  0.594 | VLoss: 0.0305 | PLoss: -0.0035 | Ent: 0.2056
+[ 35.7%] Iter 3483/9765 | SPS:  1913 | Reward:  0.063 | Value:  0.630 | VLoss: 0.0219 | PLoss: -0.0031 | Ent: 0.2077
+[ 35.7%] Iter 3484/9765 | SPS:  1914 | Reward:  0.057 | Value:  0.619 | VLoss: 0.0266 | PLoss: -0.0144 | Ent: 0.2197
+[ 35.7%] Iter 3485/9765 | SPS:  1914 | Reward:  0.059 | Value:  0.592 | VLoss: 0.0204 | PLoss: -0.0043 | Ent: 0.2072
+[ 35.7%] Iter 3486/9765 | SPS:  1914 | Reward:  0.062 | Value:  0.641 | VLoss: 0.0205 | PLoss: -0.0083 | Ent: 0.2275
+[ 35.7%] Iter 3487/9765 | SPS:  1915 | Reward:  0.070 | Value:  0.633 | VLoss: 0.0248 | PLoss: -0.0040 | Ent: 0.1974
+[ 35.7%] Iter 3488/9765 | SPS:  1915 | Reward:  0.052 | Value:  0.591 | VLoss: 0.0269 | PLoss: -0.0179 | Ent: 0.2513
+[ 35.7%] Iter 3489/9765 | SPS:  1915 | Reward:  0.044 | Value:  0.573 | VLoss: 0.0220 | PLoss: -0.0099 | Ent: 0.2509
+[ 35.7%] Iter 3490/9765 | SPS:  1915 | Reward:  0.048 | Value:  0.526 | VLoss: 0.0355 | PLoss: -0.0112 | Ent: 0.2116
+[ 35.8%] Iter 3491/9765 | SPS:  1916 | Reward:  0.061 | Value:  0.603 | VLoss: 0.0230 | PLoss: -0.0052 | Ent: 0.1658
+[ 35.8%] Iter 3492/9765 | SPS:  1916 | Reward:  0.068 | Value:  0.631 | VLoss: 0.0243 | PLoss: -0.0145 | Ent: 0.2105
+[ 35.8%] Iter 3493/9765 | SPS:  1916 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0271 | PLoss: -0.0129 | Ent: 0.1738
+[ 35.8%] Iter 3494/9765 | SPS:  1916 | Reward:  0.056 | Value:  0.712 | VLoss: 0.0207 | PLoss: -0.0033 | Ent: 0.1410
+[ 35.8%] Iter 3495/9765 | SPS:  1917 | Reward:  0.054 | Value:  0.704 | VLoss: 0.0344 | PLoss: -0.0112 | Ent: 0.1710
+[ 35.8%] Iter 3496/9765 | SPS:  1917 | Reward:  0.047 | Value:  0.651 | VLoss: 0.0244 | PLoss: -0.0102 | Ent: 0.1499
+[ 35.8%] Iter 3497/9765 | SPS:  1917 | Reward:  0.048 | Value:  0.623 | VLoss: 0.0320 | PLoss: -0.0036 | Ent: 0.1367
+[ 35.8%] Iter 3498/9765 | SPS:  1917 | Reward:  0.058 | Value:  0.574 | VLoss: 0.0320 | PLoss: -0.0036 | Ent: 0.1386
+[ 35.8%] Iter 3499/9765 | SPS:  1918 | Reward:  0.047 | Value:  0.602 | VLoss: 0.0232 | PLoss: -0.0093 | Ent: 0.1683
+[ 35.8%] Iter 3500/9765 | SPS:  1918 | Reward:  0.054 | Value:  0.615 | VLoss: 0.0303 | PLoss: -0.0095 | Ent: 0.1775
+[ 35.9%] Iter 3501/9765 | SPS:  1918 | Reward:  0.063 | Value:  0.607 | VLoss: 0.0307 | PLoss: -0.0034 | Ent: 0.2141
+[ 35.9%] Iter 3502/9765 | SPS:  1918 | Reward:  0.063 | Value:  0.623 | VLoss: 0.0267 | PLoss: -0.0056 | Ent: 0.1799
+[ 35.9%] Iter 3503/9765 | SPS:  1919 | Reward:  0.044 | Value:  0.652 | VLoss: 0.0259 | PLoss: -0.0113 | Ent: 0.2139
+[ 35.9%] Iter 3504/9765 | SPS:  1919 | Reward:  0.053 | Value:  0.667 | VLoss: 0.0230 | PLoss: -0.0086 | Ent: 0.1851
+[ 35.9%] Iter 3505/9765 | SPS:  1919 | Reward:  0.058 | Value:  0.649 | VLoss: 0.0255 | PLoss: -0.0079 | Ent: 0.2117
+[ 35.9%] Iter 3506/9765 | SPS:  1920 | Reward:  0.050 | Value:  0.586 | VLoss: 0.0257 | PLoss: -0.0094 | Ent: 0.2138
+[ 35.9%] Iter 3507/9765 | SPS:  1920 | Reward:  0.052 | Value:  0.641 | VLoss: 0.0200 | PLoss: -0.0025 | Ent: 0.2239
+[ 35.9%] Iter 3508/9765 | SPS:  1920 | Reward:  0.059 | Value:  0.628 | VLoss: 0.0226 | PLoss: -0.0064 | Ent: 0.1698
+[ 35.9%] Iter 3509/9765 | SPS:  1920 | Reward:  0.067 | Value:  0.669 | VLoss: 0.0305 | PLoss: -0.0081 | Ent: 0.1924
+[ 35.9%] Iter 3510/9765 | SPS:  1921 | Reward:  0.050 | Value:  0.687 | VLoss: 0.0307 | PLoss: -0.0057 | Ent: 0.2306
+[ 36.0%] Iter 3511/9765 | SPS:  1921 | Reward:  0.056 | Value:  0.676 | VLoss: 0.0270 | PLoss: -0.0078 | Ent: 0.2010
+[ 36.0%] Iter 3512/9765 | SPS:  1921 | Reward:  0.049 | Value:  0.678 | VLoss: 0.0191 | PLoss: -0.0075 | Ent: 0.2174
+[ 36.0%] Iter 3513/9765 | SPS:  1921 | Reward:  0.074 | Value:  0.672 | VLoss: 0.0313 | PLoss: -0.0129 | Ent: 0.2196
+[ 36.0%] Iter 3514/9765 | SPS:  1922 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0293 | PLoss: -0.0078 | Ent: 0.1754
+[ 36.0%] Iter 3515/9765 | SPS:  1922 | Reward:  0.052 | Value:  0.674 | VLoss: 0.0227 | PLoss: -0.0070 | Ent: 0.1979
+[ 36.0%] Iter 3516/9765 | SPS:  1922 | Reward:  0.056 | Value:  0.669 | VLoss: 0.0358 | PLoss: -0.0046 | Ent: 0.1866
+[ 36.0%] Iter 3517/9765 | SPS:  1922 | Reward:  0.061 | Value:  0.624 | VLoss: 0.0282 | PLoss: -0.0067 | Ent: 0.1870
+[ 36.0%] Iter 3518/9765 | SPS:  1923 | Reward:  0.064 | Value:  0.713 | VLoss: 0.0255 | PLoss: -0.0111 | Ent: 0.1710
+[ 36.0%] Iter 3519/9765 | SPS:  1923 | Reward:  0.058 | Value:  0.662 | VLoss: 0.0297 | PLoss: -0.0032 | Ent: 0.1871
+[ 36.0%] Iter 3520/9765 | SPS:  1923 | Reward:  0.073 | Value:  0.645 | VLoss: 0.0281 | PLoss: -0.0036 | Ent: 0.1837
+[ 36.1%] Iter 3521/9765 | SPS:  1923 | Reward:  0.076 | Value:  0.663 | VLoss: 0.0267 | PLoss: -0.0133 | Ent: 0.1594
+[ 36.1%] Iter 3522/9765 | SPS:  1924 | Reward:  0.058 | Value:  0.706 | VLoss: 0.0271 | PLoss: -0.0100 | Ent: 0.1787
+[ 36.1%] Iter 3523/9765 | SPS:  1924 | Reward:  0.069 | Value:  0.745 | VLoss: 0.0310 | PLoss: -0.0112 | Ent: 0.1919
+[ 36.1%] Iter 3524/9765 | SPS:  1924 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0268 | PLoss: -0.0074 | Ent: 0.1912
+[ 36.1%] Iter 3525/9765 | SPS:  1925 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0211 | PLoss: -0.0048 | Ent: 0.1983
+[ 36.1%] Iter 3526/9765 | SPS:  1925 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0255 | PLoss: -0.0048 | Ent: 0.1935
+[ 36.1%] Iter 3527/9765 | SPS:  1925 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0343 | PLoss: -0.0200 | Ent: 0.2300
+[ 36.1%] Iter 3528/9765 | SPS:  1925 | Reward:  0.056 | Value:  0.661 | VLoss: 0.0239 | PLoss: -0.0080 | Ent: 0.2138
+[ 36.1%] Iter 3529/9765 | SPS:  1926 | Reward:  0.072 | Value:  0.698 | VLoss: 0.0225 | PLoss: -0.0114 | Ent: 0.1697
+[ 36.1%] Iter 3530/9765 | SPS:  1926 | Reward:  0.076 | Value:  0.725 | VLoss: 0.0372 | PLoss: -0.0042 | Ent: 0.1781
+[ 36.2%] Iter 3531/9765 | SPS:  1926 | Reward:  0.070 | Value:  0.679 | VLoss: 0.0357 | PLoss: -0.0099 | Ent: 0.2050
+[ 36.2%] Iter 3532/9765 | SPS:  1926 | Reward:  0.059 | Value:  0.692 | VLoss: 0.0351 | PLoss: -0.0096 | Ent: 0.1811
+[ 36.2%] Iter 3533/9765 | SPS:  1927 | Reward:  0.071 | Value:  0.679 | VLoss: 0.0296 | PLoss: -0.0089 | Ent: 0.2117
+[ 36.2%] Iter 3534/9765 | SPS:  1927 | Reward:  0.055 | Value:  0.673 | VLoss: 0.0352 | PLoss: -0.0120 | Ent: 0.1830
+[ 36.2%] Iter 3535/9765 | SPS:  1927 | Reward:  0.058 | Value:  0.638 | VLoss: 0.0313 | PLoss: -0.0110 | Ent: 0.1738
+[ 36.2%] Iter 3536/9765 | SPS:  1927 | Reward:  0.067 | Value:  0.634 | VLoss: 0.0228 | PLoss: -0.0074 | Ent: 0.1769
+[ 36.2%] Iter 3537/9765 | SPS:  1928 | Reward:  0.083 | Value:  0.694 | VLoss: 0.0245 | PLoss: -0.0151 | Ent: 0.1997
+[ 36.2%] Iter 3538/9765 | SPS:  1928 | Reward:  0.055 | Value:  0.722 | VLoss: 0.0211 | PLoss: -0.0140 | Ent: 0.1805
+[ 36.2%] Iter 3539/9765 | SPS:  1928 | Reward:  0.058 | Value:  0.740 | VLoss: 0.0211 | PLoss: -0.0106 | Ent: 0.1610
+[ 36.3%] Iter 3540/9765 | SPS:  1928 | Reward:  0.062 | Value:  0.748 | VLoss: 0.0204 | PLoss: -0.0011 | Ent: 0.1719
+[ 36.3%] Iter 3541/9765 | SPS:  1929 | Reward:  0.061 | Value:  0.742 | VLoss: 0.0142 | PLoss: -0.0067 | Ent: 0.1805
+[ 36.3%] Iter 3542/9765 | SPS:  1929 | Reward:  0.060 | Value:  0.728 | VLoss: 0.0241 | PLoss: -0.0113 | Ent: 0.2333
+[ 36.3%] Iter 3543/9765 | SPS:  1929 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0289 | PLoss: -0.0007 | Ent: 0.2119
+[ 36.3%] Iter 3544/9765 | SPS:  1929 | Reward:  0.062 | Value:  0.604 | VLoss: 0.0354 | PLoss: -0.0066 | Ent: 0.1829
+[ 36.3%] Iter 3545/9765 | SPS:  1930 | Reward:  0.061 | Value:  0.649 | VLoss: 0.0285 | PLoss: -0.0079 | Ent: 0.1867
+[ 36.3%] Iter 3546/9765 | SPS:  1930 | Reward:  0.078 | Value:  0.647 | VLoss: 0.0352 | PLoss: -0.0072 | Ent: 0.1992
+[ 36.3%] Iter 3547/9765 | SPS:  1930 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0190 | PLoss: -0.0160 | Ent: 0.1870
+[ 36.3%] Iter 3548/9765 | SPS:  1930 | Reward:  0.051 | Value:  0.659 | VLoss: 0.0243 | PLoss: -0.0038 | Ent: 0.1715
+[ 36.3%] Iter 3549/9765 | SPS:  1931 | Reward:  0.054 | Value:  0.627 | VLoss: 0.0270 | PLoss: -0.0057 | Ent: 0.1909
+[ 36.4%] Iter 3550/9765 | SPS:  1931 | Reward:  0.052 | Value:  0.652 | VLoss: 0.0282 | PLoss: -0.0134 | Ent: 0.2299
+[ 36.4%] Iter 3551/9765 | SPS:  1931 | Reward:  0.074 | Value:  0.633 | VLoss: 0.0273 | PLoss: -0.0094 | Ent: 0.2297
+[ 36.4%] Iter 3552/9765 | SPS:  1931 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0225 | PLoss: -0.0013 | Ent: 0.1672
+[ 36.4%] Iter 3553/9765 | SPS:  1932 | Reward:  0.064 | Value:  0.683 | VLoss: 0.0274 | PLoss: -0.0067 | Ent: 0.1909
+[ 36.4%] Iter 3554/9765 | SPS:  1932 | Reward:  0.055 | Value:  0.719 | VLoss: 0.0235 | PLoss: -0.0115 | Ent: 0.1938
+[ 36.4%] Iter 3555/9765 | SPS:  1932 | Reward:  0.046 | Value:  0.712 | VLoss: 0.0230 | PLoss: -0.0053 | Ent: 0.1574
+[ 36.4%] Iter 3556/9765 | SPS:  1933 | Reward:  0.055 | Value:  0.607 | VLoss: 0.0221 | PLoss: -0.0074 | Ent: 0.1936
+[ 36.4%] Iter 3557/9765 | SPS:  1933 | Reward:  0.075 | Value:  0.656 | VLoss: 0.0369 | PLoss: -0.0053 | Ent: 0.1708
+[ 36.4%] Iter 3558/9765 | SPS:  1933 | Reward:  0.069 | Value:  0.690 | VLoss: 0.0271 | PLoss: -0.0018 | Ent: 0.1855
+[ 36.4%] Iter 3559/9765 | SPS:  1933 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0237 | PLoss: -0.0073 | Ent: 0.1948
+[ 36.5%] Iter 3560/9765 | SPS:  1934 | Reward:  0.052 | Value:  0.707 | VLoss: 0.0253 | PLoss: -0.0053 | Ent: 0.2110
+[ 36.5%] Iter 3561/9765 | SPS:  1934 | Reward:  0.057 | Value:  0.689 | VLoss: 0.0219 | PLoss: -0.0069 | Ent: 0.1837
+[ 36.5%] Iter 3562/9765 | SPS:  1934 | Reward:  0.062 | Value:  0.643 | VLoss: 0.0217 | PLoss: -0.0117 | Ent: 0.1560
+[ 36.5%] Iter 3563/9765 | SPS:  1934 | Reward:  0.060 | Value:  0.691 | VLoss: 0.0274 | PLoss: -0.0060 | Ent: 0.1426
+[ 36.5%] Iter 3564/9765 | SPS:  1935 | Reward:  0.064 | Value:  0.748 | VLoss: 0.0256 | PLoss: -0.0065 | Ent: 0.1520
+[ 36.5%] Iter 3565/9765 | SPS:  1935 | Reward:  0.066 | Value:  0.747 | VLoss: 0.0250 | PLoss: -0.0113 | Ent: 0.2024
+[ 36.5%] Iter 3566/9765 | SPS:  1935 | Reward:  0.060 | Value:  0.732 | VLoss: 0.0286 | PLoss: -0.0070 | Ent: 0.1814
+[ 36.5%] Iter 3567/9765 | SPS:  1935 | Reward:  0.050 | Value:  0.694 | VLoss: 0.0182 | PLoss: -0.0038 | Ent: 0.1759
+[ 36.5%] Iter 3568/9765 | SPS:  1936 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0269 | PLoss: -0.0087 | Ent: 0.1785
+[ 36.5%] Iter 3569/9765 | SPS:  1936 | Reward:  0.056 | Value:  0.624 | VLoss: 0.0234 | PLoss: -0.0078 | Ent: 0.1701
+[ 36.6%] Iter 3570/9765 | SPS:  1936 | Reward:  0.055 | Value:  0.673 | VLoss: 0.0340 | PLoss: -0.0066 | Ent: 0.1525
+[ 36.6%] Iter 3571/9765 | SPS:  1936 | Reward:  0.061 | Value:  0.634 | VLoss: 0.0291 | PLoss: -0.0106 | Ent: 0.1649
+[ 36.6%] Iter 3572/9765 | SPS:  1937 | Reward:  0.064 | Value:  0.683 | VLoss: 0.0227 | PLoss: -0.0139 | Ent: 0.1689
+[ 36.6%] Iter 3573/9765 | SPS:  1937 | Reward:  0.064 | Value:  0.716 | VLoss: 0.0266 | PLoss: -0.0088 | Ent: 0.2117
+[ 36.6%] Iter 3574/9765 | SPS:  1937 | Reward:  0.069 | Value:  0.714 | VLoss: 0.0240 | PLoss: -0.0030 | Ent: 0.1599
+[ 36.6%] Iter 3575/9765 | SPS:  1937 | Reward:  0.058 | Value:  0.735 | VLoss: 0.0220 | PLoss: -0.0082 | Ent: 0.1231
+[ 36.6%] Iter 3576/9765 | SPS:  1938 | Reward:  0.052 | Value:  0.590 | VLoss: 0.0300 | PLoss: -0.0101 | Ent: 0.2018
+[ 36.6%] Iter 3577/9765 | SPS:  1938 | Reward:  0.070 | Value:  0.644 | VLoss: 0.0344 | PLoss: -0.0116 | Ent: 0.2002
+[ 36.6%] Iter 3578/9765 | SPS:  1938 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0270 | PLoss: -0.0046 | Ent: 0.1701
+[ 36.7%] Iter 3579/9765 | SPS:  1939 | Reward:  0.065 | Value:  0.752 | VLoss: 0.0179 | PLoss: -0.0053 | Ent: 0.1744
+[ 36.7%] Iter 3580/9765 | SPS:  1939 | Reward:  0.065 | Value:  0.719 | VLoss: 0.0325 | PLoss: -0.0045 | Ent: 0.1586
+[ 36.7%] Iter 3581/9765 | SPS:  1939 | Reward:  0.066 | Value:  0.754 | VLoss: 0.0298 | PLoss: -0.0126 | Ent: 0.2006
+[ 36.7%] Iter 3582/9765 | SPS:  1939 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0392 | PLoss: -0.0074 | Ent: 0.1724
+[ 36.7%] Iter 3583/9765 | SPS:  1940 | Reward:  0.058 | Value:  0.699 | VLoss: 0.0250 | PLoss: -0.0014 | Ent: 0.1996
+[ 36.7%] Iter 3584/9765 | SPS:  1940 | Reward:  0.080 | Value:  0.676 | VLoss: 0.0320 | PLoss: -0.0107 | Ent: 0.1849
+[ 36.7%] Iter 3585/9765 | SPS:  1940 | Reward:  0.063 | Value:  0.673 | VLoss: 0.0286 | PLoss: -0.0041 | Ent: 0.1662
+[ 36.7%] Iter 3586/9765 | SPS:  1940 | Reward:  0.071 | Value:  0.724 | VLoss: 0.0311 | PLoss: -0.0037 | Ent: 0.1133
+[ 36.7%] Iter 3587/9765 | SPS:  1941 | Reward:  0.068 | Value:  0.726 | VLoss: 0.0226 | PLoss: -0.0092 | Ent: 0.1526
+[ 36.7%] Iter 3588/9765 | SPS:  1941 | Reward:  0.073 | Value:  0.694 | VLoss: 0.0353 | PLoss: -0.0125 | Ent: 0.1409
+[ 36.8%] Iter 3589/9765 | SPS:  1941 | Reward:  0.061 | Value:  0.675 | VLoss: 0.0248 | PLoss: -0.0080 | Ent: 0.1463
+[ 36.8%] Iter 3590/9765 | SPS:  1941 | Reward:  0.073 | Value:  0.666 | VLoss: 0.0357 | PLoss: -0.0099 | Ent: 0.1884
+[ 36.8%] Iter 3591/9765 | SPS:  1942 | Reward:  0.060 | Value:  0.762 | VLoss: 0.0206 | PLoss: -0.0058 | Ent: 0.1848
+[ 36.8%] Iter 3592/9765 | SPS:  1942 | Reward:  0.069 | Value:  0.687 | VLoss: 0.0292 | PLoss: -0.0047 | Ent: 0.1648
+[ 36.8%] Iter 3593/9765 | SPS:  1942 | Reward:  0.069 | Value:  0.708 | VLoss: 0.0338 | PLoss: -0.0092 | Ent: 0.1831
+[ 36.8%] Iter 3594/9765 | SPS:  1942 | Reward:  0.077 | Value:  0.707 | VLoss: 0.0232 | PLoss: -0.0119 | Ent: 0.1713
+[ 36.8%] Iter 3595/9765 | SPS:  1943 | Reward:  0.065 | Value:  0.745 | VLoss: 0.0271 | PLoss: -0.0037 | Ent: 0.1704
+[ 36.8%] Iter 3596/9765 | SPS:  1943 | Reward:  0.066 | Value:  0.756 | VLoss: 0.0226 | PLoss: -0.0056 | Ent: 0.1732
+[ 36.8%] Iter 3597/9765 | SPS:  1943 | Reward:  0.060 | Value:  0.676 | VLoss: 0.0302 | PLoss: -0.0072 | Ent: 0.1624
+[ 36.8%] Iter 3598/9765 | SPS:  1943 | Reward:  0.058 | Value:  0.633 | VLoss: 0.0281 | PLoss: -0.0185 | Ent: 0.2281
+[ 36.9%] Iter 3599/9765 | SPS:  1944 | Reward:  0.069 | Value:  0.671 | VLoss: 0.0246 | PLoss: -0.0112 | Ent: 0.1774
+[ 36.9%] Iter 3600/9765 | SPS:  1944 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0349 | PLoss: -0.0145 | Ent: 0.2421
+[ 36.9%] Iter 3601/9765 | SPS:  1944 | Reward:  0.058 | Value:  0.629 | VLoss: 0.0389 | PLoss: -0.0104 | Ent: 0.1758
+[ 36.9%] Iter 3602/9765 | SPS:  1944 | Reward:  0.070 | Value:  0.669 | VLoss: 0.0250 | PLoss: -0.0117 | Ent: 0.1915
+[ 36.9%] Iter 3603/9765 | SPS:  1945 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0254 | PLoss: -0.0043 | Ent: 0.1521
+[ 36.9%] Iter 3604/9765 | SPS:  1945 | Reward:  0.062 | Value:  0.636 | VLoss: 0.0261 | PLoss: -0.0038 | Ent: 0.1228
+[ 36.9%] Iter 3605/9765 | SPS:  1945 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0348 | PLoss: -0.0146 | Ent: 0.1601
+[ 36.9%] Iter 3606/9765 | SPS:  1945 | Reward:  0.071 | Value:  0.692 | VLoss: 0.0286 | PLoss: -0.0127 | Ent: 0.1739
+[ 36.9%] Iter 3607/9765 | SPS:  1946 | Reward:  0.069 | Value:  0.738 | VLoss: 0.0231 | PLoss: -0.0033 | Ent: 0.2317
+[ 36.9%] Iter 3608/9765 | SPS:  1946 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0283 | PLoss: -0.0125 | Ent: 0.1971
+[ 37.0%] Iter 3609/9765 | SPS:  1946 | Reward:  0.073 | Value:  0.722 | VLoss: 0.0271 | PLoss: -0.0104 | Ent: 0.1735
+[ 37.0%] Iter 3610/9765 | SPS:  1946 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0294 | PLoss: -0.0049 | Ent: 0.1608
+[ 37.0%] Iter 3611/9765 | SPS:  1947 | Reward:  0.069 | Value:  0.665 | VLoss: 0.0288 | PLoss: -0.0040 | Ent: 0.1440
+[ 37.0%] Iter 3612/9765 | SPS:  1947 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0351 | PLoss: -0.0038 | Ent: 0.1235
+[ 37.0%] Iter 3613/9765 | SPS:  1947 | Reward:  0.078 | Value:  0.736 | VLoss: 0.0231 | PLoss: -0.0072 | Ent: 0.1766
+[ 37.0%] Iter 3614/9765 | SPS:  1947 | Reward:  0.058 | Value:  0.715 | VLoss: 0.0252 | PLoss: -0.0071 | Ent: 0.1983
+[ 37.0%] Iter 3615/9765 | SPS:  1948 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0207 | PLoss: -0.0071 | Ent: 0.1539
+[ 37.0%] Iter 3616/9765 | SPS:  1948 | Reward:  0.051 | Value:  0.757 | VLoss: 0.0270 | PLoss: -0.0035 | Ent: 0.2009
+[ 37.0%] Iter 3617/9765 | SPS:  1948 | Reward:  0.074 | Value:  0.681 | VLoss: 0.0280 | PLoss: -0.0058 | Ent: 0.1395
+[ 37.1%] Iter 3618/9765 | SPS:  1948 | Reward:  0.053 | Value:  0.762 | VLoss: 0.0197 | PLoss: 0.0028 | Ent: 0.1542
+[ 37.1%] Iter 3619/9765 | SPS:  1949 | Reward:  0.065 | Value:  0.676 | VLoss: 0.0274 | PLoss: -0.0070 | Ent: 0.1568
+[ 37.1%] Iter 3620/9765 | SPS:  1949 | Reward:  0.072 | Value:  0.666 | VLoss: 0.0323 | PLoss: 0.0031 | Ent: 0.1575
+[ 37.1%] Iter 3621/9765 | SPS:  1949 | Reward:  0.064 | Value:  0.638 | VLoss: 0.0304 | PLoss: -0.0107 | Ent: 0.1595
+[ 37.1%] Iter 3622/9765 | SPS:  1949 | Reward:  0.054 | Value:  0.620 | VLoss: 0.0298 | PLoss: -0.0150 | Ent: 0.1762
+[ 37.1%] Iter 3623/9765 | SPS:  1950 | Reward:  0.057 | Value:  0.636 | VLoss: 0.0206 | PLoss: -0.0098 | Ent: 0.2162
+[ 37.1%] Iter 3624/9765 | SPS:  1950 | Reward:  0.074 | Value:  0.705 | VLoss: 0.0217 | PLoss: -0.0124 | Ent: 0.1938
+[ 37.1%] Iter 3625/9765 | SPS:  1950 | Reward:  0.060 | Value:  0.718 | VLoss: 0.0228 | PLoss: -0.0041 | Ent: 0.1835
+[ 37.1%] Iter 3626/9765 | SPS:  1950 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0220 | PLoss: -0.0061 | Ent: 0.1675
+[ 37.1%] Iter 3627/9765 | SPS:  1951 | Reward:  0.060 | Value:  0.730 | VLoss: 0.0274 | PLoss: -0.0131 | Ent: 0.1664
+[ 37.2%] Iter 3628/9765 | SPS:  1951 | Reward:  0.068 | Value:  0.653 | VLoss: 0.0295 | PLoss: -0.0094 | Ent: 0.1887
+[ 37.2%] Iter 3629/9765 | SPS:  1951 | Reward:  0.067 | Value:  0.731 | VLoss: 0.0228 | PLoss: -0.0025 | Ent: 0.1352
+[ 37.2%] Iter 3630/9765 | SPS:  1951 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0274 | PLoss: -0.0050 | Ent: 0.1792
+[ 37.2%] Iter 3631/9765 | SPS:  1952 | Reward:  0.069 | Value:  0.734 | VLoss: 0.0353 | PLoss: -0.0076 | Ent: 0.1844
+[ 37.2%] Iter 3632/9765 | SPS:  1952 | Reward:  0.081 | Value:  0.667 | VLoss: 0.0238 | PLoss: -0.0079 | Ent: 0.2157
+[ 37.2%] Iter 3633/9765 | SPS:  1952 | Reward:  0.078 | Value:  0.735 | VLoss: 0.0288 | PLoss: -0.0071 | Ent: 0.1476
+[ 37.2%] Iter 3634/9765 | SPS:  1952 | Reward:  0.075 | Value:  0.726 | VLoss: 0.0286 | PLoss: -0.0092 | Ent: 0.1863
+[ 37.2%] Iter 3635/9765 | SPS:  1953 | Reward:  0.051 | Value:  0.762 | VLoss: 0.0168 | PLoss: -0.0186 | Ent: 0.1533
+[ 37.2%] Iter 3636/9765 | SPS:  1953 | Reward:  0.059 | Value:  0.692 | VLoss: 0.0214 | PLoss: -0.0055 | Ent: 0.1799
+[ 37.2%] Iter 3637/9765 | SPS:  1953 | Reward:  0.051 | Value:  0.681 | VLoss: 0.0264 | PLoss: -0.0089 | Ent: 0.1238
+[ 37.3%] Iter 3638/9765 | SPS:  1953 | Reward:  0.059 | Value:  0.641 | VLoss: 0.0256 | PLoss: -0.0060 | Ent: 0.1641
+[ 37.3%] Iter 3639/9765 | SPS:  1954 | Reward:  0.060 | Value:  0.738 | VLoss: 0.0295 | PLoss: -0.0104 | Ent: 0.1719
+[ 37.3%] Iter 3640/9765 | SPS:  1954 | Reward:  0.060 | Value:  0.744 | VLoss: 0.0227 | PLoss: -0.0015 | Ent: 0.2128
+[ 37.3%] Iter 3641/9765 | SPS:  1954 | Reward:  0.059 | Value:  0.706 | VLoss: 0.0311 | PLoss: -0.0046 | Ent: 0.1864
+[ 37.3%] Iter 3642/9765 | SPS:  1954 | Reward:  0.062 | Value:  0.612 | VLoss: 0.0331 | PLoss: -0.0036 | Ent: 0.1854
+[ 37.3%] Iter 3643/9765 | SPS:  1955 | Reward:  0.075 | Value:  0.627 | VLoss: 0.0247 | PLoss: -0.0077 | Ent: 0.1607
+[ 37.3%] Iter 3644/9765 | SPS:  1955 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0258 | PLoss: -0.0077 | Ent: 0.1865
+[ 37.3%] Iter 3645/9765 | SPS:  1955 | Reward:  0.076 | Value:  0.694 | VLoss: 0.0337 | PLoss: -0.0100 | Ent: 0.1999
+[ 37.3%] Iter 3646/9765 | SPS:  1955 | Reward:  0.071 | Value:  0.750 | VLoss: 0.0229 | PLoss: -0.0080 | Ent: 0.1811
+[ 37.3%] Iter 3647/9765 | SPS:  1956 | Reward:  0.070 | Value:  0.773 | VLoss: 0.0194 | PLoss: -0.0089 | Ent: 0.1514
+[ 37.4%] Iter 3648/9765 | SPS:  1956 | Reward:  0.074 | Value:  0.746 | VLoss: 0.0263 | PLoss: -0.0148 | Ent: 0.1912
+[ 37.4%] Iter 3649/9765 | SPS:  1956 | Reward:  0.060 | Value:  0.701 | VLoss: 0.0273 | PLoss: -0.0103 | Ent: 0.1687
+[ 37.4%] Iter 3650/9765 | SPS:  1956 | Reward:  0.062 | Value:  0.698 | VLoss: 0.0210 | PLoss: -0.0052 | Ent: 0.1533
+[ 37.4%] Iter 3651/9765 | SPS:  1956 | Reward:  0.071 | Value:  0.719 | VLoss: 0.0296 | PLoss: -0.0054 | Ent: 0.1716
+[ 37.4%] Iter 3652/9765 | SPS:  1957 | Reward:  0.083 | Value:  0.752 | VLoss: 0.0296 | PLoss: -0.0155 | Ent: 0.1902
+[ 37.4%] Iter 3653/9765 | SPS:  1957 | Reward:  0.064 | Value:  0.736 | VLoss: 0.0205 | PLoss: -0.0124 | Ent: 0.1928
+[ 37.4%] Iter 3654/9765 | SPS:  1957 | Reward:  0.072 | Value:  0.689 | VLoss: 0.0378 | PLoss: -0.0037 | Ent: 0.2108
+[ 37.4%] Iter 3655/9765 | SPS:  1957 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0254 | PLoss: -0.0090 | Ent: 0.1633
+[ 37.4%] Iter 3656/9765 | SPS:  1958 | Reward:  0.066 | Value:  0.721 | VLoss: 0.0260 | PLoss: -0.0066 | Ent: 0.1948
+[ 37.5%] Iter 3657/9765 | SPS:  1958 | Reward:  0.077 | Value:  0.709 | VLoss: 0.0319 | PLoss: -0.0061 | Ent: 0.1682
+[ 37.5%] Iter 3658/9765 | SPS:  1958 | Reward:  0.070 | Value:  0.715 | VLoss: 0.0295 | PLoss: -0.0034 | Ent: 0.1824
+[ 37.5%] Iter 3659/9765 | SPS:  1958 | Reward:  0.056 | Value:  0.655 | VLoss: 0.0285 | PLoss: -0.0032 | Ent: 0.1565
+[ 37.5%] Iter 3660/9765 | SPS:  1959 | Reward:  0.060 | Value:  0.681 | VLoss: 0.0260 | PLoss: -0.0076 | Ent: 0.1464
+[ 37.5%] Iter 3661/9765 | SPS:  1959 | Reward:  0.055 | Value:  0.675 | VLoss: 0.0218 | PLoss: -0.0041 | Ent: 0.1936
+[ 37.5%] Iter 3662/9765 | SPS:  1959 | Reward:  0.067 | Value:  0.648 | VLoss: 0.0257 | PLoss: -0.0042 | Ent: 0.2069
+[ 37.5%] Iter 3663/9765 | SPS:  1959 | Reward:  0.052 | Value:  0.707 | VLoss: 0.0270 | PLoss: -0.0060 | Ent: 0.2014
+[ 37.5%] Iter 3664/9765 | SPS:  1960 | Reward:  0.060 | Value:  0.664 | VLoss: 0.0230 | PLoss: -0.0124 | Ent: 0.2066
+[ 37.5%] Iter 3665/9765 | SPS:  1960 | Reward:  0.061 | Value:  0.707 | VLoss: 0.0232 | PLoss: -0.0064 | Ent: 0.2143
+[ 37.5%] Iter 3666/9765 | SPS:  1960 | Reward:  0.062 | Value:  0.709 | VLoss: 0.0311 | PLoss: -0.0039 | Ent: 0.1908
+[ 37.6%] Iter 3667/9765 | SPS:  1960 | Reward:  0.057 | Value:  0.674 | VLoss: 0.0267 | PLoss: -0.0070 | Ent: 0.1904
+[ 37.6%] Iter 3668/9765 | SPS:  1961 | Reward:  0.076 | Value:  0.635 | VLoss: 0.0333 | PLoss: -0.0052 | Ent: 0.1787
+[ 37.6%] Iter 3669/9765 | SPS:  1961 | Reward:  0.068 | Value:  0.704 | VLoss: 0.0270 | PLoss: -0.0078 | Ent: 0.1691
+[ 37.6%] Iter 3670/9765 | SPS:  1961 | Reward:  0.051 | Value:  0.671 | VLoss: 0.0186 | PLoss: -0.0046 | Ent: 0.1673
+[ 37.6%] Iter 3671/9765 | SPS:  1961 | Reward:  0.065 | Value:  0.642 | VLoss: 0.0308 | PLoss: -0.0100 | Ent: 0.2151
+[ 37.6%] Iter 3672/9765 | SPS:  1962 | Reward:  0.064 | Value:  0.668 | VLoss: 0.0217 | PLoss: -0.0045 | Ent: 0.1576
+[ 37.6%] Iter 3673/9765 | SPS:  1962 | Reward:  0.063 | Value:  0.652 | VLoss: 0.0207 | PLoss: -0.0100 | Ent: 0.1721
+[ 37.6%] Iter 3674/9765 | SPS:  1962 | Reward:  0.070 | Value:  0.665 | VLoss: 0.0341 | PLoss: -0.0051 | Ent: 0.1919
+[ 37.6%] Iter 3675/9765 | SPS:  1962 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0294 | PLoss: -0.0082 | Ent: 0.1929
+[ 37.6%] Iter 3676/9765 | SPS:  1963 | Reward:  0.063 | Value:  0.657 | VLoss: 0.0325 | PLoss: -0.0088 | Ent: 0.2245
+[ 37.7%] Iter 3677/9765 | SPS:  1963 | Reward:  0.061 | Value:  0.679 | VLoss: 0.0279 | PLoss: -0.0184 | Ent: 0.2558
+[ 37.7%] Iter 3678/9765 | SPS:  1963 | Reward:  0.043 | Value:  0.682 | VLoss: 0.0192 | PLoss: -0.0056 | Ent: 0.2076
+[ 37.7%] Iter 3679/9765 | SPS:  1963 | Reward:  0.061 | Value:  0.687 | VLoss: 0.0165 | PLoss: -0.0072 | Ent: 0.1708
+[ 37.7%] Iter 3680/9765 | SPS:  1964 | Reward:  0.071 | Value:  0.717 | VLoss: 0.0241 | PLoss: -0.0044 | Ent: 0.2074
+[ 37.7%] Iter 3681/9765 | SPS:  1964 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0304 | PLoss: -0.0088 | Ent: 0.1721
+[ 37.7%] Iter 3682/9765 | SPS:  1964 | Reward:  0.051 | Value:  0.719 | VLoss: 0.0218 | PLoss: -0.0084 | Ent: 0.1804
+[ 37.7%] Iter 3683/9765 | SPS:  1964 | Reward:  0.053 | Value:  0.625 | VLoss: 0.0139 | PLoss: -0.0077 | Ent: 0.1415
+[ 37.7%] Iter 3684/9765 | SPS:  1965 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0278 | PLoss: -0.0064 | Ent: 0.2068
+[ 37.7%] Iter 3685/9765 | SPS:  1965 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0264 | PLoss: -0.0008 | Ent: 0.1810
+[ 37.7%] Iter 3686/9765 | SPS:  1965 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0256 | PLoss: -0.0020 | Ent: 0.1475
+[ 37.8%] Iter 3687/9765 | SPS:  1965 | Reward:  0.062 | Value:  0.612 | VLoss: 0.0346 | PLoss: -0.0058 | Ent: 0.1688
+[ 37.8%] Iter 3688/9765 | SPS:  1966 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0232 | PLoss: -0.0107 | Ent: 0.2173
+[ 37.8%] Iter 3689/9765 | SPS:  1966 | Reward:  0.068 | Value:  0.742 | VLoss: 0.0226 | PLoss: -0.0088 | Ent: 0.1556
+[ 37.8%] Iter 3690/9765 | SPS:  1966 | Reward:  0.050 | Value:  0.664 | VLoss: 0.0241 | PLoss: -0.0001 | Ent: 0.1948
+[ 37.8%] Iter 3691/9765 | SPS:  1966 | Reward:  0.070 | Value:  0.662 | VLoss: 0.0332 | PLoss: -0.0045 | Ent: 0.2016
+[ 37.8%] Iter 3692/9765 | SPS:  1967 | Reward:  0.072 | Value:  0.680 | VLoss: 0.0266 | PLoss: -0.0108 | Ent: 0.1527
+[ 37.8%] Iter 3693/9765 | SPS:  1967 | Reward:  0.059 | Value:  0.709 | VLoss: 0.0253 | PLoss: -0.0090 | Ent: 0.1829
+[ 37.8%] Iter 3694/9765 | SPS:  1967 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0198 | PLoss: -0.0058 | Ent: 0.1854
+[ 37.8%] Iter 3695/9765 | SPS:  1967 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0212 | PLoss: -0.0059 | Ent: 0.1530
+[ 37.8%] Iter 3696/9765 | SPS:  1968 | Reward:  0.056 | Value:  0.708 | VLoss: 0.0223 | PLoss: -0.0046 | Ent: 0.2169
+[ 37.9%] Iter 3697/9765 | SPS:  1968 | Reward:  0.074 | Value:  0.663 | VLoss: 0.0337 | PLoss: -0.0029 | Ent: 0.1371
+[ 37.9%] Iter 3698/9765 | SPS:  1968 | Reward:  0.059 | Value:  0.690 | VLoss: 0.0308 | PLoss: -0.0040 | Ent: 0.1799
+[ 37.9%] Iter 3699/9765 | SPS:  1968 | Reward:  0.060 | Value:  0.681 | VLoss: 0.0232 | PLoss: -0.0136 | Ent: 0.1877
+[ 37.9%] Iter 3700/9765 | SPS:  1969 | Reward:  0.048 | Value:  0.699 | VLoss: 0.0296 | PLoss: -0.0016 | Ent: 0.2150
+[ 37.9%] Iter 3701/9765 | SPS:  1969 | Reward:  0.063 | Value:  0.661 | VLoss: 0.0260 | PLoss: -0.0148 | Ent: 0.1750
+[ 37.9%] Iter 3702/9765 | SPS:  1969 | Reward:  0.061 | Value:  0.647 | VLoss: 0.0296 | PLoss: -0.0072 | Ent: 0.1842
+[ 37.9%] Iter 3703/9765 | SPS:  1969 | Reward:  0.065 | Value:  0.721 | VLoss: 0.0260 | PLoss: -0.0068 | Ent: 0.1491
+[ 37.9%] Iter 3704/9765 | SPS:  1970 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0326 | PLoss: -0.0059 | Ent: 0.1654
+[ 37.9%] Iter 3705/9765 | SPS:  1970 | Reward:  0.059 | Value:  0.620 | VLoss: 0.0310 | PLoss: -0.0073 | Ent: 0.2324
+[ 38.0%] Iter 3706/9765 | SPS:  1970 | Reward:  0.060 | Value:  0.727 | VLoss: 0.0257 | PLoss: -0.0107 | Ent: 0.1684
+[ 38.0%] Iter 3707/9765 | SPS:  1970 | Reward:  0.065 | Value:  0.643 | VLoss: 0.0252 | PLoss: -0.0119 | Ent: 0.1624
+[ 38.0%] Iter 3708/9765 | SPS:  1971 | Reward:  0.060 | Value:  0.709 | VLoss: 0.0214 | PLoss: -0.0075 | Ent: 0.1232
+[ 38.0%] Iter 3709/9765 | SPS:  1971 | Reward:  0.073 | Value:  0.724 | VLoss: 0.0205 | PLoss: -0.0057 | Ent: 0.1097
+[ 38.0%] Iter 3710/9765 | SPS:  1971 | Reward:  0.062 | Value:  0.689 | VLoss: 0.0285 | PLoss: -0.0013 | Ent: 0.1673
+[ 38.0%] Iter 3711/9765 | SPS:  1971 | Reward:  0.069 | Value:  0.668 | VLoss: 0.0317 | PLoss: -0.0080 | Ent: 0.1809
+[ 38.0%] Iter 3712/9765 | SPS:  1972 | Reward:  0.054 | Value:  0.676 | VLoss: 0.0260 | PLoss: -0.0063 | Ent: 0.1799
+[ 38.0%] Iter 3713/9765 | SPS:  1972 | Reward:  0.063 | Value:  0.673 | VLoss: 0.0198 | PLoss: -0.0069 | Ent: 0.1316
+[ 38.0%] Iter 3714/9765 | SPS:  1972 | Reward:  0.054 | Value:  0.674 | VLoss: 0.0239 | PLoss: -0.0033 | Ent: 0.1227
+[ 38.0%] Iter 3715/9765 | SPS:  1972 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0194 | PLoss: -0.0084 | Ent: 0.1528
+[ 38.1%] Iter 3716/9765 | SPS:  1973 | Reward:  0.051 | Value:  0.696 | VLoss: 0.0246 | PLoss: -0.0089 | Ent: 0.1442
+[ 38.1%] Iter 3717/9765 | SPS:  1973 | Reward:  0.046 | Value:  0.632 | VLoss: 0.0226 | PLoss: -0.0086 | Ent: 0.1433
+[ 38.1%] Iter 3718/9765 | SPS:  1973 | Reward:  0.069 | Value:  0.643 | VLoss: 0.0342 | PLoss: -0.0099 | Ent: 0.1733
+[ 38.1%] Iter 3719/9765 | SPS:  1973 | Reward:  0.057 | Value:  0.701 | VLoss: 0.0220 | PLoss: -0.0099 | Ent: 0.1656
+[ 38.1%] Iter 3720/9765 | SPS:  1974 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0287 | PLoss: -0.0034 | Ent: 0.1436
+[ 38.1%] Iter 3721/9765 | SPS:  1974 | Reward:  0.071 | Value:  0.658 | VLoss: 0.0348 | PLoss: -0.0081 | Ent: 0.1685
+[ 38.1%] Iter 3722/9765 | SPS:  1974 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0224 | PLoss: -0.0111 | Ent: 0.1797
+[ 38.1%] Iter 3723/9765 | SPS:  1974 | Reward:  0.069 | Value:  0.692 | VLoss: 0.0275 | PLoss: -0.0128 | Ent: 0.1627
+[ 38.1%] Iter 3724/9765 | SPS:  1975 | Reward:  0.073 | Value:  0.706 | VLoss: 0.0256 | PLoss: -0.0014 | Ent: 0.2073
+[ 38.1%] Iter 3725/9765 | SPS:  1975 | Reward:  0.056 | Value:  0.709 | VLoss: 0.0229 | PLoss: -0.0080 | Ent: 0.1788
+[ 38.2%] Iter 3726/9765 | SPS:  1975 | Reward:  0.073 | Value:  0.760 | VLoss: 0.0238 | PLoss: -0.0088 | Ent: 0.1732
+[ 38.2%] Iter 3727/9765 | SPS:  1975 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0243 | PLoss: -0.0060 | Ent: 0.1618
+[ 38.2%] Iter 3728/9765 | SPS:  1976 | Reward:  0.050 | Value:  0.704 | VLoss: 0.0220 | PLoss: -0.0041 | Ent: 0.1842
+[ 38.2%] Iter 3729/9765 | SPS:  1976 | Reward:  0.074 | Value:  0.669 | VLoss: 0.0271 | PLoss: -0.0093 | Ent: 0.2283
+[ 38.2%] Iter 3730/9765 | SPS:  1976 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0216 | PLoss: -0.0046 | Ent: 0.1973
+[ 38.2%] Iter 3731/9765 | SPS:  1976 | Reward:  0.069 | Value:  0.690 | VLoss: 0.0385 | PLoss: -0.0069 | Ent: 0.2192
+[ 38.2%] Iter 3732/9765 | SPS:  1977 | Reward:  0.065 | Value:  0.696 | VLoss: 0.0241 | PLoss: -0.0032 | Ent: 0.1830
+[ 38.2%] Iter 3733/9765 | SPS:  1977 | Reward:  0.065 | Value:  0.711 | VLoss: 0.0201 | PLoss: -0.0062 | Ent: 0.1782
+[ 38.2%] Iter 3734/9765 | SPS:  1977 | Reward:  0.052 | Value:  0.725 | VLoss: 0.0263 | PLoss: -0.0142 | Ent: 0.1651
+[ 38.2%] Iter 3735/9765 | SPS:  1977 | Reward:  0.043 | Value:  0.665 | VLoss: 0.0196 | PLoss: -0.0001 | Ent: 0.1635
+[ 38.3%] Iter 3736/9765 | SPS:  1978 | Reward:  0.063 | Value:  0.664 | VLoss: 0.0284 | PLoss: -0.0106 | Ent: 0.2025
+[ 38.3%] Iter 3737/9765 | SPS:  1978 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0321 | PLoss: -0.0127 | Ent: 0.1617
+[ 38.3%] Iter 3738/9765 | SPS:  1978 | Reward:  0.072 | Value:  0.727 | VLoss: 0.0231 | PLoss: -0.0013 | Ent: 0.1617
+[ 38.3%] Iter 3739/9765 | SPS:  1978 | Reward:  0.073 | Value:  0.707 | VLoss: 0.0312 | PLoss: -0.0056 | Ent: 0.1669
+[ 38.3%] Iter 3740/9765 | SPS:  1979 | Reward:  0.077 | Value:  0.746 | VLoss: 0.0248 | PLoss: -0.0118 | Ent: 0.1551
+[ 38.3%] Iter 3741/9765 | SPS:  1979 | Reward:  0.072 | Value:  0.722 | VLoss: 0.0310 | PLoss: -0.0118 | Ent: 0.1719
+[ 38.3%] Iter 3742/9765 | SPS:  1979 | Reward:  0.055 | Value:  0.681 | VLoss: 0.0249 | PLoss: -0.0132 | Ent: 0.1946
+[ 38.3%] Iter 3743/9765 | SPS:  1979 | Reward:  0.049 | Value:  0.672 | VLoss: 0.0243 | PLoss: -0.0077 | Ent: 0.1712
+[ 38.3%] Iter 3744/9765 | SPS:  1980 | Reward:  0.055 | Value:  0.676 | VLoss: 0.0158 | PLoss: -0.0122 | Ent: 0.1635
+[ 38.4%] Iter 3745/9765 | SPS:  1980 | Reward:  0.062 | Value:  0.600 | VLoss: 0.0259 | PLoss: -0.0109 | Ent: 0.1823
+[ 38.4%] Iter 3746/9765 | SPS:  1980 | Reward:  0.062 | Value:  0.774 | VLoss: 0.0185 | PLoss: -0.0099 | Ent: 0.1589
+[ 38.4%] Iter 3747/9765 | SPS:  1980 | Reward:  0.079 | Value:  0.693 | VLoss: 0.0277 | PLoss: -0.0057 | Ent: 0.1957
+[ 38.4%] Iter 3748/9765 | SPS:  1981 | Reward:  0.057 | Value:  0.685 | VLoss: 0.0262 | PLoss: -0.0049 | Ent: 0.1930
+[ 38.4%] Iter 3749/9765 | SPS:  1981 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0291 | PLoss: -0.0066 | Ent: 0.1320
+[ 38.4%] Iter 3750/9765 | SPS:  1981 | Reward:  0.062 | Value:  0.627 | VLoss: 0.0338 | PLoss: -0.0121 | Ent: 0.2154
+[ 38.4%] Iter 3751/9765 | SPS:  1981 | Reward:  0.079 | Value:  0.699 | VLoss: 0.0266 | PLoss: -0.0056 | Ent: 0.2119
+[ 38.4%] Iter 3752/9765 | SPS:  1982 | Reward:  0.055 | Value:  0.713 | VLoss: 0.0304 | PLoss: -0.0066 | Ent: 0.2020
+[ 38.4%] Iter 3753/9765 | SPS:  1982 | Reward:  0.063 | Value:  0.677 | VLoss: 0.0186 | PLoss: -0.0083 | Ent: 0.1917
+[ 38.4%] Iter 3754/9765 | SPS:  1982 | Reward:  0.059 | Value:  0.688 | VLoss: 0.0214 | PLoss: -0.0114 | Ent: 0.1712
+[ 38.5%] Iter 3755/9765 | SPS:  1982 | Reward:  0.071 | Value:  0.731 | VLoss: 0.0213 | PLoss: -0.0036 | Ent: 0.1495
+[ 38.5%] Iter 3756/9765 | SPS:  1982 | Reward:  0.068 | Value:  0.723 | VLoss: 0.0263 | PLoss: -0.0057 | Ent: 0.1894
+[ 38.5%] Iter 3757/9765 | SPS:  1983 | Reward:  0.056 | Value:  0.764 | VLoss: 0.0205 | PLoss: -0.0048 | Ent: 0.1972
+[ 38.5%] Iter 3758/9765 | SPS:  1983 | Reward:  0.048 | Value:  0.634 | VLoss: 0.0302 | PLoss: -0.0034 | Ent: 0.2270
+[ 38.5%] Iter 3759/9765 | SPS:  1983 | Reward:  0.058 | Value:  0.719 | VLoss: 0.0211 | PLoss: -0.0162 | Ent: 0.1869
+[ 38.5%] Iter 3760/9765 | SPS:  1984 | Reward:  0.043 | Value:  0.653 | VLoss: 0.0225 | PLoss: -0.0042 | Ent: 0.1743
+[ 38.5%] Iter 3761/9765 | SPS:  1984 | Reward:  0.061 | Value:  0.713 | VLoss: 0.0226 | PLoss: -0.0005 | Ent: 0.1612
+[ 38.5%] Iter 3762/9765 | SPS:  1984 | Reward:  0.069 | Value:  0.672 | VLoss: 0.0313 | PLoss: -0.0049 | Ent: 0.2108
+[ 38.5%] Iter 3763/9765 | SPS:  1984 | Reward:  0.063 | Value:  0.702 | VLoss: 0.0249 | PLoss: -0.0064 | Ent: 0.1993
+[ 38.5%] Iter 3764/9765 | SPS:  1984 | Reward:  0.071 | Value:  0.698 | VLoss: 0.0286 | PLoss: -0.0114 | Ent: 0.1559
+[ 38.6%] Iter 3765/9765 | SPS:  1985 | Reward:  0.067 | Value:  0.748 | VLoss: 0.0244 | PLoss: -0.0063 | Ent: 0.1387
+[ 38.6%] Iter 3766/9765 | SPS:  1985 | Reward:  0.063 | Value:  0.750 | VLoss: 0.0270 | PLoss: -0.0088 | Ent: 0.1722
+[ 38.6%] Iter 3767/9765 | SPS:  1985 | Reward:  0.076 | Value:  0.727 | VLoss: 0.0249 | PLoss: -0.0082 | Ent: 0.1497
+[ 38.6%] Iter 3768/9765 | SPS:  1985 | Reward:  0.053 | Value:  0.729 | VLoss: 0.0195 | PLoss: -0.0044 | Ent: 0.1761
+[ 38.6%] Iter 3769/9765 | SPS:  1986 | Reward:  0.069 | Value:  0.724 | VLoss: 0.0273 | PLoss: -0.0057 | Ent: 0.1517
+[ 38.6%] Iter 3770/9765 | SPS:  1986 | Reward:  0.077 | Value:  0.707 | VLoss: 0.0265 | PLoss: -0.0075 | Ent: 0.2074
+[ 38.6%] Iter 3771/9765 | SPS:  1986 | Reward:  0.060 | Value:  0.750 | VLoss: 0.0267 | PLoss: -0.0087 | Ent: 0.1786
+[ 38.6%] Iter 3772/9765 | SPS:  1986 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0308 | PLoss: -0.0075 | Ent: 0.1698
+[ 38.6%] Iter 3773/9765 | SPS:  1987 | Reward:  0.068 | Value:  0.729 | VLoss: 0.0268 | PLoss: -0.0056 | Ent: 0.1595
+[ 38.6%] Iter 3774/9765 | SPS:  1987 | Reward:  0.075 | Value:  0.714 | VLoss: 0.0295 | PLoss: -0.0045 | Ent: 0.1883
+[ 38.7%] Iter 3775/9765 | SPS:  1987 | Reward:  0.057 | Value:  0.748 | VLoss: 0.0308 | PLoss: -0.0120 | Ent: 0.1472
+[ 38.7%] Iter 3776/9765 | SPS:  1987 | Reward:  0.061 | Value:  0.640 | VLoss: 0.0319 | PLoss: -0.0024 | Ent: 0.1608
+[ 38.7%] Iter 3777/9765 | SPS:  1988 | Reward:  0.058 | Value:  0.728 | VLoss: 0.0240 | PLoss: -0.0040 | Ent: 0.2073
+[ 38.7%] Iter 3778/9765 | SPS:  1988 | Reward:  0.057 | Value:  0.660 | VLoss: 0.0233 | PLoss: -0.0088 | Ent: 0.1600
+[ 38.7%] Iter 3779/9765 | SPS:  1988 | Reward:  0.063 | Value:  0.699 | VLoss: 0.0336 | PLoss: -0.0109 | Ent: 0.1668
+[ 38.7%] Iter 3780/9765 | SPS:  1988 | Reward:  0.068 | Value:  0.672 | VLoss: 0.0269 | PLoss: -0.0037 | Ent: 0.1630
+[ 38.7%] Iter 3781/9765 | SPS:  1989 | Reward:  0.057 | Value:  0.703 | VLoss: 0.0259 | PLoss: -0.0120 | Ent: 0.1741
+[ 38.7%] Iter 3782/9765 | SPS:  1989 | Reward:  0.067 | Value:  0.683 | VLoss: 0.0354 | PLoss: -0.0089 | Ent: 0.1911
+[ 38.7%] Iter 3783/9765 | SPS:  1989 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0280 | PLoss: -0.0072 | Ent: 0.1842
+[ 38.8%] Iter 3784/9765 | SPS:  1989 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0291 | PLoss: -0.0085 | Ent: 0.1510
+[ 38.8%] Iter 3785/9765 | SPS:  1990 | Reward:  0.067 | Value:  0.658 | VLoss: 0.0336 | PLoss: -0.0076 | Ent: 0.1775
+[ 38.8%] Iter 3786/9765 | SPS:  1990 | Reward:  0.055 | Value:  0.728 | VLoss: 0.0297 | PLoss: -0.0059 | Ent: 0.2069
+[ 38.8%] Iter 3787/9765 | SPS:  1990 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0236 | PLoss: -0.0079 | Ent: 0.2148
+[ 38.8%] Iter 3788/9765 | SPS:  1990 | Reward:  0.066 | Value:  0.695 | VLoss: 0.0292 | PLoss: -0.0033 | Ent: 0.2123
+[ 38.8%] Iter 3789/9765 | SPS:  1991 | Reward:  0.053 | Value:  0.691 | VLoss: 0.0232 | PLoss: -0.0075 | Ent: 0.2287
+[ 38.8%] Iter 3790/9765 | SPS:  1991 | Reward:  0.070 | Value:  0.639 | VLoss: 0.0309 | PLoss: -0.0062 | Ent: 0.1854
+[ 38.8%] Iter 3791/9765 | SPS:  1991 | Reward:  0.067 | Value:  0.753 | VLoss: 0.0285 | PLoss: -0.0045 | Ent: 0.1794
+[ 38.8%] Iter 3792/9765 | SPS:  1991 | Reward:  0.072 | Value:  0.712 | VLoss: 0.0258 | PLoss: -0.0040 | Ent: 0.1789
+[ 38.8%] Iter 3793/9765 | SPS:  1992 | Reward:  0.073 | Value:  0.677 | VLoss: 0.0327 | PLoss: -0.0124 | Ent: 0.1700
+[ 38.9%] Iter 3794/9765 | SPS:  1992 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0265 | PLoss: -0.0076 | Ent: 0.1964
+[ 38.9%] Iter 3795/9765 | SPS:  1992 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0201 | PLoss: -0.0066 | Ent: 0.1787
+[ 38.9%] Iter 3796/9765 | SPS:  1992 | Reward:  0.054 | Value:  0.715 | VLoss: 0.0333 | PLoss: -0.0063 | Ent: 0.1680
+[ 38.9%] Iter 3797/9765 | SPS:  1993 | Reward:  0.074 | Value:  0.699 | VLoss: 0.0266 | PLoss: -0.0121 | Ent: 0.1559
+[ 38.9%] Iter 3798/9765 | SPS:  1993 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0224 | PLoss: -0.0046 | Ent: 0.1582
+[ 38.9%] Iter 3799/9765 | SPS:  1993 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0341 | PLoss: -0.0031 | Ent: 0.1413
+[ 38.9%] Iter 3800/9765 | SPS:  1993 | Reward:  0.055 | Value:  0.685 | VLoss: 0.0255 | PLoss: -0.0099 | Ent: 0.1852
+[ 38.9%] Iter 3801/9765 | SPS:  1993 | Reward:  0.054 | Value:  0.625 | VLoss: 0.0277 | PLoss: -0.0066 | Ent: 0.1941
+[ 38.9%] Iter 3802/9765 | SPS:  1994 | Reward:  0.054 | Value:  0.725 | VLoss: 0.0211 | PLoss: -0.0061 | Ent: 0.1681
+[ 38.9%] Iter 3803/9765 | SPS:  1994 | Reward:  0.068 | Value:  0.676 | VLoss: 0.0279 | PLoss: -0.0117 | Ent: 0.2101
+[ 39.0%] Iter 3804/9765 | SPS:  1994 | Reward:  0.062 | Value:  0.622 | VLoss: 0.0352 | PLoss: -0.0072 | Ent: 0.1604
+[ 39.0%] Iter 3805/9765 | SPS:  1994 | Reward:  0.043 | Value:  0.729 | VLoss: 0.0139 | PLoss: -0.0033 | Ent: 0.1569
+[ 39.0%] Iter 3806/9765 | SPS:  1995 | Reward:  0.074 | Value:  0.651 | VLoss: 0.0272 | PLoss: -0.0116 | Ent: 0.1952
+[ 39.0%] Iter 3807/9765 | SPS:  1995 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0245 | PLoss: -0.0044 | Ent: 0.1713
+[ 39.0%] Iter 3808/9765 | SPS:  1995 | Reward:  0.053 | Value:  0.711 | VLoss: 0.0185 | PLoss: -0.0025 | Ent: 0.1221
+[ 39.0%] Iter 3809/9765 | SPS:  1995 | Reward:  0.069 | Value:  0.706 | VLoss: 0.0254 | PLoss: -0.0157 | Ent: 0.1595
+[ 39.0%] Iter 3810/9765 | SPS:  1996 | Reward:  0.054 | Value:  0.746 | VLoss: 0.0201 | PLoss: -0.0058 | Ent: 0.1919
+[ 39.0%] Iter 3811/9765 | SPS:  1996 | Reward:  0.067 | Value:  0.716 | VLoss: 0.0282 | PLoss: -0.0006 | Ent: 0.1848
+[ 39.0%] Iter 3812/9765 | SPS:  1996 | Reward:  0.057 | Value:  0.744 | VLoss: 0.0217 | PLoss: -0.0040 | Ent: 0.1903
+[ 39.0%] Iter 3813/9765 | SPS:  1996 | Reward:  0.049 | Value:  0.710 | VLoss: 0.0200 | PLoss: -0.0076 | Ent: 0.2343
+[ 39.1%] Iter 3814/9765 | SPS:  1997 | Reward:  0.062 | Value:  0.744 | VLoss: 0.0368 | PLoss: -0.0054 | Ent: 0.1855
+[ 39.1%] Iter 3815/9765 | SPS:  1997 | Reward:  0.077 | Value:  0.642 | VLoss: 0.0395 | PLoss: -0.0122 | Ent: 0.1749
+[ 39.1%] Iter 3816/9765 | SPS:  1997 | Reward:  0.067 | Value:  0.667 | VLoss: 0.0269 | PLoss: -0.0078 | Ent: 0.2025
+[ 39.1%] Iter 3817/9765 | SPS:  1997 | Reward:  0.048 | Value:  0.787 | VLoss: 0.0218 | PLoss: -0.0042 | Ent: 0.1492
+[ 39.1%] Iter 3818/9765 | SPS:  1998 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0229 | PLoss: -0.0098 | Ent: 0.1649
+[ 39.1%] Iter 3819/9765 | SPS:  1998 | Reward:  0.072 | Value:  0.735 | VLoss: 0.0250 | PLoss: -0.0072 | Ent: 0.2005
+[ 39.1%] Iter 3820/9765 | SPS:  1998 | Reward:  0.055 | Value:  0.626 | VLoss: 0.0268 | PLoss: -0.0116 | Ent: 0.1671
+[ 39.1%] Iter 3821/9765 | SPS:  1998 | Reward:  0.071 | Value:  0.665 | VLoss: 0.0398 | PLoss: -0.0081 | Ent: 0.1970
+[ 39.1%] Iter 3822/9765 | SPS:  1999 | Reward:  0.060 | Value:  0.624 | VLoss: 0.0284 | PLoss: -0.0117 | Ent: 0.1550
+[ 39.2%] Iter 3823/9765 | SPS:  1999 | Reward:  0.064 | Value:  0.653 | VLoss: 0.0262 | PLoss: -0.0102 | Ent: 0.1600
+[ 39.2%] Iter 3824/9765 | SPS:  1999 | Reward:  0.076 | Value:  0.738 | VLoss: 0.0323 | PLoss: -0.0049 | Ent: 0.1411
+[ 39.2%] Iter 3825/9765 | SPS:  1999 | Reward:  0.075 | Value:  0.780 | VLoss: 0.0184 | PLoss: -0.0055 | Ent: 0.1677
+[ 39.2%] Iter 3826/9765 | SPS:  1999 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0343 | PLoss: -0.0086 | Ent: 0.1767
+[ 39.2%] Iter 3827/9765 | SPS:  2000 | Reward:  0.064 | Value:  0.695 | VLoss: 0.0209 | PLoss: -0.0089 | Ent: 0.1641
+[ 39.2%] Iter 3828/9765 | SPS:  2000 | Reward:  0.064 | Value:  0.661 | VLoss: 0.0236 | PLoss: -0.0085 | Ent: 0.1646
+[ 39.2%] Iter 3829/9765 | SPS:  2000 | Reward:  0.070 | Value:  0.712 | VLoss: 0.0255 | PLoss: -0.0030 | Ent: 0.1753
+[ 39.2%] Iter 3830/9765 | SPS:  2000 | Reward:  0.065 | Value:  0.717 | VLoss: 0.0238 | PLoss: -0.0108 | Ent: 0.1440
+[ 39.2%] Iter 3831/9765 | SPS:  2001 | Reward:  0.076 | Value:  0.691 | VLoss: 0.0257 | PLoss: -0.0015 | Ent: 0.1783
+[ 39.2%] Iter 3832/9765 | SPS:  2001 | Reward:  0.070 | Value:  0.712 | VLoss: 0.0331 | PLoss: -0.0062 | Ent: 0.2041
+[ 39.3%] Iter 3833/9765 | SPS:  2001 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0245 | PLoss: -0.0047 | Ent: 0.1505
+[ 39.3%] Iter 3834/9765 | SPS:  2001 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0261 | PLoss: -0.0044 | Ent: 0.1800
+[ 39.3%] Iter 3835/9765 | SPS:  2002 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0220 | PLoss: -0.0066 | Ent: 0.1644
+[ 39.3%] Iter 3836/9765 | SPS:  2002 | Reward:  0.059 | Value:  0.737 | VLoss: 0.0246 | PLoss: -0.0075 | Ent: 0.1943
+[ 39.3%] Iter 3837/9765 | SPS:  2002 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0277 | PLoss: -0.0043 | Ent: 0.1362
+[ 39.3%] Iter 3838/9765 | SPS:  2002 | Reward:  0.069 | Value:  0.731 | VLoss: 0.0229 | PLoss: -0.0120 | Ent: 0.1746
+[ 39.3%] Iter 3839/9765 | SPS:  2003 | Reward:  0.048 | Value:  0.700 | VLoss: 0.0152 | PLoss: 0.0004 | Ent: 0.1735
+[ 39.3%] Iter 3840/9765 | SPS:  2003 | Reward:  0.057 | Value:  0.737 | VLoss: 0.0246 | PLoss: -0.0084 | Ent: 0.1805
+[ 39.3%] Iter 3841/9765 | SPS:  2003 | Reward:  0.066 | Value:  0.718 | VLoss: 0.0228 | PLoss: -0.0136 | Ent: 0.2360
+[ 39.3%] Iter 3842/9765 | SPS:  2003 | Reward:  0.060 | Value:  0.730 | VLoss: 0.0244 | PLoss: -0.0074 | Ent: 0.1514
+[ 39.4%] Iter 3843/9765 | SPS:  2003 | Reward:  0.058 | Value:  0.724 | VLoss: 0.0219 | PLoss: -0.0019 | Ent: 0.1559
+[ 39.4%] Iter 3844/9765 | SPS:  2004 | Reward:  0.066 | Value:  0.675 | VLoss: 0.0319 | PLoss: -0.0078 | Ent: 0.1498
+[ 39.4%] Iter 3845/9765 | SPS:  2004 | Reward:  0.059 | Value:  0.645 | VLoss: 0.0256 | PLoss: -0.0095 | Ent: 0.1769
+[ 39.4%] Iter 3846/9765 | SPS:  2004 | Reward:  0.063 | Value:  0.657 | VLoss: 0.0275 | PLoss: -0.0072 | Ent: 0.1690
+[ 39.4%] Iter 3847/9765 | SPS:  2004 | Reward:  0.062 | Value:  0.729 | VLoss: 0.0227 | PLoss: -0.0060 | Ent: 0.1523
+[ 39.4%] Iter 3848/9765 | SPS:  2005 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0292 | PLoss: -0.0078 | Ent: 0.1412
+[ 39.4%] Iter 3849/9765 | SPS:  2005 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0327 | PLoss: -0.0180 | Ent: 0.1740
+[ 39.4%] Iter 3850/9765 | SPS:  2005 | Reward:  0.074 | Value:  0.640 | VLoss: 0.0284 | PLoss: -0.0027 | Ent: 0.1376
+[ 39.4%] Iter 3851/9765 | SPS:  2005 | Reward:  0.053 | Value:  0.688 | VLoss: 0.0325 | PLoss: -0.0039 | Ent: 0.1570
+[ 39.4%] Iter 3852/9765 | SPS:  2006 | Reward:  0.048 | Value:  0.704 | VLoss: 0.0167 | PLoss: -0.0141 | Ent: 0.1909
+[ 39.5%] Iter 3853/9765 | SPS:  2006 | Reward:  0.060 | Value:  0.720 | VLoss: 0.0209 | PLoss: -0.0100 | Ent: 0.1677
+[ 39.5%] Iter 3854/9765 | SPS:  2006 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0199 | PLoss: -0.0039 | Ent: 0.1468
+[ 39.5%] Iter 3855/9765 | SPS:  2006 | Reward:  0.066 | Value:  0.752 | VLoss: 0.0276 | PLoss: -0.0086 | Ent: 0.1636
+[ 39.5%] Iter 3856/9765 | SPS:  2007 | Reward:  0.061 | Value:  0.799 | VLoss: 0.0180 | PLoss: -0.0078 | Ent: 0.1739
+[ 39.5%] Iter 3857/9765 | SPS:  2007 | Reward:  0.056 | Value:  0.651 | VLoss: 0.0303 | PLoss: -0.0145 | Ent: 0.1818
+[ 39.5%] Iter 3858/9765 | SPS:  2007 | Reward:  0.059 | Value:  0.630 | VLoss: 0.0223 | PLoss: -0.0066 | Ent: 0.1535
+[ 39.5%] Iter 3859/9765 | SPS:  2007 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0242 | PLoss: -0.0088 | Ent: 0.1413
+[ 39.5%] Iter 3860/9765 | SPS:  2008 | Reward:  0.062 | Value:  0.785 | VLoss: 0.0168 | PLoss: -0.0081 | Ent: 0.1808
+[ 39.5%] Iter 3861/9765 | SPS:  2008 | Reward:  0.057 | Value:  0.709 | VLoss: 0.0288 | PLoss: -0.0127 | Ent: 0.1487
+[ 39.5%] Iter 3862/9765 | SPS:  2008 | Reward:  0.067 | Value:  0.753 | VLoss: 0.0310 | PLoss: -0.0035 | Ent: 0.2078
+[ 39.6%] Iter 3863/9765 | SPS:  2008 | Reward:  0.053 | Value:  0.705 | VLoss: 0.0261 | PLoss: -0.0144 | Ent: 0.1894
+[ 39.6%] Iter 3864/9765 | SPS:  2009 | Reward:  0.053 | Value:  0.665 | VLoss: 0.0228 | PLoss: -0.0125 | Ent: 0.2078
+[ 39.6%] Iter 3865/9765 | SPS:  2009 | Reward:  0.072 | Value:  0.709 | VLoss: 0.0258 | PLoss: -0.0085 | Ent: 0.1714
+[ 39.6%] Iter 3866/9765 | SPS:  2009 | Reward:  0.065 | Value:  0.719 | VLoss: 0.0261 | PLoss: -0.0073 | Ent: 0.1680
+[ 39.6%] Iter 3867/9765 | SPS:  2009 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0291 | PLoss: -0.0072 | Ent: 0.1430
+[ 39.6%] Iter 3868/9765 | SPS:  2009 | Reward:  0.055 | Value:  0.652 | VLoss: 0.0213 | PLoss: -0.0029 | Ent: 0.1354
+[ 39.6%] Iter 3869/9765 | SPS:  2010 | Reward:  0.074 | Value:  0.686 | VLoss: 0.0300 | PLoss: -0.0058 | Ent: 0.1827
+[ 39.6%] Iter 3870/9765 | SPS:  2010 | Reward:  0.068 | Value:  0.752 | VLoss: 0.0242 | PLoss: -0.0064 | Ent: 0.2288
+[ 39.6%] Iter 3871/9765 | SPS:  2010 | Reward:  0.065 | Value:  0.675 | VLoss: 0.0358 | PLoss: -0.0088 | Ent: 0.1801
+[ 39.7%] Iter 3872/9765 | SPS:  2010 | Reward:  0.078 | Value:  0.703 | VLoss: 0.0223 | PLoss: -0.0123 | Ent: 0.1726
+[ 39.7%] Iter 3873/9765 | SPS:  2011 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0261 | PLoss: -0.0002 | Ent: 0.1711
+[ 39.7%] Iter 3874/9765 | SPS:  2011 | Reward:  0.066 | Value:  0.651 | VLoss: 0.0283 | PLoss: -0.0052 | Ent: 0.1835
+[ 39.7%] Iter 3875/9765 | SPS:  2011 | Reward:  0.071 | Value:  0.754 | VLoss: 0.0234 | PLoss: -0.0008 | Ent: 0.1852
+[ 39.7%] Iter 3876/9765 | SPS:  2011 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0322 | PLoss: -0.0052 | Ent: 0.1966
+[ 39.7%] Iter 3877/9765 | SPS:  2012 | Reward:  0.048 | Value:  0.690 | VLoss: 0.0278 | PLoss: -0.0091 | Ent: 0.2189
+[ 39.7%] Iter 3878/9765 | SPS:  2012 | Reward:  0.084 | Value:  0.653 | VLoss: 0.0229 | PLoss: -0.0070 | Ent: 0.1876
+[ 39.7%] Iter 3879/9765 | SPS:  2012 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0197 | PLoss: -0.0025 | Ent: 0.1984
+[ 39.7%] Iter 3880/9765 | SPS:  2012 | Reward:  0.073 | Value:  0.769 | VLoss: 0.0262 | PLoss: -0.0138 | Ent: 0.1767
+[ 39.7%] Iter 3881/9765 | SPS:  2012 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0145 | PLoss: -0.0106 | Ent: 0.1755
+[ 39.8%] Iter 3882/9765 | SPS:  2013 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0280 | PLoss: -0.0108 | Ent: 0.1701
+[ 39.8%] Iter 3883/9765 | SPS:  2013 | Reward:  0.060 | Value:  0.708 | VLoss: 0.0217 | PLoss: -0.0085 | Ent: 0.1683
+[ 39.8%] Iter 3884/9765 | SPS:  2013 | Reward:  0.054 | Value:  0.704 | VLoss: 0.0267 | PLoss: -0.0121 | Ent: 0.1739
+[ 39.8%] Iter 3885/9765 | SPS:  2013 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0317 | PLoss: -0.0073 | Ent: 0.2117
+[ 39.8%] Iter 3886/9765 | SPS:  2014 | Reward:  0.063 | Value:  0.650 | VLoss: 0.0339 | PLoss: -0.0092 | Ent: 0.1833
+[ 39.8%] Iter 3887/9765 | SPS:  2014 | Reward:  0.059 | Value:  0.658 | VLoss: 0.0253 | PLoss: -0.0043 | Ent: 0.1306
+[ 39.8%] Iter 3888/9765 | SPS:  2014 | Reward:  0.076 | Value:  0.680 | VLoss: 0.0266 | PLoss: -0.0056 | Ent: 0.1918
+[ 39.8%] Iter 3889/9765 | SPS:  2014 | Reward:  0.082 | Value:  0.731 | VLoss: 0.0230 | PLoss: -0.0105 | Ent: 0.1894
+[ 39.8%] Iter 3890/9765 | SPS:  2015 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0214 | PLoss: -0.0038 | Ent: 0.2192
+[ 39.8%] Iter 3891/9765 | SPS:  2015 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0245 | PLoss: -0.0057 | Ent: 0.1800
+[ 39.9%] Iter 3892/9765 | SPS:  2015 | Reward:  0.062 | Value:  0.740 | VLoss: 0.0273 | PLoss: -0.0079 | Ent: 0.2088
+[ 39.9%] Iter 3893/9765 | SPS:  2015 | Reward:  0.049 | Value:  0.686 | VLoss: 0.0291 | PLoss: -0.0164 | Ent: 0.2088
+[ 39.9%] Iter 3894/9765 | SPS:  2015 | Reward:  0.052 | Value:  0.660 | VLoss: 0.0274 | PLoss: -0.0028 | Ent: 0.1683
+[ 39.9%] Iter 3895/9765 | SPS:  2016 | Reward:  0.064 | Value:  0.651 | VLoss: 0.0289 | PLoss: -0.0033 | Ent: 0.1650
+[ 39.9%] Iter 3896/9765 | SPS:  2016 | Reward:  0.074 | Value:  0.683 | VLoss: 0.0256 | PLoss: -0.0114 | Ent: 0.1902
+[ 39.9%] Iter 3897/9765 | SPS:  2016 | Reward:  0.066 | Value:  0.690 | VLoss: 0.0291 | PLoss: -0.0117 | Ent: 0.1792
+[ 39.9%] Iter 3898/9765 | SPS:  2016 | Reward:  0.059 | Value:  0.624 | VLoss: 0.0296 | PLoss: -0.0041 | Ent: 0.1339
+[ 39.9%] Iter 3899/9765 | SPS:  2017 | Reward:  0.075 | Value:  0.650 | VLoss: 0.0305 | PLoss: -0.0133 | Ent: 0.1581
+[ 39.9%] Iter 3900/9765 | SPS:  2017 | Reward:  0.067 | Value:  0.699 | VLoss: 0.0303 | PLoss: -0.0050 | Ent: 0.1765
+[ 39.9%] Iter 3901/9765 | SPS:  2017 | Reward:  0.059 | Value:  0.709 | VLoss: 0.0166 | PLoss: -0.0073 | Ent: 0.1854
+[ 40.0%] Iter 3902/9765 | SPS:  2017 | Reward:  0.074 | Value:  0.751 | VLoss: 0.0327 | PLoss: -0.0128 | Ent: 0.1686
+[ 40.0%] Iter 3903/9765 | SPS:  2018 | Reward:  0.065 | Value:  0.672 | VLoss: 0.0345 | PLoss: -0.0091 | Ent: 0.1863
+[ 40.0%] Iter 3904/9765 | SPS:  2018 | Reward:  0.063 | Value:  0.638 | VLoss: 0.0297 | PLoss: -0.0012 | Ent: 0.1917
+[ 40.0%] Iter 3905/9765 | SPS:  2018 | Reward:  0.063 | Value:  0.672 | VLoss: 0.0248 | PLoss: -0.0085 | Ent: 0.1943
+[ 40.0%] Iter 3906/9765 | SPS:  2018 | Reward:  0.058 | Value:  0.678 | VLoss: 0.0304 | PLoss: -0.0065 | Ent: 0.1867
+[ 40.0%] Iter 3907/9765 | SPS:  2018 | Reward:  0.064 | Value:  0.676 | VLoss: 0.0249 | PLoss: -0.0067 | Ent: 0.1657
+[ 40.0%] Iter 3908/9765 | SPS:  2019 | Reward:  0.075 | Value:  0.659 | VLoss: 0.0346 | PLoss: -0.0120 | Ent: 0.1562
+[ 40.0%] Iter 3909/9765 | SPS:  2019 | Reward:  0.067 | Value:  0.686 | VLoss: 0.0276 | PLoss: -0.0102 | Ent: 0.2001
+[ 40.0%] Iter 3910/9765 | SPS:  2019 | Reward:  0.057 | Value:  0.605 | VLoss: 0.0292 | PLoss: -0.0140 | Ent: 0.2252
+[ 40.1%] Iter 3911/9765 | SPS:  2019 | Reward:  0.060 | Value:  0.655 | VLoss: 0.0220 | PLoss: -0.0059 | Ent: 0.1760
+[ 40.1%] Iter 3912/9765 | SPS:  2020 | Reward:  0.069 | Value:  0.667 | VLoss: 0.0215 | PLoss: -0.0089 | Ent: 0.1843
+[ 40.1%] Iter 3913/9765 | SPS:  2020 | Reward:  0.071 | Value:  0.710 | VLoss: 0.0351 | PLoss: -0.0066 | Ent: 0.1411
+[ 40.1%] Iter 3914/9765 | SPS:  2020 | Reward:  0.076 | Value:  0.772 | VLoss: 0.0309 | PLoss: -0.0092 | Ent: 0.1503
+[ 40.1%] Iter 3915/9765 | SPS:  2020 | Reward:  0.068 | Value:  0.744 | VLoss: 0.0258 | PLoss: -0.0033 | Ent: 0.2166
+[ 40.1%] Iter 3916/9765 | SPS:  2020 | Reward:  0.066 | Value:  0.708 | VLoss: 0.0295 | PLoss: -0.0120 | Ent: 0.1966
+[ 40.1%] Iter 3917/9765 | SPS:  2021 | Reward:  0.048 | Value:  0.695 | VLoss: 0.0204 | PLoss: -0.0033 | Ent: 0.1897
+[ 40.1%] Iter 3918/9765 | SPS:  2021 | Reward:  0.053 | Value:  0.634 | VLoss: 0.0326 | PLoss: -0.0030 | Ent: 0.1879
+[ 40.1%] Iter 3919/9765 | SPS:  2021 | Reward:  0.061 | Value:  0.653 | VLoss: 0.0312 | PLoss: -0.0075 | Ent: 0.2214
+[ 40.1%] Iter 3920/9765 | SPS:  2021 | Reward:  0.064 | Value:  0.639 | VLoss: 0.0315 | PLoss: -0.0060 | Ent: 0.1719
+[ 40.2%] Iter 3921/9765 | SPS:  2022 | Reward:  0.056 | Value:  0.661 | VLoss: 0.0330 | PLoss: -0.0081 | Ent: 0.1790
+[ 40.2%] Iter 3922/9765 | SPS:  2022 | Reward:  0.069 | Value:  0.697 | VLoss: 0.0271 | PLoss: -0.0041 | Ent: 0.1597
+[ 40.2%] Iter 3923/9765 | SPS:  2022 | Reward:  0.062 | Value:  0.726 | VLoss: 0.0195 | PLoss: -0.0121 | Ent: 0.1673
+[ 40.2%] Iter 3924/9765 | SPS:  2022 | Reward:  0.058 | Value:  0.729 | VLoss: 0.0257 | PLoss: -0.0030 | Ent: 0.1901
+[ 40.2%] Iter 3925/9765 | SPS:  2023 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0217 | PLoss: -0.0024 | Ent: 0.1681
+[ 40.2%] Iter 3926/9765 | SPS:  2023 | Reward:  0.060 | Value:  0.635 | VLoss: 0.0421 | PLoss: -0.0073 | Ent: 0.2211
+[ 40.2%] Iter 3927/9765 | SPS:  2023 | Reward:  0.065 | Value:  0.625 | VLoss: 0.0187 | PLoss: -0.0089 | Ent: 0.2063
+[ 40.2%] Iter 3928/9765 | SPS:  2023 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0174 | PLoss: -0.0048 | Ent: 0.1787
+[ 40.2%] Iter 3929/9765 | SPS:  2024 | Reward:  0.060 | Value:  0.631 | VLoss: 0.0387 | PLoss: -0.0152 | Ent: 0.2205
+[ 40.2%] Iter 3930/9765 | SPS:  2024 | Reward:  0.048 | Value:  0.604 | VLoss: 0.0234 | PLoss: -0.0183 | Ent: 0.2007
+[ 40.3%] Iter 3931/9765 | SPS:  2024 | Reward:  0.043 | Value:  0.554 | VLoss: 0.0185 | PLoss: -0.0085 | Ent: 0.2084
+[ 40.3%] Iter 3932/9765 | SPS:  2024 | Reward:  0.047 | Value:  0.600 | VLoss: 0.0302 | PLoss: -0.0066 | Ent: 0.2315
+[ 40.3%] Iter 3933/9765 | SPS:  2025 | Reward:  0.050 | Value:  0.580 | VLoss: 0.0131 | PLoss: -0.0030 | Ent: 0.2198
+[ 40.3%] Iter 3934/9765 | SPS:  2025 | Reward:  0.052 | Value:  0.661 | VLoss: 0.0196 | PLoss: -0.0112 | Ent: 0.2211
+[ 40.3%] Iter 3935/9765 | SPS:  2025 | Reward:  0.058 | Value:  0.689 | VLoss: 0.0227 | PLoss: -0.0083 | Ent: 0.1943
+[ 40.3%] Iter 3936/9765 | SPS:  2025 | Reward:  0.069 | Value:  0.731 | VLoss: 0.0213 | PLoss: -0.0032 | Ent: 0.1568
+[ 40.3%] Iter 3937/9765 | SPS:  2025 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0297 | PLoss: -0.0026 | Ent: 0.1603
+[ 40.3%] Iter 3938/9765 | SPS:  2026 | Reward:  0.081 | Value:  0.680 | VLoss: 0.0333 | PLoss: -0.0120 | Ent: 0.1973
+[ 40.3%] Iter 3939/9765 | SPS:  2026 | Reward:  0.067 | Value:  0.669 | VLoss: 0.0207 | PLoss: -0.0068 | Ent: 0.1852
+[ 40.3%] Iter 3940/9765 | SPS:  2026 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0262 | PLoss: -0.0082 | Ent: 0.1616
+[ 40.4%] Iter 3941/9765 | SPS:  2026 | Reward:  0.076 | Value:  0.708 | VLoss: 0.0259 | PLoss: -0.0094 | Ent: 0.1921
+[ 40.4%] Iter 3942/9765 | SPS:  2027 | Reward:  0.055 | Value:  0.745 | VLoss: 0.0340 | PLoss: -0.0091 | Ent: 0.1665
+[ 40.4%] Iter 3943/9765 | SPS:  2027 | Reward:  0.063 | Value:  0.629 | VLoss: 0.0328 | PLoss: -0.0181 | Ent: 0.1825
+[ 40.4%] Iter 3944/9765 | SPS:  2027 | Reward:  0.071 | Value:  0.660 | VLoss: 0.0338 | PLoss: -0.0142 | Ent: 0.2024
+[ 40.4%] Iter 3945/9765 | SPS:  2027 | Reward:  0.054 | Value:  0.741 | VLoss: 0.0302 | PLoss: -0.0063 | Ent: 0.1732
+[ 40.4%] Iter 3946/9765 | SPS:  2028 | Reward:  0.068 | Value:  0.667 | VLoss: 0.0311 | PLoss: -0.0049 | Ent: 0.1795
+[ 40.4%] Iter 3947/9765 | SPS:  2028 | Reward:  0.063 | Value:  0.729 | VLoss: 0.0185 | PLoss: -0.0033 | Ent: 0.1590
+[ 40.4%] Iter 3948/9765 | SPS:  2028 | Reward:  0.060 | Value:  0.711 | VLoss: 0.0288 | PLoss: -0.0119 | Ent: 0.2069
+[ 40.4%] Iter 3949/9765 | SPS:  2028 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0278 | PLoss: -0.0027 | Ent: 0.1917
+[ 40.5%] Iter 3950/9765 | SPS:  2028 | Reward:  0.048 | Value:  0.625 | VLoss: 0.0229 | PLoss: -0.0017 | Ent: 0.1646
+[ 40.5%] Iter 3951/9765 | SPS:  2029 | Reward:  0.058 | Value:  0.651 | VLoss: 0.0224 | PLoss: -0.0102 | Ent: 0.1810
+[ 40.5%] Iter 3952/9765 | SPS:  2029 | Reward:  0.064 | Value:  0.696 | VLoss: 0.0386 | PLoss: -0.0091 | Ent: 0.1676
+[ 40.5%] Iter 3953/9765 | SPS:  2029 | Reward:  0.062 | Value:  0.644 | VLoss: 0.0278 | PLoss: -0.0103 | Ent: 0.1800
+[ 40.5%] Iter 3954/9765 | SPS:  2029 | Reward:  0.070 | Value:  0.642 | VLoss: 0.0241 | PLoss: -0.0052 | Ent: 0.1259
+[ 40.5%] Iter 3955/9765 | SPS:  2030 | Reward:  0.058 | Value:  0.712 | VLoss: 0.0324 | PLoss: -0.0108 | Ent: 0.1743
+[ 40.5%] Iter 3956/9765 | SPS:  2030 | Reward:  0.068 | Value:  0.696 | VLoss: 0.0258 | PLoss: -0.0078 | Ent: 0.2153
+[ 40.5%] Iter 3957/9765 | SPS:  2030 | Reward:  0.061 | Value:  0.683 | VLoss: 0.0305 | PLoss: -0.0108 | Ent: 0.1869
+[ 40.5%] Iter 3958/9765 | SPS:  2030 | Reward:  0.068 | Value:  0.664 | VLoss: 0.0298 | PLoss: -0.0066 | Ent: 0.2078
+[ 40.5%] Iter 3959/9765 | SPS:  2031 | Reward:  0.066 | Value:  0.711 | VLoss: 0.0247 | PLoss: -0.0075 | Ent: 0.1510
+[ 40.6%] Iter 3960/9765 | SPS:  2031 | Reward:  0.069 | Value:  0.785 | VLoss: 0.0133 | PLoss: -0.0075 | Ent: 0.1479
+[ 40.6%] Iter 3961/9765 | SPS:  2031 | Reward:  0.066 | Value:  0.724 | VLoss: 0.0303 | PLoss: -0.0027 | Ent: 0.1996
+[ 40.6%] Iter 3962/9765 | SPS:  2031 | Reward:  0.067 | Value:  0.659 | VLoss: 0.0257 | PLoss: -0.0037 | Ent: 0.1561
+[ 40.6%] Iter 3963/9765 | SPS:  2031 | Reward:  0.051 | Value:  0.646 | VLoss: 0.0312 | PLoss: -0.0082 | Ent: 0.1444
+[ 40.6%] Iter 3964/9765 | SPS:  2032 | Reward:  0.057 | Value:  0.705 | VLoss: 0.0192 | PLoss: -0.0108 | Ent: 0.1748
+[ 40.6%] Iter 3965/9765 | SPS:  2032 | Reward:  0.054 | Value:  0.660 | VLoss: 0.0257 | PLoss: -0.0034 | Ent: 0.2403
+[ 40.6%] Iter 3966/9765 | SPS:  2032 | Reward:  0.077 | Value:  0.674 | VLoss: 0.0248 | PLoss: -0.0050 | Ent: 0.1587
+[ 40.6%] Iter 3967/9765 | SPS:  2032 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0276 | PLoss: -0.0062 | Ent: 0.1671
+[ 40.6%] Iter 3968/9765 | SPS:  2033 | Reward:  0.058 | Value:  0.687 | VLoss: 0.0269 | PLoss: -0.0055 | Ent: 0.2082
+[ 40.6%] Iter 3969/9765 | SPS:  2033 | Reward:  0.047 | Value:  0.656 | VLoss: 0.0243 | PLoss: -0.0118 | Ent: 0.2503
+[ 40.7%] Iter 3970/9765 | SPS:  2033 | Reward:  0.057 | Value:  0.701 | VLoss: 0.0270 | PLoss: -0.0124 | Ent: 0.2076
+[ 40.7%] Iter 3971/9765 | SPS:  2033 | Reward:  0.067 | Value:  0.679 | VLoss: 0.0231 | PLoss: -0.0018 | Ent: 0.1616
+[ 40.7%] Iter 3972/9765 | SPS:  2034 | Reward:  0.064 | Value:  0.750 | VLoss: 0.0249 | PLoss: -0.0042 | Ent: 0.1666
+[ 40.7%] Iter 3973/9765 | SPS:  2034 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0221 | PLoss: -0.0048 | Ent: 0.1639
+[ 40.7%] Iter 3974/9765 | SPS:  2034 | Reward:  0.059 | Value:  0.671 | VLoss: 0.0289 | PLoss: -0.0075 | Ent: 0.1772
+[ 40.7%] Iter 3975/9765 | SPS:  2034 | Reward:  0.056 | Value:  0.690 | VLoss: 0.0266 | PLoss: -0.0047 | Ent: 0.1962
+[ 40.7%] Iter 3976/9765 | SPS:  2034 | Reward:  0.052 | Value:  0.674 | VLoss: 0.0281 | PLoss: -0.0099 | Ent: 0.1462
+[ 40.7%] Iter 3977/9765 | SPS:  2035 | Reward:  0.060 | Value:  0.659 | VLoss: 0.0316 | PLoss: -0.0065 | Ent: 0.1585
+[ 40.7%] Iter 3978/9765 | SPS:  2035 | Reward:  0.079 | Value:  0.680 | VLoss: 0.0198 | PLoss: -0.0053 | Ent: 0.1259
+[ 40.7%] Iter 3979/9765 | SPS:  2035 | Reward:  0.058 | Value:  0.717 | VLoss: 0.0261 | PLoss: -0.0049 | Ent: 0.2054
+[ 40.8%] Iter 3980/9765 | SPS:  2035 | Reward:  0.059 | Value:  0.790 | VLoss: 0.0228 | PLoss: -0.0087 | Ent: 0.1924
+[ 40.8%] Iter 3981/9765 | SPS:  2036 | Reward:  0.062 | Value:  0.668 | VLoss: 0.0196 | PLoss: -0.0087 | Ent: 0.1898
+[ 40.8%] Iter 3982/9765 | SPS:  2036 | Reward:  0.058 | Value:  0.717 | VLoss: 0.0232 | PLoss: -0.0054 | Ent: 0.2108
+[ 40.8%] Iter 3983/9765 | SPS:  2036 | Reward:  0.067 | Value:  0.696 | VLoss: 0.0246 | PLoss: -0.0053 | Ent: 0.1673
+[ 40.8%] Iter 3984/9765 | SPS:  2036 | Reward:  0.063 | Value:  0.713 | VLoss: 0.0280 | PLoss: -0.0151 | Ent: 0.1752
+[ 40.8%] Iter 3985/9765 | SPS:  2036 | Reward:  0.066 | Value:  0.646 | VLoss: 0.0245 | PLoss: -0.0153 | Ent: 0.1912
+[ 40.8%] Iter 3986/9765 | SPS:  2037 | Reward:  0.079 | Value:  0.712 | VLoss: 0.0270 | PLoss: -0.0108 | Ent: 0.2093
+[ 40.8%] Iter 3987/9765 | SPS:  2037 | Reward:  0.045 | Value:  0.744 | VLoss: 0.0226 | PLoss: -0.0162 | Ent: 0.2079
+[ 40.8%] Iter 3988/9765 | SPS:  2037 | Reward:  0.063 | Value:  0.639 | VLoss: 0.0261 | PLoss: 0.0012 | Ent: 0.1918
+[ 40.8%] Iter 3989/9765 | SPS:  2037 | Reward:  0.048 | Value:  0.662 | VLoss: 0.0253 | PLoss: -0.0147 | Ent: 0.1873
+[ 40.9%] Iter 3990/9765 | SPS:  2038 | Reward:  0.073 | Value:  0.680 | VLoss: 0.0255 | PLoss: -0.0113 | Ent: 0.2046
+[ 40.9%] Iter 3991/9765 | SPS:  2038 | Reward:  0.071 | Value:  0.689 | VLoss: 0.0281 | PLoss: -0.0054 | Ent: 0.1349
+[ 40.9%] Iter 3992/9765 | SPS:  2038 | Reward:  0.070 | Value:  0.721 | VLoss: 0.0400 | PLoss: -0.0096 | Ent: 0.1582
+[ 40.9%] Iter 3993/9765 | SPS:  2038 | Reward:  0.053 | Value:  0.636 | VLoss: 0.0184 | PLoss: -0.0122 | Ent: 0.1857
+[ 40.9%] Iter 3994/9765 | SPS:  2038 | Reward:  0.071 | Value:  0.722 | VLoss: 0.0206 | PLoss: -0.0147 | Ent: 0.1685
+[ 40.9%] Iter 3995/9765 | SPS:  2039 | Reward:  0.057 | Value:  0.618 | VLoss: 0.0307 | PLoss: -0.0053 | Ent: 0.2065
+[ 40.9%] Iter 3996/9765 | SPS:  2039 | Reward:  0.058 | Value:  0.673 | VLoss: 0.0194 | PLoss: -0.0051 | Ent: 0.1514
+[ 40.9%] Iter 3997/9765 | SPS:  2039 | Reward:  0.073 | Value:  0.665 | VLoss: 0.0363 | PLoss: -0.0055 | Ent: 0.1504
+[ 40.9%] Iter 3998/9765 | SPS:  2039 | Reward:  0.061 | Value:  0.715 | VLoss: 0.0254 | PLoss: -0.0067 | Ent: 0.2052
+[ 41.0%] Iter 3999/9765 | SPS:  2040 | Reward:  0.047 | Value:  0.702 | VLoss: 0.0199 | PLoss: -0.0068 | Ent: 0.1796
+[ 41.0%] Iter 4000/9765 | SPS:  2040 | Reward:  0.046 | Value:  0.692 | VLoss: 0.0139 | PLoss: -0.0067 | Ent: 0.2244
+[ 41.0%] Iter 4001/9765 | SPS:  2040 | Reward:  0.056 | Value:  0.690 | VLoss: 0.0282 | PLoss: 0.0028 | Ent: 0.1785
+[ 41.0%] Iter 4002/9765 | SPS:  2040 | Reward:  0.064 | Value:  0.695 | VLoss: 0.0208 | PLoss: -0.0026 | Ent: 0.1963
+[ 41.0%] Iter 4003/9765 | SPS:  2041 | Reward:  0.065 | Value:  0.661 | VLoss: 0.0309 | PLoss: -0.0060 | Ent: 0.1804
+[ 41.0%] Iter 4004/9765 | SPS:  2041 | Reward:  0.076 | Value:  0.692 | VLoss: 0.0291 | PLoss: -0.0108 | Ent: 0.1646
+[ 41.0%] Iter 4005/9765 | SPS:  2041 | Reward:  0.067 | Value:  0.657 | VLoss: 0.0313 | PLoss: -0.0169 | Ent: 0.1854
+[ 41.0%] Iter 4006/9765 | SPS:  2041 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0294 | PLoss: -0.0104 | Ent: 0.1982
+[ 41.0%] Iter 4007/9765 | SPS:  2041 | Reward:  0.065 | Value:  0.712 | VLoss: 0.0238 | PLoss: -0.0046 | Ent: 0.1816
+[ 41.0%] Iter 4008/9765 | SPS:  2042 | Reward:  0.044 | Value:  0.668 | VLoss: 0.0195 | PLoss: -0.0067 | Ent: 0.2198
+[ 41.1%] Iter 4009/9765 | SPS:  2042 | Reward:  0.055 | Value:  0.651 | VLoss: 0.0277 | PLoss: -0.0066 | Ent: 0.1682
+[ 41.1%] Iter 4010/9765 | SPS:  2042 | Reward:  0.056 | Value:  0.618 | VLoss: 0.0325 | PLoss: -0.0035 | Ent: 0.1923
+[ 41.1%] Iter 4011/9765 | SPS:  2042 | Reward:  0.061 | Value:  0.670 | VLoss: 0.0258 | PLoss: -0.0100 | Ent: 0.2044
+[ 41.1%] Iter 4012/9765 | SPS:  2043 | Reward:  0.049 | Value:  0.643 | VLoss: 0.0286 | PLoss: -0.0092 | Ent: 0.1680
+[ 41.1%] Iter 4013/9765 | SPS:  2043 | Reward:  0.059 | Value:  0.612 | VLoss: 0.0254 | PLoss: -0.0097 | Ent: 0.2220
+[ 41.1%] Iter 4014/9765 | SPS:  2043 | Reward:  0.055 | Value:  0.634 | VLoss: 0.0287 | PLoss: -0.0083 | Ent: 0.1739
+[ 41.1%] Iter 4015/9765 | SPS:  2043 | Reward:  0.050 | Value:  0.639 | VLoss: 0.0277 | PLoss: -0.0090 | Ent: 0.2213
+[ 41.1%] Iter 4016/9765 | SPS:  2044 | Reward:  0.062 | Value:  0.752 | VLoss: 0.0233 | PLoss: -0.0121 | Ent: 0.1873
+[ 41.1%] Iter 4017/9765 | SPS:  2044 | Reward:  0.080 | Value:  0.732 | VLoss: 0.0335 | PLoss: -0.0055 | Ent: 0.2032
+[ 41.1%] Iter 4018/9765 | SPS:  2044 | Reward:  0.053 | Value:  0.766 | VLoss: 0.0195 | PLoss: -0.0064 | Ent: 0.2030
+[ 41.2%] Iter 4019/9765 | SPS:  2044 | Reward:  0.070 | Value:  0.729 | VLoss: 0.0223 | PLoss: -0.0092 | Ent: 0.1609
+[ 41.2%] Iter 4020/9765 | SPS:  2044 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0276 | PLoss: -0.0055 | Ent: 0.2050
+[ 41.2%] Iter 4021/9765 | SPS:  2045 | Reward:  0.072 | Value:  0.690 | VLoss: 0.0262 | PLoss: -0.0027 | Ent: 0.2138
+[ 41.2%] Iter 4022/9765 | SPS:  2045 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0277 | PLoss: -0.0126 | Ent: 0.1735
+[ 41.2%] Iter 4023/9765 | SPS:  2045 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0227 | PLoss: -0.0123 | Ent: 0.2178
+[ 41.2%] Iter 4024/9765 | SPS:  2045 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0201 | PLoss: 0.0005 | Ent: 0.1689
+[ 41.2%] Iter 4025/9765 | SPS:  2046 | Reward:  0.054 | Value:  0.608 | VLoss: 0.0274 | PLoss: -0.0029 | Ent: 0.1891
+[ 41.2%] Iter 4026/9765 | SPS:  2046 | Reward:  0.055 | Value:  0.641 | VLoss: 0.0357 | PLoss: -0.0094 | Ent: 0.2202
+[ 41.2%] Iter 4027/9765 | SPS:  2046 | Reward:  0.060 | Value:  0.652 | VLoss: 0.0285 | PLoss: -0.0058 | Ent: 0.1684
+[ 41.2%] Iter 4028/9765 | SPS:  2046 | Reward:  0.055 | Value:  0.718 | VLoss: 0.0160 | PLoss: -0.0040 | Ent: 0.1916
+[ 41.3%] Iter 4029/9765 | SPS:  2046 | Reward:  0.047 | Value:  0.613 | VLoss: 0.0264 | PLoss: -0.0064 | Ent: 0.2142
+[ 41.3%] Iter 4030/9765 | SPS:  2047 | Reward:  0.061 | Value:  0.676 | VLoss: 0.0266 | PLoss: -0.0052 | Ent: 0.1958
+[ 41.3%] Iter 4031/9765 | SPS:  2047 | Reward:  0.064 | Value:  0.753 | VLoss: 0.0240 | PLoss: -0.0091 | Ent: 0.1643
+[ 41.3%] Iter 4032/9765 | SPS:  2047 | Reward:  0.061 | Value:  0.729 | VLoss: 0.0186 | PLoss: -0.0083 | Ent: 0.1718
+[ 41.3%] Iter 4033/9765 | SPS:  2047 | Reward:  0.073 | Value:  0.717 | VLoss: 0.0294 | PLoss: -0.0154 | Ent: 0.1604
+[ 41.3%] Iter 4034/9765 | SPS:  2048 | Reward:  0.063 | Value:  0.660 | VLoss: 0.0288 | PLoss: -0.0083 | Ent: 0.1960
+[ 41.3%] Iter 4035/9765 | SPS:  2048 | Reward:  0.059 | Value:  0.620 | VLoss: 0.0220 | PLoss: -0.0102 | Ent: 0.1921
+[ 41.3%] Iter 4036/9765 | SPS:  2048 | Reward:  0.045 | Value:  0.594 | VLoss: 0.0138 | PLoss: -0.0080 | Ent: 0.2124
+[ 41.3%] Iter 4037/9765 | SPS:  2048 | Reward:  0.049 | Value:  0.685 | VLoss: 0.0140 | PLoss: -0.0079 | Ent: 0.2013
+[ 41.4%] Iter 4038/9765 | SPS:  2049 | Reward:  0.061 | Value:  0.612 | VLoss: 0.0294 | PLoss: -0.0067 | Ent: 0.1818
+[ 41.4%] Iter 4039/9765 | SPS:  2049 | Reward:  0.056 | Value:  0.635 | VLoss: 0.0193 | PLoss: -0.0041 | Ent: 0.1552
+[ 41.4%] Iter 4040/9765 | SPS:  2049 | Reward:  0.048 | Value:  0.689 | VLoss: 0.0248 | PLoss: -0.0072 | Ent: 0.1986
+[ 41.4%] Iter 4041/9765 | SPS:  2049 | Reward:  0.063 | Value:  0.616 | VLoss: 0.0289 | PLoss: -0.0062 | Ent: 0.1652
+[ 41.4%] Iter 4042/9765 | SPS:  2049 | Reward:  0.061 | Value:  0.690 | VLoss: 0.0271 | PLoss: -0.0074 | Ent: 0.1786
+[ 41.4%] Iter 4043/9765 | SPS:  2050 | Reward:  0.073 | Value:  0.738 | VLoss: 0.0297 | PLoss: -0.0152 | Ent: 0.1678
+[ 41.4%] Iter 4044/9765 | SPS:  2050 | Reward:  0.068 | Value:  0.719 | VLoss: 0.0299 | PLoss: -0.0130 | Ent: 0.1881
+[ 41.4%] Iter 4045/9765 | SPS:  2050 | Reward:  0.074 | Value:  0.721 | VLoss: 0.0253 | PLoss: -0.0123 | Ent: 0.1365
+[ 41.4%] Iter 4046/9765 | SPS:  2050 | Reward:  0.071 | Value:  0.717 | VLoss: 0.0192 | PLoss: -0.0095 | Ent: 0.1695
+[ 41.4%] Iter 4047/9765 | SPS:  2051 | Reward:  0.055 | Value:  0.679 | VLoss: 0.0268 | PLoss: -0.0037 | Ent: 0.2000
+[ 41.5%] Iter 4048/9765 | SPS:  2051 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0210 | PLoss: -0.0043 | Ent: 0.1547
+[ 41.5%] Iter 4049/9765 | SPS:  2051 | Reward:  0.067 | Value:  0.730 | VLoss: 0.0290 | PLoss: -0.0054 | Ent: 0.1675
+[ 41.5%] Iter 4050/9765 | SPS:  2051 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0311 | PLoss: -0.0112 | Ent: 0.1466
+[ 41.5%] Iter 4051/9765 | SPS:  2051 | Reward:  0.063 | Value:  0.666 | VLoss: 0.0266 | PLoss: -0.0064 | Ent: 0.1588
+[ 41.5%] Iter 4052/9765 | SPS:  2052 | Reward:  0.053 | Value:  0.733 | VLoss: 0.0180 | PLoss: -0.0091 | Ent: 0.1276
+[ 41.5%] Iter 4053/9765 | SPS:  2052 | Reward:  0.057 | Value:  0.667 | VLoss: 0.0290 | PLoss: -0.0137 | Ent: 0.1810
+[ 41.5%] Iter 4054/9765 | SPS:  2052 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0235 | PLoss: -0.0134 | Ent: 0.1923
+[ 41.5%] Iter 4055/9765 | SPS:  2052 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0248 | PLoss: -0.0078 | Ent: 0.2078
+[ 41.5%] Iter 4056/9765 | SPS:  2053 | Reward:  0.067 | Value:  0.705 | VLoss: 0.0245 | PLoss: -0.0083 | Ent: 0.1715
+[ 41.5%] Iter 4057/9765 | SPS:  2053 | Reward:  0.066 | Value:  0.768 | VLoss: 0.0263 | PLoss: -0.0070 | Ent: 0.1646
+[ 41.6%] Iter 4058/9765 | SPS:  2053 | Reward:  0.058 | Value:  0.656 | VLoss: 0.0239 | PLoss: -0.0111 | Ent: 0.1659
+[ 41.6%] Iter 4059/9765 | SPS:  2053 | Reward:  0.047 | Value:  0.679 | VLoss: 0.0225 | PLoss: -0.0047 | Ent: 0.1544
+[ 41.6%] Iter 4060/9765 | SPS:  2054 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0200 | PLoss: -0.0079 | Ent: 0.2133
+[ 41.6%] Iter 4061/9765 | SPS:  2054 | Reward:  0.054 | Value:  0.675 | VLoss: 0.0214 | PLoss: -0.0127 | Ent: 0.2440
+[ 41.6%] Iter 4062/9765 | SPS:  2054 | Reward:  0.066 | Value:  0.748 | VLoss: 0.0194 | PLoss: -0.0078 | Ent: 0.1447
+[ 41.6%] Iter 4063/9765 | SPS:  2054 | Reward:  0.062 | Value:  0.774 | VLoss: 0.0223 | PLoss: -0.0131 | Ent: 0.1902
+[ 41.6%] Iter 4064/9765 | SPS:  2054 | Reward:  0.066 | Value:  0.671 | VLoss: 0.0244 | PLoss: -0.0098 | Ent: 0.1897
+[ 41.6%] Iter 4065/9765 | SPS:  2055 | Reward:  0.077 | Value:  0.684 | VLoss: 0.0301 | PLoss: -0.0071 | Ent: 0.1934
+[ 41.6%] Iter 4066/9765 | SPS:  2055 | Reward:  0.057 | Value:  0.680 | VLoss: 0.0299 | PLoss: -0.0068 | Ent: 0.1946
+[ 41.6%] Iter 4067/9765 | SPS:  2055 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0361 | PLoss: -0.0074 | Ent: 0.1819
+[ 41.7%] Iter 4068/9765 | SPS:  2055 | Reward:  0.056 | Value:  0.718 | VLoss: 0.0187 | PLoss: -0.0107 | Ent: 0.1668
+[ 41.7%] Iter 4069/9765 | SPS:  2056 | Reward:  0.062 | Value:  0.736 | VLoss: 0.0258 | PLoss: -0.0077 | Ent: 0.1676
+[ 41.7%] Iter 4070/9765 | SPS:  2056 | Reward:  0.067 | Value:  0.710 | VLoss: 0.0277 | PLoss: -0.0051 | Ent: 0.2052
+[ 41.7%] Iter 4071/9765 | SPS:  2056 | Reward:  0.061 | Value:  0.740 | VLoss: 0.0219 | PLoss: -0.0057 | Ent: 0.1708
+[ 41.7%] Iter 4072/9765 | SPS:  2056 | Reward:  0.073 | Value:  0.676 | VLoss: 0.0264 | PLoss: -0.0070 | Ent: 0.1810
+[ 41.7%] Iter 4073/9765 | SPS:  2056 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0174 | PLoss: -0.0060 | Ent: 0.1259
+[ 41.7%] Iter 4074/9765 | SPS:  2057 | Reward:  0.061 | Value:  0.722 | VLoss: 0.0221 | PLoss: -0.0080 | Ent: 0.1881
+[ 41.7%] Iter 4075/9765 | SPS:  2057 | Reward:  0.057 | Value:  0.714 | VLoss: 0.0143 | PLoss: -0.0120 | Ent: 0.1908
+[ 41.7%] Iter 4076/9765 | SPS:  2057 | Reward:  0.066 | Value:  0.680 | VLoss: 0.0222 | PLoss: -0.0058 | Ent: 0.1799
+[ 41.8%] Iter 4077/9765 | SPS:  2057 | Reward:  0.060 | Value:  0.755 | VLoss: 0.0256 | PLoss: -0.0059 | Ent: 0.1488
+[ 41.8%] Iter 4078/9765 | SPS:  2058 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0225 | PLoss: -0.0078 | Ent: 0.1538
+[ 41.8%] Iter 4079/9765 | SPS:  2058 | Reward:  0.057 | Value:  0.675 | VLoss: 0.0250 | PLoss: -0.0075 | Ent: 0.1886
+[ 41.8%] Iter 4080/9765 | SPS:  2058 | Reward:  0.063 | Value:  0.664 | VLoss: 0.0243 | PLoss: -0.0099 | Ent: 0.1872
+[ 41.8%] Iter 4081/9765 | SPS:  2058 | Reward:  0.059 | Value:  0.759 | VLoss: 0.0248 | PLoss: -0.0111 | Ent: 0.1650
+[ 41.8%] Iter 4082/9765 | SPS:  2058 | Reward:  0.063 | Value:  0.693 | VLoss: 0.0257 | PLoss: -0.0107 | Ent: 0.1619
+[ 41.8%] Iter 4083/9765 | SPS:  2059 | Reward:  0.070 | Value:  0.738 | VLoss: 0.0250 | PLoss: -0.0003 | Ent: 0.1418
+[ 41.8%] Iter 4084/9765 | SPS:  2059 | Reward:  0.057 | Value:  0.753 | VLoss: 0.0169 | PLoss: 0.0028 | Ent: 0.1914
+[ 41.8%] Iter 4085/9765 | SPS:  2059 | Reward:  0.065 | Value:  0.773 | VLoss: 0.0224 | PLoss: -0.0129 | Ent: 0.2060
+[ 41.8%] Iter 4086/9765 | SPS:  2059 | Reward:  0.051 | Value:  0.707 | VLoss: 0.0206 | PLoss: -0.0077 | Ent: 0.1688
+[ 41.9%] Iter 4087/9765 | SPS:  2060 | Reward:  0.071 | Value:  0.692 | VLoss: 0.0268 | PLoss: -0.0104 | Ent: 0.1685
+[ 41.9%] Iter 4088/9765 | SPS:  2060 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0293 | PLoss: -0.0073 | Ent: 0.2097
+[ 41.9%] Iter 4089/9765 | SPS:  2060 | Reward:  0.064 | Value:  0.630 | VLoss: 0.0251 | PLoss: -0.0037 | Ent: 0.1837
+[ 41.9%] Iter 4090/9765 | SPS:  2060 | Reward:  0.070 | Value:  0.750 | VLoss: 0.0243 | PLoss: -0.0054 | Ent: 0.1659
+[ 41.9%] Iter 4091/9765 | SPS:  2060 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0199 | PLoss: -0.0045 | Ent: 0.1717
+[ 41.9%] Iter 4092/9765 | SPS:  2061 | Reward:  0.063 | Value:  0.668 | VLoss: 0.0222 | PLoss: -0.0042 | Ent: 0.2017
+[ 41.9%] Iter 4093/9765 | SPS:  2061 | Reward:  0.064 | Value:  0.713 | VLoss: 0.0331 | PLoss: -0.0099 | Ent: 0.2235
+[ 41.9%] Iter 4094/9765 | SPS:  2061 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0216 | PLoss: -0.0072 | Ent: 0.1981
+[ 41.9%] Iter 4095/9765 | SPS:  2061 | Reward:  0.049 | Value:  0.750 | VLoss: 0.0254 | PLoss: -0.0151 | Ent: 0.2393
+[ 41.9%] Iter 4096/9765 | SPS:  2062 | Reward:  0.070 | Value:  0.704 | VLoss: 0.0212 | PLoss: -0.0104 | Ent: 0.1595
+[ 42.0%] Iter 4097/9765 | SPS:  2062 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0226 | PLoss: -0.0084 | Ent: 0.1884
+[ 42.0%] Iter 4098/9765 | SPS:  2062 | Reward:  0.053 | Value:  0.729 | VLoss: 0.0218 | PLoss: -0.0075 | Ent: 0.1825
+[ 42.0%] Iter 4099/9765 | SPS:  2062 | Reward:  0.071 | Value:  0.682 | VLoss: 0.0367 | PLoss: -0.0079 | Ent: 0.1554
+[ 42.0%] Iter 4100/9765 | SPS:  2062 | Reward:  0.070 | Value:  0.722 | VLoss: 0.0169 | PLoss: -0.0014 | Ent: 0.1503
+[ 42.0%] Iter 4101/9765 | SPS:  2063 | Reward:  0.056 | Value:  0.751 | VLoss: 0.0175 | PLoss: -0.0084 | Ent: 0.2077
+[ 42.0%] Iter 4102/9765 | SPS:  2063 | Reward:  0.064 | Value:  0.652 | VLoss: 0.0304 | PLoss: -0.0054 | Ent: 0.1570
+[ 42.0%] Iter 4103/9765 | SPS:  2063 | Reward:  0.063 | Value:  0.672 | VLoss: 0.0304 | PLoss: -0.0053 | Ent: 0.1778
+[ 42.0%] Iter 4104/9765 | SPS:  2063 | Reward:  0.065 | Value:  0.762 | VLoss: 0.0202 | PLoss: -0.0092 | Ent: 0.1917
+[ 42.0%] Iter 4105/9765 | SPS:  2064 | Reward:  0.061 | Value:  0.674 | VLoss: 0.0238 | PLoss: -0.0127 | Ent: 0.1750
+[ 42.0%] Iter 4106/9765 | SPS:  2064 | Reward:  0.056 | Value:  0.706 | VLoss: 0.0241 | PLoss: -0.0181 | Ent: 0.2386
+[ 42.1%] Iter 4107/9765 | SPS:  2064 | Reward:  0.053 | Value:  0.774 | VLoss: 0.0247 | PLoss: -0.0088 | Ent: 0.2049
+[ 42.1%] Iter 4108/9765 | SPS:  2064 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0244 | PLoss: -0.0069 | Ent: 0.1430
+[ 42.1%] Iter 4109/9765 | SPS:  2064 | Reward:  0.050 | Value:  0.673 | VLoss: 0.0259 | PLoss: -0.0034 | Ent: 0.2077
+[ 42.1%] Iter 4110/9765 | SPS:  2065 | Reward:  0.048 | Value:  0.709 | VLoss: 0.0207 | PLoss: -0.0060 | Ent: 0.2236
+[ 42.1%] Iter 4111/9765 | SPS:  2065 | Reward:  0.062 | Value:  0.700 | VLoss: 0.0224 | PLoss: -0.0030 | Ent: 0.1718
+[ 42.1%] Iter 4112/9765 | SPS:  2065 | Reward:  0.060 | Value:  0.727 | VLoss: 0.0202 | PLoss: -0.0107 | Ent: 0.2155
+[ 42.1%] Iter 4113/9765 | SPS:  2065 | Reward:  0.054 | Value:  0.756 | VLoss: 0.0205 | PLoss: -0.0091 | Ent: 0.2589
+[ 42.1%] Iter 4114/9765 | SPS:  2066 | Reward:  0.069 | Value:  0.714 | VLoss: 0.0292 | PLoss: -0.0115 | Ent: 0.2040
+[ 42.1%] Iter 4115/9765 | SPS:  2066 | Reward:  0.083 | Value:  0.707 | VLoss: 0.0242 | PLoss: -0.0105 | Ent: 0.1735
+[ 42.2%] Iter 4116/9765 | SPS:  2066 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0191 | PLoss: -0.0040 | Ent: 0.1686
+[ 42.2%] Iter 4117/9765 | SPS:  2066 | Reward:  0.072 | Value:  0.719 | VLoss: 0.0289 | PLoss: -0.0103 | Ent: 0.1870
+[ 42.2%] Iter 4118/9765 | SPS:  2066 | Reward:  0.049 | Value:  0.750 | VLoss: 0.0175 | PLoss: -0.0001 | Ent: 0.1563
+[ 42.2%] Iter 4119/9765 | SPS:  2067 | Reward:  0.065 | Value:  0.664 | VLoss: 0.0264 | PLoss: -0.0160 | Ent: 0.2214
+[ 42.2%] Iter 4120/9765 | SPS:  2067 | Reward:  0.069 | Value:  0.727 | VLoss: 0.0221 | PLoss: -0.0083 | Ent: 0.1441
+[ 42.2%] Iter 4121/9765 | SPS:  2067 | Reward:  0.065 | Value:  0.753 | VLoss: 0.0218 | PLoss: -0.0081 | Ent: 0.1570
+[ 42.2%] Iter 4122/9765 | SPS:  2067 | Reward:  0.062 | Value:  0.773 | VLoss: 0.0273 | PLoss: -0.0076 | Ent: 0.1351
+[ 42.2%] Iter 4123/9765 | SPS:  2068 | Reward:  0.054 | Value:  0.674 | VLoss: 0.0202 | PLoss: -0.0014 | Ent: 0.1698
+[ 42.2%] Iter 4124/9765 | SPS:  2068 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0205 | PLoss: -0.0119 | Ent: 0.1501
+[ 42.2%] Iter 4125/9765 | SPS:  2068 | Reward:  0.063 | Value:  0.687 | VLoss: 0.0305 | PLoss: -0.0068 | Ent: 0.1504
+[ 42.3%] Iter 4126/9765 | SPS:  2068 | Reward:  0.068 | Value:  0.679 | VLoss: 0.0267 | PLoss: -0.0091 | Ent: 0.1815
+[ 42.3%] Iter 4127/9765 | SPS:  2068 | Reward:  0.057 | Value:  0.704 | VLoss: 0.0234 | PLoss: -0.0041 | Ent: 0.1671
+[ 42.3%] Iter 4128/9765 | SPS:  2069 | Reward:  0.052 | Value:  0.706 | VLoss: 0.0169 | PLoss: -0.0037 | Ent: 0.1374
+[ 42.3%] Iter 4129/9765 | SPS:  2069 | Reward:  0.066 | Value:  0.661 | VLoss: 0.0213 | PLoss: -0.0093 | Ent: 0.1627
+[ 42.3%] Iter 4130/9765 | SPS:  2069 | Reward:  0.055 | Value:  0.739 | VLoss: 0.0267 | PLoss: -0.0094 | Ent: 0.1657
+[ 42.3%] Iter 4131/9765 | SPS:  2069 | Reward:  0.064 | Value:  0.748 | VLoss: 0.0239 | PLoss: -0.0071 | Ent: 0.1741
+[ 42.3%] Iter 4132/9765 | SPS:  2070 | Reward:  0.067 | Value:  0.745 | VLoss: 0.0192 | PLoss: -0.0073 | Ent: 0.1611
+[ 42.3%] Iter 4133/9765 | SPS:  2070 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0180 | PLoss: -0.0096 | Ent: 0.2375
+[ 42.3%] Iter 4134/9765 | SPS:  2070 | Reward:  0.066 | Value:  0.699 | VLoss: 0.0344 | PLoss: -0.0092 | Ent: 0.2085
+[ 42.3%] Iter 4135/9765 | SPS:  2070 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0228 | PLoss: -0.0035 | Ent: 0.1867
+[ 42.4%] Iter 4136/9765 | SPS:  2070 | Reward:  0.065 | Value:  0.691 | VLoss: 0.0313 | PLoss: -0.0136 | Ent: 0.1900
+[ 42.4%] Iter 4137/9765 | SPS:  2071 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0192 | PLoss: -0.0033 | Ent: 0.1757
+[ 42.4%] Iter 4138/9765 | SPS:  2071 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0319 | PLoss: -0.0160 | Ent: 0.1920
+[ 42.4%] Iter 4139/9765 | SPS:  2071 | Reward:  0.054 | Value:  0.681 | VLoss: 0.0306 | PLoss: -0.0071 | Ent: 0.1987
+[ 42.4%] Iter 4140/9765 | SPS:  2071 | Reward:  0.054 | Value:  0.697 | VLoss: 0.0214 | PLoss: -0.0146 | Ent: 0.1831
+[ 42.4%] Iter 4141/9765 | SPS:  2072 | Reward:  0.053 | Value:  0.648 | VLoss: 0.0196 | PLoss: -0.0115 | Ent: 0.1702
+[ 42.4%] Iter 4142/9765 | SPS:  2072 | Reward:  0.054 | Value:  0.668 | VLoss: 0.0356 | PLoss: -0.0055 | Ent: 0.2189
+[ 42.4%] Iter 4143/9765 | SPS:  2072 | Reward:  0.058 | Value:  0.697 | VLoss: 0.0166 | PLoss: -0.0142 | Ent: 0.2085
+[ 42.4%] Iter 4144/9765 | SPS:  2072 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0241 | PLoss: -0.0004 | Ent: 0.1468
+[ 42.4%] Iter 4145/9765 | SPS:  2072 | Reward:  0.064 | Value:  0.646 | VLoss: 0.0335 | PLoss: 0.0016 | Ent: 0.1584
+[ 42.5%] Iter 4146/9765 | SPS:  2073 | Reward:  0.064 | Value:  0.671 | VLoss: 0.0236 | PLoss: -0.0010 | Ent: 0.1707
+[ 42.5%] Iter 4147/9765 | SPS:  2073 | Reward:  0.070 | Value:  0.650 | VLoss: 0.0276 | PLoss: -0.0012 | Ent: 0.2138
+[ 42.5%] Iter 4148/9765 | SPS:  2073 | Reward:  0.069 | Value:  0.773 | VLoss: 0.0239 | PLoss: 0.0025 | Ent: 0.1843
+[ 42.5%] Iter 4149/9765 | SPS:  2073 | Reward:  0.050 | Value:  0.783 | VLoss: 0.0145 | PLoss: -0.0018 | Ent: 0.1377
+[ 42.5%] Iter 4150/9765 | SPS:  2073 | Reward:  0.061 | Value:  0.773 | VLoss: 0.0206 | PLoss: -0.0051 | Ent: 0.1453
+[ 42.5%] Iter 4151/9765 | SPS:  2074 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0251 | PLoss: -0.0056 | Ent: 0.2140
+[ 42.5%] Iter 4152/9765 | SPS:  2074 | Reward:  0.061 | Value:  0.750 | VLoss: 0.0232 | PLoss: 0.0016 | Ent: 0.2029
+[ 42.5%] Iter 4153/9765 | SPS:  2074 | Reward:  0.058 | Value:  0.667 | VLoss: 0.0179 | PLoss: -0.0052 | Ent: 0.2279
+[ 42.5%] Iter 4154/9765 | SPS:  2074 | Reward:  0.054 | Value:  0.663 | VLoss: 0.0204 | PLoss: -0.0105 | Ent: 0.1830
+[ 42.5%] Iter 4155/9765 | SPS:  2075 | Reward:  0.070 | Value:  0.745 | VLoss: 0.0311 | PLoss: -0.0114 | Ent: 0.1705
+[ 42.6%] Iter 4156/9765 | SPS:  2075 | Reward:  0.066 | Value:  0.659 | VLoss: 0.0370 | PLoss: -0.0070 | Ent: 0.1897
+[ 42.6%] Iter 4157/9765 | SPS:  2075 | Reward:  0.059 | Value:  0.693 | VLoss: 0.0276 | PLoss: -0.0014 | Ent: 0.1392
+[ 42.6%] Iter 4158/9765 | SPS:  2075 | Reward:  0.053 | Value:  0.660 | VLoss: 0.0325 | PLoss: -0.0061 | Ent: 0.1830
+[ 42.6%] Iter 4159/9765 | SPS:  2075 | Reward:  0.079 | Value:  0.693 | VLoss: 0.0261 | PLoss: -0.0070 | Ent: 0.2070
+[ 42.6%] Iter 4160/9765 | SPS:  2076 | Reward:  0.070 | Value:  0.737 | VLoss: 0.0331 | PLoss: -0.0123 | Ent: 0.1847
+[ 42.6%] Iter 4161/9765 | SPS:  2076 | Reward:  0.067 | Value:  0.665 | VLoss: 0.0337 | PLoss: -0.0057 | Ent: 0.1984
+[ 42.6%] Iter 4162/9765 | SPS:  2076 | Reward:  0.075 | Value:  0.639 | VLoss: 0.0396 | PLoss: -0.0030 | Ent: 0.1715
+[ 42.6%] Iter 4163/9765 | SPS:  2076 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0242 | PLoss: -0.0068 | Ent: 0.1743
+[ 42.6%] Iter 4164/9765 | SPS:  2076 | Reward:  0.062 | Value:  0.759 | VLoss: 0.0236 | PLoss: -0.0020 | Ent: 0.1579
+[ 42.7%] Iter 4165/9765 | SPS:  2077 | Reward:  0.055 | Value:  0.665 | VLoss: 0.0232 | PLoss: -0.0007 | Ent: 0.1973
+[ 42.7%] Iter 4166/9765 | SPS:  2077 | Reward:  0.060 | Value:  0.645 | VLoss: 0.0269 | PLoss: -0.0048 | Ent: 0.2035
+[ 42.7%] Iter 4167/9765 | SPS:  2077 | Reward:  0.064 | Value:  0.715 | VLoss: 0.0228 | PLoss: -0.0120 | Ent: 0.2184
+[ 42.7%] Iter 4168/9765 | SPS:  2077 | Reward:  0.058 | Value:  0.681 | VLoss: 0.0213 | PLoss: -0.0087 | Ent: 0.2031
+[ 42.7%] Iter 4169/9765 | SPS:  2078 | Reward:  0.067 | Value:  0.667 | VLoss: 0.0263 | PLoss: -0.0040 | Ent: 0.1746
+[ 42.7%] Iter 4170/9765 | SPS:  2078 | Reward:  0.073 | Value:  0.738 | VLoss: 0.0236 | PLoss: -0.0103 | Ent: 0.1891
+[ 42.7%] Iter 4171/9765 | SPS:  2078 | Reward:  0.058 | Value:  0.668 | VLoss: 0.0251 | PLoss: 0.0019 | Ent: 0.2292
+[ 42.7%] Iter 4172/9765 | SPS:  2078 | Reward:  0.065 | Value:  0.674 | VLoss: 0.0234 | PLoss: -0.0063 | Ent: 0.1411
+[ 42.7%] Iter 4173/9765 | SPS:  2078 | Reward:  0.066 | Value:  0.713 | VLoss: 0.0269 | PLoss: -0.0040 | Ent: 0.1314
+[ 42.7%] Iter 4174/9765 | SPS:  2079 | Reward:  0.054 | Value:  0.728 | VLoss: 0.0215 | PLoss: -0.0079 | Ent: 0.1946
+[ 42.8%] Iter 4175/9765 | SPS:  2079 | Reward:  0.074 | Value:  0.691 | VLoss: 0.0230 | PLoss: -0.0024 | Ent: 0.1765
+[ 42.8%] Iter 4176/9765 | SPS:  2079 | Reward:  0.070 | Value:  0.768 | VLoss: 0.0281 | PLoss: -0.0100 | Ent: 0.2169
+[ 42.8%] Iter 4177/9765 | SPS:  2079 | Reward:  0.067 | Value:  0.776 | VLoss: 0.0244 | PLoss: -0.0128 | Ent: 0.1384
+[ 42.8%] Iter 4178/9765 | SPS:  2079 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0210 | PLoss: -0.0082 | Ent: 0.1683
+[ 42.8%] Iter 4179/9765 | SPS:  2080 | Reward:  0.066 | Value:  0.745 | VLoss: 0.0250 | PLoss: -0.0084 | Ent: 0.1941
+[ 42.8%] Iter 4180/9765 | SPS:  2080 | Reward:  0.059 | Value:  0.701 | VLoss: 0.0201 | PLoss: 0.0003 | Ent: 0.1620
+[ 42.8%] Iter 4181/9765 | SPS:  2080 | Reward:  0.052 | Value:  0.740 | VLoss: 0.0192 | PLoss: -0.0076 | Ent: 0.1694
+[ 42.8%] Iter 4182/9765 | SPS:  2080 | Reward:  0.064 | Value:  0.696 | VLoss: 0.0242 | PLoss: -0.0072 | Ent: 0.2024
+[ 42.8%] Iter 4183/9765 | SPS:  2081 | Reward:  0.066 | Value:  0.734 | VLoss: 0.0265 | PLoss: -0.0106 | Ent: 0.1957
+[ 42.8%] Iter 4184/9765 | SPS:  2081 | Reward:  0.065 | Value:  0.710 | VLoss: 0.0184 | PLoss: -0.0066 | Ent: 0.1501
+[ 42.9%] Iter 4185/9765 | SPS:  2081 | Reward:  0.067 | Value:  0.703 | VLoss: 0.0295 | PLoss: -0.0055 | Ent: 0.2008
+[ 42.9%] Iter 4186/9765 | SPS:  2081 | Reward:  0.063 | Value:  0.756 | VLoss: 0.0242 | PLoss: -0.0091 | Ent: 0.1630
+[ 42.9%] Iter 4187/9765 | SPS:  2081 | Reward:  0.076 | Value:  0.757 | VLoss: 0.0218 | PLoss: -0.0126 | Ent: 0.1893
+[ 42.9%] Iter 4188/9765 | SPS:  2082 | Reward:  0.070 | Value:  0.741 | VLoss: 0.0289 | PLoss: -0.0029 | Ent: 0.1551
+[ 42.9%] Iter 4189/9765 | SPS:  2082 | Reward:  0.049 | Value:  0.717 | VLoss: 0.0193 | PLoss: -0.0047 | Ent: 0.1836
+[ 42.9%] Iter 4190/9765 | SPS:  2082 | Reward:  0.058 | Value:  0.665 | VLoss: 0.0266 | PLoss: -0.0106 | Ent: 0.1753
+[ 42.9%] Iter 4191/9765 | SPS:  2082 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0252 | PLoss: -0.0067 | Ent: 0.1854
+[ 42.9%] Iter 4192/9765 | SPS:  2082 | Reward:  0.050 | Value:  0.721 | VLoss: 0.0158 | PLoss: -0.0147 | Ent: 0.1653
+[ 42.9%] Iter 4193/9765 | SPS:  2083 | Reward:  0.050 | Value:  0.716 | VLoss: 0.0263 | PLoss: -0.0033 | Ent: 0.1795
+[ 42.9%] Iter 4194/9765 | SPS:  2083 | Reward:  0.070 | Value:  0.685 | VLoss: 0.0272 | PLoss: -0.0079 | Ent: 0.1582
+[ 43.0%] Iter 4195/9765 | SPS:  2083 | Reward:  0.057 | Value:  0.678 | VLoss: 0.0190 | PLoss: -0.0080 | Ent: 0.1528
+[ 43.0%] Iter 4196/9765 | SPS:  2083 | Reward:  0.045 | Value:  0.509 | VLoss: 0.0209 | PLoss: -0.0053 | Ent: 0.1633
+[ 43.0%] Iter 4197/9765 | SPS:  2084 | Reward:  0.038 | Value:  0.531 | VLoss: 0.0225 | PLoss: -0.0102 | Ent: 0.1467
+[ 43.0%] Iter 4198/9765 | SPS:  2084 | Reward:  0.048 | Value:  0.494 | VLoss: 0.0147 | PLoss: -0.0076 | Ent: 0.1600
+[ 43.0%] Iter 4199/9765 | SPS:  2084 | Reward:  0.047 | Value:  0.558 | VLoss: 0.0173 | PLoss: -0.0130 | Ent: 0.1334
+[ 43.0%] Iter 4200/9765 | SPS:  2084 | Reward:  0.049 | Value:  0.588 | VLoss: 0.0206 | PLoss: -0.0057 | Ent: 0.1855
+[ 43.0%] Iter 4201/9765 | SPS:  2085 | Reward:  0.048 | Value:  0.579 | VLoss: 0.0155 | PLoss: -0.0016 | Ent: 0.1568
+[ 43.0%] Iter 4202/9765 | SPS:  2085 | Reward:  0.048 | Value:  0.549 | VLoss: 0.0139 | PLoss: -0.0043 | Ent: 0.1411
+[ 43.0%] Iter 4203/9765 | SPS:  2085 | Reward:  0.035 | Value:  0.513 | VLoss: 0.0209 | PLoss: -0.0072 | Ent: 0.1596
+[ 43.1%] Iter 4204/9765 | SPS:  2085 | Reward:  0.030 | Value:  0.496 | VLoss: 0.0160 | PLoss: -0.0092 | Ent: 0.1687
+[ 43.1%] Iter 4205/9765 | SPS:  2085 | Reward:  0.044 | Value:  0.478 | VLoss: 0.0166 | PLoss: 0.0024 | Ent: 0.1662
+[ 43.1%] Iter 4206/9765 | SPS:  2086 | Reward:  0.049 | Value:  0.538 | VLoss: 0.0162 | PLoss: -0.0035 | Ent: 0.1182
+[ 43.1%] Iter 4207/9765 | SPS:  2086 | Reward:  0.046 | Value:  0.599 | VLoss: 0.0167 | PLoss: -0.0088 | Ent: 0.1574
+[ 43.1%] Iter 4208/9765 | SPS:  2086 | Reward:  0.050 | Value:  0.649 | VLoss: 0.0170 | PLoss: -0.0079 | Ent: 0.1691
+[ 43.1%] Iter 4209/9765 | SPS:  2086 | Reward:  0.062 | Value:  0.664 | VLoss: 0.0239 | PLoss: -0.0110 | Ent: 0.1520
+[ 43.1%] Iter 4210/9765 | SPS:  2087 | Reward:  0.055 | Value:  0.649 | VLoss: 0.0185 | PLoss: -0.0111 | Ent: 0.1507
+[ 43.1%] Iter 4211/9765 | SPS:  2087 | Reward:  0.068 | Value:  0.734 | VLoss: 0.0206 | PLoss: -0.0121 | Ent: 0.1683
+[ 43.1%] Iter 4212/9765 | SPS:  2087 | Reward:  0.059 | Value:  0.684 | VLoss: 0.0278 | PLoss: -0.0089 | Ent: 0.2089
+[ 43.1%] Iter 4213/9765 | SPS:  2087 | Reward:  0.058 | Value:  0.735 | VLoss: 0.0246 | PLoss: -0.0111 | Ent: 0.1771
+[ 43.2%] Iter 4214/9765 | SPS:  2087 | Reward:  0.055 | Value:  0.697 | VLoss: 0.0254 | PLoss: -0.0096 | Ent: 0.1822
+[ 43.2%] Iter 4215/9765 | SPS:  2088 | Reward:  0.062 | Value:  0.627 | VLoss: 0.0328 | PLoss: -0.0019 | Ent: 0.1565
+[ 43.2%] Iter 4216/9765 | SPS:  2088 | Reward:  0.074 | Value:  0.723 | VLoss: 0.0240 | PLoss: -0.0061 | Ent: 0.1207
+[ 43.2%] Iter 4217/9765 | SPS:  2088 | Reward:  0.077 | Value:  0.762 | VLoss: 0.0321 | PLoss: -0.0084 | Ent: 0.1356
+[ 43.2%] Iter 4218/9765 | SPS:  2088 | Reward:  0.045 | Value:  0.665 | VLoss: 0.0208 | PLoss: -0.0049 | Ent: 0.1897
+[ 43.2%] Iter 4219/9765 | SPS:  2088 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0246 | PLoss: -0.0045 | Ent: 0.1526
+[ 43.2%] Iter 4220/9765 | SPS:  2089 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0302 | PLoss: -0.0075 | Ent: 0.1515
+[ 43.2%] Iter 4221/9765 | SPS:  2089 | Reward:  0.054 | Value:  0.684 | VLoss: 0.0226 | PLoss: -0.0107 | Ent: 0.1834
+[ 43.2%] Iter 4222/9765 | SPS:  2089 | Reward:  0.069 | Value:  0.754 | VLoss: 0.0318 | PLoss: -0.0086 | Ent: 0.1936
+[ 43.2%] Iter 4223/9765 | SPS:  2089 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0229 | PLoss: 0.0010 | Ent: 0.1329
+[ 43.3%] Iter 4224/9765 | SPS:  2090 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0285 | PLoss: -0.0067 | Ent: 0.1860
+[ 43.3%] Iter 4225/9765 | SPS:  2090 | Reward:  0.061 | Value:  0.727 | VLoss: 0.0249 | PLoss: -0.0018 | Ent: 0.1476
+[ 43.3%] Iter 4226/9765 | SPS:  2090 | Reward:  0.056 | Value:  0.665 | VLoss: 0.0270 | PLoss: -0.0144 | Ent: 0.1908
+[ 43.3%] Iter 4227/9765 | SPS:  2090 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0238 | PLoss: -0.0133 | Ent: 0.1905
+[ 43.3%] Iter 4228/9765 | SPS:  2090 | Reward:  0.064 | Value:  0.699 | VLoss: 0.0192 | PLoss: -0.0046 | Ent: 0.2121
+[ 43.3%] Iter 4229/9765 | SPS:  2091 | Reward:  0.052 | Value:  0.686 | VLoss: 0.0238 | PLoss: -0.0069 | Ent: 0.1541
+[ 43.3%] Iter 4230/9765 | SPS:  2091 | Reward:  0.052 | Value:  0.600 | VLoss: 0.0201 | PLoss: -0.0043 | Ent: 0.1415
+[ 43.3%] Iter 4231/9765 | SPS:  2091 | Reward:  0.052 | Value:  0.665 | VLoss: 0.0224 | PLoss: -0.0044 | Ent: 0.1602
+[ 43.3%] Iter 4232/9765 | SPS:  2091 | Reward:  0.065 | Value:  0.619 | VLoss: 0.0211 | PLoss: -0.0060 | Ent: 0.1865
+[ 43.3%] Iter 4233/9765 | SPS:  2091 | Reward:  0.066 | Value:  0.741 | VLoss: 0.0220 | PLoss: -0.0094 | Ent: 0.1628
+[ 43.4%] Iter 4234/9765 | SPS:  2092 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0259 | PLoss: -0.0128 | Ent: 0.2038
+[ 43.4%] Iter 4235/9765 | SPS:  2092 | Reward:  0.062 | Value:  0.738 | VLoss: 0.0309 | PLoss: -0.0070 | Ent: 0.1721
+[ 43.4%] Iter 4236/9765 | SPS:  2092 | Reward:  0.062 | Value:  0.663 | VLoss: 0.0303 | PLoss: -0.0108 | Ent: 0.1507
+[ 43.4%] Iter 4237/9765 | SPS:  2092 | Reward:  0.063 | Value:  0.648 | VLoss: 0.0221 | PLoss: -0.0108 | Ent: 0.1867
+[ 43.4%] Iter 4238/9765 | SPS:  2093 | Reward:  0.067 | Value:  0.678 | VLoss: 0.0269 | PLoss: -0.0110 | Ent: 0.2153
+[ 43.4%] Iter 4239/9765 | SPS:  2093 | Reward:  0.066 | Value:  0.692 | VLoss: 0.0298 | PLoss: -0.0064 | Ent: 0.2308
+[ 43.4%] Iter 4240/9765 | SPS:  2093 | Reward:  0.065 | Value:  0.684 | VLoss: 0.0256 | PLoss: -0.0095 | Ent: 0.2053
+[ 43.4%] Iter 4241/9765 | SPS:  2093 | Reward:  0.071 | Value:  0.654 | VLoss: 0.0322 | PLoss: -0.0115 | Ent: 0.1762
+[ 43.4%] Iter 4242/9765 | SPS:  2093 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0207 | PLoss: -0.0060 | Ent: 0.1724
+[ 43.5%] Iter 4243/9765 | SPS:  2094 | Reward:  0.061 | Value:  0.690 | VLoss: 0.0227 | PLoss: -0.0054 | Ent: 0.1768
+[ 43.5%] Iter 4244/9765 | SPS:  2094 | Reward:  0.062 | Value:  0.723 | VLoss: 0.0278 | PLoss: -0.0101 | Ent: 0.1908
+[ 43.5%] Iter 4245/9765 | SPS:  2094 | Reward:  0.062 | Value:  0.698 | VLoss: 0.0192 | PLoss: -0.0102 | Ent: 0.1924
+[ 43.5%] Iter 4246/9765 | SPS:  2094 | Reward:  0.045 | Value:  0.791 | VLoss: 0.0143 | PLoss: -0.0061 | Ent: 0.1184
+[ 43.5%] Iter 4247/9765 | SPS:  2094 | Reward:  0.061 | Value:  0.731 | VLoss: 0.0259 | PLoss: -0.0099 | Ent: 0.1563
+[ 43.5%] Iter 4248/9765 | SPS:  2095 | Reward:  0.057 | Value:  0.671 | VLoss: 0.0235 | PLoss: -0.0090 | Ent: 0.1696
+[ 43.5%] Iter 4249/9765 | SPS:  2095 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0200 | PLoss: -0.0054 | Ent: 0.1622
+[ 43.5%] Iter 4250/9765 | SPS:  2095 | Reward:  0.064 | Value:  0.703 | VLoss: 0.0279 | PLoss: -0.0076 | Ent: 0.1329
+[ 43.5%] Iter 4251/9765 | SPS:  2095 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0216 | PLoss: -0.0067 | Ent: 0.1325
+[ 43.5%] Iter 4252/9765 | SPS:  2096 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0209 | PLoss: -0.0152 | Ent: 0.1324
+[ 43.6%] Iter 4253/9765 | SPS:  2096 | Reward:  0.060 | Value:  0.700 | VLoss: 0.0226 | PLoss: -0.0076 | Ent: 0.1592
+[ 43.6%] Iter 4254/9765 | SPS:  2096 | Reward:  0.058 | Value:  0.650 | VLoss: 0.0187 | PLoss: -0.0076 | Ent: 0.2002
+[ 43.6%] Iter 4255/9765 | SPS:  2096 | Reward:  0.060 | Value:  0.681 | VLoss: 0.0169 | PLoss: -0.0026 | Ent: 0.1422
+[ 43.6%] Iter 4256/9765 | SPS:  2096 | Reward:  0.043 | Value:  0.578 | VLoss: 0.0294 | PLoss: -0.0110 | Ent: 0.1717
+[ 43.6%] Iter 4257/9765 | SPS:  2097 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0328 | PLoss: -0.0102 | Ent: 0.1986
+[ 43.6%] Iter 4258/9765 | SPS:  2097 | Reward:  0.053 | Value:  0.700 | VLoss: 0.0298 | PLoss: -0.0060 | Ent: 0.2027
+[ 43.6%] Iter 4259/9765 | SPS:  2097 | Reward:  0.064 | Value:  0.629 | VLoss: 0.0345 | PLoss: -0.0078 | Ent: 0.1421
+[ 43.6%] Iter 4260/9765 | SPS:  2097 | Reward:  0.054 | Value:  0.601 | VLoss: 0.0262 | PLoss: -0.0063 | Ent: 0.1633
+[ 43.6%] Iter 4261/9765 | SPS:  2097 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0285 | PLoss: -0.0051 | Ent: 0.2018
+[ 43.6%] Iter 4262/9765 | SPS:  2098 | Reward:  0.060 | Value:  0.689 | VLoss: 0.0233 | PLoss: -0.0017 | Ent: 0.1806
+[ 43.7%] Iter 4263/9765 | SPS:  2098 | Reward:  0.071 | Value:  0.730 | VLoss: 0.0205 | PLoss: -0.0133 | Ent: 0.1585
+[ 43.7%] Iter 4264/9765 | SPS:  2098 | Reward:  0.060 | Value:  0.703 | VLoss: 0.0216 | PLoss: -0.0073 | Ent: 0.1724
+[ 43.7%] Iter 4265/9765 | SPS:  2098 | Reward:  0.068 | Value:  0.700 | VLoss: 0.0246 | PLoss: -0.0114 | Ent: 0.1517
+[ 43.7%] Iter 4266/9765 | SPS:  2098 | Reward:  0.067 | Value:  0.670 | VLoss: 0.0332 | PLoss: -0.0115 | Ent: 0.1508
+[ 43.7%] Iter 4267/9765 | SPS:  2099 | Reward:  0.050 | Value:  0.714 | VLoss: 0.0184 | PLoss: -0.0061 | Ent: 0.1790
+[ 43.7%] Iter 4268/9765 | SPS:  2099 | Reward:  0.061 | Value:  0.688 | VLoss: 0.0214 | PLoss: -0.0036 | Ent: 0.2036
+[ 43.7%] Iter 4269/9765 | SPS:  2099 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0271 | PLoss: -0.0098 | Ent: 0.1495
+[ 43.7%] Iter 4270/9765 | SPS:  2099 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0220 | PLoss: -0.0038 | Ent: 0.1300
+[ 43.7%] Iter 4271/9765 | SPS:  2100 | Reward:  0.043 | Value:  0.706 | VLoss: 0.0173 | PLoss: -0.0111 | Ent: 0.2800
+[ 43.7%] Iter 4272/9765 | SPS:  2100 | Reward:  0.073 | Value:  0.697 | VLoss: 0.0297 | PLoss: -0.0080 | Ent: 0.1571
+[ 43.8%] Iter 4273/9765 | SPS:  2100 | Reward:  0.066 | Value:  0.742 | VLoss: 0.0182 | PLoss: -0.0089 | Ent: 0.1580
+[ 43.8%] Iter 4274/9765 | SPS:  2100 | Reward:  0.051 | Value:  0.696 | VLoss: 0.0218 | PLoss: -0.0128 | Ent: 0.1668
+[ 43.8%] Iter 4275/9765 | SPS:  2100 | Reward:  0.064 | Value:  0.663 | VLoss: 0.0281 | PLoss: -0.0067 | Ent: 0.1323
+[ 43.8%] Iter 4276/9765 | SPS:  2101 | Reward:  0.059 | Value:  0.723 | VLoss: 0.0227 | PLoss: -0.0074 | Ent: 0.1760
+[ 43.8%] Iter 4277/9765 | SPS:  2101 | Reward:  0.062 | Value:  0.664 | VLoss: 0.0258 | PLoss: -0.0113 | Ent: 0.1878
+[ 43.8%] Iter 4278/9765 | SPS:  2101 | Reward:  0.050 | Value:  0.741 | VLoss: 0.0275 | PLoss: -0.0073 | Ent: 0.2128
+[ 43.8%] Iter 4279/9765 | SPS:  2101 | Reward:  0.077 | Value:  0.706 | VLoss: 0.0236 | PLoss: -0.0120 | Ent: 0.1632
+[ 43.8%] Iter 4280/9765 | SPS:  2101 | Reward:  0.063 | Value:  0.759 | VLoss: 0.0173 | PLoss: -0.0052 | Ent: 0.2080
+[ 43.8%] Iter 4281/9765 | SPS:  2102 | Reward:  0.062 | Value:  0.759 | VLoss: 0.0222 | PLoss: -0.0066 | Ent: 0.1797
+[ 43.9%] Iter 4282/9765 | SPS:  2102 | Reward:  0.058 | Value:  0.665 | VLoss: 0.0237 | PLoss: -0.0066 | Ent: 0.1957
+[ 43.9%] Iter 4283/9765 | SPS:  2102 | Reward:  0.070 | Value:  0.692 | VLoss: 0.0253 | PLoss: -0.0041 | Ent: 0.1435
+[ 43.9%] Iter 4284/9765 | SPS:  2102 | Reward:  0.071 | Value:  0.742 | VLoss: 0.0245 | PLoss: -0.0086 | Ent: 0.1571
+[ 43.9%] Iter 4285/9765 | SPS:  2102 | Reward:  0.057 | Value:  0.686 | VLoss: 0.0281 | PLoss: -0.0068 | Ent: 0.1483
+[ 43.9%] Iter 4286/9765 | SPS:  2103 | Reward:  0.062 | Value:  0.685 | VLoss: 0.0287 | PLoss: -0.0069 | Ent: 0.1475
+[ 43.9%] Iter 4287/9765 | SPS:  2103 | Reward:  0.055 | Value:  0.704 | VLoss: 0.0226 | PLoss: -0.0076 | Ent: 0.1617
+[ 43.9%] Iter 4288/9765 | SPS:  2103 | Reward:  0.057 | Value:  0.696 | VLoss: 0.0240 | PLoss: -0.0068 | Ent: 0.1895
+[ 43.9%] Iter 4289/9765 | SPS:  2103 | Reward:  0.073 | Value:  0.706 | VLoss: 0.0252 | PLoss: -0.0020 | Ent: 0.1578
+[ 43.9%] Iter 4290/9765 | SPS:  2104 | Reward:  0.052 | Value:  0.750 | VLoss: 0.0194 | PLoss: -0.0088 | Ent: 0.1807
+[ 43.9%] Iter 4291/9765 | SPS:  2104 | Reward:  0.062 | Value:  0.698 | VLoss: 0.0261 | PLoss: -0.0097 | Ent: 0.1727
+[ 44.0%] Iter 4292/9765 | SPS:  2104 | Reward:  0.063 | Value:  0.697 | VLoss: 0.0298 | PLoss: -0.0038 | Ent: 0.1490
+[ 44.0%] Iter 4293/9765 | SPS:  2104 | Reward:  0.063 | Value:  0.714 | VLoss: 0.0271 | PLoss: -0.0071 | Ent: 0.1736
+[ 44.0%] Iter 4294/9765 | SPS:  2104 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0236 | PLoss: -0.0082 | Ent: 0.1892
+[ 44.0%] Iter 4295/9765 | SPS:  2105 | Reward:  0.061 | Value:  0.782 | VLoss: 0.0228 | PLoss: -0.0148 | Ent: 0.1798
+[ 44.0%] Iter 4296/9765 | SPS:  2105 | Reward:  0.069 | Value:  0.734 | VLoss: 0.0235 | PLoss: -0.0081 | Ent: 0.1546
+[ 44.0%] Iter 4297/9765 | SPS:  2105 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0257 | PLoss: 0.0001 | Ent: 0.1561
+[ 44.0%] Iter 4298/9765 | SPS:  2105 | Reward:  0.079 | Value:  0.722 | VLoss: 0.0237 | PLoss: -0.0128 | Ent: 0.1876
+[ 44.0%] Iter 4299/9765 | SPS:  2105 | Reward:  0.071 | Value:  0.814 | VLoss: 0.0231 | PLoss: -0.0073 | Ent: 0.1765
+[ 44.0%] Iter 4300/9765 | SPS:  2106 | Reward:  0.066 | Value:  0.705 | VLoss: 0.0244 | PLoss: -0.0042 | Ent: 0.1478
+[ 44.0%] Iter 4301/9765 | SPS:  2106 | Reward:  0.065 | Value:  0.736 | VLoss: 0.0177 | PLoss: -0.0072 | Ent: 0.1866
+[ 44.1%] Iter 4302/9765 | SPS:  2106 | Reward:  0.067 | Value:  0.677 | VLoss: 0.0271 | PLoss: -0.0061 | Ent: 0.1377
+[ 44.1%] Iter 4303/9765 | SPS:  2106 | Reward:  0.066 | Value:  0.661 | VLoss: 0.0284 | PLoss: -0.0091 | Ent: 0.1567
+[ 44.1%] Iter 4304/9765 | SPS:  2106 | Reward:  0.075 | Value:  0.644 | VLoss: 0.0331 | PLoss: -0.0107 | Ent: 0.2110
+[ 44.1%] Iter 4305/9765 | SPS:  2107 | Reward:  0.068 | Value:  0.728 | VLoss: 0.0196 | PLoss: -0.0047 | Ent: 0.1758
+[ 44.1%] Iter 4306/9765 | SPS:  2107 | Reward:  0.063 | Value:  0.752 | VLoss: 0.0230 | PLoss: -0.0107 | Ent: 0.2005
+[ 44.1%] Iter 4307/9765 | SPS:  2107 | Reward:  0.060 | Value:  0.811 | VLoss: 0.0174 | PLoss: -0.0052 | Ent: 0.1220
+[ 44.1%] Iter 4308/9765 | SPS:  2107 | Reward:  0.075 | Value:  0.729 | VLoss: 0.0187 | PLoss: -0.0066 | Ent: 0.1610
+[ 44.1%] Iter 4309/9765 | SPS:  2107 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0223 | PLoss: -0.0075 | Ent: 0.1336
+[ 44.1%] Iter 4310/9765 | SPS:  2108 | Reward:  0.081 | Value:  0.718 | VLoss: 0.0234 | PLoss: 0.0010 | Ent: 0.1432
+[ 44.1%] Iter 4311/9765 | SPS:  2108 | Reward:  0.069 | Value:  0.763 | VLoss: 0.0186 | PLoss: -0.0101 | Ent: 0.2053
+[ 44.2%] Iter 4312/9765 | SPS:  2108 | Reward:  0.065 | Value:  0.772 | VLoss: 0.0267 | PLoss: -0.0079 | Ent: 0.1971
+[ 44.2%] Iter 4313/9765 | SPS:  2108 | Reward:  0.076 | Value:  0.731 | VLoss: 0.0232 | PLoss: -0.0155 | Ent: 0.2150
+[ 44.2%] Iter 4314/9765 | SPS:  2108 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0239 | PLoss: -0.0039 | Ent: 0.1669
+[ 44.2%] Iter 4315/9765 | SPS:  2109 | Reward:  0.080 | Value:  0.680 | VLoss: 0.0250 | PLoss: -0.0060 | Ent: 0.1669
+[ 44.2%] Iter 4316/9765 | SPS:  2109 | Reward:  0.055 | Value:  0.651 | VLoss: 0.0198 | PLoss: -0.0051 | Ent: 0.1892
+[ 44.2%] Iter 4317/9765 | SPS:  2109 | Reward:  0.060 | Value:  0.732 | VLoss: 0.0172 | PLoss: -0.0039 | Ent: 0.1814
+[ 44.2%] Iter 4318/9765 | SPS:  2109 | Reward:  0.059 | Value:  0.758 | VLoss: 0.0179 | PLoss: -0.0103 | Ent: 0.1407
+[ 44.2%] Iter 4319/9765 | SPS:  2110 | Reward:  0.053 | Value:  0.685 | VLoss: 0.0210 | PLoss: -0.0043 | Ent: 0.1724
+[ 44.2%] Iter 4320/9765 | SPS:  2110 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0275 | PLoss: -0.0044 | Ent: 0.1495
+[ 44.2%] Iter 4321/9765 | SPS:  2110 | Reward:  0.056 | Value:  0.711 | VLoss: 0.0264 | PLoss: -0.0078 | Ent: 0.2103
+[ 44.3%] Iter 4322/9765 | SPS:  2110 | Reward:  0.075 | Value:  0.737 | VLoss: 0.0248 | PLoss: 0.0001 | Ent: 0.1580
+[ 44.3%] Iter 4323/9765 | SPS:  2110 | Reward:  0.063 | Value:  0.740 | VLoss: 0.0222 | PLoss: -0.0106 | Ent: 0.1591
+[ 44.3%] Iter 4324/9765 | SPS:  2111 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0226 | PLoss: -0.0061 | Ent: 0.1668
+[ 44.3%] Iter 4325/9765 | SPS:  2111 | Reward:  0.048 | Value:  0.661 | VLoss: 0.0263 | PLoss: -0.0143 | Ent: 0.1605
+[ 44.3%] Iter 4326/9765 | SPS:  2111 | Reward:  0.055 | Value:  0.697 | VLoss: 0.0306 | PLoss: -0.0139 | Ent: 0.1877
+[ 44.3%] Iter 4327/9765 | SPS:  2111 | Reward:  0.061 | Value:  0.621 | VLoss: 0.0143 | PLoss: -0.0142 | Ent: 0.1885
+[ 44.3%] Iter 4328/9765 | SPS:  2111 | Reward:  0.070 | Value:  0.699 | VLoss: 0.0270 | PLoss: -0.0077 | Ent: 0.1541
+[ 44.3%] Iter 4329/9765 | SPS:  2112 | Reward:  0.068 | Value:  0.744 | VLoss: 0.0214 | PLoss: -0.0061 | Ent: 0.2032
+[ 44.3%] Iter 4330/9765 | SPS:  2112 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0173 | PLoss: -0.0053 | Ent: 0.1617
+[ 44.4%] Iter 4331/9765 | SPS:  2112 | Reward:  0.059 | Value:  0.765 | VLoss: 0.0265 | PLoss: -0.0089 | Ent: 0.1713
+[ 44.4%] Iter 4332/9765 | SPS:  2112 | Reward:  0.060 | Value:  0.703 | VLoss: 0.0269 | PLoss: -0.0028 | Ent: 0.1700
+[ 44.4%] Iter 4333/9765 | SPS:  2112 | Reward:  0.069 | Value:  0.725 | VLoss: 0.0224 | PLoss: -0.0093 | Ent: 0.1886
+[ 44.4%] Iter 4334/9765 | SPS:  2113 | Reward:  0.066 | Value:  0.640 | VLoss: 0.0243 | PLoss: -0.0117 | Ent: 0.1896
+[ 44.4%] Iter 4335/9765 | SPS:  2113 | Reward:  0.053 | Value:  0.638 | VLoss: 0.0203 | PLoss: -0.0121 | Ent: 0.1857
+[ 44.4%] Iter 4336/9765 | SPS:  2113 | Reward:  0.050 | Value:  0.684 | VLoss: 0.0222 | PLoss: -0.0078 | Ent: 0.1687
+[ 44.4%] Iter 4337/9765 | SPS:  2113 | Reward:  0.055 | Value:  0.749 | VLoss: 0.0213 | PLoss: -0.0141 | Ent: 0.1857
+[ 44.4%] Iter 4338/9765 | SPS:  2113 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0258 | PLoss: -0.0184 | Ent: 0.1975
+[ 44.4%] Iter 4339/9765 | SPS:  2114 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0174 | PLoss: -0.0054 | Ent: 0.1683
+[ 44.4%] Iter 4340/9765 | SPS:  2114 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0210 | PLoss: -0.0083 | Ent: 0.1761
+[ 44.5%] Iter 4341/9765 | SPS:  2114 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0292 | PLoss: -0.0066 | Ent: 0.1516
+[ 44.5%] Iter 4342/9765 | SPS:  2114 | Reward:  0.064 | Value:  0.776 | VLoss: 0.0205 | PLoss: -0.0058 | Ent: 0.1506
+[ 44.5%] Iter 4343/9765 | SPS:  2115 | Reward:  0.048 | Value:  0.737 | VLoss: 0.0207 | PLoss: -0.0056 | Ent: 0.1266
+[ 44.5%] Iter 4344/9765 | SPS:  2115 | Reward:  0.062 | Value:  0.628 | VLoss: 0.0259 | PLoss: -0.0106 | Ent: 0.1576
+[ 44.5%] Iter 4345/9765 | SPS:  2115 | Reward:  0.045 | Value:  0.670 | VLoss: 0.0284 | PLoss: -0.0067 | Ent: 0.2050
+[ 44.5%] Iter 4346/9765 | SPS:  2115 | Reward:  0.061 | Value:  0.631 | VLoss: 0.0257 | PLoss: -0.0074 | Ent: 0.2148
+[ 44.5%] Iter 4347/9765 | SPS:  2115 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0137 | PLoss: -0.0032 | Ent: 0.1491
+[ 44.5%] Iter 4348/9765 | SPS:  2116 | Reward:  0.059 | Value:  0.733 | VLoss: 0.0243 | PLoss: -0.0079 | Ent: 0.1673
+[ 44.5%] Iter 4349/9765 | SPS:  2116 | Reward:  0.059 | Value:  0.667 | VLoss: 0.0249 | PLoss: -0.0098 | Ent: 0.1946
+[ 44.5%] Iter 4350/9765 | SPS:  2116 | Reward:  0.050 | Value:  0.576 | VLoss: 0.0264 | PLoss: -0.0015 | Ent: 0.1669
+[ 44.6%] Iter 4351/9765 | SPS:  2116 | Reward:  0.069 | Value:  0.746 | VLoss: 0.0323 | PLoss: -0.0133 | Ent: 0.1690
+[ 44.6%] Iter 4352/9765 | SPS:  2116 | Reward:  0.080 | Value:  0.673 | VLoss: 0.0225 | PLoss: -0.0102 | Ent: 0.1977
+[ 44.6%] Iter 4353/9765 | SPS:  2117 | Reward:  0.056 | Value:  0.689 | VLoss: 0.0348 | PLoss: -0.0086 | Ent: 0.1767
+[ 44.6%] Iter 4354/9765 | SPS:  2117 | Reward:  0.059 | Value:  0.642 | VLoss: 0.0205 | PLoss: -0.0092 | Ent: 0.2130
+[ 44.6%] Iter 4355/9765 | SPS:  2117 | Reward:  0.057 | Value:  0.762 | VLoss: 0.0210 | PLoss: -0.0137 | Ent: 0.2001
+[ 44.6%] Iter 4356/9765 | SPS:  2117 | Reward:  0.054 | Value:  0.776 | VLoss: 0.0216 | PLoss: -0.0037 | Ent: 0.1662
+[ 44.6%] Iter 4357/9765 | SPS:  2117 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0193 | PLoss: -0.0067 | Ent: 0.1496
+[ 44.6%] Iter 4358/9765 | SPS:  2118 | Reward:  0.056 | Value:  0.624 | VLoss: 0.0314 | PLoss: -0.0004 | Ent: 0.1993
+[ 44.6%] Iter 4359/9765 | SPS:  2118 | Reward:  0.064 | Value:  0.679 | VLoss: 0.0253 | PLoss: -0.0061 | Ent: 0.2212
+[ 44.6%] Iter 4360/9765 | SPS:  2118 | Reward:  0.068 | Value:  0.682 | VLoss: 0.0277 | PLoss: -0.0116 | Ent: 0.1887
+[ 44.7%] Iter 4361/9765 | SPS:  2118 | Reward:  0.055 | Value:  0.634 | VLoss: 0.0214 | PLoss: -0.0087 | Ent: 0.1760
+[ 44.7%] Iter 4362/9765 | SPS:  2118 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0191 | PLoss: -0.0043 | Ent: 0.2217
+[ 44.7%] Iter 4363/9765 | SPS:  2119 | Reward:  0.054 | Value:  0.691 | VLoss: 0.0247 | PLoss: -0.0061 | Ent: 0.1444
+[ 44.7%] Iter 4364/9765 | SPS:  2119 | Reward:  0.075 | Value:  0.692 | VLoss: 0.0293 | PLoss: -0.0122 | Ent: 0.1846
+[ 44.7%] Iter 4365/9765 | SPS:  2119 | Reward:  0.061 | Value:  0.650 | VLoss: 0.0183 | PLoss: -0.0033 | Ent: 0.2167
+[ 44.7%] Iter 4366/9765 | SPS:  2119 | Reward:  0.071 | Value:  0.735 | VLoss: 0.0233 | PLoss: -0.0127 | Ent: 0.1930
+[ 44.7%] Iter 4367/9765 | SPS:  2120 | Reward:  0.069 | Value:  0.709 | VLoss: 0.0276 | PLoss: -0.0125 | Ent: 0.1795
+[ 44.7%] Iter 4368/9765 | SPS:  2120 | Reward:  0.052 | Value:  0.675 | VLoss: 0.0206 | PLoss: -0.0061 | Ent: 0.1716
+[ 44.7%] Iter 4369/9765 | SPS:  2120 | Reward:  0.058 | Value:  0.584 | VLoss: 0.0232 | PLoss: -0.0024 | Ent: 0.1396
+[ 44.8%] Iter 4370/9765 | SPS:  2120 | Reward:  0.050 | Value:  0.665 | VLoss: 0.0242 | PLoss: -0.0075 | Ent: 0.1415
+[ 44.8%] Iter 4371/9765 | SPS:  2120 | Reward:  0.061 | Value:  0.703 | VLoss: 0.0266 | PLoss: -0.0051 | Ent: 0.2301
+[ 44.8%] Iter 4372/9765 | SPS:  2121 | Reward:  0.061 | Value:  0.651 | VLoss: 0.0259 | PLoss: -0.0055 | Ent: 0.1707
+[ 44.8%] Iter 4373/9765 | SPS:  2121 | Reward:  0.061 | Value:  0.787 | VLoss: 0.0218 | PLoss: -0.0019 | Ent: 0.1492
+[ 44.8%] Iter 4374/9765 | SPS:  2121 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0232 | PLoss: -0.0146 | Ent: 0.1915
+[ 44.8%] Iter 4375/9765 | SPS:  2121 | Reward:  0.071 | Value:  0.709 | VLoss: 0.0338 | PLoss: -0.0069 | Ent: 0.1996
+[ 44.8%] Iter 4376/9765 | SPS:  2121 | Reward:  0.058 | Value:  0.705 | VLoss: 0.0387 | PLoss: -0.0138 | Ent: 0.1669
+[ 44.8%] Iter 4377/9765 | SPS:  2122 | Reward:  0.064 | Value:  0.688 | VLoss: 0.0208 | PLoss: -0.0098 | Ent: 0.1616
+[ 44.8%] Iter 4378/9765 | SPS:  2122 | Reward:  0.054 | Value:  0.682 | VLoss: 0.0181 | PLoss: -0.0014 | Ent: 0.1345
+[ 44.8%] Iter 4379/9765 | SPS:  2122 | Reward:  0.077 | Value:  0.694 | VLoss: 0.0430 | PLoss: -0.0057 | Ent: 0.2075
+[ 44.9%] Iter 4380/9765 | SPS:  2122 | Reward:  0.057 | Value:  0.686 | VLoss: 0.0277 | PLoss: -0.0061 | Ent: 0.1833
+[ 44.9%] Iter 4381/9765 | SPS:  2122 | Reward:  0.053 | Value:  0.692 | VLoss: 0.0354 | PLoss: -0.0030 | Ent: 0.2058
+[ 44.9%] Iter 4382/9765 | SPS:  2123 | Reward:  0.053 | Value:  0.675 | VLoss: 0.0185 | PLoss: -0.0064 | Ent: 0.2087
+[ 44.9%] Iter 4383/9765 | SPS:  2123 | Reward:  0.068 | Value:  0.657 | VLoss: 0.0300 | PLoss: -0.0020 | Ent: 0.1615
+[ 44.9%] Iter 4384/9765 | SPS:  2123 | Reward:  0.060 | Value:  0.695 | VLoss: 0.0258 | PLoss: -0.0075 | Ent: 0.1712
+[ 44.9%] Iter 4385/9765 | SPS:  2123 | Reward:  0.066 | Value:  0.734 | VLoss: 0.0337 | PLoss: -0.0133 | Ent: 0.1927
+[ 44.9%] Iter 4386/9765 | SPS:  2123 | Reward:  0.065 | Value:  0.693 | VLoss: 0.0226 | PLoss: -0.0037 | Ent: 0.1871
+[ 44.9%] Iter 4387/9765 | SPS:  2124 | Reward:  0.054 | Value:  0.735 | VLoss: 0.0175 | PLoss: -0.0007 | Ent: 0.1739
+[ 44.9%] Iter 4388/9765 | SPS:  2124 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0288 | PLoss: -0.0066 | Ent: 0.1428
+[ 44.9%] Iter 4389/9765 | SPS:  2124 | Reward:  0.065 | Value:  0.749 | VLoss: 0.0285 | PLoss: -0.0052 | Ent: 0.1473
+[ 45.0%] Iter 4390/9765 | SPS:  2124 | Reward:  0.050 | Value:  0.688 | VLoss: 0.0222 | PLoss: -0.0109 | Ent: 0.2167
+[ 45.0%] Iter 4391/9765 | SPS:  2124 | Reward:  0.069 | Value:  0.647 | VLoss: 0.0271 | PLoss: -0.0131 | Ent: 0.2351
+[ 45.0%] Iter 4392/9765 | SPS:  2125 | Reward:  0.061 | Value:  0.625 | VLoss: 0.0337 | PLoss: -0.0205 | Ent: 0.1836
+[ 45.0%] Iter 4393/9765 | SPS:  2125 | Reward:  0.060 | Value:  0.607 | VLoss: 0.0223 | PLoss: -0.0049 | Ent: 0.2022
+[ 45.0%] Iter 4394/9765 | SPS:  2125 | Reward:  0.076 | Value:  0.738 | VLoss: 0.0219 | PLoss: -0.0099 | Ent: 0.1368
+[ 45.0%] Iter 4395/9765 | SPS:  2125 | Reward:  0.070 | Value:  0.716 | VLoss: 0.0290 | PLoss: -0.0116 | Ent: 0.1576
+[ 45.0%] Iter 4396/9765 | SPS:  2125 | Reward:  0.066 | Value:  0.735 | VLoss: 0.0248 | PLoss: -0.0097 | Ent: 0.2183
+[ 45.0%] Iter 4397/9765 | SPS:  2126 | Reward:  0.073 | Value:  0.737 | VLoss: 0.0225 | PLoss: -0.0105 | Ent: 0.1503
+[ 45.0%] Iter 4398/9765 | SPS:  2126 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0329 | PLoss: -0.0075 | Ent: 0.1830
+[ 45.0%] Iter 4399/9765 | SPS:  2126 | Reward:  0.068 | Value:  0.706 | VLoss: 0.0240 | PLoss: -0.0086 | Ent: 0.2014
+[ 45.1%] Iter 4400/9765 | SPS:  2126 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0186 | PLoss: -0.0049 | Ent: 0.1628
+[ 45.1%] Iter 4401/9765 | SPS:  2126 | Reward:  0.071 | Value:  0.775 | VLoss: 0.0241 | PLoss: -0.0103 | Ent: 0.1545
+[ 45.1%] Iter 4402/9765 | SPS:  2127 | Reward:  0.060 | Value:  0.738 | VLoss: 0.0265 | PLoss: -0.0109 | Ent: 0.1726
+[ 45.1%] Iter 4403/9765 | SPS:  2127 | Reward:  0.067 | Value:  0.786 | VLoss: 0.0218 | PLoss: -0.0144 | Ent: 0.1820
+[ 45.1%] Iter 4404/9765 | SPS:  2127 | Reward:  0.070 | Value:  0.754 | VLoss: 0.0241 | PLoss: -0.0071 | Ent: 0.1614
+[ 45.1%] Iter 4405/9765 | SPS:  2127 | Reward:  0.068 | Value:  0.713 | VLoss: 0.0218 | PLoss: -0.0117 | Ent: 0.1466
+[ 45.1%] Iter 4406/9765 | SPS:  2127 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0235 | PLoss: -0.0057 | Ent: 0.1545
+[ 45.1%] Iter 4407/9765 | SPS:  2128 | Reward:  0.068 | Value:  0.758 | VLoss: 0.0284 | PLoss: -0.0174 | Ent: 0.2157
+[ 45.1%] Iter 4408/9765 | SPS:  2128 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0257 | PLoss: -0.0090 | Ent: 0.1838
+[ 45.2%] Iter 4409/9765 | SPS:  2128 | Reward:  0.062 | Value:  0.663 | VLoss: 0.0258 | PLoss: -0.0071 | Ent: 0.1620
+[ 45.2%] Iter 4410/9765 | SPS:  2128 | Reward:  0.059 | Value:  0.718 | VLoss: 0.0331 | PLoss: -0.0041 | Ent: 0.1400
+[ 45.2%] Iter 4411/9765 | SPS:  2128 | Reward:  0.071 | Value:  0.693 | VLoss: 0.0212 | PLoss: -0.0099 | Ent: 0.1665
+[ 45.2%] Iter 4412/9765 | SPS:  2129 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0274 | PLoss: -0.0080 | Ent: 0.2076
+[ 45.2%] Iter 4413/9765 | SPS:  2129 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0344 | PLoss: -0.0084 | Ent: 0.1507
+[ 45.2%] Iter 4414/9765 | SPS:  2129 | Reward:  0.069 | Value:  0.683 | VLoss: 0.0289 | PLoss: -0.0050 | Ent: 0.1894
+[ 45.2%] Iter 4415/9765 | SPS:  2129 | Reward:  0.074 | Value:  0.709 | VLoss: 0.0312 | PLoss: -0.0075 | Ent: 0.1916
+[ 45.2%] Iter 4416/9765 | SPS:  2129 | Reward:  0.054 | Value:  0.703 | VLoss: 0.0198 | PLoss: -0.0022 | Ent: 0.1702
+[ 45.2%] Iter 4417/9765 | SPS:  2130 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0270 | PLoss: -0.0092 | Ent: 0.1676
+[ 45.2%] Iter 4418/9765 | SPS:  2130 | Reward:  0.059 | Value:  0.677 | VLoss: 0.0219 | PLoss: -0.0076 | Ent: 0.1862
+[ 45.3%] Iter 4419/9765 | SPS:  2130 | Reward:  0.065 | Value:  0.687 | VLoss: 0.0302 | PLoss: -0.0112 | Ent: 0.1868
+[ 45.3%] Iter 4420/9765 | SPS:  2130 | Reward:  0.074 | Value:  0.727 | VLoss: 0.0187 | PLoss: -0.0021 | Ent: 0.1483
+[ 45.3%] Iter 4421/9765 | SPS:  2130 | Reward:  0.046 | Value:  0.677 | VLoss: 0.0186 | PLoss: -0.0022 | Ent: 0.1554
+[ 45.3%] Iter 4422/9765 | SPS:  2131 | Reward:  0.061 | Value:  0.798 | VLoss: 0.0204 | PLoss: -0.0025 | Ent: 0.1327
+[ 45.3%] Iter 4423/9765 | SPS:  2131 | Reward:  0.054 | Value:  0.685 | VLoss: 0.0276 | PLoss: -0.0097 | Ent: 0.1841
+[ 45.3%] Iter 4424/9765 | SPS:  2131 | Reward:  0.058 | Value:  0.653 | VLoss: 0.0215 | PLoss: -0.0167 | Ent: 0.1882
+[ 45.3%] Iter 4425/9765 | SPS:  2131 | Reward:  0.066 | Value:  0.668 | VLoss: 0.0289 | PLoss: -0.0060 | Ent: 0.2046
+[ 45.3%] Iter 4426/9765 | SPS:  2131 | Reward:  0.061 | Value:  0.744 | VLoss: 0.0264 | PLoss: -0.0078 | Ent: 0.1848
+[ 45.3%] Iter 4427/9765 | SPS:  2132 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0239 | PLoss: -0.0100 | Ent: 0.1933
+[ 45.3%] Iter 4428/9765 | SPS:  2132 | Reward:  0.053 | Value:  0.705 | VLoss: 0.0288 | PLoss: -0.0070 | Ent: 0.2062
+[ 45.4%] Iter 4429/9765 | SPS:  2132 | Reward:  0.065 | Value:  0.659 | VLoss: 0.0277 | PLoss: -0.0086 | Ent: 0.1527
+[ 45.4%] Iter 4430/9765 | SPS:  2132 | Reward:  0.064 | Value:  0.680 | VLoss: 0.0257 | PLoss: -0.0091 | Ent: 0.1901
+[ 45.4%] Iter 4431/9765 | SPS:  2132 | Reward:  0.073 | Value:  0.703 | VLoss: 0.0274 | PLoss: -0.0124 | Ent: 0.1584
+[ 45.4%] Iter 4432/9765 | SPS:  2133 | Reward:  0.066 | Value:  0.693 | VLoss: 0.0296 | PLoss: -0.0077 | Ent: 0.1738
+[ 45.4%] Iter 4433/9765 | SPS:  2133 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0225 | PLoss: -0.0068 | Ent: 0.1633
+[ 45.4%] Iter 4434/9765 | SPS:  2133 | Reward:  0.068 | Value:  0.700 | VLoss: 0.0239 | PLoss: -0.0109 | Ent: 0.2107
+[ 45.4%] Iter 4435/9765 | SPS:  2133 | Reward:  0.073 | Value:  0.725 | VLoss: 0.0297 | PLoss: -0.0022 | Ent: 0.1856
+[ 45.4%] Iter 4436/9765 | SPS:  2133 | Reward:  0.066 | Value:  0.709 | VLoss: 0.0226 | PLoss: -0.0069 | Ent: 0.1842
+[ 45.4%] Iter 4437/9765 | SPS:  2134 | Reward:  0.061 | Value:  0.757 | VLoss: 0.0255 | PLoss: -0.0075 | Ent: 0.1705
+[ 45.4%] Iter 4438/9765 | SPS:  2134 | Reward:  0.069 | Value:  0.703 | VLoss: 0.0262 | PLoss: -0.0092 | Ent: 0.1821
+[ 45.5%] Iter 4439/9765 | SPS:  2134 | Reward:  0.067 | Value:  0.676 | VLoss: 0.0264 | PLoss: -0.0069 | Ent: 0.1708
+[ 45.5%] Iter 4440/9765 | SPS:  2134 | Reward:  0.078 | Value:  0.715 | VLoss: 0.0411 | PLoss: -0.0025 | Ent: 0.1510
+[ 45.5%] Iter 4441/9765 | SPS:  2134 | Reward:  0.075 | Value:  0.737 | VLoss: 0.0255 | PLoss: -0.0109 | Ent: 0.1757
+[ 45.5%] Iter 4442/9765 | SPS:  2135 | Reward:  0.051 | Value:  0.698 | VLoss: 0.0174 | PLoss: -0.0129 | Ent: 0.1570
+[ 45.5%] Iter 4443/9765 | SPS:  2135 | Reward:  0.075 | Value:  0.675 | VLoss: 0.0371 | PLoss: -0.0042 | Ent: 0.1597
+[ 45.5%] Iter 4444/9765 | SPS:  2135 | Reward:  0.070 | Value:  0.639 | VLoss: 0.0298 | PLoss: -0.0009 | Ent: 0.1839
+[ 45.5%] Iter 4445/9765 | SPS:  2135 | Reward:  0.065 | Value:  0.664 | VLoss: 0.0318 | PLoss: -0.0124 | Ent: 0.1718
+[ 45.5%] Iter 4446/9765 | SPS:  2135 | Reward:  0.066 | Value:  0.755 | VLoss: 0.0211 | PLoss: -0.0092 | Ent: 0.1878
+[ 45.5%] Iter 4447/9765 | SPS:  2136 | Reward:  0.058 | Value:  0.771 | VLoss: 0.0206 | PLoss: -0.0035 | Ent: 0.1802
+[ 45.6%] Iter 4448/9765 | SPS:  2136 | Reward:  0.057 | Value:  0.702 | VLoss: 0.0209 | PLoss: -0.0054 | Ent: 0.1592
+[ 45.6%] Iter 4449/9765 | SPS:  2136 | Reward:  0.068 | Value:  0.682 | VLoss: 0.0241 | PLoss: -0.0067 | Ent: 0.2023
+[ 45.6%] Iter 4450/9765 | SPS:  2136 | Reward:  0.081 | Value:  0.734 | VLoss: 0.0300 | PLoss: -0.0047 | Ent: 0.1522
+[ 45.6%] Iter 4451/9765 | SPS:  2136 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0242 | PLoss: -0.0020 | Ent: 0.1671
+[ 45.6%] Iter 4452/9765 | SPS:  2137 | Reward:  0.059 | Value:  0.766 | VLoss: 0.0202 | PLoss: -0.0047 | Ent: 0.1248
+[ 45.6%] Iter 4453/9765 | SPS:  2137 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0337 | PLoss: -0.0093 | Ent: 0.1912
+[ 45.6%] Iter 4454/9765 | SPS:  2137 | Reward:  0.077 | Value:  0.721 | VLoss: 0.0259 | PLoss: -0.0053 | Ent: 0.1772
+[ 45.6%] Iter 4455/9765 | SPS:  2137 | Reward:  0.069 | Value:  0.744 | VLoss: 0.0261 | PLoss: -0.0082 | Ent: 0.1456
+[ 45.6%] Iter 4456/9765 | SPS:  2137 | Reward:  0.052 | Value:  0.768 | VLoss: 0.0180 | PLoss: -0.0117 | Ent: 0.1608
+[ 45.6%] Iter 4457/9765 | SPS:  2138 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0224 | PLoss: -0.0079 | Ent: 0.1511
+[ 45.7%] Iter 4458/9765 | SPS:  2138 | Reward:  0.047 | Value:  0.729 | VLoss: 0.0177 | PLoss: -0.0036 | Ent: 0.1919
+[ 45.7%] Iter 4459/9765 | SPS:  2138 | Reward:  0.054 | Value:  0.662 | VLoss: 0.0282 | PLoss: -0.0089 | Ent: 0.1583
+[ 45.7%] Iter 4460/9765 | SPS:  2138 | Reward:  0.058 | Value:  0.701 | VLoss: 0.0211 | PLoss: -0.0111 | Ent: 0.1878
+[ 45.7%] Iter 4461/9765 | SPS:  2138 | Reward:  0.057 | Value:  0.643 | VLoss: 0.0273 | PLoss: -0.0027 | Ent: 0.1507
+[ 45.7%] Iter 4462/9765 | SPS:  2139 | Reward:  0.066 | Value:  0.607 | VLoss: 0.0372 | PLoss: -0.0114 | Ent: 0.1893
+[ 45.7%] Iter 4463/9765 | SPS:  2139 | Reward:  0.078 | Value:  0.700 | VLoss: 0.0380 | PLoss: -0.0101 | Ent: 0.1595
+[ 45.7%] Iter 4464/9765 | SPS:  2139 | Reward:  0.076 | Value:  0.724 | VLoss: 0.0363 | PLoss: -0.0195 | Ent: 0.1783
+[ 45.7%] Iter 4465/9765 | SPS:  2139 | Reward:  0.065 | Value:  0.718 | VLoss: 0.0250 | PLoss: -0.0055 | Ent: 0.1725
+[ 45.7%] Iter 4466/9765 | SPS:  2139 | Reward:  0.060 | Value:  0.711 | VLoss: 0.0271 | PLoss: -0.0096 | Ent: 0.1516
+[ 45.7%] Iter 4467/9765 | SPS:  2140 | Reward:  0.056 | Value:  0.683 | VLoss: 0.0226 | PLoss: -0.0077 | Ent: 0.2002
+[ 45.8%] Iter 4468/9765 | SPS:  2140 | Reward:  0.058 | Value:  0.609 | VLoss: 0.0248 | PLoss: -0.0063 | Ent: 0.1864
+[ 45.8%] Iter 4469/9765 | SPS:  2140 | Reward:  0.068 | Value:  0.740 | VLoss: 0.0272 | PLoss: -0.0089 | Ent: 0.1731
+[ 45.8%] Iter 4470/9765 | SPS:  2140 | Reward:  0.066 | Value:  0.744 | VLoss: 0.0269 | PLoss: 0.0031 | Ent: 0.1427
+[ 45.8%] Iter 4471/9765 | SPS:  2140 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0176 | PLoss: -0.0051 | Ent: 0.1462
+[ 45.8%] Iter 4472/9765 | SPS:  2141 | Reward:  0.062 | Value:  0.776 | VLoss: 0.0215 | PLoss: -0.0079 | Ent: 0.2218
+[ 45.8%] Iter 4473/9765 | SPS:  2141 | Reward:  0.070 | Value:  0.751 | VLoss: 0.0292 | PLoss: -0.0125 | Ent: 0.1590
+[ 45.8%] Iter 4474/9765 | SPS:  2141 | Reward:  0.072 | Value:  0.733 | VLoss: 0.0273 | PLoss: -0.0101 | Ent: 0.1584
+[ 45.8%] Iter 4475/9765 | SPS:  2141 | Reward:  0.062 | Value:  0.743 | VLoss: 0.0294 | PLoss: -0.0076 | Ent: 0.1434
+[ 45.8%] Iter 4476/9765 | SPS:  2141 | Reward:  0.067 | Value:  0.717 | VLoss: 0.0284 | PLoss: -0.0110 | Ent: 0.1729
+[ 45.8%] Iter 4477/9765 | SPS:  2142 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0262 | PLoss: -0.0089 | Ent: 0.1816
+[ 45.9%] Iter 4478/9765 | SPS:  2142 | Reward:  0.064 | Value:  0.661 | VLoss: 0.0301 | PLoss: -0.0053 | Ent: 0.1524
+[ 45.9%] Iter 4479/9765 | SPS:  2142 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0157 | PLoss: -0.0064 | Ent: 0.1748
+[ 45.9%] Iter 4480/9765 | SPS:  2142 | Reward:  0.062 | Value:  0.740 | VLoss: 0.0194 | PLoss: -0.0031 | Ent: 0.1622
+[ 45.9%] Iter 4481/9765 | SPS:  2142 | Reward:  0.068 | Value:  0.735 | VLoss: 0.0289 | PLoss: -0.0152 | Ent: 0.1774
+[ 45.9%] Iter 4482/9765 | SPS:  2143 | Reward:  0.068 | Value:  0.720 | VLoss: 0.0196 | PLoss: -0.0058 | Ent: 0.1956
+[ 45.9%] Iter 4483/9765 | SPS:  2143 | Reward:  0.051 | Value:  0.711 | VLoss: 0.0251 | PLoss: -0.0065 | Ent: 0.1852
+[ 45.9%] Iter 4484/9765 | SPS:  2143 | Reward:  0.060 | Value:  0.679 | VLoss: 0.0234 | PLoss: -0.0093 | Ent: 0.1702
+[ 45.9%] Iter 4485/9765 | SPS:  2143 | Reward:  0.052 | Value:  0.652 | VLoss: 0.0235 | PLoss: -0.0088 | Ent: 0.1604
+[ 45.9%] Iter 4486/9765 | SPS:  2143 | Reward:  0.063 | Value:  0.730 | VLoss: 0.0271 | PLoss: -0.0082 | Ent: 0.1539
+[ 45.9%] Iter 4487/9765 | SPS:  2144 | Reward:  0.068 | Value:  0.673 | VLoss: 0.0275 | PLoss: -0.0167 | Ent: 0.2036
+[ 46.0%] Iter 4488/9765 | SPS:  2144 | Reward:  0.062 | Value:  0.771 | VLoss: 0.0198 | PLoss: -0.0077 | Ent: 0.1705
+[ 46.0%] Iter 4489/9765 | SPS:  2144 | Reward:  0.064 | Value:  0.690 | VLoss: 0.0326 | PLoss: -0.0095 | Ent: 0.1824
+[ 46.0%] Iter 4490/9765 | SPS:  2144 | Reward:  0.066 | Value:  0.704 | VLoss: 0.0299 | PLoss: -0.0083 | Ent: 0.1932
+[ 46.0%] Iter 4491/9765 | SPS:  2144 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0201 | PLoss: -0.0060 | Ent: 0.1638
+[ 46.0%] Iter 4492/9765 | SPS:  2145 | Reward:  0.043 | Value:  0.652 | VLoss: 0.0235 | PLoss: -0.0105 | Ent: 0.1834
+[ 46.0%] Iter 4493/9765 | SPS:  2145 | Reward:  0.060 | Value:  0.725 | VLoss: 0.0204 | PLoss: -0.0013 | Ent: 0.1872
+[ 46.0%] Iter 4494/9765 | SPS:  2145 | Reward:  0.069 | Value:  0.705 | VLoss: 0.0206 | PLoss: -0.0048 | Ent: 0.1458
+[ 46.0%] Iter 4495/9765 | SPS:  2145 | Reward:  0.068 | Value:  0.710 | VLoss: 0.0263 | PLoss: -0.0074 | Ent: 0.1945
+[ 46.0%] Iter 4496/9765 | SPS:  2145 | Reward:  0.050 | Value:  0.676 | VLoss: 0.0212 | PLoss: -0.0037 | Ent: 0.2067
+[ 46.1%] Iter 4497/9765 | SPS:  2146 | Reward:  0.060 | Value:  0.625 | VLoss: 0.0242 | PLoss: -0.0071 | Ent: 0.2142
+[ 46.1%] Iter 4498/9765 | SPS:  2146 | Reward:  0.056 | Value:  0.626 | VLoss: 0.0280 | PLoss: -0.0090 | Ent: 0.1923
+[ 46.1%] Iter 4499/9765 | SPS:  2146 | Reward:  0.057 | Value:  0.693 | VLoss: 0.0253 | PLoss: -0.0136 | Ent: 0.1974
+[ 46.1%] Iter 4500/9765 | SPS:  2146 | Reward:  0.052 | Value:  0.756 | VLoss: 0.0146 | PLoss: -0.0048 | Ent: 0.1627
+[ 46.1%] Iter 4501/9765 | SPS:  2146 | Reward:  0.064 | Value:  0.642 | VLoss: 0.0279 | PLoss: -0.0060 | Ent: 0.1696
+[ 46.1%] Iter 4502/9765 | SPS:  2147 | Reward:  0.058 | Value:  0.665 | VLoss: 0.0334 | PLoss: -0.0194 | Ent: 0.2198
+[ 46.1%] Iter 4503/9765 | SPS:  2147 | Reward:  0.064 | Value:  0.682 | VLoss: 0.0229 | PLoss: -0.0063 | Ent: 0.2143
+[ 46.1%] Iter 4504/9765 | SPS:  2147 | Reward:  0.048 | Value:  0.638 | VLoss: 0.0261 | PLoss: -0.0038 | Ent: 0.2578
+[ 46.1%] Iter 4505/9765 | SPS:  2147 | Reward:  0.067 | Value:  0.724 | VLoss: 0.0212 | PLoss: -0.0046 | Ent: 0.1687
+[ 46.1%] Iter 4506/9765 | SPS:  2147 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0262 | PLoss: -0.0115 | Ent: 0.1550
+[ 46.2%] Iter 4507/9765 | SPS:  2148 | Reward:  0.063 | Value:  0.667 | VLoss: 0.0314 | PLoss: -0.0023 | Ent: 0.1920
+[ 46.2%] Iter 4508/9765 | SPS:  2148 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0313 | PLoss: -0.0057 | Ent: 0.1816
+[ 46.2%] Iter 4509/9765 | SPS:  2148 | Reward:  0.061 | Value:  0.666 | VLoss: 0.0404 | PLoss: -0.0067 | Ent: 0.2145
+[ 46.2%] Iter 4510/9765 | SPS:  2148 | Reward:  0.065 | Value:  0.650 | VLoss: 0.0289 | PLoss: -0.0093 | Ent: 0.1780
+[ 46.2%] Iter 4511/9765 | SPS:  2148 | Reward:  0.071 | Value:  0.685 | VLoss: 0.0231 | PLoss: -0.0068 | Ent: 0.2076
+[ 46.2%] Iter 4512/9765 | SPS:  2149 | Reward:  0.065 | Value:  0.713 | VLoss: 0.0196 | PLoss: -0.0041 | Ent: 0.2189
+[ 46.2%] Iter 4513/9765 | SPS:  2149 | Reward:  0.066 | Value:  0.727 | VLoss: 0.0241 | PLoss: -0.0057 | Ent: 0.2122
+[ 46.2%] Iter 4514/9765 | SPS:  2149 | Reward:  0.062 | Value:  0.768 | VLoss: 0.0223 | PLoss: -0.0104 | Ent: 0.2163
+[ 46.2%] Iter 4515/9765 | SPS:  2149 | Reward:  0.055 | Value:  0.739 | VLoss: 0.0204 | PLoss: -0.0128 | Ent: 0.2061
+[ 46.2%] Iter 4516/9765 | SPS:  2149 | Reward:  0.057 | Value:  0.620 | VLoss: 0.0266 | PLoss: -0.0098 | Ent: 0.1820
+[ 46.3%] Iter 4517/9765 | SPS:  2150 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0190 | PLoss: -0.0105 | Ent: 0.1719
+[ 46.3%] Iter 4518/9765 | SPS:  2150 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0170 | PLoss: -0.0103 | Ent: 0.1758
+[ 46.3%] Iter 4519/9765 | SPS:  2150 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0243 | PLoss: -0.0096 | Ent: 0.1854
+[ 46.3%] Iter 4520/9765 | SPS:  2150 | Reward:  0.072 | Value:  0.744 | VLoss: 0.0169 | PLoss: -0.0062 | Ent: 0.1968
+[ 46.3%] Iter 4521/9765 | SPS:  2150 | Reward:  0.062 | Value:  0.785 | VLoss: 0.0284 | PLoss: -0.0062 | Ent: 0.1488
+[ 46.3%] Iter 4522/9765 | SPS:  2151 | Reward:  0.069 | Value:  0.683 | VLoss: 0.0276 | PLoss: -0.0129 | Ent: 0.1833
+[ 46.3%] Iter 4523/9765 | SPS:  2151 | Reward:  0.072 | Value:  0.733 | VLoss: 0.0228 | PLoss: -0.0025 | Ent: 0.1622
+[ 46.3%] Iter 4524/9765 | SPS:  2151 | Reward:  0.058 | Value:  0.721 | VLoss: 0.0156 | PLoss: -0.0160 | Ent: 0.2270
+[ 46.3%] Iter 4525/9765 | SPS:  2151 | Reward:  0.068 | Value:  0.627 | VLoss: 0.0257 | PLoss: -0.0084 | Ent: 0.1883
+[ 46.3%] Iter 4526/9765 | SPS:  2151 | Reward:  0.072 | Value:  0.743 | VLoss: 0.0215 | PLoss: -0.0072 | Ent: 0.1922
+[ 46.4%] Iter 4527/9765 | SPS:  2152 | Reward:  0.071 | Value:  0.797 | VLoss: 0.0230 | PLoss: -0.0043 | Ent: 0.1672
+[ 46.4%] Iter 4528/9765 | SPS:  2152 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0229 | PLoss: -0.0115 | Ent: 0.1728
+[ 46.4%] Iter 4529/9765 | SPS:  2152 | Reward:  0.040 | Value:  0.653 | VLoss: 0.0181 | PLoss: -0.0045 | Ent: 0.1524
+[ 46.4%] Iter 4530/9765 | SPS:  2152 | Reward:  0.062 | Value:  0.650 | VLoss: 0.0201 | PLoss: -0.0023 | Ent: 0.1633
+[ 46.4%] Iter 4531/9765 | SPS:  2152 | Reward:  0.067 | Value:  0.656 | VLoss: 0.0236 | PLoss: -0.0035 | Ent: 0.1737
+[ 46.4%] Iter 4532/9765 | SPS:  2153 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0251 | PLoss: -0.0109 | Ent: 0.2053
+[ 46.4%] Iter 4533/9765 | SPS:  2153 | Reward:  0.060 | Value:  0.814 | VLoss: 0.0202 | PLoss: -0.0084 | Ent: 0.1530
+[ 46.4%] Iter 4534/9765 | SPS:  2153 | Reward:  0.069 | Value:  0.716 | VLoss: 0.0217 | PLoss: -0.0090 | Ent: 0.2110
+[ 46.4%] Iter 4535/9765 | SPS:  2153 | Reward:  0.065 | Value:  0.678 | VLoss: 0.0303 | PLoss: -0.0072 | Ent: 0.2013
+[ 46.5%] Iter 4536/9765 | SPS:  2153 | Reward:  0.065 | Value:  0.726 | VLoss: 0.0265 | PLoss: 0.0015 | Ent: 0.1827
+[ 46.5%] Iter 4537/9765 | SPS:  2154 | Reward:  0.068 | Value:  0.737 | VLoss: 0.0214 | PLoss: -0.0027 | Ent: 0.1476
+[ 46.5%] Iter 4538/9765 | SPS:  2154 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0283 | PLoss: -0.0149 | Ent: 0.1813
+[ 46.5%] Iter 4539/9765 | SPS:  2154 | Reward:  0.073 | Value:  0.701 | VLoss: 0.0216 | PLoss: -0.0097 | Ent: 0.1828
+[ 46.5%] Iter 4540/9765 | SPS:  2154 | Reward:  0.054 | Value:  0.778 | VLoss: 0.0190 | PLoss: -0.0104 | Ent: 0.2145
+[ 46.5%] Iter 4541/9765 | SPS:  2154 | Reward:  0.061 | Value:  0.677 | VLoss: 0.0262 | PLoss: -0.0055 | Ent: 0.1838
+[ 46.5%] Iter 4542/9765 | SPS:  2155 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0209 | PLoss: -0.0038 | Ent: 0.1954
+[ 46.5%] Iter 4543/9765 | SPS:  2155 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0340 | PLoss: -0.0047 | Ent: 0.1762
+[ 46.5%] Iter 4544/9765 | SPS:  2155 | Reward:  0.068 | Value:  0.722 | VLoss: 0.0288 | PLoss: -0.0051 | Ent: 0.1980
+[ 46.5%] Iter 4545/9765 | SPS:  2155 | Reward:  0.073 | Value:  0.703 | VLoss: 0.0265 | PLoss: 0.0012 | Ent: 0.2013
+[ 46.6%] Iter 4546/9765 | SPS:  2155 | Reward:  0.071 | Value:  0.700 | VLoss: 0.0184 | PLoss: 0.0039 | Ent: 0.1630
+[ 46.6%] Iter 4547/9765 | SPS:  2156 | Reward:  0.062 | Value:  0.762 | VLoss: 0.0208 | PLoss: -0.0099 | Ent: 0.1779
+[ 46.6%] Iter 4548/9765 | SPS:  2156 | Reward:  0.066 | Value:  0.711 | VLoss: 0.0247 | PLoss: -0.0105 | Ent: 0.2243
+[ 46.6%] Iter 4549/9765 | SPS:  2156 | Reward:  0.070 | Value:  0.719 | VLoss: 0.0322 | PLoss: -0.0103 | Ent: 0.1867
+[ 46.6%] Iter 4550/9765 | SPS:  2156 | Reward:  0.054 | Value:  0.684 | VLoss: 0.0198 | PLoss: -0.0043 | Ent: 0.1643
+[ 46.6%] Iter 4551/9765 | SPS:  2156 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0206 | PLoss: -0.0035 | Ent: 0.1791
+[ 46.6%] Iter 4552/9765 | SPS:  2157 | Reward:  0.065 | Value:  0.684 | VLoss: 0.0206 | PLoss: -0.0178 | Ent: 0.2173
+[ 46.6%] Iter 4553/9765 | SPS:  2157 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0255 | PLoss: -0.0049 | Ent: 0.1739
+[ 46.6%] Iter 4554/9765 | SPS:  2157 | Reward:  0.053 | Value:  0.708 | VLoss: 0.0238 | PLoss: -0.0210 | Ent: 0.1683
+[ 46.6%] Iter 4555/9765 | SPS:  2157 | Reward:  0.053 | Value:  0.672 | VLoss: 0.0241 | PLoss: -0.0067 | Ent: 0.2196
+[ 46.7%] Iter 4556/9765 | SPS:  2157 | Reward:  0.060 | Value:  0.691 | VLoss: 0.0226 | PLoss: -0.0089 | Ent: 0.1519
+[ 46.7%] Iter 4557/9765 | SPS:  2158 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0222 | PLoss: -0.0023 | Ent: 0.1894
+[ 46.7%] Iter 4558/9765 | SPS:  2158 | Reward:  0.055 | Value:  0.743 | VLoss: 0.0164 | PLoss: -0.0063 | Ent: 0.1165
+[ 46.7%] Iter 4559/9765 | SPS:  2158 | Reward:  0.076 | Value:  0.711 | VLoss: 0.0241 | PLoss: -0.0017 | Ent: 0.1705
+[ 46.7%] Iter 4560/9765 | SPS:  2158 | Reward:  0.059 | Value:  0.736 | VLoss: 0.0235 | PLoss: -0.0110 | Ent: 0.2089
+[ 46.7%] Iter 4561/9765 | SPS:  2158 | Reward:  0.061 | Value:  0.737 | VLoss: 0.0216 | PLoss: -0.0042 | Ent: 0.1716
+[ 46.7%] Iter 4562/9765 | SPS:  2159 | Reward:  0.083 | Value:  0.759 | VLoss: 0.0320 | PLoss: -0.0073 | Ent: 0.1752
+[ 46.7%] Iter 4563/9765 | SPS:  2159 | Reward:  0.071 | Value:  0.735 | VLoss: 0.0299 | PLoss: -0.0025 | Ent: 0.2067
+[ 46.7%] Iter 4564/9765 | SPS:  2159 | Reward:  0.067 | Value:  0.653 | VLoss: 0.0272 | PLoss: -0.0043 | Ent: 0.1935
+[ 46.7%] Iter 4565/9765 | SPS:  2159 | Reward:  0.062 | Value:  0.797 | VLoss: 0.0183 | PLoss: -0.0066 | Ent: 0.1795
+[ 46.8%] Iter 4566/9765 | SPS:  2159 | Reward:  0.064 | Value:  0.747 | VLoss: 0.0241 | PLoss: -0.0085 | Ent: 0.1840
+[ 46.8%] Iter 4567/9765 | SPS:  2159 | Reward:  0.068 | Value:  0.700 | VLoss: 0.0260 | PLoss: -0.0094 | Ent: 0.1708
+[ 46.8%] Iter 4568/9765 | SPS:  2160 | Reward:  0.050 | Value:  0.737 | VLoss: 0.0226 | PLoss: -0.0028 | Ent: 0.1915
+[ 46.8%] Iter 4569/9765 | SPS:  2160 | Reward:  0.065 | Value:  0.679 | VLoss: 0.0208 | PLoss: -0.0024 | Ent: 0.1624
+[ 46.8%] Iter 4570/9765 | SPS:  2160 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0270 | PLoss: -0.0079 | Ent: 0.1942
+[ 46.8%] Iter 4571/9765 | SPS:  2160 | Reward:  0.079 | Value:  0.666 | VLoss: 0.0353 | PLoss: -0.0046 | Ent: 0.1953
+[ 46.8%] Iter 4572/9765 | SPS:  2160 | Reward:  0.062 | Value:  0.755 | VLoss: 0.0165 | PLoss: -0.0071 | Ent: 0.2284
+[ 46.8%] Iter 4573/9765 | SPS:  2161 | Reward:  0.080 | Value:  0.714 | VLoss: 0.0193 | PLoss: -0.0034 | Ent: 0.1986
+[ 46.8%] Iter 4574/9765 | SPS:  2161 | Reward:  0.056 | Value:  0.734 | VLoss: 0.0229 | PLoss: -0.0135 | Ent: 0.1727
+[ 46.9%] Iter 4575/9765 | SPS:  2161 | Reward:  0.058 | Value:  0.678 | VLoss: 0.0214 | PLoss: -0.0084 | Ent: 0.2159
+[ 46.9%] Iter 4576/9765 | SPS:  2161 | Reward:  0.057 | Value:  0.703 | VLoss: 0.0181 | PLoss: -0.0144 | Ent: 0.1989
+[ 46.9%] Iter 4577/9765 | SPS:  2161 | Reward:  0.069 | Value:  0.630 | VLoss: 0.0248 | PLoss: -0.0099 | Ent: 0.1902
+[ 46.9%] Iter 4578/9765 | SPS:  2162 | Reward:  0.068 | Value:  0.731 | VLoss: 0.0223 | PLoss: -0.0090 | Ent: 0.1538
+[ 46.9%] Iter 4579/9765 | SPS:  2162 | Reward:  0.067 | Value:  0.727 | VLoss: 0.0306 | PLoss: -0.0046 | Ent: 0.1534
+[ 46.9%] Iter 4580/9765 | SPS:  2162 | Reward:  0.066 | Value:  0.702 | VLoss: 0.0259 | PLoss: -0.0107 | Ent: 0.1671
+[ 46.9%] Iter 4581/9765 | SPS:  2162 | Reward:  0.056 | Value:  0.665 | VLoss: 0.0318 | PLoss: -0.0134 | Ent: 0.1935
+[ 46.9%] Iter 4582/9765 | SPS:  2162 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0233 | PLoss: 0.0009 | Ent: 0.1737
+[ 46.9%] Iter 4583/9765 | SPS:  2162 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0257 | PLoss: -0.0074 | Ent: 0.2148
+[ 46.9%] Iter 4584/9765 | SPS:  2163 | Reward:  0.067 | Value:  0.682 | VLoss: 0.0252 | PLoss: -0.0013 | Ent: 0.1903
+[ 47.0%] Iter 4585/9765 | SPS:  2163 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0359 | PLoss: -0.0146 | Ent: 0.2148
+[ 47.0%] Iter 4586/9765 | SPS:  2163 | Reward:  0.056 | Value:  0.653 | VLoss: 0.0266 | PLoss: -0.0063 | Ent: 0.1327
+[ 47.0%] Iter 4587/9765 | SPS:  2163 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0243 | PLoss: -0.0077 | Ent: 0.1720
+[ 47.0%] Iter 4588/9765 | SPS:  2163 | Reward:  0.055 | Value:  0.656 | VLoss: 0.0299 | PLoss: -0.0084 | Ent: 0.1644
+[ 47.0%] Iter 4589/9765 | SPS:  2164 | Reward:  0.072 | Value:  0.677 | VLoss: 0.0232 | PLoss: -0.0044 | Ent: 0.1961
+[ 47.0%] Iter 4590/9765 | SPS:  2164 | Reward:  0.058 | Value:  0.740 | VLoss: 0.0274 | PLoss: -0.0112 | Ent: 0.2299
+[ 47.0%] Iter 4591/9765 | SPS:  2164 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0236 | PLoss: -0.0036 | Ent: 0.1982
+[ 47.0%] Iter 4592/9765 | SPS:  2164 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0253 | PLoss: -0.0074 | Ent: 0.1913
+[ 47.0%] Iter 4593/9765 | SPS:  2164 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0289 | PLoss: -0.0104 | Ent: 0.1896
+[ 47.0%] Iter 4594/9765 | SPS:  2165 | Reward:  0.056 | Value:  0.597 | VLoss: 0.0290 | PLoss: -0.0093 | Ent: 0.1847
+[ 47.1%] Iter 4595/9765 | SPS:  2165 | Reward:  0.078 | Value:  0.716 | VLoss: 0.0241 | PLoss: -0.0020 | Ent: 0.1739
+[ 47.1%] Iter 4596/9765 | SPS:  2165 | Reward:  0.056 | Value:  0.812 | VLoss: 0.0223 | PLoss: -0.0094 | Ent: 0.1902
+[ 47.1%] Iter 4597/9765 | SPS:  2165 | Reward:  0.062 | Value:  0.765 | VLoss: 0.0262 | PLoss: -0.0123 | Ent: 0.1930
+[ 47.1%] Iter 4598/9765 | SPS:  2165 | Reward:  0.065 | Value:  0.655 | VLoss: 0.0250 | PLoss: -0.0086 | Ent: 0.1362
+[ 47.1%] Iter 4599/9765 | SPS:  2166 | Reward:  0.073 | Value:  0.732 | VLoss: 0.0241 | PLoss: -0.0162 | Ent: 0.1944
+[ 47.1%] Iter 4600/9765 | SPS:  2166 | Reward:  0.064 | Value:  0.752 | VLoss: 0.0210 | PLoss: -0.0051 | Ent: 0.1337
+[ 47.1%] Iter 4601/9765 | SPS:  2166 | Reward:  0.060 | Value:  0.648 | VLoss: 0.0277 | PLoss: -0.0023 | Ent: 0.1985
+[ 47.1%] Iter 4602/9765 | SPS:  2166 | Reward:  0.055 | Value:  0.633 | VLoss: 0.0185 | PLoss: -0.0062 | Ent: 0.1800
+[ 47.1%] Iter 4603/9765 | SPS:  2166 | Reward:  0.073 | Value:  0.746 | VLoss: 0.0249 | PLoss: -0.0011 | Ent: 0.2182
+[ 47.1%] Iter 4604/9765 | SPS:  2166 | Reward:  0.074 | Value:  0.657 | VLoss: 0.0299 | PLoss: -0.0163 | Ent: 0.2118
+[ 47.2%] Iter 4605/9765 | SPS:  2166 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0391 | PLoss: -0.0109 | Ent: 0.1811
+[ 47.2%] Iter 4606/9765 | SPS:  2167 | Reward:  0.064 | Value:  0.684 | VLoss: 0.0295 | PLoss: -0.0052 | Ent: 0.2174
+[ 47.2%] Iter 4607/9765 | SPS:  2167 | Reward:  0.068 | Value:  0.617 | VLoss: 0.0276 | PLoss: -0.0017 | Ent: 0.1997
+[ 47.2%] Iter 4608/9765 | SPS:  2167 | Reward:  0.057 | Value:  0.714 | VLoss: 0.0248 | PLoss: -0.0065 | Ent: 0.1725
+[ 47.2%] Iter 4609/9765 | SPS:  2167 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0245 | PLoss: -0.0071 | Ent: 0.1992
+[ 47.2%] Iter 4610/9765 | SPS:  2167 | Reward:  0.073 | Value:  0.721 | VLoss: 0.0312 | PLoss: -0.0034 | Ent: 0.1423
+[ 47.2%] Iter 4611/9765 | SPS:  2168 | Reward:  0.077 | Value:  0.705 | VLoss: 0.0265 | PLoss: -0.0016 | Ent: 0.2256
+[ 47.2%] Iter 4612/9765 | SPS:  2168 | Reward:  0.072 | Value:  0.772 | VLoss: 0.0303 | PLoss: -0.0029 | Ent: 0.1616
+[ 47.2%] Iter 4613/9765 | SPS:  2168 | Reward:  0.062 | Value:  0.780 | VLoss: 0.0266 | PLoss: -0.0111 | Ent: 0.1785
+[ 47.3%] Iter 4614/9765 | SPS:  2168 | Reward:  0.069 | Value:  0.678 | VLoss: 0.0282 | PLoss: -0.0048 | Ent: 0.1690
+[ 47.3%] Iter 4615/9765 | SPS:  2168 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0253 | PLoss: -0.0081 | Ent: 0.1987
+[ 47.3%] Iter 4616/9765 | SPS:  2169 | Reward:  0.068 | Value:  0.680 | VLoss: 0.0254 | PLoss: -0.0040 | Ent: 0.2006
+[ 47.3%] Iter 4617/9765 | SPS:  2169 | Reward:  0.052 | Value:  0.706 | VLoss: 0.0244 | PLoss: -0.0066 | Ent: 0.1676
+[ 47.3%] Iter 4618/9765 | SPS:  2169 | Reward:  0.056 | Value:  0.685 | VLoss: 0.0268 | PLoss: -0.0025 | Ent: 0.1754
+[ 47.3%] Iter 4619/9765 | SPS:  2169 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0304 | PLoss: -0.0135 | Ent: 0.1845
+[ 47.3%] Iter 4620/9765 | SPS:  2169 | Reward:  0.078 | Value:  0.675 | VLoss: 0.0284 | PLoss: -0.0147 | Ent: 0.1776
+[ 47.3%] Iter 4621/9765 | SPS:  2170 | Reward:  0.079 | Value:  0.732 | VLoss: 0.0335 | PLoss: -0.0122 | Ent: 0.1599
+[ 47.3%] Iter 4622/9765 | SPS:  2170 | Reward:  0.054 | Value:  0.759 | VLoss: 0.0210 | PLoss: -0.0100 | Ent: 0.2116
+[ 47.3%] Iter 4623/9765 | SPS:  2170 | Reward:  0.047 | Value:  0.656 | VLoss: 0.0249 | PLoss: -0.0122 | Ent: 0.2000
+[ 47.4%] Iter 4624/9765 | SPS:  2170 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0169 | PLoss: -0.0030 | Ent: 0.1990
+[ 47.4%] Iter 4625/9765 | SPS:  2170 | Reward:  0.060 | Value:  0.743 | VLoss: 0.0188 | PLoss: -0.0088 | Ent: 0.2147
+[ 47.4%] Iter 4626/9765 | SPS:  2171 | Reward:  0.070 | Value:  0.760 | VLoss: 0.0222 | PLoss: -0.0122 | Ent: 0.1450
+[ 47.4%] Iter 4627/9765 | SPS:  2171 | Reward:  0.070 | Value:  0.776 | VLoss: 0.0207 | PLoss: -0.0073 | Ent: 0.2306
+[ 47.4%] Iter 4628/9765 | SPS:  2171 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0244 | PLoss: -0.0087 | Ent: 0.1721
+[ 47.4%] Iter 4629/9765 | SPS:  2171 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0191 | PLoss: -0.0077 | Ent: 0.1917
+[ 47.4%] Iter 4630/9765 | SPS:  2171 | Reward:  0.075 | Value:  0.741 | VLoss: 0.0240 | PLoss: -0.0077 | Ent: 0.1492
+[ 47.4%] Iter 4631/9765 | SPS:  2171 | Reward:  0.064 | Value:  0.704 | VLoss: 0.0300 | PLoss: -0.0007 | Ent: 0.1580
+[ 47.4%] Iter 4632/9765 | SPS:  2172 | Reward:  0.062 | Value:  0.666 | VLoss: 0.0217 | PLoss: -0.0142 | Ent: 0.1952
+[ 47.4%] Iter 4633/9765 | SPS:  2172 | Reward:  0.064 | Value:  0.739 | VLoss: 0.0214 | PLoss: -0.0057 | Ent: 0.1538
+[ 47.5%] Iter 4634/9765 | SPS:  2172 | Reward:  0.071 | Value:  0.786 | VLoss: 0.0216 | PLoss: -0.0068 | Ent: 0.1795
+[ 47.5%] Iter 4635/9765 | SPS:  2172 | Reward:  0.074 | Value:  0.721 | VLoss: 0.0277 | PLoss: -0.0078 | Ent: 0.2191
+[ 47.5%] Iter 4636/9765 | SPS:  2172 | Reward:  0.067 | Value:  0.747 | VLoss: 0.0268 | PLoss: -0.0072 | Ent: 0.1562
+[ 47.5%] Iter 4637/9765 | SPS:  2173 | Reward:  0.076 | Value:  0.711 | VLoss: 0.0255 | PLoss: -0.0058 | Ent: 0.1729
+[ 47.5%] Iter 4638/9765 | SPS:  2173 | Reward:  0.061 | Value:  0.756 | VLoss: 0.0196 | PLoss: -0.0112 | Ent: 0.1823
+[ 47.5%] Iter 4639/9765 | SPS:  2173 | Reward:  0.069 | Value:  0.741 | VLoss: 0.0268 | PLoss: -0.0047 | Ent: 0.1772
+[ 47.5%] Iter 4640/9765 | SPS:  2173 | Reward:  0.078 | Value:  0.701 | VLoss: 0.0345 | PLoss: -0.0091 | Ent: 0.1885
+[ 47.5%] Iter 4641/9765 | SPS:  2173 | Reward:  0.073 | Value:  0.697 | VLoss: 0.0195 | PLoss: -0.0091 | Ent: 0.1789
+[ 47.5%] Iter 4642/9765 | SPS:  2174 | Reward:  0.067 | Value:  0.742 | VLoss: 0.0215 | PLoss: -0.0061 | Ent: 0.1990
+[ 47.5%] Iter 4643/9765 | SPS:  2174 | Reward:  0.062 | Value:  0.752 | VLoss: 0.0152 | PLoss: -0.0073 | Ent: 0.2199
+[ 47.6%] Iter 4644/9765 | SPS:  2174 | Reward:  0.062 | Value:  0.749 | VLoss: 0.0297 | PLoss: -0.0046 | Ent: 0.1600
+[ 47.6%] Iter 4645/9765 | SPS:  2174 | Reward:  0.070 | Value:  0.728 | VLoss: 0.0325 | PLoss: -0.0108 | Ent: 0.2292
+[ 47.6%] Iter 4646/9765 | SPS:  2174 | Reward:  0.064 | Value:  0.685 | VLoss: 0.0257 | PLoss: -0.0118 | Ent: 0.1758
+[ 47.6%] Iter 4647/9765 | SPS:  2174 | Reward:  0.071 | Value:  0.729 | VLoss: 0.0253 | PLoss: -0.0073 | Ent: 0.1759
+[ 47.6%] Iter 4648/9765 | SPS:  2175 | Reward:  0.057 | Value:  0.724 | VLoss: 0.0192 | PLoss: -0.0093 | Ent: 0.1675
+[ 47.6%] Iter 4649/9765 | SPS:  2175 | Reward:  0.068 | Value:  0.736 | VLoss: 0.0297 | PLoss: -0.0092 | Ent: 0.2015
+[ 47.6%] Iter 4650/9765 | SPS:  2175 | Reward:  0.062 | Value:  0.714 | VLoss: 0.0283 | PLoss: -0.0058 | Ent: 0.1634
+[ 47.6%] Iter 4651/9765 | SPS:  2175 | Reward:  0.067 | Value:  0.677 | VLoss: 0.0278 | PLoss: -0.0018 | Ent: 0.1793
+[ 47.6%] Iter 4652/9765 | SPS:  2175 | Reward:  0.058 | Value:  0.633 | VLoss: 0.0239 | PLoss: -0.0042 | Ent: 0.1725
+[ 47.6%] Iter 4653/9765 | SPS:  2176 | Reward:  0.065 | Value:  0.734 | VLoss: 0.0286 | PLoss: -0.0012 | Ent: 0.1577
+[ 47.7%] Iter 4654/9765 | SPS:  2176 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0265 | PLoss: -0.0087 | Ent: 0.2036
+[ 47.7%] Iter 4655/9765 | SPS:  2176 | Reward:  0.070 | Value:  0.699 | VLoss: 0.0305 | PLoss: -0.0042 | Ent: 0.1731
+[ 47.7%] Iter 4656/9765 | SPS:  2176 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0198 | PLoss: -0.0064 | Ent: 0.1382
+[ 47.7%] Iter 4657/9765 | SPS:  2176 | Reward:  0.068 | Value:  0.753 | VLoss: 0.0209 | PLoss: -0.0101 | Ent: 0.1622
+[ 47.7%] Iter 4658/9765 | SPS:  2177 | Reward:  0.061 | Value:  0.763 | VLoss: 0.0239 | PLoss: -0.0085 | Ent: 0.1918
+[ 47.7%] Iter 4659/9765 | SPS:  2177 | Reward:  0.065 | Value:  0.709 | VLoss: 0.0355 | PLoss: -0.0069 | Ent: 0.2036
+[ 47.7%] Iter 4660/9765 | SPS:  2177 | Reward:  0.058 | Value:  0.729 | VLoss: 0.0170 | PLoss: -0.0082 | Ent: 0.1826
+[ 47.7%] Iter 4661/9765 | SPS:  2177 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0321 | PLoss: -0.0096 | Ent: 0.1838
+[ 47.7%] Iter 4662/9765 | SPS:  2177 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0252 | PLoss: -0.0096 | Ent: 0.1895
+[ 47.8%] Iter 4663/9765 | SPS:  2178 | Reward:  0.071 | Value:  0.689 | VLoss: 0.0271 | PLoss: -0.0049 | Ent: 0.1498
+[ 47.8%] Iter 4664/9765 | SPS:  2178 | Reward:  0.057 | Value:  0.761 | VLoss: 0.0146 | PLoss: -0.0060 | Ent: 0.1599
+[ 47.8%] Iter 4665/9765 | SPS:  2178 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0350 | PLoss: -0.0107 | Ent: 0.1747
+[ 47.8%] Iter 4666/9765 | SPS:  2178 | Reward:  0.061 | Value:  0.717 | VLoss: 0.0248 | PLoss: -0.0154 | Ent: 0.1866
+[ 47.8%] Iter 4667/9765 | SPS:  2178 | Reward:  0.049 | Value:  0.681 | VLoss: 0.0279 | PLoss: -0.0156 | Ent: 0.1903
+[ 47.8%] Iter 4668/9765 | SPS:  2178 | Reward:  0.045 | Value:  0.689 | VLoss: 0.0212 | PLoss: 0.0010 | Ent: 0.2182
+[ 47.8%] Iter 4669/9765 | SPS:  2179 | Reward:  0.067 | Value:  0.625 | VLoss: 0.0236 | PLoss: -0.0108 | Ent: 0.2123
+[ 47.8%] Iter 4670/9765 | SPS:  2179 | Reward:  0.068 | Value:  0.684 | VLoss: 0.0356 | PLoss: -0.0161 | Ent: 0.2150
+[ 47.8%] Iter 4671/9765 | SPS:  2179 | Reward:  0.066 | Value:  0.762 | VLoss: 0.0248 | PLoss: -0.0075 | Ent: 0.1626
+[ 47.8%] Iter 4672/9765 | SPS:  2179 | Reward:  0.067 | Value:  0.747 | VLoss: 0.0242 | PLoss: -0.0066 | Ent: 0.1756
+[ 47.9%] Iter 4673/9765 | SPS:  2179 | Reward:  0.063 | Value:  0.783 | VLoss: 0.0229 | PLoss: -0.0087 | Ent: 0.1475
+[ 47.9%] Iter 4674/9765 | SPS:  2180 | Reward:  0.053 | Value:  0.688 | VLoss: 0.0297 | PLoss: -0.0025 | Ent: 0.1919
+[ 47.9%] Iter 4675/9765 | SPS:  2180 | Reward:  0.054 | Value:  0.722 | VLoss: 0.0158 | PLoss: -0.0079 | Ent: 0.1435
+[ 47.9%] Iter 4676/9765 | SPS:  2180 | Reward:  0.051 | Value:  0.752 | VLoss: 0.0266 | PLoss: -0.0140 | Ent: 0.2135
+[ 47.9%] Iter 4677/9765 | SPS:  2180 | Reward:  0.051 | Value:  0.683 | VLoss: 0.0266 | PLoss: -0.0085 | Ent: 0.2088
+[ 47.9%] Iter 4678/9765 | SPS:  2180 | Reward:  0.063 | Value:  0.693 | VLoss: 0.0248 | PLoss: -0.0087 | Ent: 0.1794
+[ 47.9%] Iter 4679/9765 | SPS:  2181 | Reward:  0.056 | Value:  0.696 | VLoss: 0.0251 | PLoss: -0.0061 | Ent: 0.2009
+[ 47.9%] Iter 4680/9765 | SPS:  2181 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0203 | PLoss: -0.0027 | Ent: 0.1868
+[ 47.9%] Iter 4681/9765 | SPS:  2181 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0262 | PLoss: -0.0161 | Ent: 0.1886
+[ 47.9%] Iter 4682/9765 | SPS:  2181 | Reward:  0.059 | Value:  0.703 | VLoss: 0.0172 | PLoss: -0.0070 | Ent: 0.1774
+[ 48.0%] Iter 4683/9765 | SPS:  2181 | Reward:  0.057 | Value:  0.743 | VLoss: 0.0186 | PLoss: -0.0034 | Ent: 0.1688
+[ 48.0%] Iter 4684/9765 | SPS:  2182 | Reward:  0.077 | Value:  0.755 | VLoss: 0.0266 | PLoss: -0.0082 | Ent: 0.1618
+[ 48.0%] Iter 4685/9765 | SPS:  2182 | Reward:  0.073 | Value:  0.775 | VLoss: 0.0234 | PLoss: -0.0129 | Ent: 0.2212
+[ 48.0%] Iter 4686/9765 | SPS:  2182 | Reward:  0.070 | Value:  0.704 | VLoss: 0.0299 | PLoss: -0.0130 | Ent: 0.2017
+[ 48.0%] Iter 4687/9765 | SPS:  2182 | Reward:  0.055 | Value:  0.727 | VLoss: 0.0253 | PLoss: -0.0040 | Ent: 0.1773
+[ 48.0%] Iter 4688/9765 | SPS:  2182 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0244 | PLoss: -0.0086 | Ent: 0.1819
+[ 48.0%] Iter 4689/9765 | SPS:  2183 | Reward:  0.063 | Value:  0.689 | VLoss: 0.0249 | PLoss: -0.0087 | Ent: 0.1603
+[ 48.0%] Iter 4690/9765 | SPS:  2183 | Reward:  0.044 | Value:  0.705 | VLoss: 0.0261 | PLoss: -0.0049 | Ent: 0.2255
+[ 48.0%] Iter 4691/9765 | SPS:  2183 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0251 | PLoss: -0.0110 | Ent: 0.1795
+[ 48.0%] Iter 4692/9765 | SPS:  2183 | Reward:  0.053 | Value:  0.668 | VLoss: 0.0283 | PLoss: -0.0196 | Ent: 0.2331
+[ 48.1%] Iter 4693/9765 | SPS:  2183 | Reward:  0.067 | Value:  0.664 | VLoss: 0.0292 | PLoss: -0.0075 | Ent: 0.1563
+[ 48.1%] Iter 4694/9765 | SPS:  2183 | Reward:  0.067 | Value:  0.730 | VLoss: 0.0297 | PLoss: -0.0106 | Ent: 0.1916
+[ 48.1%] Iter 4695/9765 | SPS:  2184 | Reward:  0.059 | Value:  0.752 | VLoss: 0.0226 | PLoss: -0.0111 | Ent: 0.2226
+[ 48.1%] Iter 4696/9765 | SPS:  2184 | Reward:  0.053 | Value:  0.701 | VLoss: 0.0239 | PLoss: -0.0017 | Ent: 0.1754
+[ 48.1%] Iter 4697/9765 | SPS:  2184 | Reward:  0.052 | Value:  0.674 | VLoss: 0.0190 | PLoss: -0.0076 | Ent: 0.1839
+[ 48.1%] Iter 4698/9765 | SPS:  2184 | Reward:  0.050 | Value:  0.655 | VLoss: 0.0234 | PLoss: -0.0151 | Ent: 0.2292
+[ 48.1%] Iter 4699/9765 | SPS:  2184 | Reward:  0.073 | Value:  0.679 | VLoss: 0.0314 | PLoss: -0.0064 | Ent: 0.1562
+[ 48.1%] Iter 4700/9765 | SPS:  2185 | Reward:  0.058 | Value:  0.782 | VLoss: 0.0212 | PLoss: -0.0074 | Ent: 0.2079
+[ 48.1%] Iter 4701/9765 | SPS:  2185 | Reward:  0.050 | Value:  0.710 | VLoss: 0.0205 | PLoss: -0.0141 | Ent: 0.2040
+[ 48.2%] Iter 4702/9765 | SPS:  2185 | Reward:  0.063 | Value:  0.670 | VLoss: 0.0218 | PLoss: -0.0052 | Ent: 0.1937
+[ 48.2%] Iter 4703/9765 | SPS:  2185 | Reward:  0.070 | Value:  0.728 | VLoss: 0.0278 | PLoss: -0.0137 | Ent: 0.1709
+[ 48.2%] Iter 4704/9765 | SPS:  2185 | Reward:  0.065 | Value:  0.736 | VLoss: 0.0259 | PLoss: -0.0094 | Ent: 0.1841
+[ 48.2%] Iter 4705/9765 | SPS:  2186 | Reward:  0.060 | Value:  0.680 | VLoss: 0.0322 | PLoss: -0.0068 | Ent: 0.1581
+[ 48.2%] Iter 4706/9765 | SPS:  2186 | Reward:  0.057 | Value:  0.691 | VLoss: 0.0234 | PLoss: -0.0098 | Ent: 0.1566
+[ 48.2%] Iter 4707/9765 | SPS:  2186 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0198 | PLoss: -0.0071 | Ent: 0.1591
+[ 48.2%] Iter 4708/9765 | SPS:  2186 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0253 | PLoss: -0.0069 | Ent: 0.1966
+[ 48.2%] Iter 4709/9765 | SPS:  2186 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0227 | PLoss: -0.0090 | Ent: 0.2097
+[ 48.2%] Iter 4710/9765 | SPS:  2187 | Reward:  0.059 | Value:  0.711 | VLoss: 0.0273 | PLoss: -0.0068 | Ent: 0.2033
+[ 48.2%] Iter 4711/9765 | SPS:  2187 | Reward:  0.053 | Value:  0.690 | VLoss: 0.0208 | PLoss: -0.0018 | Ent: 0.1858
+[ 48.3%] Iter 4712/9765 | SPS:  2187 | Reward:  0.066 | Value:  0.682 | VLoss: 0.0193 | PLoss: -0.0089 | Ent: 0.1916
+[ 48.3%] Iter 4713/9765 | SPS:  2187 | Reward:  0.068 | Value:  0.724 | VLoss: 0.0309 | PLoss: -0.0009 | Ent: 0.1894
+[ 48.3%] Iter 4714/9765 | SPS:  2187 | Reward:  0.069 | Value:  0.730 | VLoss: 0.0229 | PLoss: -0.0087 | Ent: 0.1837
+[ 48.3%] Iter 4715/9765 | SPS:  2187 | Reward:  0.059 | Value:  0.767 | VLoss: 0.0167 | PLoss: -0.0081 | Ent: 0.2117
+[ 48.3%] Iter 4716/9765 | SPS:  2188 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0239 | PLoss: -0.0038 | Ent: 0.1802
+[ 48.3%] Iter 4717/9765 | SPS:  2188 | Reward:  0.061 | Value:  0.720 | VLoss: 0.0155 | PLoss: -0.0055 | Ent: 0.1534
+[ 48.3%] Iter 4718/9765 | SPS:  2188 | Reward:  0.066 | Value:  0.769 | VLoss: 0.0248 | PLoss: -0.0065 | Ent: 0.1625
+[ 48.3%] Iter 4719/9765 | SPS:  2188 | Reward:  0.051 | Value:  0.649 | VLoss: 0.0222 | PLoss: -0.0051 | Ent: 0.1871
+[ 48.3%] Iter 4720/9765 | SPS:  2188 | Reward:  0.076 | Value:  0.609 | VLoss: 0.0307 | PLoss: -0.0137 | Ent: 0.1946
+[ 48.3%] Iter 4721/9765 | SPS:  2189 | Reward:  0.060 | Value:  0.722 | VLoss: 0.0173 | PLoss: -0.0088 | Ent: 0.1937
+[ 48.4%] Iter 4722/9765 | SPS:  2189 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0220 | PLoss: -0.0027 | Ent: 0.1662
+[ 48.4%] Iter 4723/9765 | SPS:  2189 | Reward:  0.080 | Value:  0.743 | VLoss: 0.0202 | PLoss: -0.0073 | Ent: 0.1690
+[ 48.4%] Iter 4724/9765 | SPS:  2189 | Reward:  0.077 | Value:  0.769 | VLoss: 0.0251 | PLoss: -0.0079 | Ent: 0.1574
+[ 48.4%] Iter 4725/9765 | SPS:  2189 | Reward:  0.075 | Value:  0.734 | VLoss: 0.0263 | PLoss: -0.0003 | Ent: 0.1588
+[ 48.4%] Iter 4726/9765 | SPS:  2190 | Reward:  0.063 | Value:  0.652 | VLoss: 0.0245 | PLoss: -0.0062 | Ent: 0.2400
+[ 48.4%] Iter 4727/9765 | SPS:  2190 | Reward:  0.059 | Value:  0.690 | VLoss: 0.0293 | PLoss: -0.0012 | Ent: 0.1534
+[ 48.4%] Iter 4728/9765 | SPS:  2190 | Reward:  0.071 | Value:  0.646 | VLoss: 0.0345 | PLoss: -0.0101 | Ent: 0.1710
+[ 48.4%] Iter 4729/9765 | SPS:  2190 | Reward:  0.067 | Value:  0.738 | VLoss: 0.0188 | PLoss: -0.0070 | Ent: 0.1616
+[ 48.4%] Iter 4730/9765 | SPS:  2190 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0266 | PLoss: -0.0095 | Ent: 0.1927
+[ 48.4%] Iter 4731/9765 | SPS:  2190 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0166 | PLoss: -0.0042 | Ent: 0.1906
+[ 48.5%] Iter 4732/9765 | SPS:  2191 | Reward:  0.046 | Value:  0.716 | VLoss: 0.0309 | PLoss: -0.0113 | Ent: 0.1684
+[ 48.5%] Iter 4733/9765 | SPS:  2191 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0251 | PLoss: -0.0049 | Ent: 0.1709
+[ 48.5%] Iter 4734/9765 | SPS:  2191 | Reward:  0.048 | Value:  0.751 | VLoss: 0.0193 | PLoss: -0.0040 | Ent: 0.2592
+[ 48.5%] Iter 4735/9765 | SPS:  2191 | Reward:  0.062 | Value:  0.755 | VLoss: 0.0178 | PLoss: -0.0013 | Ent: 0.2108
+[ 48.5%] Iter 4736/9765 | SPS:  2191 | Reward:  0.079 | Value:  0.754 | VLoss: 0.0274 | PLoss: -0.0092 | Ent: 0.1950
+[ 48.5%] Iter 4737/9765 | SPS:  2192 | Reward:  0.061 | Value:  0.698 | VLoss: 0.0290 | PLoss: -0.0089 | Ent: 0.1822
+[ 48.5%] Iter 4738/9765 | SPS:  2192 | Reward:  0.064 | Value:  0.746 | VLoss: 0.0183 | PLoss: -0.0110 | Ent: 0.1801
+[ 48.5%] Iter 4739/9765 | SPS:  2192 | Reward:  0.065 | Value:  0.721 | VLoss: 0.0236 | PLoss: -0.0050 | Ent: 0.2119
+[ 48.5%] Iter 4740/9765 | SPS:  2192 | Reward:  0.056 | Value:  0.774 | VLoss: 0.0222 | PLoss: -0.0096 | Ent: 0.1382
+[ 48.6%] Iter 4741/9765 | SPS:  2192 | Reward:  0.061 | Value:  0.689 | VLoss: 0.0327 | PLoss: -0.0127 | Ent: 0.1301
+[ 48.6%] Iter 4742/9765 | SPS:  2193 | Reward:  0.060 | Value:  0.724 | VLoss: 0.0207 | PLoss: -0.0056 | Ent: 0.1321
+[ 48.6%] Iter 4743/9765 | SPS:  2193 | Reward:  0.057 | Value:  0.743 | VLoss: 0.0284 | PLoss: -0.0019 | Ent: 0.1999
+[ 48.6%] Iter 4744/9765 | SPS:  2193 | Reward:  0.062 | Value:  0.644 | VLoss: 0.0271 | PLoss: -0.0060 | Ent: 0.1850
+[ 48.6%] Iter 4745/9765 | SPS:  2193 | Reward:  0.055 | Value:  0.691 | VLoss: 0.0297 | PLoss: -0.0074 | Ent: 0.1934
+[ 48.6%] Iter 4746/9765 | SPS:  2193 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0172 | PLoss: -0.0060 | Ent: 0.1661
+[ 48.6%] Iter 4747/9765 | SPS:  2194 | Reward:  0.064 | Value:  0.756 | VLoss: 0.0246 | PLoss: -0.0075 | Ent: 0.1755
+[ 48.6%] Iter 4748/9765 | SPS:  2194 | Reward:  0.072 | Value:  0.717 | VLoss: 0.0221 | PLoss: -0.0096 | Ent: 0.2082
+[ 48.6%] Iter 4749/9765 | SPS:  2194 | Reward:  0.063 | Value:  0.716 | VLoss: 0.0223 | PLoss: -0.0140 | Ent: 0.1597
+[ 48.6%] Iter 4750/9765 | SPS:  2194 | Reward:  0.066 | Value:  0.678 | VLoss: 0.0281 | PLoss: -0.0077 | Ent: 0.1954
+[ 48.7%] Iter 4751/9765 | SPS:  2194 | Reward:  0.064 | Value:  0.669 | VLoss: 0.0260 | PLoss: -0.0054 | Ent: 0.1681
+[ 48.7%] Iter 4752/9765 | SPS:  2194 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0193 | PLoss: -0.0041 | Ent: 0.1797
+[ 48.7%] Iter 4753/9765 | SPS:  2195 | Reward:  0.061 | Value:  0.776 | VLoss: 0.0239 | PLoss: -0.0011 | Ent: 0.1674
+[ 48.7%] Iter 4754/9765 | SPS:  2195 | Reward:  0.051 | Value:  0.734 | VLoss: 0.0273 | PLoss: -0.0061 | Ent: 0.2026
+[ 48.7%] Iter 4755/9765 | SPS:  2195 | Reward:  0.068 | Value:  0.666 | VLoss: 0.0311 | PLoss: -0.0072 | Ent: 0.2135
+[ 48.7%] Iter 4756/9765 | SPS:  2195 | Reward:  0.079 | Value:  0.745 | VLoss: 0.0207 | PLoss: -0.0124 | Ent: 0.1628
+[ 48.7%] Iter 4757/9765 | SPS:  2195 | Reward:  0.064 | Value:  0.812 | VLoss: 0.0231 | PLoss: 0.0021 | Ent: 0.1944
+[ 48.7%] Iter 4758/9765 | SPS:  2196 | Reward:  0.061 | Value:  0.750 | VLoss: 0.0199 | PLoss: -0.0076 | Ent: 0.1997
+[ 48.7%] Iter 4759/9765 | SPS:  2196 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0242 | PLoss: -0.0060 | Ent: 0.2160
+[ 48.7%] Iter 4760/9765 | SPS:  2196 | Reward:  0.064 | Value:  0.731 | VLoss: 0.0201 | PLoss: -0.0108 | Ent: 0.1841
+[ 48.8%] Iter 4761/9765 | SPS:  2196 | Reward:  0.051 | Value:  0.757 | VLoss: 0.0330 | PLoss: -0.0102 | Ent: 0.1526
+[ 48.8%] Iter 4762/9765 | SPS:  2196 | Reward:  0.066 | Value:  0.655 | VLoss: 0.0325 | PLoss: -0.0132 | Ent: 0.1892
+[ 48.8%] Iter 4763/9765 | SPS:  2196 | Reward:  0.060 | Value:  0.688 | VLoss: 0.0225 | PLoss: -0.0069 | Ent: 0.2084
+[ 48.8%] Iter 4764/9765 | SPS:  2197 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0240 | PLoss: -0.0096 | Ent: 0.1812
+[ 48.8%] Iter 4765/9765 | SPS:  2197 | Reward:  0.065 | Value:  0.638 | VLoss: 0.0246 | PLoss: -0.0090 | Ent: 0.1812
+[ 48.8%] Iter 4766/9765 | SPS:  2197 | Reward:  0.071 | Value:  0.768 | VLoss: 0.0234 | PLoss: -0.0130 | Ent: 0.2003
+[ 48.8%] Iter 4767/9765 | SPS:  2197 | Reward:  0.063 | Value:  0.696 | VLoss: 0.0239 | PLoss: -0.0111 | Ent: 0.2260
+[ 48.8%] Iter 4768/9765 | SPS:  2197 | Reward:  0.054 | Value:  0.762 | VLoss: 0.0249 | PLoss: -0.0107 | Ent: 0.1459
+[ 48.8%] Iter 4769/9765 | SPS:  2198 | Reward:  0.061 | Value:  0.658 | VLoss: 0.0228 | PLoss: -0.0052 | Ent: 0.2145
+[ 48.8%] Iter 4770/9765 | SPS:  2198 | Reward:  0.069 | Value:  0.643 | VLoss: 0.0316 | PLoss: -0.0077 | Ent: 0.2183
+[ 48.9%] Iter 4771/9765 | SPS:  2198 | Reward:  0.073 | Value:  0.728 | VLoss: 0.0281 | PLoss: -0.0034 | Ent: 0.1878
+[ 48.9%] Iter 4772/9765 | SPS:  2198 | Reward:  0.050 | Value:  0.757 | VLoss: 0.0168 | PLoss: -0.0055 | Ent: 0.1679
+[ 48.9%] Iter 4773/9765 | SPS:  2198 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0173 | PLoss: -0.0016 | Ent: 0.1988
+[ 48.9%] Iter 4774/9765 | SPS:  2199 | Reward:  0.059 | Value:  0.779 | VLoss: 0.0163 | PLoss: -0.0051 | Ent: 0.2163
+[ 48.9%] Iter 4775/9765 | SPS:  2199 | Reward:  0.061 | Value:  0.707 | VLoss: 0.0184 | PLoss: -0.0083 | Ent: 0.1670
+[ 48.9%] Iter 4776/9765 | SPS:  2199 | Reward:  0.062 | Value:  0.666 | VLoss: 0.0304 | PLoss: -0.0055 | Ent: 0.1815
+[ 48.9%] Iter 4777/9765 | SPS:  2199 | Reward:  0.068 | Value:  0.738 | VLoss: 0.0260 | PLoss: -0.0151 | Ent: 0.2009
+[ 48.9%] Iter 4778/9765 | SPS:  2199 | Reward:  0.062 | Value:  0.784 | VLoss: 0.0238 | PLoss: -0.0030 | Ent: 0.2046
+[ 48.9%] Iter 4779/9765 | SPS:  2199 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0214 | PLoss: -0.0110 | Ent: 0.1747
+[ 49.0%] Iter 4780/9765 | SPS:  2200 | Reward:  0.053 | Value:  0.675 | VLoss: 0.0243 | PLoss: -0.0103 | Ent: 0.1962
+[ 49.0%] Iter 4781/9765 | SPS:  2200 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0259 | PLoss: -0.0100 | Ent: 0.2325
+[ 49.0%] Iter 4782/9765 | SPS:  2200 | Reward:  0.069 | Value:  0.721 | VLoss: 0.0196 | PLoss: -0.0086 | Ent: 0.1617
+[ 49.0%] Iter 4783/9765 | SPS:  2200 | Reward:  0.055 | Value:  0.727 | VLoss: 0.0190 | PLoss: -0.0090 | Ent: 0.1375
+[ 49.0%] Iter 4784/9765 | SPS:  2200 | Reward:  0.072 | Value:  0.692 | VLoss: 0.0349 | PLoss: -0.0112 | Ent: 0.1603
+[ 49.0%] Iter 4785/9765 | SPS:  2201 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0203 | PLoss: -0.0048 | Ent: 0.1908
+[ 49.0%] Iter 4786/9765 | SPS:  2201 | Reward:  0.064 | Value:  0.707 | VLoss: 0.0285 | PLoss: -0.0058 | Ent: 0.2170
+[ 49.0%] Iter 4787/9765 | SPS:  2201 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0197 | PLoss: -0.0061 | Ent: 0.2293
+[ 49.0%] Iter 4788/9765 | SPS:  2201 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0163 | PLoss: -0.0088 | Ent: 0.1996
+[ 49.0%] Iter 4789/9765 | SPS:  2201 | Reward:  0.060 | Value:  0.704 | VLoss: 0.0187 | PLoss: 0.0013 | Ent: 0.1843
+[ 49.1%] Iter 4790/9765 | SPS:  2201 | Reward:  0.060 | Value:  0.677 | VLoss: 0.0240 | PLoss: -0.0038 | Ent: 0.2165
+[ 49.1%] Iter 4791/9765 | SPS:  2202 | Reward:  0.055 | Value:  0.704 | VLoss: 0.0202 | PLoss: -0.0151 | Ent: 0.1954
+[ 49.1%] Iter 4792/9765 | SPS:  2202 | Reward:  0.055 | Value:  0.695 | VLoss: 0.0199 | PLoss: -0.0067 | Ent: 0.1985
+[ 49.1%] Iter 4793/9765 | SPS:  2202 | Reward:  0.060 | Value:  0.760 | VLoss: 0.0164 | PLoss: -0.0078 | Ent: 0.1747
+[ 49.1%] Iter 4794/9765 | SPS:  2202 | Reward:  0.060 | Value:  0.817 | VLoss: 0.0137 | PLoss: -0.0054 | Ent: 0.1824
+[ 49.1%] Iter 4795/9765 | SPS:  2202 | Reward:  0.056 | Value:  0.685 | VLoss: 0.0243 | PLoss: -0.0080 | Ent: 0.1898
+[ 49.1%] Iter 4796/9765 | SPS:  2203 | Reward:  0.070 | Value:  0.737 | VLoss: 0.0348 | PLoss: -0.0090 | Ent: 0.1913
+[ 49.1%] Iter 4797/9765 | SPS:  2203 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0340 | PLoss: -0.0044 | Ent: 0.1796
+[ 49.1%] Iter 4798/9765 | SPS:  2203 | Reward:  0.069 | Value:  0.693 | VLoss: 0.0267 | PLoss: -0.0025 | Ent: 0.1646
+[ 49.1%] Iter 4799/9765 | SPS:  2203 | Reward:  0.055 | Value:  0.630 | VLoss: 0.0191 | PLoss: -0.0060 | Ent: 0.1902
+[ 49.2%] Iter 4800/9765 | SPS:  2203 | Reward:  0.082 | Value:  0.699 | VLoss: 0.0240 | PLoss: -0.0079 | Ent: 0.1427
+[ 49.2%] Iter 4801/9765 | SPS:  2204 | Reward:  0.058 | Value:  0.766 | VLoss: 0.0167 | PLoss: -0.0063 | Ent: 0.1678
+[ 49.2%] Iter 4802/9765 | SPS:  2204 | Reward:  0.058 | Value:  0.743 | VLoss: 0.0267 | PLoss: -0.0096 | Ent: 0.1990
+[ 49.2%] Iter 4803/9765 | SPS:  2204 | Reward:  0.069 | Value:  0.661 | VLoss: 0.0215 | PLoss: -0.0102 | Ent: 0.2117
+[ 49.2%] Iter 4804/9765 | SPS:  2204 | Reward:  0.061 | Value:  0.645 | VLoss: 0.0271 | PLoss: -0.0115 | Ent: 0.1652
+[ 49.2%] Iter 4805/9765 | SPS:  2204 | Reward:  0.070 | Value:  0.691 | VLoss: 0.0319 | PLoss: -0.0026 | Ent: 0.1708
+[ 49.2%] Iter 4806/9765 | SPS:  2204 | Reward:  0.060 | Value:  0.713 | VLoss: 0.0205 | PLoss: -0.0058 | Ent: 0.1774
+[ 49.2%] Iter 4807/9765 | SPS:  2205 | Reward:  0.076 | Value:  0.739 | VLoss: 0.0209 | PLoss: -0.0118 | Ent: 0.2237
+[ 49.2%] Iter 4808/9765 | SPS:  2205 | Reward:  0.062 | Value:  0.700 | VLoss: 0.0187 | PLoss: -0.0055 | Ent: 0.2367
+[ 49.2%] Iter 4809/9765 | SPS:  2205 | Reward:  0.068 | Value:  0.684 | VLoss: 0.0332 | PLoss: -0.0132 | Ent: 0.2081
+[ 49.3%] Iter 4810/9765 | SPS:  2205 | Reward:  0.063 | Value:  0.653 | VLoss: 0.0263 | PLoss: -0.0161 | Ent: 0.1990
+[ 49.3%] Iter 4811/9765 | SPS:  2205 | Reward:  0.057 | Value:  0.702 | VLoss: 0.0214 | PLoss: -0.0097 | Ent: 0.2007
+[ 49.3%] Iter 4812/9765 | SPS:  2206 | Reward:  0.068 | Value:  0.723 | VLoss: 0.0264 | PLoss: -0.0076 | Ent: 0.2134
+[ 49.3%] Iter 4813/9765 | SPS:  2206 | Reward:  0.056 | Value:  0.731 | VLoss: 0.0223 | PLoss: -0.0025 | Ent: 0.1680
+[ 49.3%] Iter 4814/9765 | SPS:  2206 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0294 | PLoss: -0.0037 | Ent: 0.1702
+[ 49.3%] Iter 4815/9765 | SPS:  2206 | Reward:  0.070 | Value:  0.713 | VLoss: 0.0215 | PLoss: -0.0083 | Ent: 0.1896
+[ 49.3%] Iter 4816/9765 | SPS:  2206 | Reward:  0.048 | Value:  0.765 | VLoss: 0.0141 | PLoss: -0.0055 | Ent: 0.1815
+[ 49.3%] Iter 4817/9765 | SPS:  2207 | Reward:  0.072 | Value:  0.756 | VLoss: 0.0193 | PLoss: -0.0109 | Ent: 0.1172
+[ 49.3%] Iter 4818/9765 | SPS:  2207 | Reward:  0.057 | Value:  0.677 | VLoss: 0.0177 | PLoss: -0.0089 | Ent: 0.1645
+[ 49.3%] Iter 4819/9765 | SPS:  2207 | Reward:  0.074 | Value:  0.668 | VLoss: 0.0216 | PLoss: 0.0023 | Ent: 0.1614
+[ 49.4%] Iter 4820/9765 | SPS:  2207 | Reward:  0.066 | Value:  0.757 | VLoss: 0.0236 | PLoss: -0.0041 | Ent: 0.1647
+[ 49.4%] Iter 4821/9765 | SPS:  2207 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0304 | PLoss: -0.0089 | Ent: 0.1608
+[ 49.4%] Iter 4822/9765 | SPS:  2207 | Reward:  0.055 | Value:  0.691 | VLoss: 0.0194 | PLoss: -0.0108 | Ent: 0.1823
+[ 49.4%] Iter 4823/9765 | SPS:  2208 | Reward:  0.066 | Value:  0.766 | VLoss: 0.0172 | PLoss: -0.0054 | Ent: 0.1565
+[ 49.4%] Iter 4824/9765 | SPS:  2208 | Reward:  0.063 | Value:  0.792 | VLoss: 0.0244 | PLoss: -0.0131 | Ent: 0.1884
+[ 49.4%] Iter 4825/9765 | SPS:  2208 | Reward:  0.056 | Value:  0.718 | VLoss: 0.0233 | PLoss: -0.0060 | Ent: 0.1683
+[ 49.4%] Iter 4826/9765 | SPS:  2208 | Reward:  0.068 | Value:  0.718 | VLoss: 0.0250 | PLoss: -0.0122 | Ent: 0.2029
+[ 49.4%] Iter 4827/9765 | SPS:  2208 | Reward:  0.058 | Value:  0.738 | VLoss: 0.0262 | PLoss: -0.0110 | Ent: 0.1768
+[ 49.4%] Iter 4828/9765 | SPS:  2209 | Reward:  0.070 | Value:  0.683 | VLoss: 0.0243 | PLoss: -0.0040 | Ent: 0.1647
+[ 49.5%] Iter 4829/9765 | SPS:  2209 | Reward:  0.051 | Value:  0.757 | VLoss: 0.0274 | PLoss: -0.0096 | Ent: 0.1940
+[ 49.5%] Iter 4830/9765 | SPS:  2209 | Reward:  0.057 | Value:  0.681 | VLoss: 0.0239 | PLoss: -0.0020 | Ent: 0.1921
+[ 49.5%] Iter 4831/9765 | SPS:  2209 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0248 | PLoss: -0.0137 | Ent: 0.1829
+[ 49.5%] Iter 4832/9765 | SPS:  2209 | Reward:  0.060 | Value:  0.697 | VLoss: 0.0251 | PLoss: -0.0078 | Ent: 0.1666
+[ 49.5%] Iter 4833/9765 | SPS:  2210 | Reward:  0.066 | Value:  0.768 | VLoss: 0.0213 | PLoss: -0.0068 | Ent: 0.2273
+[ 49.5%] Iter 4834/9765 | SPS:  2210 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0197 | PLoss: -0.0136 | Ent: 0.1861
+[ 49.5%] Iter 4835/9765 | SPS:  2210 | Reward:  0.060 | Value:  0.730 | VLoss: 0.0226 | PLoss: -0.0052 | Ent: 0.1813
+[ 49.5%] Iter 4836/9765 | SPS:  2210 | Reward:  0.052 | Value:  0.768 | VLoss: 0.0200 | PLoss: -0.0045 | Ent: 0.1849
+[ 49.5%] Iter 4837/9765 | SPS:  2210 | Reward:  0.049 | Value:  0.654 | VLoss: 0.0160 | PLoss: -0.0120 | Ent: 0.2307
+[ 49.5%] Iter 4838/9765 | SPS:  2210 | Reward:  0.056 | Value:  0.680 | VLoss: 0.0212 | PLoss: -0.0147 | Ent: 0.1902
+[ 49.6%] Iter 4839/9765 | SPS:  2211 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0181 | PLoss: -0.0060 | Ent: 0.1737
+[ 49.6%] Iter 4840/9765 | SPS:  2211 | Reward:  0.062 | Value:  0.733 | VLoss: 0.0272 | PLoss: -0.0172 | Ent: 0.1680
+[ 49.6%] Iter 4841/9765 | SPS:  2211 | Reward:  0.057 | Value:  0.713 | VLoss: 0.0159 | PLoss: -0.0121 | Ent: 0.2109
+[ 49.6%] Iter 4842/9765 | SPS:  2211 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0269 | PLoss: -0.0130 | Ent: 0.1962
+[ 49.6%] Iter 4843/9765 | SPS:  2211 | Reward:  0.065 | Value:  0.781 | VLoss: 0.0223 | PLoss: -0.0067 | Ent: 0.1592
+[ 49.6%] Iter 4844/9765 | SPS:  2212 | Reward:  0.056 | Value:  0.726 | VLoss: 0.0194 | PLoss: 0.0000 | Ent: 0.1814
+[ 49.6%] Iter 4845/9765 | SPS:  2212 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0270 | PLoss: -0.0095 | Ent: 0.1579
+[ 49.6%] Iter 4846/9765 | SPS:  2212 | Reward:  0.051 | Value:  0.644 | VLoss: 0.0266 | PLoss: -0.0137 | Ent: 0.2381
+[ 49.6%] Iter 4847/9765 | SPS:  2212 | Reward:  0.056 | Value:  0.749 | VLoss: 0.0223 | PLoss: -0.0065 | Ent: 0.1329
+[ 49.6%] Iter 4848/9765 | SPS:  2212 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0306 | PLoss: -0.0065 | Ent: 0.2113
+[ 49.7%] Iter 4849/9765 | SPS:  2213 | Reward:  0.067 | Value:  0.628 | VLoss: 0.0290 | PLoss: -0.0114 | Ent: 0.1641
+[ 49.7%] Iter 4850/9765 | SPS:  2213 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0308 | PLoss: -0.0078 | Ent: 0.1735
+[ 49.7%] Iter 4851/9765 | SPS:  2213 | Reward:  0.068 | Value:  0.713 | VLoss: 0.0246 | PLoss: -0.0071 | Ent: 0.1601
+[ 49.7%] Iter 4852/9765 | SPS:  2213 | Reward:  0.069 | Value:  0.721 | VLoss: 0.0339 | PLoss: -0.0129 | Ent: 0.1944
+[ 49.7%] Iter 4853/9765 | SPS:  2213 | Reward:  0.060 | Value:  0.726 | VLoss: 0.0257 | PLoss: -0.0058 | Ent: 0.1798
+[ 49.7%] Iter 4854/9765 | SPS:  2213 | Reward:  0.059 | Value:  0.710 | VLoss: 0.0205 | PLoss: -0.0051 | Ent: 0.2030
+[ 49.7%] Iter 4855/9765 | SPS:  2214 | Reward:  0.063 | Value:  0.628 | VLoss: 0.0264 | PLoss: -0.0084 | Ent: 0.2183
+[ 49.7%] Iter 4856/9765 | SPS:  2214 | Reward:  0.067 | Value:  0.682 | VLoss: 0.0253 | PLoss: -0.0078 | Ent: 0.1734
+[ 49.7%] Iter 4857/9765 | SPS:  2214 | Reward:  0.040 | Value:  0.691 | VLoss: 0.0183 | PLoss: -0.0112 | Ent: 0.2206
+[ 49.7%] Iter 4858/9765 | SPS:  2214 | Reward:  0.053 | Value:  0.682 | VLoss: 0.0223 | PLoss: -0.0105 | Ent: 0.1558
+[ 49.8%] Iter 4859/9765 | SPS:  2214 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0260 | PLoss: -0.0082 | Ent: 0.1822
+[ 49.8%] Iter 4860/9765 | SPS:  2215 | Reward:  0.054 | Value:  0.804 | VLoss: 0.0163 | PLoss: -0.0070 | Ent: 0.1623
+[ 49.8%] Iter 4861/9765 | SPS:  2215 | Reward:  0.062 | Value:  0.763 | VLoss: 0.0222 | PLoss: -0.0108 | Ent: 0.1597
+[ 49.8%] Iter 4862/9765 | SPS:  2215 | Reward:  0.062 | Value:  0.773 | VLoss: 0.0168 | PLoss: -0.0053 | Ent: 0.1442
+[ 49.8%] Iter 4863/9765 | SPS:  2215 | Reward:  0.065 | Value:  0.700 | VLoss: 0.0254 | PLoss: -0.0110 | Ent: 0.1981
+[ 49.8%] Iter 4864/9765 | SPS:  2215 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0218 | PLoss: -0.0137 | Ent: 0.1740
+[ 49.8%] Iter 4865/9765 | SPS:  2215 | Reward:  0.056 | Value:  0.737 | VLoss: 0.0176 | PLoss: -0.0011 | Ent: 0.1764
+[ 49.8%] Iter 4866/9765 | SPS:  2216 | Reward:  0.059 | Value:  0.694 | VLoss: 0.0161 | PLoss: -0.0080 | Ent: 0.1960
+[ 49.8%] Iter 4867/9765 | SPS:  2216 | Reward:  0.072 | Value:  0.747 | VLoss: 0.0253 | PLoss: -0.0087 | Ent: 0.1716
+[ 49.9%] Iter 4868/9765 | SPS:  2216 | Reward:  0.077 | Value:  0.716 | VLoss: 0.0264 | PLoss: -0.0130 | Ent: 0.2045
+[ 49.9%] Iter 4869/9765 | SPS:  2216 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0204 | PLoss: -0.0031 | Ent: 0.1834
+[ 49.9%] Iter 4870/9765 | SPS:  2216 | Reward:  0.049 | Value:  0.655 | VLoss: 0.0225 | PLoss: -0.0026 | Ent: 0.1794
+[ 49.9%] Iter 4871/9765 | SPS:  2217 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0247 | PLoss: -0.0115 | Ent: 0.2106
+[ 49.9%] Iter 4872/9765 | SPS:  2217 | Reward:  0.060 | Value:  0.640 | VLoss: 0.0264 | PLoss: -0.0127 | Ent: 0.2386
+[ 49.9%] Iter 4873/9765 | SPS:  2217 | Reward:  0.045 | Value:  0.631 | VLoss: 0.0279 | PLoss: -0.0148 | Ent: 0.2604
+[ 49.9%] Iter 4874/9765 | SPS:  2217 | Reward:  0.068 | Value:  0.704 | VLoss: 0.0192 | PLoss: -0.0026 | Ent: 0.1950
+[ 49.9%] Iter 4875/9765 | SPS:  2217 | Reward:  0.064 | Value:  0.726 | VLoss: 0.0245 | PLoss: -0.0102 | Ent: 0.2273
+[ 49.9%] Iter 4876/9765 | SPS:  2218 | Reward:  0.066 | Value:  0.703 | VLoss: 0.0248 | PLoss: -0.0035 | Ent: 0.1623
+[ 49.9%] Iter 4877/9765 | SPS:  2218 | Reward:  0.078 | Value:  0.714 | VLoss: 0.0188 | PLoss: -0.0002 | Ent: 0.1932
+[ 50.0%] Iter 4878/9765 | SPS:  2218 | Reward:  0.068 | Value:  0.725 | VLoss: 0.0277 | PLoss: -0.0029 | Ent: 0.1791
+[ 50.0%] Iter 4879/9765 | SPS:  2218 | Reward:  0.070 | Value:  0.696 | VLoss: 0.0210 | PLoss: -0.0084 | Ent: 0.1863
+[ 50.0%] Iter 4880/9765 | SPS:  2218 | Reward:  0.072 | Value:  0.735 | VLoss: 0.0236 | PLoss: -0.0002 | Ent: 0.1788
+[ 50.0%] Iter 4881/9765 | SPS:  2218 | Reward:  0.060 | Value:  0.720 | VLoss: 0.0308 | PLoss: -0.0108 | Ent: 0.1842
+[ 50.0%] Iter 4882/9765 | SPS:  2219 | Reward:  0.070 | Value:  0.686 | VLoss: 0.0333 | PLoss: -0.0096 | Ent: 0.1966
+[ 50.0%] Iter 4883/9765 | SPS:  2219 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0239 | PLoss: -0.0098 | Ent: 0.2048
+[ 50.0%] Iter 4884/9765 | SPS:  2219 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0253 | PLoss: -0.0059 | Ent: 0.1985
+[ 50.0%] Iter 4885/9765 | SPS:  2219 | Reward:  0.067 | Value:  0.761 | VLoss: 0.0230 | PLoss: -0.0048 | Ent: 0.1692
+[ 50.0%] Iter 4886/9765 | SPS:  2219 | Reward:  0.056 | Value:  0.746 | VLoss: 0.0167 | PLoss: -0.0028 | Ent: 0.1407
+[ 50.0%] Iter 4887/9765 | SPS:  2219 | Reward:  0.070 | Value:  0.720 | VLoss: 0.0245 | PLoss: -0.0028 | Ent: 0.1528
+[ 50.1%] Iter 4888/9765 | SPS:  2220 | Reward:  0.067 | Value:  0.743 | VLoss: 0.0298 | PLoss: -0.0041 | Ent: 0.1877
+[ 50.1%] Iter 4889/9765 | SPS:  2220 | Reward:  0.057 | Value:  0.693 | VLoss: 0.0339 | PLoss: -0.0020 | Ent: 0.1900
+[ 50.1%] Iter 4890/9765 | SPS:  2220 | Reward:  0.066 | Value:  0.674 | VLoss: 0.0218 | PLoss: -0.0122 | Ent: 0.1939
+[ 50.1%] Iter 4891/9765 | SPS:  2220 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0247 | PLoss: -0.0214 | Ent: 0.2612
+[ 50.1%] Iter 4892/9765 | SPS:  2220 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0235 | PLoss: -0.0042 | Ent: 0.2039
+[ 50.1%] Iter 4893/9765 | SPS:  2221 | Reward:  0.053 | Value:  0.737 | VLoss: 0.0205 | PLoss: -0.0102 | Ent: 0.2148
+[ 50.1%] Iter 4894/9765 | SPS:  2221 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0228 | PLoss: -0.0030 | Ent: 0.1588
+[ 50.1%] Iter 4895/9765 | SPS:  2221 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0230 | PLoss: -0.0082 | Ent: 0.1849
+[ 50.1%] Iter 4896/9765 | SPS:  2221 | Reward:  0.073 | Value:  0.748 | VLoss: 0.0301 | PLoss: -0.0138 | Ent: 0.1874
+[ 50.1%] Iter 4897/9765 | SPS:  2221 | Reward:  0.075 | Value:  0.712 | VLoss: 0.0259 | PLoss: -0.0088 | Ent: 0.1578
+[ 50.2%] Iter 4898/9765 | SPS:  2221 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0254 | PLoss: -0.0090 | Ent: 0.1627
+[ 50.2%] Iter 4899/9765 | SPS:  2222 | Reward:  0.067 | Value:  0.744 | VLoss: 0.0198 | PLoss: -0.0052 | Ent: 0.2209
+[ 50.2%] Iter 4900/9765 | SPS:  2222 | Reward:  0.053 | Value:  0.783 | VLoss: 0.0166 | PLoss: -0.0076 | Ent: 0.2258
+[ 50.2%] Iter 4901/9765 | SPS:  2222 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0238 | PLoss: -0.0133 | Ent: 0.1857
+[ 50.2%] Iter 4902/9765 | SPS:  2222 | Reward:  0.060 | Value:  0.753 | VLoss: 0.0177 | PLoss: -0.0113 | Ent: 0.1923
+[ 50.2%] Iter 4903/9765 | SPS:  2222 | Reward:  0.054 | Value:  0.652 | VLoss: 0.0200 | PLoss: -0.0070 | Ent: 0.1822
+[ 50.2%] Iter 4904/9765 | SPS:  2223 | Reward:  0.073 | Value:  0.666 | VLoss: 0.0255 | PLoss: -0.0125 | Ent: 0.1689
+[ 50.2%] Iter 4905/9765 | SPS:  2223 | Reward:  0.055 | Value:  0.771 | VLoss: 0.0151 | PLoss: -0.0078 | Ent: 0.1561
+[ 50.2%] Iter 4906/9765 | SPS:  2223 | Reward:  0.044 | Value:  0.695 | VLoss: 0.0137 | PLoss: -0.0103 | Ent: 0.2228
+[ 50.3%] Iter 4907/9765 | SPS:  2223 | Reward:  0.038 | Value:  0.676 | VLoss: 0.0147 | PLoss: -0.0002 | Ent: 0.1833
+[ 50.3%] Iter 4908/9765 | SPS:  2223 | Reward:  0.051 | Value:  0.621 | VLoss: 0.0316 | PLoss: -0.0018 | Ent: 0.2001
+[ 50.3%] Iter 4909/9765 | SPS:  2224 | Reward:  0.059 | Value:  0.663 | VLoss: 0.0216 | PLoss: -0.0082 | Ent: 0.1735
+[ 50.3%] Iter 4910/9765 | SPS:  2224 | Reward:  0.073 | Value:  0.679 | VLoss: 0.0241 | PLoss: -0.0007 | Ent: 0.1550
+[ 50.3%] Iter 4911/9765 | SPS:  2224 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0269 | PLoss: -0.0049 | Ent: 0.1943
+[ 50.3%] Iter 4912/9765 | SPS:  2224 | Reward:  0.071 | Value:  0.693 | VLoss: 0.0263 | PLoss: -0.0170 | Ent: 0.2151
+[ 50.3%] Iter 4913/9765 | SPS:  2224 | Reward:  0.062 | Value:  0.783 | VLoss: 0.0242 | PLoss: -0.0118 | Ent: 0.2163
+[ 50.3%] Iter 4914/9765 | SPS:  2224 | Reward:  0.055 | Value:  0.707 | VLoss: 0.0242 | PLoss: -0.0031 | Ent: 0.2333
+[ 50.3%] Iter 4915/9765 | SPS:  2225 | Reward:  0.044 | Value:  0.676 | VLoss: 0.0169 | PLoss: 0.0009 | Ent: 0.1653
+[ 50.3%] Iter 4916/9765 | SPS:  2225 | Reward:  0.039 | Value:  0.668 | VLoss: 0.0148 | PLoss: -0.0114 | Ent: 0.2051
+[ 50.4%] Iter 4917/9765 | SPS:  2225 | Reward:  0.043 | Value:  0.630 | VLoss: 0.0214 | PLoss: -0.0065 | Ent: 0.2325
+[ 50.4%] Iter 4918/9765 | SPS:  2225 | Reward:  0.058 | Value:  0.654 | VLoss: 0.0269 | PLoss: -0.0075 | Ent: 0.2250
+[ 50.4%] Iter 4919/9765 | SPS:  2225 | Reward:  0.048 | Value:  0.629 | VLoss: 0.0220 | PLoss: -0.0117 | Ent: 0.2051
+[ 50.4%] Iter 4920/9765 | SPS:  2226 | Reward:  0.055 | Value:  0.679 | VLoss: 0.0150 | PLoss: -0.0043 | Ent: 0.2166
+[ 50.4%] Iter 4921/9765 | SPS:  2226 | Reward:  0.065 | Value:  0.797 | VLoss: 0.0227 | PLoss: -0.0055 | Ent: 0.1656
+[ 50.4%] Iter 4922/9765 | SPS:  2226 | Reward:  0.069 | Value:  0.701 | VLoss: 0.0334 | PLoss: -0.0082 | Ent: 0.1777
+[ 50.4%] Iter 4923/9765 | SPS:  2226 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0186 | PLoss: -0.0088 | Ent: 0.1829
+[ 50.4%] Iter 4924/9765 | SPS:  2226 | Reward:  0.052 | Value:  0.766 | VLoss: 0.0195 | PLoss: -0.0160 | Ent: 0.2162
+[ 50.4%] Iter 4925/9765 | SPS:  2227 | Reward:  0.054 | Value:  0.713 | VLoss: 0.0181 | PLoss: -0.0073 | Ent: 0.2204
+[ 50.4%] Iter 4926/9765 | SPS:  2227 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0285 | PLoss: -0.0132 | Ent: 0.2207
+[ 50.5%] Iter 4927/9765 | SPS:  2227 | Reward:  0.075 | Value:  0.734 | VLoss: 0.0264 | PLoss: -0.0071 | Ent: 0.1658
+[ 50.5%] Iter 4928/9765 | SPS:  2227 | Reward:  0.054 | Value:  0.706 | VLoss: 0.0266 | PLoss: -0.0104 | Ent: 0.1776
+[ 50.5%] Iter 4929/9765 | SPS:  2227 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0251 | PLoss: -0.0065 | Ent: 0.1825
+[ 50.5%] Iter 4930/9765 | SPS:  2227 | Reward:  0.057 | Value:  0.734 | VLoss: 0.0281 | PLoss: -0.0075 | Ent: 0.1990
+[ 50.5%] Iter 4931/9765 | SPS:  2228 | Reward:  0.057 | Value:  0.685 | VLoss: 0.0178 | PLoss: -0.0045 | Ent: 0.1196
+[ 50.5%] Iter 4932/9765 | SPS:  2228 | Reward:  0.067 | Value:  0.720 | VLoss: 0.0238 | PLoss: -0.0108 | Ent: 0.1965
+[ 50.5%] Iter 4933/9765 | SPS:  2228 | Reward:  0.063 | Value:  0.791 | VLoss: 0.0150 | PLoss: -0.0054 | Ent: 0.1642
+[ 50.5%] Iter 4934/9765 | SPS:  2228 | Reward:  0.067 | Value:  0.819 | VLoss: 0.0191 | PLoss: -0.0122 | Ent: 0.2291
+[ 50.5%] Iter 4935/9765 | SPS:  2228 | Reward:  0.070 | Value:  0.747 | VLoss: 0.0268 | PLoss: -0.0086 | Ent: 0.1682
+[ 50.5%] Iter 4936/9765 | SPS:  2229 | Reward:  0.052 | Value:  0.722 | VLoss: 0.0236 | PLoss: -0.0163 | Ent: 0.2236
+[ 50.6%] Iter 4937/9765 | SPS:  2229 | Reward:  0.060 | Value:  0.631 | VLoss: 0.0309 | PLoss: -0.0021 | Ent: 0.1662
+[ 50.6%] Iter 4938/9765 | SPS:  2229 | Reward:  0.059 | Value:  0.665 | VLoss: 0.0250 | PLoss: -0.0084 | Ent: 0.1891
+[ 50.6%] Iter 4939/9765 | SPS:  2229 | Reward:  0.059 | Value:  0.681 | VLoss: 0.0224 | PLoss: -0.0057 | Ent: 0.1856
+[ 50.6%] Iter 4940/9765 | SPS:  2229 | Reward:  0.066 | Value:  0.664 | VLoss: 0.0288 | PLoss: -0.0113 | Ent: 0.1845
+[ 50.6%] Iter 4941/9765 | SPS:  2229 | Reward:  0.053 | Value:  0.703 | VLoss: 0.0254 | PLoss: -0.0116 | Ent: 0.1673
+[ 50.6%] Iter 4942/9765 | SPS:  2230 | Reward:  0.071 | Value:  0.734 | VLoss: 0.0279 | PLoss: -0.0097 | Ent: 0.1945
+[ 50.6%] Iter 4943/9765 | SPS:  2230 | Reward:  0.064 | Value:  0.764 | VLoss: 0.0292 | PLoss: -0.0033 | Ent: 0.1733
+[ 50.6%] Iter 4944/9765 | SPS:  2230 | Reward:  0.069 | Value:  0.764 | VLoss: 0.0225 | PLoss: -0.0078 | Ent: 0.1603
+[ 50.6%] Iter 4945/9765 | SPS:  2230 | Reward:  0.062 | Value:  0.767 | VLoss: 0.0213 | PLoss: 0.0002 | Ent: 0.1655
+[ 50.7%] Iter 4946/9765 | SPS:  2230 | Reward:  0.066 | Value:  0.701 | VLoss: 0.0229 | PLoss: -0.0057 | Ent: 0.1709
+[ 50.7%] Iter 4947/9765 | SPS:  2231 | Reward:  0.050 | Value:  0.763 | VLoss: 0.0144 | PLoss: -0.0095 | Ent: 0.1583
+[ 50.7%] Iter 4948/9765 | SPS:  2231 | Reward:  0.063 | Value:  0.702 | VLoss: 0.0226 | PLoss: -0.0186 | Ent: 0.2253
+[ 50.7%] Iter 4949/9765 | SPS:  2231 | Reward:  0.085 | Value:  0.765 | VLoss: 0.0324 | PLoss: -0.0079 | Ent: 0.1983
+[ 50.7%] Iter 4950/9765 | SPS:  2231 | Reward:  0.073 | Value:  0.780 | VLoss: 0.0162 | PLoss: -0.0056 | Ent: 0.1847
+[ 50.7%] Iter 4951/9765 | SPS:  2231 | Reward:  0.066 | Value:  0.751 | VLoss: 0.0268 | PLoss: -0.0040 | Ent: 0.1846
+[ 50.7%] Iter 4952/9765 | SPS:  2231 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0269 | PLoss: -0.0123 | Ent: 0.1972
+[ 50.7%] Iter 4953/9765 | SPS:  2232 | Reward:  0.059 | Value:  0.753 | VLoss: 0.0364 | PLoss: -0.0012 | Ent: 0.1976
+[ 50.7%] Iter 4954/9765 | SPS:  2232 | Reward:  0.062 | Value:  0.740 | VLoss: 0.0211 | PLoss: -0.0017 | Ent: 0.1969
+[ 50.7%] Iter 4955/9765 | SPS:  2232 | Reward:  0.068 | Value:  0.714 | VLoss: 0.0290 | PLoss: -0.0094 | Ent: 0.2054
+[ 50.8%] Iter 4956/9765 | SPS:  2232 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0254 | PLoss: -0.0079 | Ent: 0.1658
+[ 50.8%] Iter 4957/9765 | SPS:  2232 | Reward:  0.052 | Value:  0.767 | VLoss: 0.0208 | PLoss: -0.0150 | Ent: 0.1852
+[ 50.8%] Iter 4958/9765 | SPS:  2233 | Reward:  0.069 | Value:  0.735 | VLoss: 0.0272 | PLoss: -0.0124 | Ent: 0.2276
+[ 50.8%] Iter 4959/9765 | SPS:  2233 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0242 | PLoss: -0.0002 | Ent: 0.1516
+[ 50.8%] Iter 4960/9765 | SPS:  2233 | Reward:  0.065 | Value:  0.740 | VLoss: 0.0237 | PLoss: -0.0185 | Ent: 0.2021
+[ 50.8%] Iter 4961/9765 | SPS:  2233 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0238 | PLoss: -0.0117 | Ent: 0.2381
+[ 50.8%] Iter 4962/9765 | SPS:  2233 | Reward:  0.075 | Value:  0.706 | VLoss: 0.0254 | PLoss: -0.0082 | Ent: 0.1273
+[ 50.8%] Iter 4963/9765 | SPS:  2233 | Reward:  0.076 | Value:  0.730 | VLoss: 0.0265 | PLoss: -0.0033 | Ent: 0.2278
+[ 50.8%] Iter 4964/9765 | SPS:  2234 | Reward:  0.072 | Value:  0.759 | VLoss: 0.0209 | PLoss: -0.0032 | Ent: 0.1732
+[ 50.8%] Iter 4965/9765 | SPS:  2234 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0233 | PLoss: -0.0051 | Ent: 0.2035
+[ 50.9%] Iter 4966/9765 | SPS:  2234 | Reward:  0.062 | Value:  0.756 | VLoss: 0.0149 | PLoss: -0.0093 | Ent: 0.2122
+[ 50.9%] Iter 4967/9765 | SPS:  2234 | Reward:  0.066 | Value:  0.760 | VLoss: 0.0258 | PLoss: -0.0093 | Ent: 0.1763
+[ 50.9%] Iter 4968/9765 | SPS:  2234 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0160 | PLoss: -0.0046 | Ent: 0.1641
+[ 50.9%] Iter 4969/9765 | SPS:  2235 | Reward:  0.056 | Value:  0.770 | VLoss: 0.0255 | PLoss: -0.0079 | Ent: 0.1848
+[ 50.9%] Iter 4970/9765 | SPS:  2235 | Reward:  0.070 | Value:  0.606 | VLoss: 0.0320 | PLoss: -0.0045 | Ent: 0.1864
+[ 50.9%] Iter 4971/9765 | SPS:  2235 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0178 | PLoss: -0.0013 | Ent: 0.1940
+[ 50.9%] Iter 4972/9765 | SPS:  2235 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0257 | PLoss: -0.0088 | Ent: 0.2227
+[ 50.9%] Iter 4973/9765 | SPS:  2235 | Reward:  0.068 | Value:  0.697 | VLoss: 0.0298 | PLoss: -0.0048 | Ent: 0.1538
+[ 50.9%] Iter 4974/9765 | SPS:  2235 | Reward:  0.061 | Value:  0.706 | VLoss: 0.0234 | PLoss: -0.0071 | Ent: 0.1359
+[ 50.9%] Iter 4975/9765 | SPS:  2236 | Reward:  0.066 | Value:  0.760 | VLoss: 0.0189 | PLoss: -0.0144 | Ent: 0.1984
+[ 51.0%] Iter 4976/9765 | SPS:  2236 | Reward:  0.042 | Value:  0.739 | VLoss: 0.0208 | PLoss: -0.0099 | Ent: 0.1759
+[ 51.0%] Iter 4977/9765 | SPS:  2236 | Reward:  0.072 | Value:  0.720 | VLoss: 0.0296 | PLoss: -0.0062 | Ent: 0.2035
+[ 51.0%] Iter 4978/9765 | SPS:  2236 | Reward:  0.069 | Value:  0.756 | VLoss: 0.0266 | PLoss: -0.0120 | Ent: 0.1905
+[ 51.0%] Iter 4979/9765 | SPS:  2236 | Reward:  0.072 | Value:  0.721 | VLoss: 0.0223 | PLoss: -0.0105 | Ent: 0.1965
+[ 51.0%] Iter 4980/9765 | SPS:  2237 | Reward:  0.064 | Value:  0.683 | VLoss: 0.0261 | PLoss: -0.0094 | Ent: 0.1621
+[ 51.0%] Iter 4981/9765 | SPS:  2237 | Reward:  0.083 | Value:  0.682 | VLoss: 0.0344 | PLoss: -0.0100 | Ent: 0.1725
+[ 51.0%] Iter 4982/9765 | SPS:  2237 | Reward:  0.065 | Value:  0.694 | VLoss: 0.0278 | PLoss: -0.0149 | Ent: 0.2063
+[ 51.0%] Iter 4983/9765 | SPS:  2237 | Reward:  0.064 | Value:  0.773 | VLoss: 0.0098 | PLoss: -0.0043 | Ent: 0.2418
+[ 51.0%] Iter 4984/9765 | SPS:  2237 | Reward:  0.060 | Value:  0.762 | VLoss: 0.0304 | PLoss: -0.0116 | Ent: 0.1913
+[ 51.0%] Iter 4985/9765 | SPS:  2237 | Reward:  0.077 | Value:  0.750 | VLoss: 0.0248 | PLoss: -0.0071 | Ent: 0.1514
+[ 51.1%] Iter 4986/9765 | SPS:  2238 | Reward:  0.052 | Value:  0.685 | VLoss: 0.0343 | PLoss: -0.0088 | Ent: 0.1678
+[ 51.1%] Iter 4987/9765 | SPS:  2238 | Reward:  0.081 | Value:  0.694 | VLoss: 0.0251 | PLoss: -0.0066 | Ent: 0.1522
+[ 51.1%] Iter 4988/9765 | SPS:  2238 | Reward:  0.052 | Value:  0.711 | VLoss: 0.0256 | PLoss: -0.0061 | Ent: 0.1891
+[ 51.1%] Iter 4989/9765 | SPS:  2238 | Reward:  0.080 | Value:  0.753 | VLoss: 0.0188 | PLoss: -0.0050 | Ent: 0.1719
+[ 51.1%] Iter 4990/9765 | SPS:  2238 | Reward:  0.069 | Value:  0.776 | VLoss: 0.0268 | PLoss: -0.0118 | Ent: 0.2105
+[ 51.1%] Iter 4991/9765 | SPS:  2238 | Reward:  0.068 | Value:  0.741 | VLoss: 0.0175 | PLoss: -0.0063 | Ent: 0.1678
+[ 51.1%] Iter 4992/9765 | SPS:  2239 | Reward:  0.078 | Value:  0.756 | VLoss: 0.0202 | PLoss: -0.0030 | Ent: 0.1831
+[ 51.1%] Iter 4993/9765 | SPS:  2239 | Reward:  0.057 | Value:  0.768 | VLoss: 0.0238 | PLoss: -0.0077 | Ent: 0.1691
+[ 51.1%] Iter 4994/9765 | SPS:  2239 | Reward:  0.067 | Value:  0.667 | VLoss: 0.0267 | PLoss: -0.0059 | Ent: 0.1902
+[ 51.2%] Iter 4995/9765 | SPS:  2239 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0169 | PLoss: -0.0025 | Ent: 0.1951
+[ 51.2%] Iter 4996/9765 | SPS:  2239 | Reward:  0.069 | Value:  0.716 | VLoss: 0.0287 | PLoss: -0.0101 | Ent: 0.2303
+[ 51.2%] Iter 4997/9765 | SPS:  2239 | Reward:  0.070 | Value:  0.629 | VLoss: 0.0263 | PLoss: -0.0120 | Ent: 0.2279
+[ 51.2%] Iter 4998/9765 | SPS:  2240 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0225 | PLoss: -0.0087 | Ent: 0.1978
+[ 51.2%] Iter 4999/9765 | SPS:  2240 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0258 | PLoss: -0.0066 | Ent: 0.2004
+[ 51.2%] Iter 5000/9765 | SPS:  2240 | Reward:  0.058 | Value:  0.743 | VLoss: 0.0158 | PLoss: -0.0071 | Ent: 0.1859
+[ 51.2%] Iter 5001/9765 | SPS:  2240 | Reward:  0.085 | Value:  0.771 | VLoss: 0.0275 | PLoss: -0.0175 | Ent: 0.2079
+[ 51.2%] Iter 5002/9765 | SPS:  2240 | Reward:  0.071 | Value:  0.767 | VLoss: 0.0265 | PLoss: -0.0097 | Ent: 0.1953
+[ 51.2%] Iter 5003/9765 | SPS:  2241 | Reward:  0.065 | Value:  0.747 | VLoss: 0.0294 | PLoss: -0.0077 | Ent: 0.1831
+[ 51.2%] Iter 5004/9765 | SPS:  2241 | Reward:  0.069 | Value:  0.688 | VLoss: 0.0247 | PLoss: -0.0038 | Ent: 0.1606
+[ 51.3%] Iter 5005/9765 | SPS:  2241 | Reward:  0.056 | Value:  0.715 | VLoss: 0.0254 | PLoss: -0.0003 | Ent: 0.1521
+[ 51.3%] Iter 5006/9765 | SPS:  2241 | Reward:  0.052 | Value:  0.699 | VLoss: 0.0209 | PLoss: -0.0089 | Ent: 0.2228
+[ 51.3%] Iter 5007/9765 | SPS:  2241 | Reward:  0.059 | Value:  0.711 | VLoss: 0.0260 | PLoss: -0.0109 | Ent: 0.2253
+[ 51.3%] Iter 5008/9765 | SPS:  2241 | Reward:  0.069 | Value:  0.663 | VLoss: 0.0280 | PLoss: -0.0116 | Ent: 0.1978
+[ 51.3%] Iter 5009/9765 | SPS:  2242 | Reward:  0.066 | Value:  0.705 | VLoss: 0.0340 | PLoss: -0.0060 | Ent: 0.1788
+[ 51.3%] Iter 5010/9765 | SPS:  2242 | Reward:  0.066 | Value:  0.747 | VLoss: 0.0256 | PLoss: -0.0141 | Ent: 0.1613
+[ 51.3%] Iter 5011/9765 | SPS:  2242 | Reward:  0.057 | Value:  0.753 | VLoss: 0.0177 | PLoss: -0.0056 | Ent: 0.1888
+[ 51.3%] Iter 5012/9765 | SPS:  2242 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0226 | PLoss: -0.0088 | Ent: 0.1912
+[ 51.3%] Iter 5013/9765 | SPS:  2242 | Reward:  0.055 | Value:  0.664 | VLoss: 0.0197 | PLoss: -0.0094 | Ent: 0.1848
+[ 51.3%] Iter 5014/9765 | SPS:  2242 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0268 | PLoss: -0.0093 | Ent: 0.1929
+[ 51.4%] Iter 5015/9765 | SPS:  2243 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0205 | PLoss: -0.0079 | Ent: 0.1781
+[ 51.4%] Iter 5016/9765 | SPS:  2243 | Reward:  0.071 | Value:  0.766 | VLoss: 0.0199 | PLoss: -0.0041 | Ent: 0.1285
+[ 51.4%] Iter 5017/9765 | SPS:  2243 | Reward:  0.071 | Value:  0.708 | VLoss: 0.0283 | PLoss: -0.0082 | Ent: 0.1514
+[ 51.4%] Iter 5018/9765 | SPS:  2243 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0237 | PLoss: -0.0152 | Ent: 0.1763
+[ 51.4%] Iter 5019/9765 | SPS:  2243 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0288 | PLoss: -0.0178 | Ent: 0.2117
+[ 51.4%] Iter 5020/9765 | SPS:  2244 | Reward:  0.065 | Value:  0.669 | VLoss: 0.0196 | PLoss: -0.0099 | Ent: 0.2191
+[ 51.4%] Iter 5021/9765 | SPS:  2244 | Reward:  0.059 | Value:  0.683 | VLoss: 0.0282 | PLoss: -0.0037 | Ent: 0.2159
+[ 51.4%] Iter 5022/9765 | SPS:  2244 | Reward:  0.053 | Value:  0.757 | VLoss: 0.0286 | PLoss: -0.0087 | Ent: 0.1745
+[ 51.4%] Iter 5023/9765 | SPS:  2244 | Reward:  0.070 | Value:  0.743 | VLoss: 0.0235 | PLoss: -0.0048 | Ent: 0.1901
+[ 51.4%] Iter 5024/9765 | SPS:  2244 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0248 | PLoss: -0.0126 | Ent: 0.2156
+[ 51.5%] Iter 5025/9765 | SPS:  2244 | Reward:  0.061 | Value:  0.757 | VLoss: 0.0230 | PLoss: -0.0036 | Ent: 0.1610
+[ 51.5%] Iter 5026/9765 | SPS:  2245 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0304 | PLoss: -0.0056 | Ent: 0.1516
+[ 51.5%] Iter 5027/9765 | SPS:  2245 | Reward:  0.075 | Value:  0.711 | VLoss: 0.0231 | PLoss: -0.0011 | Ent: 0.1704
+[ 51.5%] Iter 5028/9765 | SPS:  2245 | Reward:  0.070 | Value:  0.729 | VLoss: 0.0246 | PLoss: -0.0008 | Ent: 0.1808
+[ 51.5%] Iter 5029/9765 | SPS:  2245 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0200 | PLoss: -0.0024 | Ent: 0.1789
+[ 51.5%] Iter 5030/9765 | SPS:  2245 | Reward:  0.076 | Value:  0.738 | VLoss: 0.0242 | PLoss: -0.0087 | Ent: 0.1808
+[ 51.5%] Iter 5031/9765 | SPS:  2246 | Reward:  0.061 | Value:  0.718 | VLoss: 0.0235 | PLoss: -0.0057 | Ent: 0.2085
+[ 51.5%] Iter 5032/9765 | SPS:  2246 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0219 | PLoss: -0.0096 | Ent: 0.1682
+[ 51.5%] Iter 5033/9765 | SPS:  2246 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0216 | PLoss: -0.0038 | Ent: 0.1813
+[ 51.6%] Iter 5034/9765 | SPS:  2246 | Reward:  0.063 | Value:  0.769 | VLoss: 0.0203 | PLoss: -0.0089 | Ent: 0.1752
+[ 51.6%] Iter 5035/9765 | SPS:  2246 | Reward:  0.066 | Value:  0.703 | VLoss: 0.0282 | PLoss: -0.0112 | Ent: 0.2212
+[ 51.6%] Iter 5036/9765 | SPS:  2246 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0272 | PLoss: -0.0132 | Ent: 0.2395
+[ 51.6%] Iter 5037/9765 | SPS:  2247 | Reward:  0.057 | Value:  0.633 | VLoss: 0.0230 | PLoss: -0.0071 | Ent: 0.1657
+[ 51.6%] Iter 5038/9765 | SPS:  2247 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0274 | PLoss: -0.0006 | Ent: 0.2254
+[ 51.6%] Iter 5039/9765 | SPS:  2247 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0207 | PLoss: -0.0069 | Ent: 0.1382
+[ 51.6%] Iter 5040/9765 | SPS:  2247 | Reward:  0.080 | Value:  0.747 | VLoss: 0.0283 | PLoss: -0.0093 | Ent: 0.1643
+[ 51.6%] Iter 5041/9765 | SPS:  2247 | Reward:  0.059 | Value:  0.763 | VLoss: 0.0237 | PLoss: -0.0030 | Ent: 0.2185
+[ 51.6%] Iter 5042/9765 | SPS:  2247 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0198 | PLoss: -0.0136 | Ent: 0.1789
+[ 51.6%] Iter 5043/9765 | SPS:  2248 | Reward:  0.065 | Value:  0.757 | VLoss: 0.0193 | PLoss: -0.0088 | Ent: 0.1907
+[ 51.7%] Iter 5044/9765 | SPS:  2248 | Reward:  0.068 | Value:  0.736 | VLoss: 0.0257 | PLoss: -0.0160 | Ent: 0.2090
+[ 51.7%] Iter 5045/9765 | SPS:  2248 | Reward:  0.063 | Value:  0.611 | VLoss: 0.0253 | PLoss: -0.0097 | Ent: 0.1906
+[ 51.7%] Iter 5046/9765 | SPS:  2248 | Reward:  0.067 | Value:  0.707 | VLoss: 0.0272 | PLoss: -0.0153 | Ent: 0.1753
+[ 51.7%] Iter 5047/9765 | SPS:  2248 | Reward:  0.068 | Value:  0.705 | VLoss: 0.0274 | PLoss: -0.0027 | Ent: 0.1638
+[ 51.7%] Iter 5048/9765 | SPS:  2248 | Reward:  0.067 | Value:  0.730 | VLoss: 0.0305 | PLoss: -0.0030 | Ent: 0.2021
+[ 51.7%] Iter 5049/9765 | SPS:  2249 | Reward:  0.060 | Value:  0.754 | VLoss: 0.0285 | PLoss: -0.0028 | Ent: 0.2082
+[ 51.7%] Iter 5050/9765 | SPS:  2249 | Reward:  0.056 | Value:  0.685 | VLoss: 0.0273 | PLoss: -0.0087 | Ent: 0.1630
+[ 51.7%] Iter 5051/9765 | SPS:  2249 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0228 | PLoss: -0.0090 | Ent: 0.2259
+[ 51.7%] Iter 5052/9765 | SPS:  2249 | Reward:  0.052 | Value:  0.705 | VLoss: 0.0287 | PLoss: -0.0049 | Ent: 0.2229
+[ 51.7%] Iter 5053/9765 | SPS:  2249 | Reward:  0.071 | Value:  0.664 | VLoss: 0.0320 | PLoss: -0.0143 | Ent: 0.1854
+[ 51.8%] Iter 5054/9765 | SPS:  2250 | Reward:  0.062 | Value:  0.664 | VLoss: 0.0224 | PLoss: -0.0126 | Ent: 0.1789
+[ 51.8%] Iter 5055/9765 | SPS:  2250 | Reward:  0.059 | Value:  0.728 | VLoss: 0.0232 | PLoss: -0.0074 | Ent: 0.1761
+[ 51.8%] Iter 5056/9765 | SPS:  2250 | Reward:  0.071 | Value:  0.723 | VLoss: 0.0289 | PLoss: -0.0036 | Ent: 0.1578
+[ 51.8%] Iter 5057/9765 | SPS:  2250 | Reward:  0.060 | Value:  0.673 | VLoss: 0.0301 | PLoss: -0.0101 | Ent: 0.1880
+[ 51.8%] Iter 5058/9765 | SPS:  2250 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0272 | PLoss: -0.0028 | Ent: 0.1820
+[ 51.8%] Iter 5059/9765 | SPS:  2250 | Reward:  0.067 | Value:  0.669 | VLoss: 0.0278 | PLoss: -0.0127 | Ent: 0.1333
+[ 51.8%] Iter 5060/9765 | SPS:  2251 | Reward:  0.050 | Value:  0.709 | VLoss: 0.0243 | PLoss: -0.0192 | Ent: 0.1604
+[ 51.8%] Iter 5061/9765 | SPS:  2251 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0203 | PLoss: -0.0094 | Ent: 0.1918
+[ 51.8%] Iter 5062/9765 | SPS:  2251 | Reward:  0.074 | Value:  0.702 | VLoss: 0.0248 | PLoss: -0.0060 | Ent: 0.2184
+[ 51.8%] Iter 5063/9765 | SPS:  2251 | Reward:  0.068 | Value:  0.745 | VLoss: 0.0256 | PLoss: -0.0109 | Ent: 0.1987
+[ 51.9%] Iter 5064/9765 | SPS:  2251 | Reward:  0.056 | Value:  0.772 | VLoss: 0.0319 | PLoss: -0.0121 | Ent: 0.1975
+[ 51.9%] Iter 5065/9765 | SPS:  2251 | Reward:  0.061 | Value:  0.603 | VLoss: 0.0190 | PLoss: -0.0061 | Ent: 0.1536
+[ 51.9%] Iter 5066/9765 | SPS:  2252 | Reward:  0.060 | Value:  0.670 | VLoss: 0.0227 | PLoss: -0.0089 | Ent: 0.1897
+[ 51.9%] Iter 5067/9765 | SPS:  2252 | Reward:  0.056 | Value:  0.728 | VLoss: 0.0228 | PLoss: -0.0107 | Ent: 0.1822
+[ 51.9%] Iter 5068/9765 | SPS:  2252 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0263 | PLoss: -0.0048 | Ent: 0.1986
+[ 51.9%] Iter 5069/9765 | SPS:  2252 | Reward:  0.060 | Value:  0.707 | VLoss: 0.0239 | PLoss: -0.0087 | Ent: 0.2353
+[ 51.9%] Iter 5070/9765 | SPS:  2252 | Reward:  0.072 | Value:  0.749 | VLoss: 0.0228 | PLoss: 0.0006 | Ent: 0.1764
+[ 51.9%] Iter 5071/9765 | SPS:  2253 | Reward:  0.057 | Value:  0.702 | VLoss: 0.0251 | PLoss: -0.0053 | Ent: 0.2309
+[ 51.9%] Iter 5072/9765 | SPS:  2253 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0328 | PLoss: 0.0059 | Ent: 0.1947
+[ 52.0%] Iter 5073/9765 | SPS:  2253 | Reward:  0.058 | Value:  0.756 | VLoss: 0.0219 | PLoss: -0.0054 | Ent: 0.1704
+[ 52.0%] Iter 5074/9765 | SPS:  2253 | Reward:  0.050 | Value:  0.718 | VLoss: 0.0174 | PLoss: -0.0029 | Ent: 0.1727
+[ 52.0%] Iter 5075/9765 | SPS:  2253 | Reward:  0.062 | Value:  0.658 | VLoss: 0.0242 | PLoss: -0.0135 | Ent: 0.1917
+[ 52.0%] Iter 5076/9765 | SPS:  2253 | Reward:  0.063 | Value:  0.666 | VLoss: 0.0257 | PLoss: -0.0071 | Ent: 0.1323
+[ 52.0%] Iter 5077/9765 | SPS:  2254 | Reward:  0.072 | Value:  0.710 | VLoss: 0.0264 | PLoss: -0.0054 | Ent: 0.1571
+[ 52.0%] Iter 5078/9765 | SPS:  2254 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0259 | PLoss: -0.0064 | Ent: 0.1680
+[ 52.0%] Iter 5079/9765 | SPS:  2254 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0223 | PLoss: -0.0129 | Ent: 0.2066
+[ 52.0%] Iter 5080/9765 | SPS:  2254 | Reward:  0.056 | Value:  0.630 | VLoss: 0.0297 | PLoss: -0.0099 | Ent: 0.2047
+[ 52.0%] Iter 5081/9765 | SPS:  2254 | Reward:  0.066 | Value:  0.682 | VLoss: 0.0196 | PLoss: -0.0079 | Ent: 0.1732
+[ 52.0%] Iter 5082/9765 | SPS:  2254 | Reward:  0.055 | Value:  0.738 | VLoss: 0.0129 | PLoss: -0.0117 | Ent: 0.1804
+[ 52.1%] Iter 5083/9765 | SPS:  2255 | Reward:  0.072 | Value:  0.690 | VLoss: 0.0326 | PLoss: -0.0099 | Ent: 0.2039
+[ 52.1%] Iter 5084/9765 | SPS:  2255 | Reward:  0.061 | Value:  0.727 | VLoss: 0.0293 | PLoss: -0.0097 | Ent: 0.1654
+[ 52.1%] Iter 5085/9765 | SPS:  2255 | Reward:  0.071 | Value:  0.698 | VLoss: 0.0197 | PLoss: -0.0050 | Ent: 0.1860
+[ 52.1%] Iter 5086/9765 | SPS:  2255 | Reward:  0.050 | Value:  0.650 | VLoss: 0.0223 | PLoss: -0.0062 | Ent: 0.1869
+[ 52.1%] Iter 5087/9765 | SPS:  2255 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0241 | PLoss: -0.0117 | Ent: 0.2322
+[ 52.1%] Iter 5088/9765 | SPS:  2255 | Reward:  0.055 | Value:  0.667 | VLoss: 0.0306 | PLoss: -0.0155 | Ent: 0.2308
+[ 52.1%] Iter 5089/9765 | SPS:  2256 | Reward:  0.053 | Value:  0.644 | VLoss: 0.0283 | PLoss: -0.0128 | Ent: 0.2250
+[ 52.1%] Iter 5090/9765 | SPS:  2256 | Reward:  0.072 | Value:  0.692 | VLoss: 0.0193 | PLoss: -0.0065 | Ent: 0.1947
+[ 52.1%] Iter 5091/9765 | SPS:  2256 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0183 | PLoss: -0.0064 | Ent: 0.1429
+[ 52.1%] Iter 5092/9765 | SPS:  2256 | Reward:  0.054 | Value:  0.730 | VLoss: 0.0212 | PLoss: -0.0022 | Ent: 0.2097
+[ 52.2%] Iter 5093/9765 | SPS:  2256 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0199 | PLoss: -0.0072 | Ent: 0.1669
+[ 52.2%] Iter 5094/9765 | SPS:  2257 | Reward:  0.053 | Value:  0.702 | VLoss: 0.0180 | PLoss: -0.0102 | Ent: 0.1679
+[ 52.2%] Iter 5095/9765 | SPS:  2257 | Reward:  0.063 | Value:  0.738 | VLoss: 0.0270 | PLoss: -0.0018 | Ent: 0.1743
+[ 52.2%] Iter 5096/9765 | SPS:  2257 | Reward:  0.075 | Value:  0.729 | VLoss: 0.0294 | PLoss: -0.0062 | Ent: 0.1675
+[ 52.2%] Iter 5097/9765 | SPS:  2257 | Reward:  0.077 | Value:  0.669 | VLoss: 0.0270 | PLoss: -0.0116 | Ent: 0.1855
+[ 52.2%] Iter 5098/9765 | SPS:  2257 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0253 | PLoss: -0.0083 | Ent: 0.2060
+[ 52.2%] Iter 5099/9765 | SPS:  2257 | Reward:  0.059 | Value:  0.717 | VLoss: 0.0269 | PLoss: -0.0086 | Ent: 0.2267
+[ 52.2%] Iter 5100/9765 | SPS:  2258 | Reward:  0.061 | Value:  0.669 | VLoss: 0.0274 | PLoss: -0.0067 | Ent: 0.1648
+[ 52.2%] Iter 5101/9765 | SPS:  2258 | Reward:  0.059 | Value:  0.729 | VLoss: 0.0248 | PLoss: -0.0101 | Ent: 0.1594
+[ 52.2%] Iter 5102/9765 | SPS:  2258 | Reward:  0.056 | Value:  0.603 | VLoss: 0.0199 | PLoss: -0.0105 | Ent: 0.2239
+[ 52.3%] Iter 5103/9765 | SPS:  2258 | Reward:  0.061 | Value:  0.623 | VLoss: 0.0229 | PLoss: -0.0139 | Ent: 0.1631
+[ 52.3%] Iter 5104/9765 | SPS:  2258 | Reward:  0.059 | Value:  0.666 | VLoss: 0.0217 | PLoss: -0.0133 | Ent: 0.2157
+[ 52.3%] Iter 5105/9765 | SPS:  2258 | Reward:  0.047 | Value:  0.621 | VLoss: 0.0200 | PLoss: -0.0086 | Ent: 0.2097
+[ 52.3%] Iter 5106/9765 | SPS:  2259 | Reward:  0.056 | Value:  0.663 | VLoss: 0.0249 | PLoss: -0.0149 | Ent: 0.2078
+[ 52.3%] Iter 5107/9765 | SPS:  2259 | Reward:  0.079 | Value:  0.705 | VLoss: 0.0353 | PLoss: -0.0092 | Ent: 0.1934
+[ 52.3%] Iter 5108/9765 | SPS:  2259 | Reward:  0.062 | Value:  0.766 | VLoss: 0.0180 | PLoss: -0.0017 | Ent: 0.2072
+[ 52.3%] Iter 5109/9765 | SPS:  2259 | Reward:  0.065 | Value:  0.719 | VLoss: 0.0271 | PLoss: -0.0046 | Ent: 0.2131
+[ 52.3%] Iter 5110/9765 | SPS:  2259 | Reward:  0.068 | Value:  0.763 | VLoss: 0.0187 | PLoss: -0.0110 | Ent: 0.2127
+[ 52.3%] Iter 5111/9765 | SPS:  2259 | Reward:  0.066 | Value:  0.728 | VLoss: 0.0226 | PLoss: -0.0037 | Ent: 0.2239
+[ 52.4%] Iter 5112/9765 | SPS:  2260 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0299 | PLoss: -0.0085 | Ent: 0.2145
+[ 52.4%] Iter 5113/9765 | SPS:  2260 | Reward:  0.057 | Value:  0.725 | VLoss: 0.0205 | PLoss: -0.0067 | Ent: 0.1689
+[ 52.4%] Iter 5114/9765 | SPS:  2260 | Reward:  0.065 | Value:  0.670 | VLoss: 0.0276 | PLoss: -0.0097 | Ent: 0.1613
+[ 52.4%] Iter 5115/9765 | SPS:  2260 | Reward:  0.055 | Value:  0.732 | VLoss: 0.0291 | PLoss: -0.0046 | Ent: 0.1986
+[ 52.4%] Iter 5116/9765 | SPS:  2260 | Reward:  0.071 | Value:  0.680 | VLoss: 0.0263 | PLoss: -0.0028 | Ent: 0.1719
+[ 52.4%] Iter 5117/9765 | SPS:  2260 | Reward:  0.067 | Value:  0.722 | VLoss: 0.0291 | PLoss: -0.0097 | Ent: 0.1910
+[ 52.4%] Iter 5118/9765 | SPS:  2261 | Reward:  0.057 | Value:  0.722 | VLoss: 0.0225 | PLoss: -0.0035 | Ent: 0.2204
+[ 52.4%] Iter 5119/9765 | SPS:  2261 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0323 | PLoss: -0.0068 | Ent: 0.2346
+[ 52.4%] Iter 5120/9765 | SPS:  2261 | Reward:  0.060 | Value:  0.669 | VLoss: 0.0240 | PLoss: -0.0093 | Ent: 0.2050
+[ 52.4%] Iter 5121/9765 | SPS:  2261 | Reward:  0.069 | Value:  0.752 | VLoss: 0.0256 | PLoss: -0.0092 | Ent: 0.1822
+[ 52.5%] Iter 5122/9765 | SPS:  2261 | Reward:  0.058 | Value:  0.655 | VLoss: 0.0312 | PLoss: -0.0089 | Ent: 0.1659
+[ 52.5%] Iter 5123/9765 | SPS:  2261 | Reward:  0.061 | Value:  0.632 | VLoss: 0.0217 | PLoss: -0.0039 | Ent: 0.1824
+[ 52.5%] Iter 5124/9765 | SPS:  2262 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0267 | PLoss: -0.0130 | Ent: 0.2071
+[ 52.5%] Iter 5125/9765 | SPS:  2262 | Reward:  0.065 | Value:  0.722 | VLoss: 0.0240 | PLoss: -0.0009 | Ent: 0.1578
+[ 52.5%] Iter 5126/9765 | SPS:  2262 | Reward:  0.058 | Value:  0.695 | VLoss: 0.0342 | PLoss: -0.0101 | Ent: 0.1841
+[ 52.5%] Iter 5127/9765 | SPS:  2262 | Reward:  0.070 | Value:  0.717 | VLoss: 0.0312 | PLoss: -0.0065 | Ent: 0.1906
+[ 52.5%] Iter 5128/9765 | SPS:  2262 | Reward:  0.060 | Value:  0.760 | VLoss: 0.0266 | PLoss: -0.0048 | Ent: 0.2144
+[ 52.5%] Iter 5129/9765 | SPS:  2262 | Reward:  0.073 | Value:  0.742 | VLoss: 0.0181 | PLoss: -0.0099 | Ent: 0.1321
+[ 52.5%] Iter 5130/9765 | SPS:  2263 | Reward:  0.061 | Value:  0.742 | VLoss: 0.0242 | PLoss: -0.0050 | Ent: 0.1809
+[ 52.5%] Iter 5131/9765 | SPS:  2263 | Reward:  0.064 | Value:  0.691 | VLoss: 0.0307 | PLoss: -0.0046 | Ent: 0.2273
+[ 52.6%] Iter 5132/9765 | SPS:  2263 | Reward:  0.070 | Value:  0.716 | VLoss: 0.0226 | PLoss: -0.0062 | Ent: 0.1894
+[ 52.6%] Iter 5133/9765 | SPS:  2263 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0260 | PLoss: -0.0080 | Ent: 0.1661
+[ 52.6%] Iter 5134/9765 | SPS:  2263 | Reward:  0.059 | Value:  0.735 | VLoss: 0.0213 | PLoss: -0.0029 | Ent: 0.1769
+[ 52.6%] Iter 5135/9765 | SPS:  2264 | Reward:  0.070 | Value:  0.688 | VLoss: 0.0436 | PLoss: -0.0121 | Ent: 0.1616
+[ 52.6%] Iter 5136/9765 | SPS:  2264 | Reward:  0.079 | Value:  0.699 | VLoss: 0.0256 | PLoss: -0.0077 | Ent: 0.1515
+[ 52.6%] Iter 5137/9765 | SPS:  2264 | Reward:  0.066 | Value:  0.783 | VLoss: 0.0312 | PLoss: -0.0079 | Ent: 0.1480
+[ 52.6%] Iter 5138/9765 | SPS:  2264 | Reward:  0.072 | Value:  0.734 | VLoss: 0.0233 | PLoss: -0.0085 | Ent: 0.1930
+[ 52.6%] Iter 5139/9765 | SPS:  2264 | Reward:  0.067 | Value:  0.739 | VLoss: 0.0243 | PLoss: -0.0011 | Ent: 0.1813
+[ 52.6%] Iter 5140/9765 | SPS:  2264 | Reward:  0.058 | Value:  0.761 | VLoss: 0.0243 | PLoss: -0.0031 | Ent: 0.1624
+[ 52.6%] Iter 5141/9765 | SPS:  2265 | Reward:  0.063 | Value:  0.747 | VLoss: 0.0214 | PLoss: -0.0079 | Ent: 0.1943
+[ 52.7%] Iter 5142/9765 | SPS:  2265 | Reward:  0.071 | Value:  0.748 | VLoss: 0.0179 | PLoss: -0.0042 | Ent: 0.1784
+[ 52.7%] Iter 5143/9765 | SPS:  2265 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0239 | PLoss: -0.0119 | Ent: 0.2173
+[ 52.7%] Iter 5144/9765 | SPS:  2265 | Reward:  0.056 | Value:  0.722 | VLoss: 0.0245 | PLoss: -0.0036 | Ent: 0.2144
+[ 52.7%] Iter 5145/9765 | SPS:  2265 | Reward:  0.067 | Value:  0.669 | VLoss: 0.0296 | PLoss: -0.0133 | Ent: 0.1693
+[ 52.7%] Iter 5146/9765 | SPS:  2265 | Reward:  0.061 | Value:  0.722 | VLoss: 0.0232 | PLoss: -0.0069 | Ent: 0.2091
+[ 52.7%] Iter 5147/9765 | SPS:  2266 | Reward:  0.058 | Value:  0.721 | VLoss: 0.0244 | PLoss: -0.0031 | Ent: 0.1992
+[ 52.7%] Iter 5148/9765 | SPS:  2266 | Reward:  0.060 | Value:  0.680 | VLoss: 0.0260 | PLoss: -0.0037 | Ent: 0.2167
+[ 52.7%] Iter 5149/9765 | SPS:  2266 | Reward:  0.070 | Value:  0.732 | VLoss: 0.0192 | PLoss: -0.0056 | Ent: 0.2254
+[ 52.7%] Iter 5150/9765 | SPS:  2266 | Reward:  0.064 | Value:  0.666 | VLoss: 0.0273 | PLoss: -0.0074 | Ent: 0.1685
+[ 52.7%] Iter 5151/9765 | SPS:  2266 | Reward:  0.072 | Value:  0.721 | VLoss: 0.0247 | PLoss: -0.0065 | Ent: 0.1730
+[ 52.8%] Iter 5152/9765 | SPS:  2266 | Reward:  0.060 | Value:  0.751 | VLoss: 0.0291 | PLoss: -0.0134 | Ent: 0.1853
+[ 52.8%] Iter 5153/9765 | SPS:  2267 | Reward:  0.071 | Value:  0.680 | VLoss: 0.0292 | PLoss: -0.0080 | Ent: 0.2064
+[ 52.8%] Iter 5154/9765 | SPS:  2267 | Reward:  0.074 | Value:  0.665 | VLoss: 0.0340 | PLoss: -0.0075 | Ent: 0.1960
+[ 52.8%] Iter 5155/9765 | SPS:  2267 | Reward:  0.063 | Value:  0.674 | VLoss: 0.0221 | PLoss: -0.0036 | Ent: 0.2245
+[ 52.8%] Iter 5156/9765 | SPS:  2267 | Reward:  0.071 | Value:  0.700 | VLoss: 0.0237 | PLoss: -0.0057 | Ent: 0.1915
+[ 52.8%] Iter 5157/9765 | SPS:  2267 | Reward:  0.067 | Value:  0.788 | VLoss: 0.0240 | PLoss: -0.0102 | Ent: 0.2356
+[ 52.8%] Iter 5158/9765 | SPS:  2267 | Reward:  0.059 | Value:  0.744 | VLoss: 0.0319 | PLoss: -0.0138 | Ent: 0.2040
+[ 52.8%] Iter 5159/9765 | SPS:  2268 | Reward:  0.068 | Value:  0.660 | VLoss: 0.0248 | PLoss: -0.0057 | Ent: 0.1745
+[ 52.8%] Iter 5160/9765 | SPS:  2268 | Reward:  0.048 | Value:  0.717 | VLoss: 0.0203 | PLoss: -0.0033 | Ent: 0.2078
+[ 52.9%] Iter 5161/9765 | SPS:  2268 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0278 | PLoss: -0.0112 | Ent: 0.2253
+[ 52.9%] Iter 5162/9765 | SPS:  2268 | Reward:  0.071 | Value:  0.758 | VLoss: 0.0215 | PLoss: -0.0070 | Ent: 0.2113
+[ 52.9%] Iter 5163/9765 | SPS:  2268 | Reward:  0.060 | Value:  0.735 | VLoss: 0.0232 | PLoss: -0.0043 | Ent: 0.1561
+[ 52.9%] Iter 5164/9765 | SPS:  2268 | Reward:  0.057 | Value:  0.617 | VLoss: 0.0229 | PLoss: -0.0115 | Ent: 0.2568
+[ 52.9%] Iter 5165/9765 | SPS:  2269 | Reward:  0.062 | Value:  0.748 | VLoss: 0.0214 | PLoss: -0.0076 | Ent: 0.1986
+[ 52.9%] Iter 5166/9765 | SPS:  2269 | Reward:  0.070 | Value:  0.725 | VLoss: 0.0215 | PLoss: -0.0107 | Ent: 0.2062
+[ 52.9%] Iter 5167/9765 | SPS:  2269 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0274 | PLoss: -0.0047 | Ent: 0.1771
+[ 52.9%] Iter 5168/9765 | SPS:  2269 | Reward:  0.072 | Value:  0.710 | VLoss: 0.0354 | PLoss: -0.0073 | Ent: 0.1838
+[ 52.9%] Iter 5169/9765 | SPS:  2269 | Reward:  0.061 | Value:  0.776 | VLoss: 0.0261 | PLoss: -0.0037 | Ent: 0.1853
+[ 52.9%] Iter 5170/9765 | SPS:  2269 | Reward:  0.076 | Value:  0.717 | VLoss: 0.0213 | PLoss: -0.0096 | Ent: 0.2355
+[ 53.0%] Iter 5171/9765 | SPS:  2270 | Reward:  0.068 | Value:  0.734 | VLoss: 0.0276 | PLoss: -0.0129 | Ent: 0.1632
+[ 53.0%] Iter 5172/9765 | SPS:  2270 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0173 | PLoss: -0.0003 | Ent: 0.1865
+[ 53.0%] Iter 5173/9765 | SPS:  2270 | Reward:  0.071 | Value:  0.768 | VLoss: 0.0201 | PLoss: -0.0075 | Ent: 0.1761
+[ 53.0%] Iter 5174/9765 | SPS:  2270 | Reward:  0.054 | Value:  0.721 | VLoss: 0.0304 | PLoss: -0.0059 | Ent: 0.1808
+[ 53.0%] Iter 5175/9765 | SPS:  2270 | Reward:  0.045 | Value:  0.703 | VLoss: 0.0209 | PLoss: -0.0147 | Ent: 0.1792
+[ 53.0%] Iter 5176/9765 | SPS:  2270 | Reward:  0.049 | Value:  0.649 | VLoss: 0.0165 | PLoss: -0.0116 | Ent: 0.1689
+[ 53.0%] Iter 5177/9765 | SPS:  2271 | Reward:  0.058 | Value:  0.663 | VLoss: 0.0316 | PLoss: -0.0069 | Ent: 0.2137
+[ 53.0%] Iter 5178/9765 | SPS:  2271 | Reward:  0.077 | Value:  0.726 | VLoss: 0.0297 | PLoss: -0.0188 | Ent: 0.2026
+[ 53.0%] Iter 5179/9765 | SPS:  2271 | Reward:  0.057 | Value:  0.700 | VLoss: 0.0327 | PLoss: -0.0088 | Ent: 0.2103
+[ 53.0%] Iter 5180/9765 | SPS:  2271 | Reward:  0.047 | Value:  0.670 | VLoss: 0.0224 | PLoss: -0.0053 | Ent: 0.1555
+[ 53.1%] Iter 5181/9765 | SPS:  2271 | Reward:  0.065 | Value:  0.689 | VLoss: 0.0310 | PLoss: -0.0085 | Ent: 0.1686
+[ 53.1%] Iter 5182/9765 | SPS:  2271 | Reward:  0.066 | Value:  0.660 | VLoss: 0.0188 | PLoss: -0.0094 | Ent: 0.1776
+[ 53.1%] Iter 5183/9765 | SPS:  2272 | Reward:  0.071 | Value:  0.763 | VLoss: 0.0260 | PLoss: -0.0185 | Ent: 0.1742
+[ 53.1%] Iter 5184/9765 | SPS:  2272 | Reward:  0.058 | Value:  0.771 | VLoss: 0.0207 | PLoss: -0.0124 | Ent: 0.1879
+[ 53.1%] Iter 5185/9765 | SPS:  2272 | Reward:  0.055 | Value:  0.628 | VLoss: 0.0280 | PLoss: -0.0060 | Ent: 0.1687
+[ 53.1%] Iter 5186/9765 | SPS:  2272 | Reward:  0.059 | Value:  0.728 | VLoss: 0.0224 | PLoss: -0.0091 | Ent: 0.1822
+[ 53.1%] Iter 5187/9765 | SPS:  2272 | Reward:  0.074 | Value:  0.717 | VLoss: 0.0228 | PLoss: -0.0117 | Ent: 0.2118
+[ 53.1%] Iter 5188/9765 | SPS:  2272 | Reward:  0.065 | Value:  0.765 | VLoss: 0.0231 | PLoss: -0.0141 | Ent: 0.1670
+[ 53.1%] Iter 5189/9765 | SPS:  2273 | Reward:  0.068 | Value:  0.747 | VLoss: 0.0168 | PLoss: -0.0049 | Ent: 0.1595
+[ 53.1%] Iter 5190/9765 | SPS:  2273 | Reward:  0.061 | Value:  0.792 | VLoss: 0.0191 | PLoss: -0.0176 | Ent: 0.1718
+[ 53.2%] Iter 5191/9765 | SPS:  2273 | Reward:  0.058 | Value:  0.729 | VLoss: 0.0217 | PLoss: -0.0045 | Ent: 0.1715
+[ 53.2%] Iter 5192/9765 | SPS:  2273 | Reward:  0.065 | Value:  0.647 | VLoss: 0.0241 | PLoss: -0.0031 | Ent: 0.2006
+[ 53.2%] Iter 5193/9765 | SPS:  2273 | Reward:  0.061 | Value:  0.778 | VLoss: 0.0185 | PLoss: -0.0087 | Ent: 0.1633
+[ 53.2%] Iter 5194/9765 | SPS:  2273 | Reward:  0.064 | Value:  0.724 | VLoss: 0.0187 | PLoss: -0.0067 | Ent: 0.1546
+[ 53.2%] Iter 5195/9765 | SPS:  2274 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0225 | PLoss: -0.0053 | Ent: 0.1520
+[ 53.2%] Iter 5196/9765 | SPS:  2274 | Reward:  0.063 | Value:  0.727 | VLoss: 0.0270 | PLoss: -0.0138 | Ent: 0.2163
+[ 53.2%] Iter 5197/9765 | SPS:  2274 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0195 | PLoss: -0.0092 | Ent: 0.1822
+[ 53.2%] Iter 5198/9765 | SPS:  2274 | Reward:  0.066 | Value:  0.761 | VLoss: 0.0141 | PLoss: -0.0045 | Ent: 0.2139
+[ 53.2%] Iter 5199/9765 | SPS:  2274 | Reward:  0.070 | Value:  0.712 | VLoss: 0.0245 | PLoss: -0.0067 | Ent: 0.1679
+[ 53.3%] Iter 5200/9765 | SPS:  2274 | Reward:  0.069 | Value:  0.742 | VLoss: 0.0231 | PLoss: -0.0097 | Ent: 0.2421
+[ 53.3%] Iter 5201/9765 | SPS:  2275 | Reward:  0.067 | Value:  0.718 | VLoss: 0.0227 | PLoss: -0.0126 | Ent: 0.1673
+[ 53.3%] Iter 5202/9765 | SPS:  2275 | Reward:  0.067 | Value:  0.703 | VLoss: 0.0229 | PLoss: -0.0031 | Ent: 0.1933
+[ 53.3%] Iter 5203/9765 | SPS:  2275 | Reward:  0.066 | Value:  0.699 | VLoss: 0.0237 | PLoss: -0.0127 | Ent: 0.2096
+[ 53.3%] Iter 5204/9765 | SPS:  2275 | Reward:  0.062 | Value:  0.780 | VLoss: 0.0228 | PLoss: -0.0105 | Ent: 0.1733
+[ 53.3%] Iter 5205/9765 | SPS:  2275 | Reward:  0.060 | Value:  0.659 | VLoss: 0.0235 | PLoss: -0.0148 | Ent: 0.2227
+[ 53.3%] Iter 5206/9765 | SPS:  2275 | Reward:  0.064 | Value:  0.732 | VLoss: 0.0191 | PLoss: -0.0029 | Ent: 0.1587
+[ 53.3%] Iter 5207/9765 | SPS:  2276 | Reward:  0.055 | Value:  0.666 | VLoss: 0.0324 | PLoss: -0.0116 | Ent: 0.2181
+[ 53.3%] Iter 5208/9765 | SPS:  2276 | Reward:  0.067 | Value:  0.705 | VLoss: 0.0237 | PLoss: -0.0112 | Ent: 0.1772
+[ 53.3%] Iter 5209/9765 | SPS:  2276 | Reward:  0.073 | Value:  0.736 | VLoss: 0.0235 | PLoss: -0.0037 | Ent: 0.2093
+[ 53.4%] Iter 5210/9765 | SPS:  2276 | Reward:  0.059 | Value:  0.753 | VLoss: 0.0261 | PLoss: -0.0042 | Ent: 0.1826
+[ 53.4%] Iter 5211/9765 | SPS:  2276 | Reward:  0.055 | Value:  0.712 | VLoss: 0.0247 | PLoss: -0.0039 | Ent: 0.2030
+[ 53.4%] Iter 5212/9765 | SPS:  2276 | Reward:  0.068 | Value:  0.684 | VLoss: 0.0308 | PLoss: -0.0137 | Ent: 0.2096
+[ 53.4%] Iter 5213/9765 | SPS:  2276 | Reward:  0.080 | Value:  0.749 | VLoss: 0.0295 | PLoss: -0.0244 | Ent: 0.1653
+[ 53.4%] Iter 5214/9765 | SPS:  2277 | Reward:  0.061 | Value:  0.742 | VLoss: 0.0268 | PLoss: 0.0028 | Ent: 0.1527
+[ 53.4%] Iter 5215/9765 | SPS:  2277 | Reward:  0.051 | Value:  0.771 | VLoss: 0.0201 | PLoss: -0.0068 | Ent: 0.1856
+[ 53.4%] Iter 5216/9765 | SPS:  2277 | Reward:  0.067 | Value:  0.640 | VLoss: 0.0206 | PLoss: -0.0039 | Ent: 0.1721
+[ 53.4%] Iter 5217/9765 | SPS:  2277 | Reward:  0.069 | Value:  0.729 | VLoss: 0.0422 | PLoss: -0.0153 | Ent: 0.2143
+[ 53.4%] Iter 5218/9765 | SPS:  2277 | Reward:  0.068 | Value:  0.724 | VLoss: 0.0283 | PLoss: -0.0077 | Ent: 0.1483
+[ 53.4%] Iter 5219/9765 | SPS:  2277 | Reward:  0.063 | Value:  0.738 | VLoss: 0.0267 | PLoss: -0.0060 | Ent: 0.1858
+[ 53.5%] Iter 5220/9765 | SPS:  2278 | Reward:  0.056 | Value:  0.719 | VLoss: 0.0240 | PLoss: -0.0124 | Ent: 0.1810
+[ 53.5%] Iter 5221/9765 | SPS:  2278 | Reward:  0.059 | Value:  0.756 | VLoss: 0.0241 | PLoss: -0.0121 | Ent: 0.2033
+[ 53.5%] Iter 5222/9765 | SPS:  2278 | Reward:  0.056 | Value:  0.736 | VLoss: 0.0260 | PLoss: -0.0105 | Ent: 0.1871
+[ 53.5%] Iter 5223/9765 | SPS:  2278 | Reward:  0.060 | Value:  0.767 | VLoss: 0.0241 | PLoss: -0.0148 | Ent: 0.1993
+[ 53.5%] Iter 5224/9765 | SPS:  2278 | Reward:  0.057 | Value:  0.761 | VLoss: 0.0239 | PLoss: -0.0074 | Ent: 0.1883
+[ 53.5%] Iter 5225/9765 | SPS:  2278 | Reward:  0.056 | Value:  0.731 | VLoss: 0.0222 | PLoss: -0.0053 | Ent: 0.1840
+[ 53.5%] Iter 5226/9765 | SPS:  2279 | Reward:  0.065 | Value:  0.695 | VLoss: 0.0241 | PLoss: -0.0053 | Ent: 0.1708
+[ 53.5%] Iter 5227/9765 | SPS:  2279 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0265 | PLoss: -0.0096 | Ent: 0.2068
+[ 53.5%] Iter 5228/9765 | SPS:  2279 | Reward:  0.065 | Value:  0.706 | VLoss: 0.0211 | PLoss: -0.0047 | Ent: 0.1699
+[ 53.5%] Iter 5229/9765 | SPS:  2279 | Reward:  0.061 | Value:  0.703 | VLoss: 0.0302 | PLoss: -0.0097 | Ent: 0.1877
+[ 53.6%] Iter 5230/9765 | SPS:  2279 | Reward:  0.059 | Value:  0.716 | VLoss: 0.0218 | PLoss: -0.0040 | Ent: 0.2063
+[ 53.6%] Iter 5231/9765 | SPS:  2279 | Reward:  0.058 | Value:  0.646 | VLoss: 0.0311 | PLoss: -0.0047 | Ent: 0.2053
+[ 53.6%] Iter 5232/9765 | SPS:  2280 | Reward:  0.062 | Value:  0.651 | VLoss: 0.0344 | PLoss: -0.0048 | Ent: 0.1884
+[ 53.6%] Iter 5233/9765 | SPS:  2280 | Reward:  0.068 | Value:  0.726 | VLoss: 0.0232 | PLoss: -0.0062 | Ent: 0.1797
+[ 53.6%] Iter 5234/9765 | SPS:  2280 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0252 | PLoss: -0.0102 | Ent: 0.1777
+[ 53.6%] Iter 5235/9765 | SPS:  2280 | Reward:  0.059 | Value:  0.712 | VLoss: 0.0307 | PLoss: -0.0152 | Ent: 0.1562
+[ 53.6%] Iter 5236/9765 | SPS:  2280 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0204 | PLoss: -0.0072 | Ent: 0.1768
+[ 53.6%] Iter 5237/9765 | SPS:  2280 | Reward:  0.064 | Value:  0.692 | VLoss: 0.0321 | PLoss: -0.0102 | Ent: 0.1967
+[ 53.6%] Iter 5238/9765 | SPS:  2281 | Reward:  0.062 | Value:  0.644 | VLoss: 0.0292 | PLoss: -0.0068 | Ent: 0.1877
+[ 53.7%] Iter 5239/9765 | SPS:  2281 | Reward:  0.069 | Value:  0.729 | VLoss: 0.0258 | PLoss: -0.0107 | Ent: 0.1660
+[ 53.7%] Iter 5240/9765 | SPS:  2281 | Reward:  0.056 | Value:  0.759 | VLoss: 0.0310 | PLoss: -0.0071 | Ent: 0.1574
+[ 53.7%] Iter 5241/9765 | SPS:  2281 | Reward:  0.074 | Value:  0.708 | VLoss: 0.0385 | PLoss: -0.0069 | Ent: 0.2007
+[ 53.7%] Iter 5242/9765 | SPS:  2281 | Reward:  0.055 | Value:  0.714 | VLoss: 0.0179 | PLoss: -0.0135 | Ent: 0.2159
+[ 53.7%] Iter 5243/9765 | SPS:  2281 | Reward:  0.068 | Value:  0.711 | VLoss: 0.0253 | PLoss: -0.0130 | Ent: 0.1605
+[ 53.7%] Iter 5244/9765 | SPS:  2282 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0249 | PLoss: -0.0103 | Ent: 0.1626
+[ 53.7%] Iter 5245/9765 | SPS:  2282 | Reward:  0.052 | Value:  0.697 | VLoss: 0.0253 | PLoss: -0.0036 | Ent: 0.2179
+[ 53.7%] Iter 5246/9765 | SPS:  2282 | Reward:  0.074 | Value:  0.649 | VLoss: 0.0215 | PLoss: -0.0045 | Ent: 0.2089
+[ 53.7%] Iter 5247/9765 | SPS:  2282 | Reward:  0.050 | Value:  0.676 | VLoss: 0.0213 | PLoss: -0.0020 | Ent: 0.2002
+[ 53.7%] Iter 5248/9765 | SPS:  2282 | Reward:  0.057 | Value:  0.778 | VLoss: 0.0209 | PLoss: -0.0125 | Ent: 0.1937
+[ 53.8%] Iter 5249/9765 | SPS:  2282 | Reward:  0.052 | Value:  0.737 | VLoss: 0.0245 | PLoss: -0.0104 | Ent: 0.2098
+[ 53.8%] Iter 5250/9765 | SPS:  2283 | Reward:  0.045 | Value:  0.641 | VLoss: 0.0168 | PLoss: -0.0057 | Ent: 0.1905
+[ 53.8%] Iter 5251/9765 | SPS:  2283 | Reward:  0.060 | Value:  0.677 | VLoss: 0.0227 | PLoss: -0.0054 | Ent: 0.1663
+[ 53.8%] Iter 5252/9765 | SPS:  2283 | Reward:  0.055 | Value:  0.635 | VLoss: 0.0283 | PLoss: -0.0010 | Ent: 0.1922
+[ 53.8%] Iter 5253/9765 | SPS:  2283 | Reward:  0.054 | Value:  0.676 | VLoss: 0.0242 | PLoss: -0.0108 | Ent: 0.2093
+[ 53.8%] Iter 5254/9765 | SPS:  2283 | Reward:  0.064 | Value:  0.733 | VLoss: 0.0247 | PLoss: -0.0100 | Ent: 0.2121
+[ 53.8%] Iter 5255/9765 | SPS:  2283 | Reward:  0.072 | Value:  0.710 | VLoss: 0.0274 | PLoss: -0.0005 | Ent: 0.1806
+[ 53.8%] Iter 5256/9765 | SPS:  2284 | Reward:  0.062 | Value:  0.665 | VLoss: 0.0198 | PLoss: -0.0033 | Ent: 0.2191
+[ 53.8%] Iter 5257/9765 | SPS:  2284 | Reward:  0.066 | Value:  0.643 | VLoss: 0.0286 | PLoss: -0.0049 | Ent: 0.2003
+[ 53.8%] Iter 5258/9765 | SPS:  2284 | Reward:  0.062 | Value:  0.724 | VLoss: 0.0258 | PLoss: -0.0110 | Ent: 0.2165
+[ 53.9%] Iter 5259/9765 | SPS:  2284 | Reward:  0.062 | Value:  0.668 | VLoss: 0.0311 | PLoss: -0.0092 | Ent: 0.1603
+[ 53.9%] Iter 5260/9765 | SPS:  2284 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0199 | PLoss: -0.0049 | Ent: 0.1824
+[ 53.9%] Iter 5261/9765 | SPS:  2284 | Reward:  0.046 | Value:  0.694 | VLoss: 0.0164 | PLoss: -0.0081 | Ent: 0.1736
+[ 53.9%] Iter 5262/9765 | SPS:  2285 | Reward:  0.061 | Value:  0.729 | VLoss: 0.0193 | PLoss: -0.0096 | Ent: 0.2080
+[ 53.9%] Iter 5263/9765 | SPS:  2285 | Reward:  0.053 | Value:  0.678 | VLoss: 0.0251 | PLoss: -0.0010 | Ent: 0.1703
+[ 53.9%] Iter 5264/9765 | SPS:  2285 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0251 | PLoss: -0.0109 | Ent: 0.2253
+[ 53.9%] Iter 5265/9765 | SPS:  2285 | Reward:  0.060 | Value:  0.646 | VLoss: 0.0210 | PLoss: -0.0055 | Ent: 0.1589
+[ 53.9%] Iter 5266/9765 | SPS:  2285 | Reward:  0.054 | Value:  0.714 | VLoss: 0.0208 | PLoss: -0.0036 | Ent: 0.1745
+[ 53.9%] Iter 5267/9765 | SPS:  2285 | Reward:  0.046 | Value:  0.612 | VLoss: 0.0280 | PLoss: -0.0093 | Ent: 0.2354
+[ 53.9%] Iter 5268/9765 | SPS:  2286 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0192 | PLoss: -0.0063 | Ent: 0.1888
+[ 54.0%] Iter 5269/9765 | SPS:  2286 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0258 | PLoss: -0.0019 | Ent: 0.1629
+[ 54.0%] Iter 5270/9765 | SPS:  2286 | Reward:  0.057 | Value:  0.714 | VLoss: 0.0182 | PLoss: -0.0023 | Ent: 0.1726
+[ 54.0%] Iter 5271/9765 | SPS:  2286 | Reward:  0.061 | Value:  0.734 | VLoss: 0.0263 | PLoss: -0.0081 | Ent: 0.1648
+[ 54.0%] Iter 5272/9765 | SPS:  2286 | Reward:  0.072 | Value:  0.702 | VLoss: 0.0284 | PLoss: -0.0043 | Ent: 0.1697
+[ 54.0%] Iter 5273/9765 | SPS:  2286 | Reward:  0.070 | Value:  0.660 | VLoss: 0.0296 | PLoss: -0.0074 | Ent: 0.2148
+[ 54.0%] Iter 5274/9765 | SPS:  2287 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0224 | PLoss: -0.0077 | Ent: 0.1772
+[ 54.0%] Iter 5275/9765 | SPS:  2287 | Reward:  0.051 | Value:  0.664 | VLoss: 0.0184 | PLoss: -0.0066 | Ent: 0.1609
+[ 54.0%] Iter 5276/9765 | SPS:  2287 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0287 | PLoss: -0.0173 | Ent: 0.1950
+[ 54.0%] Iter 5277/9765 | SPS:  2287 | Reward:  0.056 | Value:  0.692 | VLoss: 0.0189 | PLoss: -0.0047 | Ent: 0.2107
+[ 54.1%] Iter 5278/9765 | SPS:  2287 | Reward:  0.061 | Value:  0.742 | VLoss: 0.0237 | PLoss: -0.0017 | Ent: 0.1571
+[ 54.1%] Iter 5279/9765 | SPS:  2287 | Reward:  0.053 | Value:  0.703 | VLoss: 0.0268 | PLoss: -0.0116 | Ent: 0.1780
+[ 54.1%] Iter 5280/9765 | SPS:  2288 | Reward:  0.063 | Value:  0.741 | VLoss: 0.0245 | PLoss: -0.0082 | Ent: 0.1757
+[ 54.1%] Iter 5281/9765 | SPS:  2288 | Reward:  0.064 | Value:  0.703 | VLoss: 0.0232 | PLoss: -0.0064 | Ent: 0.2044
+[ 54.1%] Iter 5282/9765 | SPS:  2288 | Reward:  0.075 | Value:  0.751 | VLoss: 0.0250 | PLoss: -0.0001 | Ent: 0.1780
+[ 54.1%] Iter 5283/9765 | SPS:  2288 | Reward:  0.063 | Value:  0.715 | VLoss: 0.0201 | PLoss: -0.0101 | Ent: 0.1653
+[ 54.1%] Iter 5284/9765 | SPS:  2288 | Reward:  0.075 | Value:  0.737 | VLoss: 0.0312 | PLoss: -0.0160 | Ent: 0.1729
+[ 54.1%] Iter 5285/9765 | SPS:  2288 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0222 | PLoss: -0.0075 | Ent: 0.2112
+[ 54.1%] Iter 5286/9765 | SPS:  2289 | Reward:  0.069 | Value:  0.626 | VLoss: 0.0236 | PLoss: -0.0032 | Ent: 0.1558
+[ 54.1%] Iter 5287/9765 | SPS:  2289 | Reward:  0.066 | Value:  0.723 | VLoss: 0.0280 | PLoss: -0.0060 | Ent: 0.1692
+[ 54.2%] Iter 5288/9765 | SPS:  2289 | Reward:  0.059 | Value:  0.736 | VLoss: 0.0222 | PLoss: -0.0117 | Ent: 0.2378
+[ 54.2%] Iter 5289/9765 | SPS:  2289 | Reward:  0.052 | Value:  0.695 | VLoss: 0.0187 | PLoss: -0.0041 | Ent: 0.1854
+[ 54.2%] Iter 5290/9765 | SPS:  2289 | Reward:  0.067 | Value:  0.681 | VLoss: 0.0212 | PLoss: -0.0029 | Ent: 0.2099
+[ 54.2%] Iter 5291/9765 | SPS:  2289 | Reward:  0.069 | Value:  0.791 | VLoss: 0.0227 | PLoss: -0.0094 | Ent: 0.2232
+[ 54.2%] Iter 5292/9765 | SPS:  2290 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0254 | PLoss: -0.0047 | Ent: 0.2384
+[ 54.2%] Iter 5293/9765 | SPS:  2290 | Reward:  0.071 | Value:  0.645 | VLoss: 0.0340 | PLoss: -0.0100 | Ent: 0.2421
+[ 54.2%] Iter 5294/9765 | SPS:  2290 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0291 | PLoss: -0.0103 | Ent: 0.1892
+[ 54.2%] Iter 5295/9765 | SPS:  2290 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0227 | PLoss: -0.0014 | Ent: 0.1692
+[ 54.2%] Iter 5296/9765 | SPS:  2290 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0343 | PLoss: -0.0026 | Ent: 0.1910
+[ 54.2%] Iter 5297/9765 | SPS:  2290 | Reward:  0.072 | Value:  0.712 | VLoss: 0.0281 | PLoss: -0.0023 | Ent: 0.2176
+[ 54.3%] Iter 5298/9765 | SPS:  2290 | Reward:  0.058 | Value:  0.691 | VLoss: 0.0205 | PLoss: -0.0137 | Ent: 0.1869
+[ 54.3%] Iter 5299/9765 | SPS:  2291 | Reward:  0.071 | Value:  0.692 | VLoss: 0.0315 | PLoss: -0.0111 | Ent: 0.2078
+[ 54.3%] Iter 5300/9765 | SPS:  2291 | Reward:  0.058 | Value:  0.683 | VLoss: 0.0251 | PLoss: -0.0062 | Ent: 0.1988
+[ 54.3%] Iter 5301/9765 | SPS:  2291 | Reward:  0.057 | Value:  0.711 | VLoss: 0.0284 | PLoss: -0.0142 | Ent: 0.2012
+[ 54.3%] Iter 5302/9765 | SPS:  2291 | Reward:  0.065 | Value:  0.765 | VLoss: 0.0218 | PLoss: -0.0077 | Ent: 0.1929
+[ 54.3%] Iter 5303/9765 | SPS:  2291 | Reward:  0.076 | Value:  0.736 | VLoss: 0.0228 | PLoss: -0.0069 | Ent: 0.1679
+[ 54.3%] Iter 5304/9765 | SPS:  2291 | Reward:  0.071 | Value:  0.696 | VLoss: 0.0267 | PLoss: -0.0133 | Ent: 0.1948
+[ 54.3%] Iter 5305/9765 | SPS:  2292 | Reward:  0.066 | Value:  0.648 | VLoss: 0.0211 | PLoss: -0.0109 | Ent: 0.1923
+[ 54.3%] Iter 5306/9765 | SPS:  2292 | Reward:  0.051 | Value:  0.720 | VLoss: 0.0228 | PLoss: -0.0066 | Ent: 0.1955
+[ 54.3%] Iter 5307/9765 | SPS:  2292 | Reward:  0.068 | Value:  0.709 | VLoss: 0.0342 | PLoss: -0.0047 | Ent: 0.1820
+[ 54.4%] Iter 5308/9765 | SPS:  2292 | Reward:  0.048 | Value:  0.726 | VLoss: 0.0218 | PLoss: -0.0088 | Ent: 0.1605
+[ 54.4%] Iter 5309/9765 | SPS:  2292 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0283 | PLoss: -0.0081 | Ent: 0.1974
+[ 54.4%] Iter 5310/9765 | SPS:  2292 | Reward:  0.077 | Value:  0.730 | VLoss: 0.0270 | PLoss: -0.0103 | Ent: 0.2196
+[ 54.4%] Iter 5311/9765 | SPS:  2293 | Reward:  0.071 | Value:  0.749 | VLoss: 0.0237 | PLoss: -0.0091 | Ent: 0.2018
+[ 54.4%] Iter 5312/9765 | SPS:  2293 | Reward:  0.061 | Value:  0.721 | VLoss: 0.0197 | PLoss: -0.0062 | Ent: 0.2064
+[ 54.4%] Iter 5313/9765 | SPS:  2293 | Reward:  0.050 | Value:  0.705 | VLoss: 0.0200 | PLoss: -0.0067 | Ent: 0.1824
+[ 54.4%] Iter 5314/9765 | SPS:  2293 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0265 | PLoss: -0.0063 | Ent: 0.1827
+[ 54.4%] Iter 5315/9765 | SPS:  2293 | Reward:  0.066 | Value:  0.771 | VLoss: 0.0279 | PLoss: -0.0141 | Ent: 0.2025
+[ 54.4%] Iter 5316/9765 | SPS:  2293 | Reward:  0.070 | Value:  0.699 | VLoss: 0.0166 | PLoss: -0.0034 | Ent: 0.1832
+[ 54.4%] Iter 5317/9765 | SPS:  2294 | Reward:  0.045 | Value:  0.637 | VLoss: 0.0297 | PLoss: -0.0055 | Ent: 0.2323
+[ 54.5%] Iter 5318/9765 | SPS:  2294 | Reward:  0.072 | Value:  0.664 | VLoss: 0.0280 | PLoss: -0.0131 | Ent: 0.1902
+[ 54.5%] Iter 5319/9765 | SPS:  2294 | Reward:  0.069 | Value:  0.764 | VLoss: 0.0252 | PLoss: -0.0083 | Ent: 0.1896
+[ 54.5%] Iter 5320/9765 | SPS:  2294 | Reward:  0.071 | Value:  0.719 | VLoss: 0.0256 | PLoss: -0.0110 | Ent: 0.2045
+[ 54.5%] Iter 5321/9765 | SPS:  2294 | Reward:  0.068 | Value:  0.686 | VLoss: 0.0273 | PLoss: -0.0014 | Ent: 0.1713
+[ 54.5%] Iter 5322/9765 | SPS:  2294 | Reward:  0.058 | Value:  0.718 | VLoss: 0.0157 | PLoss: -0.0152 | Ent: 0.2586
+[ 54.5%] Iter 5323/9765 | SPS:  2295 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0281 | PLoss: -0.0080 | Ent: 0.2235
+[ 54.5%] Iter 5324/9765 | SPS:  2295 | Reward:  0.062 | Value:  0.662 | VLoss: 0.0260 | PLoss: -0.0018 | Ent: 0.1836
+[ 54.5%] Iter 5325/9765 | SPS:  2295 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0229 | PLoss: 0.0011 | Ent: 0.1594
+[ 54.5%] Iter 5326/9765 | SPS:  2295 | Reward:  0.056 | Value:  0.634 | VLoss: 0.0255 | PLoss: -0.0058 | Ent: 0.1800
+[ 54.6%] Iter 5327/9765 | SPS:  2295 | Reward:  0.063 | Value:  0.645 | VLoss: 0.0180 | PLoss: -0.0127 | Ent: 0.1746
+[ 54.6%] Iter 5328/9765 | SPS:  2295 | Reward:  0.060 | Value:  0.716 | VLoss: 0.0223 | PLoss: -0.0008 | Ent: 0.1234
+[ 54.6%] Iter 5329/9765 | SPS:  2296 | Reward:  0.051 | Value:  0.635 | VLoss: 0.0243 | PLoss: -0.0091 | Ent: 0.1913
+[ 54.6%] Iter 5330/9765 | SPS:  2296 | Reward:  0.050 | Value:  0.609 | VLoss: 0.0170 | PLoss: -0.0058 | Ent: 0.1637
+[ 54.6%] Iter 5331/9765 | SPS:  2296 | Reward:  0.051 | Value:  0.694 | VLoss: 0.0180 | PLoss: -0.0021 | Ent: 0.1858
+[ 54.6%] Iter 5332/9765 | SPS:  2296 | Reward:  0.051 | Value:  0.575 | VLoss: 0.0330 | PLoss: -0.0068 | Ent: 0.1666
+[ 54.6%] Iter 5333/9765 | SPS:  2296 | Reward:  0.056 | Value:  0.645 | VLoss: 0.0163 | PLoss: -0.0082 | Ent: 0.1681
+[ 54.6%] Iter 5334/9765 | SPS:  2296 | Reward:  0.050 | Value:  0.675 | VLoss: 0.0172 | PLoss: -0.0126 | Ent: 0.1985
+[ 54.6%] Iter 5335/9765 | SPS:  2297 | Reward:  0.056 | Value:  0.658 | VLoss: 0.0171 | PLoss: -0.0017 | Ent: 0.1714
+[ 54.6%] Iter 5336/9765 | SPS:  2297 | Reward:  0.053 | Value:  0.570 | VLoss: 0.0234 | PLoss: -0.0035 | Ent: 0.2134
+[ 54.7%] Iter 5337/9765 | SPS:  2297 | Reward:  0.066 | Value:  0.559 | VLoss: 0.0293 | PLoss: -0.0070 | Ent: 0.1728
+[ 54.7%] Iter 5338/9765 | SPS:  2297 | Reward:  0.055 | Value:  0.553 | VLoss: 0.0208 | PLoss: -0.0073 | Ent: 0.2489
+[ 54.7%] Iter 5339/9765 | SPS:  2297 | Reward:  0.067 | Value:  0.672 | VLoss: 0.0237 | PLoss: -0.0061 | Ent: 0.1708
+[ 54.7%] Iter 5340/9765 | SPS:  2297 | Reward:  0.070 | Value:  0.749 | VLoss: 0.0305 | PLoss: -0.0049 | Ent: 0.1816
+[ 54.7%] Iter 5341/9765 | SPS:  2298 | Reward:  0.053 | Value:  0.758 | VLoss: 0.0166 | PLoss: -0.0083 | Ent: 0.2165
+[ 54.7%] Iter 5342/9765 | SPS:  2298 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0172 | PLoss: -0.0067 | Ent: 0.1656
+[ 54.7%] Iter 5343/9765 | SPS:  2298 | Reward:  0.055 | Value:  0.686 | VLoss: 0.0225 | PLoss: -0.0076 | Ent: 0.1520
+[ 54.7%] Iter 5344/9765 | SPS:  2298 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0167 | PLoss: -0.0013 | Ent: 0.1742
+[ 54.7%] Iter 5345/9765 | SPS:  2298 | Reward:  0.072 | Value:  0.709 | VLoss: 0.0280 | PLoss: -0.0071 | Ent: 0.2578
+[ 54.7%] Iter 5346/9765 | SPS:  2298 | Reward:  0.049 | Value:  0.732 | VLoss: 0.0165 | PLoss: -0.0127 | Ent: 0.1524
+[ 54.8%] Iter 5347/9765 | SPS:  2299 | Reward:  0.053 | Value:  0.754 | VLoss: 0.0180 | PLoss: -0.0113 | Ent: 0.1833
+[ 54.8%] Iter 5348/9765 | SPS:  2299 | Reward:  0.057 | Value:  0.712 | VLoss: 0.0230 | PLoss: -0.0065 | Ent: 0.1440
+[ 54.8%] Iter 5349/9765 | SPS:  2299 | Reward:  0.042 | Value:  0.683 | VLoss: 0.0212 | PLoss: -0.0034 | Ent: 0.1721
+[ 54.8%] Iter 5350/9765 | SPS:  2299 | Reward:  0.062 | Value:  0.611 | VLoss: 0.0277 | PLoss: -0.0096 | Ent: 0.1904
+[ 54.8%] Iter 5351/9765 | SPS:  2299 | Reward:  0.072 | Value:  0.681 | VLoss: 0.0270 | PLoss: -0.0049 | Ent: 0.1849
+[ 54.8%] Iter 5352/9765 | SPS:  2299 | Reward:  0.075 | Value:  0.772 | VLoss: 0.0248 | PLoss: -0.0047 | Ent: 0.2009
+[ 54.8%] Iter 5353/9765 | SPS:  2300 | Reward:  0.064 | Value:  0.779 | VLoss: 0.0235 | PLoss: -0.0102 | Ent: 0.2047
+[ 54.8%] Iter 5354/9765 | SPS:  2300 | Reward:  0.052 | Value:  0.658 | VLoss: 0.0192 | PLoss: -0.0048 | Ent: 0.2231
+[ 54.8%] Iter 5355/9765 | SPS:  2300 | Reward:  0.061 | Value:  0.759 | VLoss: 0.0257 | PLoss: -0.0062 | Ent: 0.1785
+[ 54.8%] Iter 5356/9765 | SPS:  2300 | Reward:  0.044 | Value:  0.770 | VLoss: 0.0178 | PLoss: -0.0104 | Ent: 0.1738
+[ 54.9%] Iter 5357/9765 | SPS:  2300 | Reward:  0.067 | Value:  0.689 | VLoss: 0.0219 | PLoss: -0.0031 | Ent: 0.1656
+[ 54.9%] Iter 5358/9765 | SPS:  2300 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0192 | PLoss: -0.0085 | Ent: 0.1500
+[ 54.9%] Iter 5359/9765 | SPS:  2301 | Reward:  0.051 | Value:  0.694 | VLoss: 0.0216 | PLoss: -0.0071 | Ent: 0.1254
+[ 54.9%] Iter 5360/9765 | SPS:  2301 | Reward:  0.061 | Value:  0.733 | VLoss: 0.0201 | PLoss: -0.0028 | Ent: 0.1384
+[ 54.9%] Iter 5361/9765 | SPS:  2301 | Reward:  0.063 | Value:  0.752 | VLoss: 0.0356 | PLoss: -0.0041 | Ent: 0.1711
+[ 54.9%] Iter 5362/9765 | SPS:  2301 | Reward:  0.049 | Value:  0.712 | VLoss: 0.0184 | PLoss: -0.0019 | Ent: 0.1882
+[ 54.9%] Iter 5363/9765 | SPS:  2301 | Reward:  0.071 | Value:  0.650 | VLoss: 0.0369 | PLoss: -0.0134 | Ent: 0.1982
+[ 54.9%] Iter 5364/9765 | SPS:  2301 | Reward:  0.068 | Value:  0.713 | VLoss: 0.0275 | PLoss: -0.0071 | Ent: 0.1823
+[ 54.9%] Iter 5365/9765 | SPS:  2302 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0249 | PLoss: -0.0068 | Ent: 0.1387
+[ 55.0%] Iter 5366/9765 | SPS:  2302 | Reward:  0.054 | Value:  0.723 | VLoss: 0.0168 | PLoss: -0.0097 | Ent: 0.1691
+[ 55.0%] Iter 5367/9765 | SPS:  2302 | Reward:  0.060 | Value:  0.703 | VLoss: 0.0284 | PLoss: -0.0107 | Ent: 0.1724
+[ 55.0%] Iter 5368/9765 | SPS:  2302 | Reward:  0.065 | Value:  0.766 | VLoss: 0.0178 | PLoss: -0.0109 | Ent: 0.1550
+[ 55.0%] Iter 5369/9765 | SPS:  2302 | Reward:  0.055 | Value:  0.676 | VLoss: 0.0295 | PLoss: -0.0033 | Ent: 0.2098
+[ 55.0%] Iter 5370/9765 | SPS:  2302 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0226 | PLoss: -0.0060 | Ent: 0.2251
+[ 55.0%] Iter 5371/9765 | SPS:  2302 | Reward:  0.064 | Value:  0.701 | VLoss: 0.0244 | PLoss: -0.0167 | Ent: 0.2288
+[ 55.0%] Iter 5372/9765 | SPS:  2303 | Reward:  0.063 | Value:  0.785 | VLoss: 0.0207 | PLoss: -0.0045 | Ent: 0.2219
+[ 55.0%] Iter 5373/9765 | SPS:  2303 | Reward:  0.074 | Value:  0.757 | VLoss: 0.0281 | PLoss: -0.0061 | Ent: 0.1745
+[ 55.0%] Iter 5374/9765 | SPS:  2303 | Reward:  0.063 | Value:  0.754 | VLoss: 0.0162 | PLoss: -0.0091 | Ent: 0.2050
+[ 55.0%] Iter 5375/9765 | SPS:  2303 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0271 | PLoss: -0.0072 | Ent: 0.1737
+[ 55.1%] Iter 5376/9765 | SPS:  2303 | Reward:  0.067 | Value:  0.730 | VLoss: 0.0254 | PLoss: -0.0090 | Ent: 0.2042
+[ 55.1%] Iter 5377/9765 | SPS:  2303 | Reward:  0.072 | Value:  0.716 | VLoss: 0.0256 | PLoss: -0.0095 | Ent: 0.1860
+[ 55.1%] Iter 5378/9765 | SPS:  2304 | Reward:  0.062 | Value:  0.751 | VLoss: 0.0323 | PLoss: -0.0140 | Ent: 0.2306
+[ 55.1%] Iter 5379/9765 | SPS:  2304 | Reward:  0.073 | Value:  0.668 | VLoss: 0.0251 | PLoss: -0.0060 | Ent: 0.1987
+[ 55.1%] Iter 5380/9765 | SPS:  2304 | Reward:  0.064 | Value:  0.773 | VLoss: 0.0197 | PLoss: -0.0160 | Ent: 0.1740
+[ 55.1%] Iter 5381/9765 | SPS:  2304 | Reward:  0.057 | Value:  0.665 | VLoss: 0.0249 | PLoss: -0.0075 | Ent: 0.1969
+[ 55.1%] Iter 5382/9765 | SPS:  2304 | Reward:  0.065 | Value:  0.688 | VLoss: 0.0228 | PLoss: -0.0050 | Ent: 0.1957
+[ 55.1%] Iter 5383/9765 | SPS:  2304 | Reward:  0.069 | Value:  0.745 | VLoss: 0.0238 | PLoss: -0.0119 | Ent: 0.1872
+[ 55.1%] Iter 5384/9765 | SPS:  2305 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0326 | PLoss: -0.0085 | Ent: 0.1776
+[ 55.1%] Iter 5385/9765 | SPS:  2305 | Reward:  0.058 | Value:  0.715 | VLoss: 0.0311 | PLoss: -0.0097 | Ent: 0.1774
+[ 55.2%] Iter 5386/9765 | SPS:  2305 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0246 | PLoss: -0.0020 | Ent: 0.2014
+[ 55.2%] Iter 5387/9765 | SPS:  2305 | Reward:  0.057 | Value:  0.712 | VLoss: 0.0196 | PLoss: -0.0144 | Ent: 0.1586
+[ 55.2%] Iter 5388/9765 | SPS:  2305 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0241 | PLoss: -0.0057 | Ent: 0.2159
+[ 55.2%] Iter 5389/9765 | SPS:  2305 | Reward:  0.074 | Value:  0.622 | VLoss: 0.0300 | PLoss: -0.0096 | Ent: 0.2013
+[ 55.2%] Iter 5390/9765 | SPS:  2306 | Reward:  0.057 | Value:  0.672 | VLoss: 0.0314 | PLoss: -0.0028 | Ent: 0.1688
+[ 55.2%] Iter 5391/9765 | SPS:  2306 | Reward:  0.066 | Value:  0.699 | VLoss: 0.0260 | PLoss: -0.0081 | Ent: 0.1947
+[ 55.2%] Iter 5392/9765 | SPS:  2306 | Reward:  0.069 | Value:  0.715 | VLoss: 0.0229 | PLoss: -0.0049 | Ent: 0.1847
+[ 55.2%] Iter 5393/9765 | SPS:  2306 | Reward:  0.055 | Value:  0.744 | VLoss: 0.0229 | PLoss: -0.0157 | Ent: 0.2123
+[ 55.2%] Iter 5394/9765 | SPS:  2306 | Reward:  0.073 | Value:  0.723 | VLoss: 0.0217 | PLoss: -0.0107 | Ent: 0.1662
+[ 55.2%] Iter 5395/9765 | SPS:  2306 | Reward:  0.050 | Value:  0.687 | VLoss: 0.0253 | PLoss: -0.0147 | Ent: 0.1795
+[ 55.3%] Iter 5396/9765 | SPS:  2306 | Reward:  0.066 | Value:  0.642 | VLoss: 0.0220 | PLoss: -0.0045 | Ent: 0.1689
+[ 55.3%] Iter 5397/9765 | SPS:  2307 | Reward:  0.067 | Value:  0.853 | VLoss: 0.0221 | PLoss: -0.0074 | Ent: 0.1603
+[ 55.3%] Iter 5398/9765 | SPS:  2307 | Reward:  0.068 | Value:  0.789 | VLoss: 0.0219 | PLoss: -0.0055 | Ent: 0.1731
+[ 55.3%] Iter 5399/9765 | SPS:  2307 | Reward:  0.076 | Value:  0.723 | VLoss: 0.0174 | PLoss: -0.0039 | Ent: 0.2175
+[ 55.3%] Iter 5400/9765 | SPS:  2307 | Reward:  0.059 | Value:  0.757 | VLoss: 0.0262 | PLoss: -0.0032 | Ent: 0.1901
+[ 55.3%] Iter 5401/9765 | SPS:  2307 | Reward:  0.057 | Value:  0.759 | VLoss: 0.0248 | PLoss: -0.0080 | Ent: 0.1963
+[ 55.3%] Iter 5402/9765 | SPS:  2307 | Reward:  0.065 | Value:  0.692 | VLoss: 0.0284 | PLoss: -0.0089 | Ent: 0.1878
+[ 55.3%] Iter 5403/9765 | SPS:  2308 | Reward:  0.056 | Value:  0.651 | VLoss: 0.0240 | PLoss: -0.0108 | Ent: 0.1648
+[ 55.3%] Iter 5404/9765 | SPS:  2308 | Reward:  0.063 | Value:  0.755 | VLoss: 0.0225 | PLoss: -0.0038 | Ent: 0.2063
+[ 55.4%] Iter 5405/9765 | SPS:  2308 | Reward:  0.053 | Value:  0.751 | VLoss: 0.0253 | PLoss: -0.0089 | Ent: 0.2178
+[ 55.4%] Iter 5406/9765 | SPS:  2308 | Reward:  0.065 | Value:  0.704 | VLoss: 0.0255 | PLoss: -0.0013 | Ent: 0.1903
+[ 55.4%] Iter 5407/9765 | SPS:  2308 | Reward:  0.064 | Value:  0.762 | VLoss: 0.0191 | PLoss: -0.0077 | Ent: 0.2103
+[ 55.4%] Iter 5408/9765 | SPS:  2308 | Reward:  0.056 | Value:  0.698 | VLoss: 0.0185 | PLoss: -0.0074 | Ent: 0.1783
+[ 55.4%] Iter 5409/9765 | SPS:  2309 | Reward:  0.057 | Value:  0.679 | VLoss: 0.0261 | PLoss: -0.0105 | Ent: 0.2334
+[ 55.4%] Iter 5410/9765 | SPS:  2309 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0276 | PLoss: -0.0022 | Ent: 0.1862
+[ 55.4%] Iter 5411/9765 | SPS:  2309 | Reward:  0.047 | Value:  0.682 | VLoss: 0.0147 | PLoss: -0.0059 | Ent: 0.1753
+[ 55.4%] Iter 5412/9765 | SPS:  2309 | Reward:  0.059 | Value:  0.633 | VLoss: 0.0214 | PLoss: -0.0075 | Ent: 0.1775
+[ 55.4%] Iter 5413/9765 | SPS:  2309 | Reward:  0.047 | Value:  0.698 | VLoss: 0.0184 | PLoss: -0.0070 | Ent: 0.1618
+[ 55.4%] Iter 5414/9765 | SPS:  2309 | Reward:  0.062 | Value:  0.750 | VLoss: 0.0212 | PLoss: -0.0046 | Ent: 0.1898
+[ 55.5%] Iter 5415/9765 | SPS:  2310 | Reward:  0.062 | Value:  0.744 | VLoss: 0.0233 | PLoss: -0.0055 | Ent: 0.1904
+[ 55.5%] Iter 5416/9765 | SPS:  2310 | Reward:  0.054 | Value:  0.673 | VLoss: 0.0191 | PLoss: -0.0070 | Ent: 0.1925
+[ 55.5%] Iter 5417/9765 | SPS:  2310 | Reward:  0.069 | Value:  0.699 | VLoss: 0.0235 | PLoss: -0.0062 | Ent: 0.1618
+[ 55.5%] Iter 5418/9765 | SPS:  2310 | Reward:  0.062 | Value:  0.635 | VLoss: 0.0274 | PLoss: -0.0115 | Ent: 0.1857
+[ 55.5%] Iter 5419/9765 | SPS:  2310 | Reward:  0.056 | Value:  0.760 | VLoss: 0.0279 | PLoss: -0.0090 | Ent: 0.2063
+[ 55.5%] Iter 5420/9765 | SPS:  2310 | Reward:  0.069 | Value:  0.718 | VLoss: 0.0276 | PLoss: -0.0006 | Ent: 0.1837
+[ 55.5%] Iter 5421/9765 | SPS:  2310 | Reward:  0.062 | Value:  0.749 | VLoss: 0.0244 | PLoss: -0.0083 | Ent: 0.2068
+[ 55.5%] Iter 5422/9765 | SPS:  2311 | Reward:  0.053 | Value:  0.663 | VLoss: 0.0185 | PLoss: -0.0104 | Ent: 0.1627
+[ 55.5%] Iter 5423/9765 | SPS:  2311 | Reward:  0.066 | Value:  0.676 | VLoss: 0.0256 | PLoss: -0.0016 | Ent: 0.1993
+[ 55.5%] Iter 5424/9765 | SPS:  2311 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0232 | PLoss: -0.0091 | Ent: 0.2279
+[ 55.6%] Iter 5425/9765 | SPS:  2311 | Reward:  0.066 | Value:  0.757 | VLoss: 0.0165 | PLoss: -0.0002 | Ent: 0.1647
+[ 55.6%] Iter 5426/9765 | SPS:  2311 | Reward:  0.059 | Value:  0.733 | VLoss: 0.0270 | PLoss: -0.0060 | Ent: 0.1821
+[ 55.6%] Iter 5427/9765 | SPS:  2311 | Reward:  0.065 | Value:  0.713 | VLoss: 0.0233 | PLoss: -0.0043 | Ent: 0.2025
+[ 55.6%] Iter 5428/9765 | SPS:  2312 | Reward:  0.062 | Value:  0.788 | VLoss: 0.0172 | PLoss: -0.0054 | Ent: 0.1875
+[ 55.6%] Iter 5429/9765 | SPS:  2312 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0185 | PLoss: -0.0068 | Ent: 0.1990
+[ 55.6%] Iter 5430/9765 | SPS:  2312 | Reward:  0.063 | Value:  0.689 | VLoss: 0.0251 | PLoss: -0.0093 | Ent: 0.2035
+[ 55.6%] Iter 5431/9765 | SPS:  2312 | Reward:  0.057 | Value:  0.739 | VLoss: 0.0191 | PLoss: -0.0066 | Ent: 0.1721
+[ 55.6%] Iter 5432/9765 | SPS:  2312 | Reward:  0.051 | Value:  0.681 | VLoss: 0.0182 | PLoss: -0.0115 | Ent: 0.2009
+[ 55.6%] Iter 5433/9765 | SPS:  2312 | Reward:  0.063 | Value:  0.734 | VLoss: 0.0227 | PLoss: -0.0071 | Ent: 0.2086
+[ 55.6%] Iter 5434/9765 | SPS:  2313 | Reward:  0.075 | Value:  0.683 | VLoss: 0.0278 | PLoss: -0.0077 | Ent: 0.1876
+[ 55.7%] Iter 5435/9765 | SPS:  2313 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0198 | PLoss: -0.0052 | Ent: 0.2100
+[ 55.7%] Iter 5436/9765 | SPS:  2313 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0342 | PLoss: -0.0040 | Ent: 0.2042
+[ 55.7%] Iter 5437/9765 | SPS:  2313 | Reward:  0.062 | Value:  0.723 | VLoss: 0.0294 | PLoss: -0.0056 | Ent: 0.1604
+[ 55.7%] Iter 5438/9765 | SPS:  2313 | Reward:  0.062 | Value:  0.626 | VLoss: 0.0225 | PLoss: -0.0027 | Ent: 0.1794
+[ 55.7%] Iter 5439/9765 | SPS:  2313 | Reward:  0.050 | Value:  0.711 | VLoss: 0.0225 | PLoss: -0.0077 | Ent: 0.1818
+[ 55.7%] Iter 5440/9765 | SPS:  2313 | Reward:  0.050 | Value:  0.642 | VLoss: 0.0216 | PLoss: -0.0062 | Ent: 0.1892
+[ 55.7%] Iter 5441/9765 | SPS:  2314 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0277 | PLoss: -0.0054 | Ent: 0.2022
+[ 55.7%] Iter 5442/9765 | SPS:  2314 | Reward:  0.062 | Value:  0.752 | VLoss: 0.0289 | PLoss: -0.0048 | Ent: 0.2092
+[ 55.7%] Iter 5443/9765 | SPS:  2314 | Reward:  0.060 | Value:  0.701 | VLoss: 0.0294 | PLoss: -0.0011 | Ent: 0.2798
+[ 55.8%] Iter 5444/9765 | SPS:  2314 | Reward:  0.062 | Value:  0.630 | VLoss: 0.0203 | PLoss: -0.0087 | Ent: 0.1937
+[ 55.8%] Iter 5445/9765 | SPS:  2314 | Reward:  0.060 | Value:  0.645 | VLoss: 0.0285 | PLoss: -0.0041 | Ent: 0.1844
+[ 55.8%] Iter 5446/9765 | SPS:  2314 | Reward:  0.053 | Value:  0.723 | VLoss: 0.0193 | PLoss: -0.0089 | Ent: 0.1809
+[ 55.8%] Iter 5447/9765 | SPS:  2315 | Reward:  0.058 | Value:  0.696 | VLoss: 0.0188 | PLoss: -0.0021 | Ent: 0.2071
+[ 55.8%] Iter 5448/9765 | SPS:  2315 | Reward:  0.052 | Value:  0.693 | VLoss: 0.0243 | PLoss: -0.0036 | Ent: 0.1646
+[ 55.8%] Iter 5449/9765 | SPS:  2315 | Reward:  0.055 | Value:  0.618 | VLoss: 0.0206 | PLoss: -0.0056 | Ent: 0.1718
+[ 55.8%] Iter 5450/9765 | SPS:  2315 | Reward:  0.060 | Value:  0.664 | VLoss: 0.0217 | PLoss: -0.0102 | Ent: 0.2257
+[ 55.8%] Iter 5451/9765 | SPS:  2315 | Reward:  0.052 | Value:  0.750 | VLoss: 0.0230 | PLoss: -0.0050 | Ent: 0.1723
+[ 55.8%] Iter 5452/9765 | SPS:  2315 | Reward:  0.057 | Value:  0.679 | VLoss: 0.0252 | PLoss: -0.0071 | Ent: 0.2037
+[ 55.8%] Iter 5453/9765 | SPS:  2316 | Reward:  0.074 | Value:  0.729 | VLoss: 0.0242 | PLoss: -0.0114 | Ent: 0.1719
+[ 55.9%] Iter 5454/9765 | SPS:  2316 | Reward:  0.058 | Value:  0.690 | VLoss: 0.0269 | PLoss: -0.0103 | Ent: 0.1677
+[ 55.9%] Iter 5455/9765 | SPS:  2316 | Reward:  0.072 | Value:  0.731 | VLoss: 0.0184 | PLoss: -0.0090 | Ent: 0.2092
+[ 55.9%] Iter 5456/9765 | SPS:  2316 | Reward:  0.070 | Value:  0.737 | VLoss: 0.0235 | PLoss: -0.0051 | Ent: 0.2440
+[ 55.9%] Iter 5457/9765 | SPS:  2316 | Reward:  0.061 | Value:  0.743 | VLoss: 0.0279 | PLoss: -0.0109 | Ent: 0.2113
+[ 55.9%] Iter 5458/9765 | SPS:  2316 | Reward:  0.050 | Value:  0.697 | VLoss: 0.0184 | PLoss: 0.0017 | Ent: 0.1343
+[ 55.9%] Iter 5459/9765 | SPS:  2317 | Reward:  0.059 | Value:  0.648 | VLoss: 0.0215 | PLoss: -0.0105 | Ent: 0.2037
+[ 55.9%] Iter 5460/9765 | SPS:  2317 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0231 | PLoss: -0.0051 | Ent: 0.2749
+[ 55.9%] Iter 5461/9765 | SPS:  2317 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0205 | PLoss: -0.0092 | Ent: 0.2034
+[ 55.9%] Iter 5462/9765 | SPS:  2317 | Reward:  0.056 | Value:  0.706 | VLoss: 0.0225 | PLoss: -0.0109 | Ent: 0.2534
+[ 55.9%] Iter 5463/9765 | SPS:  2317 | Reward:  0.055 | Value:  0.754 | VLoss: 0.0234 | PLoss: -0.0093 | Ent: 0.2138
+[ 56.0%] Iter 5464/9765 | SPS:  2317 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0162 | PLoss: -0.0027 | Ent: 0.2039
+[ 56.0%] Iter 5465/9765 | SPS:  2318 | Reward:  0.045 | Value:  0.686 | VLoss: 0.0215 | PLoss: -0.0113 | Ent: 0.1654
+[ 56.0%] Iter 5466/9765 | SPS:  2318 | Reward:  0.070 | Value:  0.714 | VLoss: 0.0242 | PLoss: -0.0078 | Ent: 0.2133
+[ 56.0%] Iter 5467/9765 | SPS:  2318 | Reward:  0.056 | Value:  0.701 | VLoss: 0.0251 | PLoss: -0.0121 | Ent: 0.2492
+[ 56.0%] Iter 5468/9765 | SPS:  2318 | Reward:  0.061 | Value:  0.713 | VLoss: 0.0241 | PLoss: -0.0057 | Ent: 0.1598
+[ 56.0%] Iter 5469/9765 | SPS:  2318 | Reward:  0.060 | Value:  0.731 | VLoss: 0.0161 | PLoss: -0.0043 | Ent: 0.1801
+[ 56.0%] Iter 5470/9765 | SPS:  2318 | Reward:  0.066 | Value:  0.726 | VLoss: 0.0279 | PLoss: -0.0104 | Ent: 0.2014
+[ 56.0%] Iter 5471/9765 | SPS:  2318 | Reward:  0.066 | Value:  0.694 | VLoss: 0.0257 | PLoss: -0.0040 | Ent: 0.1827
+[ 56.0%] Iter 5472/9765 | SPS:  2319 | Reward:  0.065 | Value:  0.675 | VLoss: 0.0259 | PLoss: -0.0050 | Ent: 0.1851
+[ 56.0%] Iter 5473/9765 | SPS:  2319 | Reward:  0.069 | Value:  0.723 | VLoss: 0.0225 | PLoss: -0.0076 | Ent: 0.1583
+[ 56.1%] Iter 5474/9765 | SPS:  2319 | Reward:  0.067 | Value:  0.740 | VLoss: 0.0269 | PLoss: -0.0071 | Ent: 0.1943
+[ 56.1%] Iter 5475/9765 | SPS:  2319 | Reward:  0.066 | Value:  0.735 | VLoss: 0.0203 | PLoss: -0.0072 | Ent: 0.1829
+[ 56.1%] Iter 5476/9765 | SPS:  2319 | Reward:  0.086 | Value:  0.752 | VLoss: 0.0306 | PLoss: -0.0042 | Ent: 0.1439
+[ 56.1%] Iter 5477/9765 | SPS:  2319 | Reward:  0.071 | Value:  0.782 | VLoss: 0.0278 | PLoss: -0.0186 | Ent: 0.1902
+[ 56.1%] Iter 5478/9765 | SPS:  2319 | Reward:  0.065 | Value:  0.735 | VLoss: 0.0204 | PLoss: -0.0030 | Ent: 0.1763
+[ 56.1%] Iter 5479/9765 | SPS:  2320 | Reward:  0.062 | Value:  0.748 | VLoss: 0.0185 | PLoss: -0.0089 | Ent: 0.1957
+[ 56.1%] Iter 5480/9765 | SPS:  2320 | Reward:  0.062 | Value:  0.763 | VLoss: 0.0209 | PLoss: -0.0055 | Ent: 0.1515
+[ 56.1%] Iter 5481/9765 | SPS:  2320 | Reward:  0.063 | Value:  0.725 | VLoss: 0.0221 | PLoss: -0.0012 | Ent: 0.2037
+[ 56.1%] Iter 5482/9765 | SPS:  2320 | Reward:  0.053 | Value:  0.690 | VLoss: 0.0279 | PLoss: -0.0053 | Ent: 0.1660
+[ 56.1%] Iter 5483/9765 | SPS:  2320 | Reward:  0.070 | Value:  0.694 | VLoss: 0.0257 | PLoss: -0.0050 | Ent: 0.1884
+[ 56.2%] Iter 5484/9765 | SPS:  2320 | Reward:  0.061 | Value:  0.725 | VLoss: 0.0258 | PLoss: -0.0023 | Ent: 0.2236
+[ 56.2%] Iter 5485/9765 | SPS:  2321 | Reward:  0.075 | Value:  0.769 | VLoss: 0.0250 | PLoss: -0.0070 | Ent: 0.1897
+[ 56.2%] Iter 5486/9765 | SPS:  2321 | Reward:  0.069 | Value:  0.700 | VLoss: 0.0214 | PLoss: -0.0023 | Ent: 0.2295
+[ 56.2%] Iter 5487/9765 | SPS:  2321 | Reward:  0.073 | Value:  0.768 | VLoss: 0.0225 | PLoss: 0.0030 | Ent: 0.2142
+[ 56.2%] Iter 5488/9765 | SPS:  2321 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0201 | PLoss: 0.0012 | Ent: 0.1613
+[ 56.2%] Iter 5489/9765 | SPS:  2321 | Reward:  0.055 | Value:  0.656 | VLoss: 0.0244 | PLoss: -0.0101 | Ent: 0.2064
+[ 56.2%] Iter 5490/9765 | SPS:  2321 | Reward:  0.059 | Value:  0.703 | VLoss: 0.0212 | PLoss: -0.0017 | Ent: 0.2285
+[ 56.2%] Iter 5491/9765 | SPS:  2321 | Reward:  0.089 | Value:  0.704 | VLoss: 0.0254 | PLoss: -0.0134 | Ent: 0.1957
+[ 56.2%] Iter 5492/9765 | SPS:  2322 | Reward:  0.062 | Value:  0.771 | VLoss: 0.0211 | PLoss: -0.0108 | Ent: 0.2046
+[ 56.3%] Iter 5493/9765 | SPS:  2322 | Reward:  0.059 | Value:  0.747 | VLoss: 0.0235 | PLoss: -0.0105 | Ent: 0.1980
+[ 56.3%] Iter 5494/9765 | SPS:  2322 | Reward:  0.063 | Value:  0.672 | VLoss: 0.0256 | PLoss: -0.0002 | Ent: 0.1977
+[ 56.3%] Iter 5495/9765 | SPS:  2322 | Reward:  0.080 | Value:  0.692 | VLoss: 0.0293 | PLoss: -0.0118 | Ent: 0.2311
+[ 56.3%] Iter 5496/9765 | SPS:  2322 | Reward:  0.056 | Value:  0.765 | VLoss: 0.0225 | PLoss: -0.0168 | Ent: 0.2245
+[ 56.3%] Iter 5497/9765 | SPS:  2322 | Reward:  0.076 | Value:  0.721 | VLoss: 0.0328 | PLoss: -0.0136 | Ent: 0.2180
+[ 56.3%] Iter 5498/9765 | SPS:  2323 | Reward:  0.061 | Value:  0.678 | VLoss: 0.0247 | PLoss: -0.0059 | Ent: 0.1837
+[ 56.3%] Iter 5499/9765 | SPS:  2323 | Reward:  0.052 | Value:  0.716 | VLoss: 0.0257 | PLoss: -0.0062 | Ent: 0.1743
+[ 56.3%] Iter 5500/9765 | SPS:  2323 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0241 | PLoss: -0.0103 | Ent: 0.1783
+[ 56.3%] Iter 5501/9765 | SPS:  2323 | Reward:  0.067 | Value:  0.689 | VLoss: 0.0258 | PLoss: -0.0116 | Ent: 0.1952
+[ 56.3%] Iter 5502/9765 | SPS:  2323 | Reward:  0.067 | Value:  0.710 | VLoss: 0.0298 | PLoss: -0.0076 | Ent: 0.1843
+[ 56.4%] Iter 5503/9765 | SPS:  2323 | Reward:  0.053 | Value:  0.755 | VLoss: 0.0249 | PLoss: -0.0060 | Ent: 0.1930
+[ 56.4%] Iter 5504/9765 | SPS:  2323 | Reward:  0.069 | Value:  0.663 | VLoss: 0.0224 | PLoss: -0.0144 | Ent: 0.2049
+[ 56.4%] Iter 5505/9765 | SPS:  2324 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0215 | PLoss: -0.0100 | Ent: 0.1933
+[ 56.4%] Iter 5506/9765 | SPS:  2324 | Reward:  0.054 | Value:  0.725 | VLoss: 0.0213 | PLoss: -0.0040 | Ent: 0.1639
+[ 56.4%] Iter 5507/9765 | SPS:  2324 | Reward:  0.070 | Value:  0.673 | VLoss: 0.0260 | PLoss: -0.0041 | Ent: 0.2132
+[ 56.4%] Iter 5508/9765 | SPS:  2324 | Reward:  0.058 | Value:  0.728 | VLoss: 0.0235 | PLoss: -0.0123 | Ent: 0.2495
+[ 56.4%] Iter 5509/9765 | SPS:  2324 | Reward:  0.060 | Value:  0.681 | VLoss: 0.0257 | PLoss: -0.0053 | Ent: 0.2011
+[ 56.4%] Iter 5510/9765 | SPS:  2324 | Reward:  0.054 | Value:  0.630 | VLoss: 0.0228 | PLoss: -0.0174 | Ent: 0.2212
+[ 56.4%] Iter 5511/9765 | SPS:  2325 | Reward:  0.056 | Value:  0.698 | VLoss: 0.0229 | PLoss: -0.0090 | Ent: 0.1815
+[ 56.4%] Iter 5512/9765 | SPS:  2325 | Reward:  0.054 | Value:  0.632 | VLoss: 0.0202 | PLoss: -0.0046 | Ent: 0.1977
+[ 56.5%] Iter 5513/9765 | SPS:  2325 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0212 | PLoss: -0.0120 | Ent: 0.1977
+[ 56.5%] Iter 5514/9765 | SPS:  2325 | Reward:  0.069 | Value:  0.742 | VLoss: 0.0219 | PLoss: -0.0089 | Ent: 0.2117
+[ 56.5%] Iter 5515/9765 | SPS:  2325 | Reward:  0.066 | Value:  0.801 | VLoss: 0.0293 | PLoss: -0.0110 | Ent: 0.1864
+[ 56.5%] Iter 5516/9765 | SPS:  2325 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0168 | PLoss: -0.0006 | Ent: 0.2639
+[ 56.5%] Iter 5517/9765 | SPS:  2325 | Reward:  0.073 | Value:  0.653 | VLoss: 0.0324 | PLoss: -0.0037 | Ent: 0.1955
+[ 56.5%] Iter 5518/9765 | SPS:  2326 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0198 | PLoss: -0.0055 | Ent: 0.1900
+[ 56.5%] Iter 5519/9765 | SPS:  2326 | Reward:  0.061 | Value:  0.776 | VLoss: 0.0217 | PLoss: -0.0062 | Ent: 0.1970
+[ 56.5%] Iter 5520/9765 | SPS:  2326 | Reward:  0.059 | Value:  0.695 | VLoss: 0.0164 | PLoss: -0.0041 | Ent: 0.2021
+[ 56.5%] Iter 5521/9765 | SPS:  2326 | Reward:  0.047 | Value:  0.707 | VLoss: 0.0190 | PLoss: -0.0063 | Ent: 0.2085
+[ 56.5%] Iter 5522/9765 | SPS:  2326 | Reward:  0.051 | Value:  0.792 | VLoss: 0.0144 | PLoss: -0.0072 | Ent: 0.2267
+[ 56.6%] Iter 5523/9765 | SPS:  2326 | Reward:  0.058 | Value:  0.709 | VLoss: 0.0242 | PLoss: -0.0089 | Ent: 0.1898
+[ 56.6%] Iter 5524/9765 | SPS:  2327 | Reward:  0.071 | Value:  0.687 | VLoss: 0.0281 | PLoss: -0.0102 | Ent: 0.1878
+[ 56.6%] Iter 5525/9765 | SPS:  2327 | Reward:  0.059 | Value:  0.654 | VLoss: 0.0232 | PLoss: -0.0068 | Ent: 0.1938
+[ 56.6%] Iter 5526/9765 | SPS:  2327 | Reward:  0.072 | Value:  0.694 | VLoss: 0.0210 | PLoss: -0.0052 | Ent: 0.2178
+[ 56.6%] Iter 5527/9765 | SPS:  2327 | Reward:  0.073 | Value:  0.733 | VLoss: 0.0277 | PLoss: -0.0126 | Ent: 0.2358
+[ 56.6%] Iter 5528/9765 | SPS:  2327 | Reward:  0.073 | Value:  0.744 | VLoss: 0.0294 | PLoss: -0.0055 | Ent: 0.2060
+[ 56.6%] Iter 5529/9765 | SPS:  2327 | Reward:  0.071 | Value:  0.710 | VLoss: 0.0280 | PLoss: -0.0124 | Ent: 0.2420
+[ 56.6%] Iter 5530/9765 | SPS:  2328 | Reward:  0.054 | Value:  0.674 | VLoss: 0.0265 | PLoss: -0.0078 | Ent: 0.1951
+[ 56.6%] Iter 5531/9765 | SPS:  2328 | Reward:  0.058 | Value:  0.741 | VLoss: 0.0219 | PLoss: -0.0041 | Ent: 0.1825
+[ 56.7%] Iter 5532/9765 | SPS:  2328 | Reward:  0.071 | Value:  0.707 | VLoss: 0.0246 | PLoss: -0.0032 | Ent: 0.2039
+[ 56.7%] Iter 5533/9765 | SPS:  2328 | Reward:  0.052 | Value:  0.820 | VLoss: 0.0155 | PLoss: -0.0079 | Ent: 0.2004
+[ 56.7%] Iter 5534/9765 | SPS:  2328 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0278 | PLoss: -0.0040 | Ent: 0.1931
+[ 56.7%] Iter 5535/9765 | SPS:  2328 | Reward:  0.072 | Value:  0.772 | VLoss: 0.0250 | PLoss: -0.0092 | Ent: 0.2030
+[ 56.7%] Iter 5536/9765 | SPS:  2328 | Reward:  0.072 | Value:  0.750 | VLoss: 0.0209 | PLoss: -0.0046 | Ent: 0.2149
+[ 56.7%] Iter 5537/9765 | SPS:  2329 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0301 | PLoss: -0.0058 | Ent: 0.2313
+[ 56.7%] Iter 5538/9765 | SPS:  2329 | Reward:  0.064 | Value:  0.759 | VLoss: 0.0225 | PLoss: -0.0069 | Ent: 0.2132
+[ 56.7%] Iter 5539/9765 | SPS:  2329 | Reward:  0.051 | Value:  0.693 | VLoss: 0.0223 | PLoss: -0.0092 | Ent: 0.2183
+[ 56.7%] Iter 5540/9765 | SPS:  2329 | Reward:  0.069 | Value:  0.678 | VLoss: 0.0269 | PLoss: -0.0078 | Ent: 0.1683
+[ 56.7%] Iter 5541/9765 | SPS:  2329 | Reward:  0.075 | Value:  0.704 | VLoss: 0.0297 | PLoss: -0.0046 | Ent: 0.1756
+[ 56.8%] Iter 5542/9765 | SPS:  2329 | Reward:  0.050 | Value:  0.711 | VLoss: 0.0240 | PLoss: -0.0021 | Ent: 0.1988
+[ 56.8%] Iter 5543/9765 | SPS:  2330 | Reward:  0.080 | Value:  0.700 | VLoss: 0.0298 | PLoss: -0.0131 | Ent: 0.2273
+[ 56.8%] Iter 5544/9765 | SPS:  2330 | Reward:  0.063 | Value:  0.701 | VLoss: 0.0264 | PLoss: -0.0078 | Ent: 0.1906
+[ 56.8%] Iter 5545/9765 | SPS:  2330 | Reward:  0.056 | Value:  0.714 | VLoss: 0.0234 | PLoss: -0.0093 | Ent: 0.1854
+[ 56.8%] Iter 5546/9765 | SPS:  2330 | Reward:  0.070 | Value:  0.645 | VLoss: 0.0197 | PLoss: -0.0053 | Ent: 0.1477
+[ 56.8%] Iter 5547/9765 | SPS:  2330 | Reward:  0.060 | Value:  0.690 | VLoss: 0.0199 | PLoss: -0.0148 | Ent: 0.2147
+[ 56.8%] Iter 5548/9765 | SPS:  2330 | Reward:  0.065 | Value:  0.768 | VLoss: 0.0246 | PLoss: -0.0071 | Ent: 0.1979
+[ 56.8%] Iter 5549/9765 | SPS:  2330 | Reward:  0.063 | Value:  0.715 | VLoss: 0.0252 | PLoss: -0.0067 | Ent: 0.2099
+[ 56.8%] Iter 5550/9765 | SPS:  2331 | Reward:  0.059 | Value:  0.752 | VLoss: 0.0190 | PLoss: -0.0011 | Ent: 0.2489
+[ 56.8%] Iter 5551/9765 | SPS:  2331 | Reward:  0.070 | Value:  0.788 | VLoss: 0.0278 | PLoss: -0.0069 | Ent: 0.2046
+[ 56.9%] Iter 5552/9765 | SPS:  2331 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0268 | PLoss: -0.0079 | Ent: 0.2233
+[ 56.9%] Iter 5553/9765 | SPS:  2331 | Reward:  0.055 | Value:  0.695 | VLoss: 0.0177 | PLoss: -0.0037 | Ent: 0.1916
+[ 56.9%] Iter 5554/9765 | SPS:  2331 | Reward:  0.066 | Value:  0.772 | VLoss: 0.0306 | PLoss: -0.0052 | Ent: 0.1718
+[ 56.9%] Iter 5555/9765 | SPS:  2331 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0245 | PLoss: -0.0095 | Ent: 0.2298
+[ 56.9%] Iter 5556/9765 | SPS:  2332 | Reward:  0.069 | Value:  0.740 | VLoss: 0.0247 | PLoss: -0.0104 | Ent: 0.2116
+[ 56.9%] Iter 5557/9765 | SPS:  2332 | Reward:  0.049 | Value:  0.762 | VLoss: 0.0213 | PLoss: -0.0068 | Ent: 0.1826
+[ 56.9%] Iter 5558/9765 | SPS:  2332 | Reward:  0.058 | Value:  0.702 | VLoss: 0.0146 | PLoss: -0.0035 | Ent: 0.1951
+[ 56.9%] Iter 5559/9765 | SPS:  2332 | Reward:  0.061 | Value:  0.702 | VLoss: 0.0189 | PLoss: -0.0095 | Ent: 0.2016
+[ 56.9%] Iter 5560/9765 | SPS:  2332 | Reward:  0.056 | Value:  0.764 | VLoss: 0.0156 | PLoss: -0.0034 | Ent: 0.1629
+[ 56.9%] Iter 5561/9765 | SPS:  2332 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0149 | PLoss: -0.0016 | Ent: 0.1479
+[ 57.0%] Iter 5562/9765 | SPS:  2333 | Reward:  0.054 | Value:  0.659 | VLoss: 0.0191 | PLoss: -0.0073 | Ent: 0.1732
+[ 57.0%] Iter 5563/9765 | SPS:  2333 | Reward:  0.056 | Value:  0.689 | VLoss: 0.0148 | PLoss: -0.0053 | Ent: 0.1903
+[ 57.0%] Iter 5564/9765 | SPS:  2333 | Reward:  0.048 | Value:  0.653 | VLoss: 0.0193 | PLoss: -0.0095 | Ent: 0.1873
+[ 57.0%] Iter 5565/9765 | SPS:  2333 | Reward:  0.056 | Value:  0.670 | VLoss: 0.0224 | PLoss: -0.0115 | Ent: 0.2203
+[ 57.0%] Iter 5566/9765 | SPS:  2333 | Reward:  0.064 | Value:  0.697 | VLoss: 0.0244 | PLoss: -0.0110 | Ent: 0.1822
+[ 57.0%] Iter 5567/9765 | SPS:  2333 | Reward:  0.065 | Value:  0.722 | VLoss: 0.0265 | PLoss: -0.0122 | Ent: 0.1767
+[ 57.0%] Iter 5568/9765 | SPS:  2334 | Reward:  0.060 | Value:  0.756 | VLoss: 0.0177 | PLoss: -0.0073 | Ent: 0.1999
+[ 57.0%] Iter 5569/9765 | SPS:  2334 | Reward:  0.069 | Value:  0.740 | VLoss: 0.0287 | PLoss: -0.0111 | Ent: 0.1701
+[ 57.0%] Iter 5570/9765 | SPS:  2334 | Reward:  0.059 | Value:  0.725 | VLoss: 0.0275 | PLoss: -0.0101 | Ent: 0.2075
+[ 57.1%] Iter 5571/9765 | SPS:  2334 | Reward:  0.077 | Value:  0.718 | VLoss: 0.0283 | PLoss: -0.0128 | Ent: 0.1855
+[ 57.1%] Iter 5572/9765 | SPS:  2334 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0265 | PLoss: -0.0070 | Ent: 0.2268
+[ 57.1%] Iter 5573/9765 | SPS:  2334 | Reward:  0.067 | Value:  0.684 | VLoss: 0.0337 | PLoss: -0.0069 | Ent: 0.1876
+[ 57.1%] Iter 5574/9765 | SPS:  2334 | Reward:  0.058 | Value:  0.647 | VLoss: 0.0215 | PLoss: -0.0096 | Ent: 0.2235
+[ 57.1%] Iter 5575/9765 | SPS:  2335 | Reward:  0.053 | Value:  0.768 | VLoss: 0.0178 | PLoss: -0.0098 | Ent: 0.2159
+[ 57.1%] Iter 5576/9765 | SPS:  2335 | Reward:  0.063 | Value:  0.707 | VLoss: 0.0273 | PLoss: -0.0092 | Ent: 0.2027
+[ 57.1%] Iter 5577/9765 | SPS:  2335 | Reward:  0.071 | Value:  0.727 | VLoss: 0.0285 | PLoss: -0.0116 | Ent: 0.1998
+[ 57.1%] Iter 5578/9765 | SPS:  2335 | Reward:  0.056 | Value:  0.708 | VLoss: 0.0253 | PLoss: -0.0070 | Ent: 0.1662
+[ 57.1%] Iter 5579/9765 | SPS:  2335 | Reward:  0.062 | Value:  0.729 | VLoss: 0.0248 | PLoss: -0.0069 | Ent: 0.1514
+[ 57.1%] Iter 5580/9765 | SPS:  2335 | Reward:  0.060 | Value:  0.718 | VLoss: 0.0220 | PLoss: -0.0080 | Ent: 0.1731
+[ 57.2%] Iter 5581/9765 | SPS:  2336 | Reward:  0.068 | Value:  0.779 | VLoss: 0.0214 | PLoss: 0.0008 | Ent: 0.1466
+[ 57.2%] Iter 5582/9765 | SPS:  2336 | Reward:  0.048 | Value:  0.703 | VLoss: 0.0261 | PLoss: -0.0068 | Ent: 0.1830
+[ 57.2%] Iter 5583/9765 | SPS:  2336 | Reward:  0.069 | Value:  0.714 | VLoss: 0.0245 | PLoss: -0.0065 | Ent: 0.1987
+[ 57.2%] Iter 5584/9765 | SPS:  2336 | Reward:  0.068 | Value:  0.672 | VLoss: 0.0226 | PLoss: -0.0111 | Ent: 0.2354
+[ 57.2%] Iter 5585/9765 | SPS:  2336 | Reward:  0.071 | Value:  0.824 | VLoss: 0.0182 | PLoss: -0.0110 | Ent: 0.1795
+[ 57.2%] Iter 5586/9765 | SPS:  2336 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0271 | PLoss: -0.0045 | Ent: 0.1933
+[ 57.2%] Iter 5587/9765 | SPS:  2336 | Reward:  0.055 | Value:  0.675 | VLoss: 0.0301 | PLoss: -0.0028 | Ent: 0.2242
+[ 57.2%] Iter 5588/9765 | SPS:  2337 | Reward:  0.069 | Value:  0.681 | VLoss: 0.0239 | PLoss: -0.0149 | Ent: 0.2403
+[ 57.2%] Iter 5589/9765 | SPS:  2337 | Reward:  0.057 | Value:  0.737 | VLoss: 0.0254 | PLoss: -0.0040 | Ent: 0.1772
+[ 57.2%] Iter 5590/9765 | SPS:  2337 | Reward:  0.070 | Value:  0.688 | VLoss: 0.0264 | PLoss: -0.0069 | Ent: 0.2123
+[ 57.3%] Iter 5591/9765 | SPS:  2337 | Reward:  0.056 | Value:  0.718 | VLoss: 0.0207 | PLoss: -0.0142 | Ent: 0.2241
+[ 57.3%] Iter 5592/9765 | SPS:  2337 | Reward:  0.053 | Value:  0.702 | VLoss: 0.0191 | PLoss: -0.0066 | Ent: 0.1663
+[ 57.3%] Iter 5593/9765 | SPS:  2337 | Reward:  0.073 | Value:  0.700 | VLoss: 0.0205 | PLoss: -0.0149 | Ent: 0.2185
+[ 57.3%] Iter 5594/9765 | SPS:  2338 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0242 | PLoss: -0.0011 | Ent: 0.1749
+[ 57.3%] Iter 5595/9765 | SPS:  2338 | Reward:  0.060 | Value:  0.719 | VLoss: 0.0265 | PLoss: -0.0085 | Ent: 0.2031
+[ 57.3%] Iter 5596/9765 | SPS:  2338 | Reward:  0.072 | Value:  0.736 | VLoss: 0.0259 | PLoss: -0.0063 | Ent: 0.2055
+[ 57.3%] Iter 5597/9765 | SPS:  2338 | Reward:  0.060 | Value:  0.764 | VLoss: 0.0216 | PLoss: -0.0073 | Ent: 0.1834
+[ 57.3%] Iter 5598/9765 | SPS:  2338 | Reward:  0.054 | Value:  0.617 | VLoss: 0.0143 | PLoss: -0.0065 | Ent: 0.1501
+[ 57.3%] Iter 5599/9765 | SPS:  2338 | Reward:  0.051 | Value:  0.706 | VLoss: 0.0176 | PLoss: -0.0055 | Ent: 0.1478
+[ 57.3%] Iter 5600/9765 | SPS:  2339 | Reward:  0.056 | Value:  0.644 | VLoss: 0.0246 | PLoss: -0.0097 | Ent: 0.1815
+[ 57.4%] Iter 5601/9765 | SPS:  2339 | Reward:  0.052 | Value:  0.599 | VLoss: 0.0241 | PLoss: -0.0099 | Ent: 0.1540
+[ 57.4%] Iter 5602/9765 | SPS:  2339 | Reward:  0.055 | Value:  0.635 | VLoss: 0.0205 | PLoss: -0.0096 | Ent: 0.1714
+[ 57.4%] Iter 5603/9765 | SPS:  2339 | Reward:  0.065 | Value:  0.663 | VLoss: 0.0225 | PLoss: -0.0094 | Ent: 0.1986
+[ 57.4%] Iter 5604/9765 | SPS:  2339 | Reward:  0.069 | Value:  0.806 | VLoss: 0.0271 | PLoss: -0.0033 | Ent: 0.2030
+[ 57.4%] Iter 5605/9765 | SPS:  2339 | Reward:  0.053 | Value:  0.741 | VLoss: 0.0176 | PLoss: -0.0054 | Ent: 0.2108
+[ 57.4%] Iter 5606/9765 | SPS:  2339 | Reward:  0.067 | Value:  0.742 | VLoss: 0.0232 | PLoss: -0.0063 | Ent: 0.1952
+[ 57.4%] Iter 5607/9765 | SPS:  2340 | Reward:  0.053 | Value:  0.644 | VLoss: 0.0187 | PLoss: -0.0086 | Ent: 0.1740
+[ 57.4%] Iter 5608/9765 | SPS:  2340 | Reward:  0.064 | Value:  0.740 | VLoss: 0.0212 | PLoss: -0.0044 | Ent: 0.2110
+[ 57.4%] Iter 5609/9765 | SPS:  2340 | Reward:  0.055 | Value:  0.786 | VLoss: 0.0203 | PLoss: -0.0075 | Ent: 0.1977
+[ 57.5%] Iter 5610/9765 | SPS:  2340 | Reward:  0.054 | Value:  0.744 | VLoss: 0.0255 | PLoss: -0.0070 | Ent: 0.2250
+[ 57.5%] Iter 5611/9765 | SPS:  2340 | Reward:  0.047 | Value:  0.679 | VLoss: 0.0178 | PLoss: -0.0027 | Ent: 0.1967
+[ 57.5%] Iter 5612/9765 | SPS:  2340 | Reward:  0.060 | Value:  0.740 | VLoss: 0.0283 | PLoss: -0.0131 | Ent: 0.1889
+[ 57.5%] Iter 5613/9765 | SPS:  2341 | Reward:  0.073 | Value:  0.666 | VLoss: 0.0259 | PLoss: -0.0073 | Ent: 0.2153
+[ 57.5%] Iter 5614/9765 | SPS:  2341 | Reward:  0.070 | Value:  0.766 | VLoss: 0.0259 | PLoss: -0.0053 | Ent: 0.2110
+[ 57.5%] Iter 5615/9765 | SPS:  2341 | Reward:  0.066 | Value:  0.759 | VLoss: 0.0190 | PLoss: -0.0073 | Ent: 0.2132
+[ 57.5%] Iter 5616/9765 | SPS:  2341 | Reward:  0.062 | Value:  0.749 | VLoss: 0.0280 | PLoss: -0.0090 | Ent: 0.1855
+[ 57.5%] Iter 5617/9765 | SPS:  2341 | Reward:  0.067 | Value:  0.637 | VLoss: 0.0215 | PLoss: -0.0053 | Ent: 0.2312
+[ 57.5%] Iter 5618/9765 | SPS:  2341 | Reward:  0.078 | Value:  0.701 | VLoss: 0.0202 | PLoss: -0.0131 | Ent: 0.2152
+[ 57.5%] Iter 5619/9765 | SPS:  2341 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0185 | PLoss: -0.0173 | Ent: 0.2097
+[ 57.6%] Iter 5620/9765 | SPS:  2342 | Reward:  0.048 | Value:  0.678 | VLoss: 0.0253 | PLoss: -0.0101 | Ent: 0.2031
+[ 57.6%] Iter 5621/9765 | SPS:  2342 | Reward:  0.066 | Value:  0.745 | VLoss: 0.0239 | PLoss: -0.0028 | Ent: 0.2122
+[ 57.6%] Iter 5622/9765 | SPS:  2342 | Reward:  0.058 | Value:  0.696 | VLoss: 0.0244 | PLoss: -0.0097 | Ent: 0.2374
+[ 57.6%] Iter 5623/9765 | SPS:  2342 | Reward:  0.050 | Value:  0.671 | VLoss: 0.0290 | PLoss: -0.0052 | Ent: 0.1875
+[ 57.6%] Iter 5624/9765 | SPS:  2342 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0284 | PLoss: -0.0111 | Ent: 0.1703
+[ 57.6%] Iter 5625/9765 | SPS:  2342 | Reward:  0.059 | Value:  0.730 | VLoss: 0.0205 | PLoss: -0.0088 | Ent: 0.2002
+[ 57.6%] Iter 5626/9765 | SPS:  2343 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0304 | PLoss: -0.0068 | Ent: 0.1986
+[ 57.6%] Iter 5627/9765 | SPS:  2343 | Reward:  0.060 | Value:  0.755 | VLoss: 0.0243 | PLoss: 0.0012 | Ent: 0.1735
+[ 57.6%] Iter 5628/9765 | SPS:  2343 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0225 | PLoss: -0.0084 | Ent: 0.1875
+[ 57.6%] Iter 5629/9765 | SPS:  2343 | Reward:  0.074 | Value:  0.746 | VLoss: 0.0260 | PLoss: -0.0093 | Ent: 0.1937
+[ 57.7%] Iter 5630/9765 | SPS:  2343 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0255 | PLoss: -0.0076 | Ent: 0.2164
+[ 57.7%] Iter 5631/9765 | SPS:  2343 | Reward:  0.065 | Value:  0.620 | VLoss: 0.0307 | PLoss: -0.0054 | Ent: 0.2130
+[ 57.7%] Iter 5632/9765 | SPS:  2344 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0291 | PLoss: -0.0131 | Ent: 0.2303
+[ 57.7%] Iter 5633/9765 | SPS:  2344 | Reward:  0.061 | Value:  0.704 | VLoss: 0.0281 | PLoss: -0.0073 | Ent: 0.2176
+[ 57.7%] Iter 5634/9765 | SPS:  2344 | Reward:  0.056 | Value:  0.671 | VLoss: 0.0164 | PLoss: -0.0061 | Ent: 0.1698
+[ 57.7%] Iter 5635/9765 | SPS:  2344 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0153 | PLoss: -0.0020 | Ent: 0.1507
+[ 57.7%] Iter 5636/9765 | SPS:  2344 | Reward:  0.073 | Value:  0.727 | VLoss: 0.0230 | PLoss: -0.0115 | Ent: 0.1759
+[ 57.7%] Iter 5637/9765 | SPS:  2344 | Reward:  0.058 | Value:  0.748 | VLoss: 0.0282 | PLoss: -0.0015 | Ent: 0.1756
+[ 57.7%] Iter 5638/9765 | SPS:  2344 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0282 | PLoss: -0.0055 | Ent: 0.2216
+[ 57.7%] Iter 5639/9765 | SPS:  2345 | Reward:  0.069 | Value:  0.737 | VLoss: 0.0198 | PLoss: -0.0050 | Ent: 0.1892
+[ 57.8%] Iter 5640/9765 | SPS:  2345 | Reward:  0.068 | Value:  0.823 | VLoss: 0.0180 | PLoss: -0.0070 | Ent: 0.2155
+[ 57.8%] Iter 5641/9765 | SPS:  2345 | Reward:  0.053 | Value:  0.705 | VLoss: 0.0267 | PLoss: -0.0077 | Ent: 0.2102
+[ 57.8%] Iter 5642/9765 | SPS:  2345 | Reward:  0.065 | Value:  0.650 | VLoss: 0.0264 | PLoss: -0.0071 | Ent: 0.1975
+[ 57.8%] Iter 5643/9765 | SPS:  2345 | Reward:  0.067 | Value:  0.683 | VLoss: 0.0304 | PLoss: -0.0092 | Ent: 0.2361
+[ 57.8%] Iter 5644/9765 | SPS:  2345 | Reward:  0.057 | Value:  0.752 | VLoss: 0.0212 | PLoss: -0.0065 | Ent: 0.1852
+[ 57.8%] Iter 5645/9765 | SPS:  2345 | Reward:  0.065 | Value:  0.655 | VLoss: 0.0243 | PLoss: -0.0056 | Ent: 0.1754
+[ 57.8%] Iter 5646/9765 | SPS:  2346 | Reward:  0.055 | Value:  0.671 | VLoss: 0.0263 | PLoss: -0.0066 | Ent: 0.2277
+[ 57.8%] Iter 5647/9765 | SPS:  2346 | Reward:  0.070 | Value:  0.740 | VLoss: 0.0194 | PLoss: -0.0061 | Ent: 0.1772
+[ 57.8%] Iter 5648/9765 | SPS:  2346 | Reward:  0.069 | Value:  0.756 | VLoss: 0.0277 | PLoss: -0.0126 | Ent: 0.1774
+[ 57.8%] Iter 5649/9765 | SPS:  2346 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0301 | PLoss: -0.0085 | Ent: 0.1719
+[ 57.9%] Iter 5650/9765 | SPS:  2346 | Reward:  0.067 | Value:  0.734 | VLoss: 0.0246 | PLoss: -0.0106 | Ent: 0.1904
+[ 57.9%] Iter 5651/9765 | SPS:  2346 | Reward:  0.072 | Value:  0.767 | VLoss: 0.0290 | PLoss: -0.0065 | Ent: 0.1745
+[ 57.9%] Iter 5652/9765 | SPS:  2347 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0285 | PLoss: -0.0019 | Ent: 0.1904
+[ 57.9%] Iter 5653/9765 | SPS:  2347 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0207 | PLoss: -0.0080 | Ent: 0.2006
+[ 57.9%] Iter 5654/9765 | SPS:  2347 | Reward:  0.059 | Value:  0.698 | VLoss: 0.0215 | PLoss: -0.0115 | Ent: 0.2207
+[ 57.9%] Iter 5655/9765 | SPS:  2347 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0189 | PLoss: -0.0054 | Ent: 0.1968
+[ 57.9%] Iter 5656/9765 | SPS:  2347 | Reward:  0.065 | Value:  0.691 | VLoss: 0.0156 | PLoss: -0.0016 | Ent: 0.1702
+[ 57.9%] Iter 5657/9765 | SPS:  2347 | Reward:  0.058 | Value:  0.803 | VLoss: 0.0197 | PLoss: -0.0052 | Ent: 0.1797
+[ 57.9%] Iter 5658/9765 | SPS:  2347 | Reward:  0.047 | Value:  0.685 | VLoss: 0.0268 | PLoss: -0.0072 | Ent: 0.2096
+[ 58.0%] Iter 5659/9765 | SPS:  2348 | Reward:  0.051 | Value:  0.597 | VLoss: 0.0155 | PLoss: -0.0145 | Ent: 0.2221
+[ 58.0%] Iter 5660/9765 | SPS:  2348 | Reward:  0.070 | Value:  0.675 | VLoss: 0.0241 | PLoss: 0.0022 | Ent: 0.2175
+[ 58.0%] Iter 5661/9765 | SPS:  2348 | Reward:  0.066 | Value:  0.756 | VLoss: 0.0269 | PLoss: -0.0055 | Ent: 0.2108
+[ 58.0%] Iter 5662/9765 | SPS:  2348 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0185 | PLoss: -0.0003 | Ent: 0.1772
+[ 58.0%] Iter 5663/9765 | SPS:  2348 | Reward:  0.065 | Value:  0.667 | VLoss: 0.0272 | PLoss: -0.0026 | Ent: 0.2263
+[ 58.0%] Iter 5664/9765 | SPS:  2348 | Reward:  0.055 | Value:  0.801 | VLoss: 0.0195 | PLoss: -0.0123 | Ent: 0.1931
+[ 58.0%] Iter 5665/9765 | SPS:  2348 | Reward:  0.052 | Value:  0.757 | VLoss: 0.0253 | PLoss: -0.0096 | Ent: 0.1868
+[ 58.0%] Iter 5666/9765 | SPS:  2349 | Reward:  0.047 | Value:  0.671 | VLoss: 0.0316 | PLoss: -0.0032 | Ent: 0.1875
+[ 58.0%] Iter 5667/9765 | SPS:  2349 | Reward:  0.061 | Value:  0.661 | VLoss: 0.0270 | PLoss: -0.0069 | Ent: 0.1808
+[ 58.0%] Iter 5668/9765 | SPS:  2349 | Reward:  0.039 | Value:  0.616 | VLoss: 0.0213 | PLoss: -0.0066 | Ent: 0.2245
+[ 58.1%] Iter 5669/9765 | SPS:  2349 | Reward:  0.051 | Value:  0.706 | VLoss: 0.0230 | PLoss: -0.0092 | Ent: 0.1833
+[ 58.1%] Iter 5670/9765 | SPS:  2349 | Reward:  0.048 | Value:  0.629 | VLoss: 0.0217 | PLoss: -0.0044 | Ent: 0.2286
+[ 58.1%] Iter 5671/9765 | SPS:  2349 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0242 | PLoss: -0.0171 | Ent: 0.1995
+[ 58.1%] Iter 5672/9765 | SPS:  2350 | Reward:  0.047 | Value:  0.643 | VLoss: 0.0175 | PLoss: -0.0032 | Ent: 0.2258
+[ 58.1%] Iter 5673/9765 | SPS:  2350 | Reward:  0.072 | Value:  0.701 | VLoss: 0.0219 | PLoss: -0.0079 | Ent: 0.2304
+[ 58.1%] Iter 5674/9765 | SPS:  2350 | Reward:  0.057 | Value:  0.738 | VLoss: 0.0281 | PLoss: -0.0063 | Ent: 0.2215
+[ 58.1%] Iter 5675/9765 | SPS:  2350 | Reward:  0.055 | Value:  0.683 | VLoss: 0.0249 | PLoss: -0.0065 | Ent: 0.2156
+[ 58.1%] Iter 5676/9765 | SPS:  2350 | Reward:  0.057 | Value:  0.668 | VLoss: 0.0188 | PLoss: -0.0036 | Ent: 0.1736
+[ 58.1%] Iter 5677/9765 | SPS:  2350 | Reward:  0.060 | Value:  0.625 | VLoss: 0.0225 | PLoss: 0.0011 | Ent: 0.2088
+[ 58.1%] Iter 5678/9765 | SPS:  2351 | Reward:  0.063 | Value:  0.746 | VLoss: 0.0273 | PLoss: -0.0038 | Ent: 0.2073
+[ 58.2%] Iter 5679/9765 | SPS:  2351 | Reward:  0.068 | Value:  0.712 | VLoss: 0.0188 | PLoss: -0.0031 | Ent: 0.1800
+[ 58.2%] Iter 5680/9765 | SPS:  2351 | Reward:  0.066 | Value:  0.743 | VLoss: 0.0216 | PLoss: -0.0035 | Ent: 0.1951
+[ 58.2%] Iter 5681/9765 | SPS:  2351 | Reward:  0.058 | Value:  0.696 | VLoss: 0.0218 | PLoss: -0.0014 | Ent: 0.2264
+[ 58.2%] Iter 5682/9765 | SPS:  2351 | Reward:  0.051 | Value:  0.606 | VLoss: 0.0242 | PLoss: -0.0098 | Ent: 0.2293
+[ 58.2%] Iter 5683/9765 | SPS:  2351 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0329 | PLoss: -0.0046 | Ent: 0.2185
+[ 58.2%] Iter 5684/9765 | SPS:  2351 | Reward:  0.065 | Value:  0.711 | VLoss: 0.0347 | PLoss: -0.0055 | Ent: 0.1835
+[ 58.2%] Iter 5685/9765 | SPS:  2352 | Reward:  0.069 | Value:  0.714 | VLoss: 0.0229 | PLoss: -0.0055 | Ent: 0.2197
+[ 58.2%] Iter 5686/9765 | SPS:  2352 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0238 | PLoss: -0.0137 | Ent: 0.2382
+[ 58.2%] Iter 5687/9765 | SPS:  2352 | Reward:  0.075 | Value:  0.792 | VLoss: 0.0197 | PLoss: -0.0134 | Ent: 0.1818
+[ 58.2%] Iter 5688/9765 | SPS:  2352 | Reward:  0.068 | Value:  0.743 | VLoss: 0.0234 | PLoss: -0.0090 | Ent: 0.1803
+[ 58.3%] Iter 5689/9765 | SPS:  2352 | Reward:  0.060 | Value:  0.715 | VLoss: 0.0213 | PLoss: -0.0033 | Ent: 0.1964
+[ 58.3%] Iter 5690/9765 | SPS:  2352 | Reward:  0.057 | Value:  0.667 | VLoss: 0.0240 | PLoss: -0.0033 | Ent: 0.1628
+[ 58.3%] Iter 5691/9765 | SPS:  2352 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0145 | PLoss: -0.0112 | Ent: 0.1867
+[ 58.3%] Iter 5692/9765 | SPS:  2353 | Reward:  0.068 | Value:  0.707 | VLoss: 0.0267 | PLoss: -0.0112 | Ent: 0.2166
+[ 58.3%] Iter 5693/9765 | SPS:  2353 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0247 | PLoss: 0.0033 | Ent: 0.1670
+[ 58.3%] Iter 5694/9765 | SPS:  2353 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0243 | PLoss: -0.0123 | Ent: 0.2041
+[ 58.3%] Iter 5695/9765 | SPS:  2353 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0188 | PLoss: -0.0164 | Ent: 0.1926
+[ 58.3%] Iter 5696/9765 | SPS:  2353 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0231 | PLoss: -0.0069 | Ent: 0.2163
+[ 58.3%] Iter 5697/9765 | SPS:  2353 | Reward:  0.049 | Value:  0.677 | VLoss: 0.0201 | PLoss: -0.0070 | Ent: 0.1951
+[ 58.4%] Iter 5698/9765 | SPS:  2354 | Reward:  0.058 | Value:  0.724 | VLoss: 0.0249 | PLoss: -0.0050 | Ent: 0.1331
+[ 58.4%] Iter 5699/9765 | SPS:  2354 | Reward:  0.061 | Value:  0.686 | VLoss: 0.0144 | PLoss: -0.0077 | Ent: 0.1776
+[ 58.4%] Iter 5700/9765 | SPS:  2354 | Reward:  0.048 | Value:  0.739 | VLoss: 0.0178 | PLoss: -0.0090 | Ent: 0.1981
+[ 58.4%] Iter 5701/9765 | SPS:  2354 | Reward:  0.073 | Value:  0.716 | VLoss: 0.0223 | PLoss: -0.0023 | Ent: 0.1779
+[ 58.4%] Iter 5702/9765 | SPS:  2354 | Reward:  0.061 | Value:  0.749 | VLoss: 0.0228 | PLoss: -0.0041 | Ent: 0.2212
+[ 58.4%] Iter 5703/9765 | SPS:  2354 | Reward:  0.057 | Value:  0.754 | VLoss: 0.0244 | PLoss: -0.0036 | Ent: 0.1861
+[ 58.4%] Iter 5704/9765 | SPS:  2354 | Reward:  0.067 | Value:  0.679 | VLoss: 0.0248 | PLoss: -0.0105 | Ent: 0.1868
+[ 58.4%] Iter 5705/9765 | SPS:  2355 | Reward:  0.074 | Value:  0.680 | VLoss: 0.0243 | PLoss: -0.0118 | Ent: 0.2531
+[ 58.4%] Iter 5706/9765 | SPS:  2355 | Reward:  0.059 | Value:  0.721 | VLoss: 0.0251 | PLoss: -0.0095 | Ent: 0.2040
+[ 58.4%] Iter 5707/9765 | SPS:  2355 | Reward:  0.054 | Value:  0.704 | VLoss: 0.0231 | PLoss: -0.0000 | Ent: 0.2054
+[ 58.5%] Iter 5708/9765 | SPS:  2355 | Reward:  0.072 | Value:  0.756 | VLoss: 0.0202 | PLoss: -0.0069 | Ent: 0.2363
+[ 58.5%] Iter 5709/9765 | SPS:  2355 | Reward:  0.059 | Value:  0.687 | VLoss: 0.0266 | PLoss: -0.0100 | Ent: 0.1676
+[ 58.5%] Iter 5710/9765 | SPS:  2355 | Reward:  0.058 | Value:  0.721 | VLoss: 0.0232 | PLoss: -0.0103 | Ent: 0.1797
+[ 58.5%] Iter 5711/9765 | SPS:  2356 | Reward:  0.056 | Value:  0.713 | VLoss: 0.0211 | PLoss: -0.0115 | Ent: 0.1834
+[ 58.5%] Iter 5712/9765 | SPS:  2356 | Reward:  0.065 | Value:  0.694 | VLoss: 0.0212 | PLoss: -0.0028 | Ent: 0.2168
+[ 58.5%] Iter 5713/9765 | SPS:  2356 | Reward:  0.057 | Value:  0.698 | VLoss: 0.0234 | PLoss: -0.0067 | Ent: 0.2029
+[ 58.5%] Iter 5714/9765 | SPS:  2356 | Reward:  0.046 | Value:  0.664 | VLoss: 0.0255 | PLoss: -0.0092 | Ent: 0.1767
+[ 58.5%] Iter 5715/9765 | SPS:  2356 | Reward:  0.072 | Value:  0.674 | VLoss: 0.0317 | PLoss: -0.0122 | Ent: 0.1837
+[ 58.5%] Iter 5716/9765 | SPS:  2356 | Reward:  0.066 | Value:  0.702 | VLoss: 0.0257 | PLoss: -0.0093 | Ent: 0.2453
+[ 58.5%] Iter 5717/9765 | SPS:  2356 | Reward:  0.066 | Value:  0.693 | VLoss: 0.0219 | PLoss: -0.0006 | Ent: 0.1971
+[ 58.6%] Iter 5718/9765 | SPS:  2357 | Reward:  0.063 | Value:  0.653 | VLoss: 0.0232 | PLoss: -0.0068 | Ent: 0.1943
+[ 58.6%] Iter 5719/9765 | SPS:  2357 | Reward:  0.069 | Value:  0.702 | VLoss: 0.0240 | PLoss: -0.0062 | Ent: 0.1450
+[ 58.6%] Iter 5720/9765 | SPS:  2357 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0245 | PLoss: -0.0067 | Ent: 0.2429
+[ 58.6%] Iter 5721/9765 | SPS:  2357 | Reward:  0.066 | Value:  0.691 | VLoss: 0.0270 | PLoss: -0.0167 | Ent: 0.1918
+[ 58.6%] Iter 5722/9765 | SPS:  2357 | Reward:  0.066 | Value:  0.739 | VLoss: 0.0233 | PLoss: -0.0008 | Ent: 0.2198
+[ 58.6%] Iter 5723/9765 | SPS:  2357 | Reward:  0.069 | Value:  0.735 | VLoss: 0.0242 | PLoss: -0.0055 | Ent: 0.1816
+[ 58.6%] Iter 5724/9765 | SPS:  2357 | Reward:  0.059 | Value:  0.821 | VLoss: 0.0220 | PLoss: -0.0104 | Ent: 0.1970
+[ 58.6%] Iter 5725/9765 | SPS:  2358 | Reward:  0.067 | Value:  0.693 | VLoss: 0.0249 | PLoss: 0.0006 | Ent: 0.1876
+[ 58.6%] Iter 5726/9765 | SPS:  2358 | Reward:  0.079 | Value:  0.738 | VLoss: 0.0325 | PLoss: -0.0031 | Ent: 0.1914
+[ 58.6%] Iter 5727/9765 | SPS:  2358 | Reward:  0.058 | Value:  0.634 | VLoss: 0.0257 | PLoss: -0.0053 | Ent: 0.2304
+[ 58.7%] Iter 5728/9765 | SPS:  2358 | Reward:  0.068 | Value:  0.689 | VLoss: 0.0239 | PLoss: -0.0066 | Ent: 0.2274
+[ 58.7%] Iter 5729/9765 | SPS:  2358 | Reward:  0.051 | Value:  0.658 | VLoss: 0.0344 | PLoss: -0.0025 | Ent: 0.2008
+[ 58.7%] Iter 5730/9765 | SPS:  2358 | Reward:  0.042 | Value:  0.639 | VLoss: 0.0232 | PLoss: -0.0119 | Ent: 0.2183
+[ 58.7%] Iter 5731/9765 | SPS:  2358 | Reward:  0.061 | Value:  0.648 | VLoss: 0.0277 | PLoss: -0.0027 | Ent: 0.2294
+[ 58.7%] Iter 5732/9765 | SPS:  2359 | Reward:  0.053 | Value:  0.676 | VLoss: 0.0269 | PLoss: -0.0111 | Ent: 0.2063
+[ 58.7%] Iter 5733/9765 | SPS:  2359 | Reward:  0.055 | Value:  0.679 | VLoss: 0.0224 | PLoss: -0.0099 | Ent: 0.2813
+[ 58.7%] Iter 5734/9765 | SPS:  2359 | Reward:  0.068 | Value:  0.706 | VLoss: 0.0338 | PLoss: 0.0042 | Ent: 0.1949
+[ 58.7%] Iter 5735/9765 | SPS:  2359 | Reward:  0.061 | Value:  0.581 | VLoss: 0.0255 | PLoss: -0.0063 | Ent: 0.2093
+[ 58.7%] Iter 5736/9765 | SPS:  2359 | Reward:  0.060 | Value:  0.626 | VLoss: 0.0252 | PLoss: -0.0054 | Ent: 0.2108
+[ 58.8%] Iter 5737/9765 | SPS:  2359 | Reward:  0.057 | Value:  0.743 | VLoss: 0.0232 | PLoss: -0.0030 | Ent: 0.2471
+[ 58.8%] Iter 5738/9765 | SPS:  2360 | Reward:  0.064 | Value:  0.738 | VLoss: 0.0282 | PLoss: -0.0071 | Ent: 0.1829
+[ 58.8%] Iter 5739/9765 | SPS:  2360 | Reward:  0.076 | Value:  0.746 | VLoss: 0.0285 | PLoss: -0.0081 | Ent: 0.1538
+[ 58.8%] Iter 5740/9765 | SPS:  2360 | Reward:  0.056 | Value:  0.680 | VLoss: 0.0199 | PLoss: -0.0084 | Ent: 0.1812
+[ 58.8%] Iter 5741/9765 | SPS:  2360 | Reward:  0.061 | Value:  0.779 | VLoss: 0.0247 | PLoss: -0.0102 | Ent: 0.2396
+[ 58.8%] Iter 5742/9765 | SPS:  2360 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0254 | PLoss: -0.0079 | Ent: 0.2620
+[ 58.8%] Iter 5743/9765 | SPS:  2360 | Reward:  0.064 | Value:  0.719 | VLoss: 0.0280 | PLoss: -0.0115 | Ent: 0.2056
+[ 58.8%] Iter 5744/9765 | SPS:  2360 | Reward:  0.057 | Value:  0.690 | VLoss: 0.0175 | PLoss: -0.0040 | Ent: 0.1740
+[ 58.8%] Iter 5745/9765 | SPS:  2361 | Reward:  0.054 | Value:  0.705 | VLoss: 0.0159 | PLoss: -0.0092 | Ent: 0.2043
+[ 58.8%] Iter 5746/9765 | SPS:  2361 | Reward:  0.071 | Value:  0.774 | VLoss: 0.0169 | PLoss: -0.0080 | Ent: 0.1849
+[ 58.9%] Iter 5747/9765 | SPS:  2361 | Reward:  0.059 | Value:  0.772 | VLoss: 0.0263 | PLoss: -0.0042 | Ent: 0.1961
+[ 58.9%] Iter 5748/9765 | SPS:  2361 | Reward:  0.061 | Value:  0.630 | VLoss: 0.0244 | PLoss: -0.0058 | Ent: 0.2179
+[ 58.9%] Iter 5749/9765 | SPS:  2361 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0240 | PLoss: -0.0052 | Ent: 0.2022
+[ 58.9%] Iter 5750/9765 | SPS:  2361 | Reward:  0.070 | Value:  0.725 | VLoss: 0.0292 | PLoss: -0.0020 | Ent: 0.1938
+[ 58.9%] Iter 5751/9765 | SPS:  2361 | Reward:  0.063 | Value:  0.714 | VLoss: 0.0224 | PLoss: -0.0057 | Ent: 0.1594
+[ 58.9%] Iter 5752/9765 | SPS:  2362 | Reward:  0.067 | Value:  0.689 | VLoss: 0.0237 | PLoss: -0.0047 | Ent: 0.2321
+[ 58.9%] Iter 5753/9765 | SPS:  2362 | Reward:  0.066 | Value:  0.741 | VLoss: 0.0252 | PLoss: -0.0108 | Ent: 0.2401
+[ 58.9%] Iter 5754/9765 | SPS:  2362 | Reward:  0.058 | Value:  0.702 | VLoss: 0.0245 | PLoss: -0.0024 | Ent: 0.2221
+[ 58.9%] Iter 5755/9765 | SPS:  2362 | Reward:  0.067 | Value:  0.736 | VLoss: 0.0229 | PLoss: -0.0098 | Ent: 0.2519
+[ 58.9%] Iter 5756/9765 | SPS:  2362 | Reward:  0.050 | Value:  0.720 | VLoss: 0.0276 | PLoss: -0.0072 | Ent: 0.2042
+[ 59.0%] Iter 5757/9765 | SPS:  2362 | Reward:  0.060 | Value:  0.685 | VLoss: 0.0376 | PLoss: -0.0069 | Ent: 0.1959
+[ 59.0%] Iter 5758/9765 | SPS:  2363 | Reward:  0.056 | Value:  0.664 | VLoss: 0.0302 | PLoss: -0.0038 | Ent: 0.1859
+[ 59.0%] Iter 5759/9765 | SPS:  2363 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0224 | PLoss: -0.0058 | Ent: 0.2277
+[ 59.0%] Iter 5760/9765 | SPS:  2363 | Reward:  0.053 | Value:  0.817 | VLoss: 0.0186 | PLoss: -0.0073 | Ent: 0.2557
+[ 59.0%] Iter 5761/9765 | SPS:  2363 | Reward:  0.087 | Value:  0.773 | VLoss: 0.0222 | PLoss: -0.0056 | Ent: 0.1930
+[ 59.0%] Iter 5762/9765 | SPS:  2363 | Reward:  0.065 | Value:  0.834 | VLoss: 0.0209 | PLoss: -0.0073 | Ent: 0.2172
+[ 59.0%] Iter 5763/9765 | SPS:  2363 | Reward:  0.070 | Value:  0.732 | VLoss: 0.0292 | PLoss: -0.0098 | Ent: 0.2485
+[ 59.0%] Iter 5764/9765 | SPS:  2363 | Reward:  0.058 | Value:  0.741 | VLoss: 0.0229 | PLoss: -0.0053 | Ent: 0.2310
+[ 59.0%] Iter 5765/9765 | SPS:  2364 | Reward:  0.058 | Value:  0.729 | VLoss: 0.0254 | PLoss: -0.0013 | Ent: 0.1882
+[ 59.0%] Iter 5766/9765 | SPS:  2364 | Reward:  0.071 | Value:  0.672 | VLoss: 0.0257 | PLoss: -0.0040 | Ent: 0.2023
+[ 59.1%] Iter 5767/9765 | SPS:  2364 | Reward:  0.071 | Value:  0.714 | VLoss: 0.0161 | PLoss: -0.0078 | Ent: 0.1806
+[ 59.1%] Iter 5768/9765 | SPS:  2364 | Reward:  0.064 | Value:  0.768 | VLoss: 0.0211 | PLoss: -0.0060 | Ent: 0.1809
+[ 59.1%] Iter 5769/9765 | SPS:  2364 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0160 | PLoss: -0.0107 | Ent: 0.2123
+[ 59.1%] Iter 5770/9765 | SPS:  2364 | Reward:  0.053 | Value:  0.787 | VLoss: 0.0185 | PLoss: -0.0067 | Ent: 0.2412
+[ 59.1%] Iter 5771/9765 | SPS:  2364 | Reward:  0.053 | Value:  0.786 | VLoss: 0.0229 | PLoss: -0.0046 | Ent: 0.1832
+[ 59.1%] Iter 5772/9765 | SPS:  2365 | Reward:  0.049 | Value:  0.708 | VLoss: 0.0164 | PLoss: -0.0062 | Ent: 0.1998
+[ 59.1%] Iter 5773/9765 | SPS:  2365 | Reward:  0.050 | Value:  0.701 | VLoss: 0.0226 | PLoss: -0.0003 | Ent: 0.2109
+[ 59.1%] Iter 5774/9765 | SPS:  2365 | Reward:  0.058 | Value:  0.680 | VLoss: 0.0257 | PLoss: -0.0117 | Ent: 0.2077
+[ 59.1%] Iter 5775/9765 | SPS:  2365 | Reward:  0.069 | Value:  0.727 | VLoss: 0.0254 | PLoss: -0.0083 | Ent: 0.2064
+[ 59.2%] Iter 5776/9765 | SPS:  2365 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0302 | PLoss: -0.0010 | Ent: 0.2022
+[ 59.2%] Iter 5777/9765 | SPS:  2365 | Reward:  0.060 | Value:  0.653 | VLoss: 0.0255 | PLoss: 0.0001 | Ent: 0.1975
+[ 59.2%] Iter 5778/9765 | SPS:  2366 | Reward:  0.074 | Value:  0.708 | VLoss: 0.0264 | PLoss: -0.0107 | Ent: 0.2101
+[ 59.2%] Iter 5779/9765 | SPS:  2366 | Reward:  0.057 | Value:  0.665 | VLoss: 0.0270 | PLoss: -0.0108 | Ent: 0.2784
+[ 59.2%] Iter 5780/9765 | SPS:  2366 | Reward:  0.057 | Value:  0.708 | VLoss: 0.0313 | PLoss: -0.0076 | Ent: 0.2035
+[ 59.2%] Iter 5781/9765 | SPS:  2366 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0259 | PLoss: -0.0028 | Ent: 0.1693
+[ 59.2%] Iter 5782/9765 | SPS:  2366 | Reward:  0.054 | Value:  0.656 | VLoss: 0.0215 | PLoss: -0.0037 | Ent: 0.2241
+[ 59.2%] Iter 5783/9765 | SPS:  2366 | Reward:  0.062 | Value:  0.629 | VLoss: 0.0315 | PLoss: -0.0098 | Ent: 0.1685
+[ 59.2%] Iter 5784/9765 | SPS:  2366 | Reward:  0.062 | Value:  0.791 | VLoss: 0.0189 | PLoss: -0.0085 | Ent: 0.2085
+[ 59.2%] Iter 5785/9765 | SPS:  2367 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0225 | PLoss: -0.0080 | Ent: 0.2231
+[ 59.3%] Iter 5786/9765 | SPS:  2367 | Reward:  0.056 | Value:  0.681 | VLoss: 0.0212 | PLoss: -0.0073 | Ent: 0.2042
+[ 59.3%] Iter 5787/9765 | SPS:  2367 | Reward:  0.061 | Value:  0.651 | VLoss: 0.0413 | PLoss: -0.0048 | Ent: 0.2074
+[ 59.3%] Iter 5788/9765 | SPS:  2367 | Reward:  0.061 | Value:  0.663 | VLoss: 0.0221 | PLoss: -0.0084 | Ent: 0.1986
+[ 59.3%] Iter 5789/9765 | SPS:  2367 | Reward:  0.060 | Value:  0.680 | VLoss: 0.0216 | PLoss: -0.0042 | Ent: 0.2046
+[ 59.3%] Iter 5790/9765 | SPS:  2367 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0208 | PLoss: -0.0003 | Ent: 0.2168
+[ 59.3%] Iter 5791/9765 | SPS:  2367 | Reward:  0.066 | Value:  0.780 | VLoss: 0.0179 | PLoss: -0.0065 | Ent: 0.1785
+[ 59.3%] Iter 5792/9765 | SPS:  2368 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0216 | PLoss: -0.0085 | Ent: 0.2114
+[ 59.3%] Iter 5793/9765 | SPS:  2368 | Reward:  0.057 | Value:  0.724 | VLoss: 0.0178 | PLoss: -0.0102 | Ent: 0.2243
+[ 59.3%] Iter 5794/9765 | SPS:  2368 | Reward:  0.066 | Value:  0.690 | VLoss: 0.0199 | PLoss: -0.0022 | Ent: 0.1810
+[ 59.3%] Iter 5795/9765 | SPS:  2368 | Reward:  0.061 | Value:  0.744 | VLoss: 0.0213 | PLoss: -0.0102 | Ent: 0.1769
+[ 59.4%] Iter 5796/9765 | SPS:  2368 | Reward:  0.060 | Value:  0.745 | VLoss: 0.0138 | PLoss: -0.0087 | Ent: 0.1773
+[ 59.4%] Iter 5797/9765 | SPS:  2368 | Reward:  0.071 | Value:  0.764 | VLoss: 0.0305 | PLoss: -0.0005 | Ent: 0.1726
+[ 59.4%] Iter 5798/9765 | SPS:  2368 | Reward:  0.061 | Value:  0.649 | VLoss: 0.0366 | PLoss: -0.0044 | Ent: 0.2444
+[ 59.4%] Iter 5799/9765 | SPS:  2369 | Reward:  0.080 | Value:  0.637 | VLoss: 0.0341 | PLoss: -0.0102 | Ent: 0.1992
+[ 59.4%] Iter 5800/9765 | SPS:  2369 | Reward:  0.075 | Value:  0.745 | VLoss: 0.0286 | PLoss: -0.0033 | Ent: 0.2101
+[ 59.4%] Iter 5801/9765 | SPS:  2369 | Reward:  0.049 | Value:  0.701 | VLoss: 0.0272 | PLoss: -0.0127 | Ent: 0.1935
+[ 59.4%] Iter 5802/9765 | SPS:  2369 | Reward:  0.067 | Value:  0.662 | VLoss: 0.0293 | PLoss: -0.0067 | Ent: 0.1753
+[ 59.4%] Iter 5803/9765 | SPS:  2369 | Reward:  0.057 | Value:  0.643 | VLoss: 0.0279 | PLoss: -0.0032 | Ent: 0.2369
+[ 59.4%] Iter 5804/9765 | SPS:  2369 | Reward:  0.063 | Value:  0.674 | VLoss: 0.0223 | PLoss: -0.0072 | Ent: 0.2069
+[ 59.4%] Iter 5805/9765 | SPS:  2369 | Reward:  0.047 | Value:  0.722 | VLoss: 0.0248 | PLoss: -0.0001 | Ent: 0.2216
+[ 59.5%] Iter 5806/9765 | SPS:  2370 | Reward:  0.053 | Value:  0.641 | VLoss: 0.0231 | PLoss: -0.0100 | Ent: 0.2452
+[ 59.5%] Iter 5807/9765 | SPS:  2370 | Reward:  0.048 | Value:  0.671 | VLoss: 0.0299 | PLoss: -0.0071 | Ent: 0.1942
+[ 59.5%] Iter 5808/9765 | SPS:  2370 | Reward:  0.055 | Value:  0.618 | VLoss: 0.0235 | PLoss: -0.0090 | Ent: 0.2077
+[ 59.5%] Iter 5809/9765 | SPS:  2370 | Reward:  0.063 | Value:  0.777 | VLoss: 0.0300 | PLoss: -0.0060 | Ent: 0.2238
+[ 59.5%] Iter 5810/9765 | SPS:  2370 | Reward:  0.056 | Value:  0.589 | VLoss: 0.0244 | PLoss: -0.0104 | Ent: 0.2552
+[ 59.5%] Iter 5811/9765 | SPS:  2370 | Reward:  0.057 | Value:  0.648 | VLoss: 0.0265 | PLoss: -0.0044 | Ent: 0.2439
+[ 59.5%] Iter 5812/9765 | SPS:  2370 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0288 | PLoss: 0.0024 | Ent: 0.2073
+[ 59.5%] Iter 5813/9765 | SPS:  2371 | Reward:  0.072 | Value:  0.769 | VLoss: 0.0256 | PLoss: -0.0017 | Ent: 0.1852
+[ 59.5%] Iter 5814/9765 | SPS:  2371 | Reward:  0.059 | Value:  0.728 | VLoss: 0.0236 | PLoss: -0.0086 | Ent: 0.1716
+[ 59.5%] Iter 5815/9765 | SPS:  2371 | Reward:  0.068 | Value:  0.716 | VLoss: 0.0209 | PLoss: -0.0071 | Ent: 0.2021
+[ 59.6%] Iter 5816/9765 | SPS:  2371 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0333 | PLoss: -0.0148 | Ent: 0.2372
+[ 59.6%] Iter 5817/9765 | SPS:  2371 | Reward:  0.058 | Value:  0.711 | VLoss: 0.0207 | PLoss: -0.0232 | Ent: 0.2336
+[ 59.6%] Iter 5818/9765 | SPS:  2371 | Reward:  0.055 | Value:  0.591 | VLoss: 0.0262 | PLoss: -0.0005 | Ent: 0.2322
+[ 59.6%] Iter 5819/9765 | SPS:  2371 | Reward:  0.071 | Value:  0.710 | VLoss: 0.0212 | PLoss: -0.0045 | Ent: 0.2169
+[ 59.6%] Iter 5820/9765 | SPS:  2372 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0220 | PLoss: -0.0060 | Ent: 0.2398
+[ 59.6%] Iter 5821/9765 | SPS:  2372 | Reward:  0.064 | Value:  0.761 | VLoss: 0.0218 | PLoss: -0.0044 | Ent: 0.2097
+[ 59.6%] Iter 5822/9765 | SPS:  2372 | Reward:  0.067 | Value:  0.749 | VLoss: 0.0180 | PLoss: -0.0042 | Ent: 0.1865
+[ 59.6%] Iter 5823/9765 | SPS:  2372 | Reward:  0.050 | Value:  0.714 | VLoss: 0.0330 | PLoss: -0.0152 | Ent: 0.2095
+[ 59.6%] Iter 5824/9765 | SPS:  2372 | Reward:  0.069 | Value:  0.676 | VLoss: 0.0294 | PLoss: -0.0091 | Ent: 0.2273
+[ 59.7%] Iter 5825/9765 | SPS:  2372 | Reward:  0.059 | Value:  0.763 | VLoss: 0.0204 | PLoss: -0.0042 | Ent: 0.2080
+[ 59.7%] Iter 5826/9765 | SPS:  2372 | Reward:  0.075 | Value:  0.751 | VLoss: 0.0305 | PLoss: -0.0083 | Ent: 0.2206
+[ 59.7%] Iter 5827/9765 | SPS:  2373 | Reward:  0.055 | Value:  0.638 | VLoss: 0.0199 | PLoss: -0.0037 | Ent: 0.2183
+[ 59.7%] Iter 5828/9765 | SPS:  2373 | Reward:  0.065 | Value:  0.746 | VLoss: 0.0216 | PLoss: -0.0052 | Ent: 0.1650
+[ 59.7%] Iter 5829/9765 | SPS:  2373 | Reward:  0.073 | Value:  0.710 | VLoss: 0.0274 | PLoss: -0.0092 | Ent: 0.2301
+[ 59.7%] Iter 5830/9765 | SPS:  2373 | Reward:  0.062 | Value:  0.700 | VLoss: 0.0167 | PLoss: -0.0115 | Ent: 0.2240
+[ 59.7%] Iter 5831/9765 | SPS:  2373 | Reward:  0.061 | Value:  0.792 | VLoss: 0.0190 | PLoss: -0.0169 | Ent: 0.2128
+[ 59.7%] Iter 5832/9765 | SPS:  2373 | Reward:  0.070 | Value:  0.735 | VLoss: 0.0293 | PLoss: -0.0058 | Ent: 0.2107
+[ 59.7%] Iter 5833/9765 | SPS:  2373 | Reward:  0.062 | Value:  0.597 | VLoss: 0.0267 | PLoss: -0.0065 | Ent: 0.2323
+[ 59.7%] Iter 5834/9765 | SPS:  2374 | Reward:  0.052 | Value:  0.673 | VLoss: 0.0248 | PLoss: -0.0084 | Ent: 0.1901
+[ 59.8%] Iter 5835/9765 | SPS:  2374 | Reward:  0.043 | Value:  0.641 | VLoss: 0.0252 | PLoss: -0.0040 | Ent: 0.2046
+[ 59.8%] Iter 5836/9765 | SPS:  2374 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0308 | PLoss: -0.0050 | Ent: 0.2131
+[ 59.8%] Iter 5837/9765 | SPS:  2374 | Reward:  0.058 | Value:  0.741 | VLoss: 0.0229 | PLoss: -0.0065 | Ent: 0.2021
+[ 59.8%] Iter 5838/9765 | SPS:  2374 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0164 | PLoss: -0.0050 | Ent: 0.1476
+[ 59.8%] Iter 5839/9765 | SPS:  2374 | Reward:  0.059 | Value:  0.695 | VLoss: 0.0216 | PLoss: -0.0068 | Ent: 0.2027
+[ 59.8%] Iter 5840/9765 | SPS:  2374 | Reward:  0.069 | Value:  0.721 | VLoss: 0.0314 | PLoss: -0.0112 | Ent: 0.2098
+[ 59.8%] Iter 5841/9765 | SPS:  2375 | Reward:  0.052 | Value:  0.729 | VLoss: 0.0169 | PLoss: -0.0129 | Ent: 0.2277
+[ 59.8%] Iter 5842/9765 | SPS:  2375 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0273 | PLoss: -0.0146 | Ent: 0.2077
+[ 59.8%] Iter 5843/9765 | SPS:  2375 | Reward:  0.068 | Value:  0.741 | VLoss: 0.0286 | PLoss: -0.0039 | Ent: 0.2657
+[ 59.8%] Iter 5844/9765 | SPS:  2375 | Reward:  0.068 | Value:  0.705 | VLoss: 0.0189 | PLoss: -0.0116 | Ent: 0.2662
+[ 59.9%] Iter 5845/9765 | SPS:  2375 | Reward:  0.043 | Value:  0.715 | VLoss: 0.0236 | PLoss: -0.0044 | Ent: 0.1768
+[ 59.9%] Iter 5846/9765 | SPS:  2375 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0315 | PLoss: -0.0115 | Ent: 0.1916
+[ 59.9%] Iter 5847/9765 | SPS:  2375 | Reward:  0.073 | Value:  0.778 | VLoss: 0.0235 | PLoss: -0.0042 | Ent: 0.1938
+[ 59.9%] Iter 5848/9765 | SPS:  2376 | Reward:  0.079 | Value:  0.692 | VLoss: 0.0272 | PLoss: -0.0135 | Ent: 0.1942
+[ 59.9%] Iter 5849/9765 | SPS:  2376 | Reward:  0.063 | Value:  0.737 | VLoss: 0.0331 | PLoss: -0.0120 | Ent: 0.2072
+[ 59.9%] Iter 5850/9765 | SPS:  2376 | Reward:  0.066 | Value:  0.700 | VLoss: 0.0263 | PLoss: -0.0021 | Ent: 0.2589
+[ 59.9%] Iter 5851/9765 | SPS:  2376 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0312 | PLoss: -0.0124 | Ent: 0.2276
+[ 59.9%] Iter 5852/9765 | SPS:  2376 | Reward:  0.059 | Value:  0.685 | VLoss: 0.0210 | PLoss: -0.0225 | Ent: 0.2393
+[ 59.9%] Iter 5853/9765 | SPS:  2376 | Reward:  0.063 | Value:  0.673 | VLoss: 0.0239 | PLoss: -0.0025 | Ent: 0.2227
+[ 59.9%] Iter 5854/9765 | SPS:  2376 | Reward:  0.081 | Value:  0.691 | VLoss: 0.0327 | PLoss: -0.0111 | Ent: 0.1549
+[ 60.0%] Iter 5855/9765 | SPS:  2377 | Reward:  0.066 | Value:  0.721 | VLoss: 0.0204 | PLoss: -0.0109 | Ent: 0.1575
+[ 60.0%] Iter 5856/9765 | SPS:  2377 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0181 | PLoss: -0.0114 | Ent: 0.2248
+[ 60.0%] Iter 5857/9765 | SPS:  2377 | Reward:  0.055 | Value:  0.754 | VLoss: 0.0203 | PLoss: -0.0081 | Ent: 0.2112
+[ 60.0%] Iter 5858/9765 | SPS:  2377 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0185 | PLoss: -0.0106 | Ent: 0.1688
+[ 60.0%] Iter 5859/9765 | SPS:  2377 | Reward:  0.060 | Value:  0.687 | VLoss: 0.0274 | PLoss: -0.0046 | Ent: 0.1974
+[ 60.0%] Iter 5860/9765 | SPS:  2377 | Reward:  0.054 | Value:  0.702 | VLoss: 0.0264 | PLoss: -0.0096 | Ent: 0.2455
+[ 60.0%] Iter 5861/9765 | SPS:  2377 | Reward:  0.059 | Value:  0.666 | VLoss: 0.0232 | PLoss: -0.0060 | Ent: 0.2094
+[ 60.0%] Iter 5862/9765 | SPS:  2378 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0255 | PLoss: -0.0041 | Ent: 0.2374
+[ 60.0%] Iter 5863/9765 | SPS:  2378 | Reward:  0.066 | Value:  0.725 | VLoss: 0.0257 | PLoss: -0.0054 | Ent: 0.2154
+[ 60.1%] Iter 5864/9765 | SPS:  2378 | Reward:  0.062 | Value:  0.751 | VLoss: 0.0329 | PLoss: -0.0124 | Ent: 0.1853
+[ 60.1%] Iter 5865/9765 | SPS:  2378 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0273 | PLoss: -0.0096 | Ent: 0.1898
+[ 60.1%] Iter 5866/9765 | SPS:  2378 | Reward:  0.060 | Value:  0.744 | VLoss: 0.0151 | PLoss: -0.0047 | Ent: 0.2253
+[ 60.1%] Iter 5867/9765 | SPS:  2378 | Reward:  0.047 | Value:  0.656 | VLoss: 0.0245 | PLoss: -0.0134 | Ent: 0.2088
+[ 60.1%] Iter 5868/9765 | SPS:  2378 | Reward:  0.050 | Value:  0.613 | VLoss: 0.0216 | PLoss: -0.0072 | Ent: 0.1910
+[ 60.1%] Iter 5869/9765 | SPS:  2379 | Reward:  0.048 | Value:  0.659 | VLoss: 0.0214 | PLoss: -0.0067 | Ent: 0.1981
+[ 60.1%] Iter 5870/9765 | SPS:  2379 | Reward:  0.062 | Value:  0.651 | VLoss: 0.0219 | PLoss: 0.0014 | Ent: 0.2085
+[ 60.1%] Iter 5871/9765 | SPS:  2379 | Reward:  0.052 | Value:  0.629 | VLoss: 0.0253 | PLoss: -0.0078 | Ent: 0.1820
+[ 60.1%] Iter 5872/9765 | SPS:  2379 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0188 | PLoss: -0.0057 | Ent: 0.2200
+[ 60.1%] Iter 5873/9765 | SPS:  2379 | Reward:  0.056 | Value:  0.682 | VLoss: 0.0149 | PLoss: -0.0110 | Ent: 0.1741
+[ 60.2%] Iter 5874/9765 | SPS:  2379 | Reward:  0.064 | Value:  0.708 | VLoss: 0.0206 | PLoss: -0.0068 | Ent: 0.2010
+[ 60.2%] Iter 5875/9765 | SPS:  2379 | Reward:  0.053 | Value:  0.665 | VLoss: 0.0229 | PLoss: -0.0030 | Ent: 0.1821
+[ 60.2%] Iter 5876/9765 | SPS:  2380 | Reward:  0.048 | Value:  0.645 | VLoss: 0.0155 | PLoss: -0.0036 | Ent: 0.1692
+[ 60.2%] Iter 5877/9765 | SPS:  2380 | Reward:  0.039 | Value:  0.645 | VLoss: 0.0339 | PLoss: -0.0152 | Ent: 0.2074
+[ 60.2%] Iter 5878/9765 | SPS:  2380 | Reward:  0.062 | Value:  0.577 | VLoss: 0.0290 | PLoss: -0.0150 | Ent: 0.2531
+[ 60.2%] Iter 5879/9765 | SPS:  2380 | Reward:  0.063 | Value:  0.637 | VLoss: 0.0302 | PLoss: -0.0107 | Ent: 0.2039
+[ 60.2%] Iter 5880/9765 | SPS:  2380 | Reward:  0.067 | Value:  0.726 | VLoss: 0.0181 | PLoss: -0.0048 | Ent: 0.2007
+[ 60.2%] Iter 5881/9765 | SPS:  2380 | Reward:  0.052 | Value:  0.738 | VLoss: 0.0295 | PLoss: -0.0147 | Ent: 0.2465
+[ 60.2%] Iter 5882/9765 | SPS:  2380 | Reward:  0.059 | Value:  0.710 | VLoss: 0.0205 | PLoss: -0.0066 | Ent: 0.2005
+[ 60.2%] Iter 5883/9765 | SPS:  2381 | Reward:  0.067 | Value:  0.754 | VLoss: 0.0234 | PLoss: -0.0092 | Ent: 0.2057
+[ 60.3%] Iter 5884/9765 | SPS:  2381 | Reward:  0.062 | Value:  0.754 | VLoss: 0.0238 | PLoss: -0.0102 | Ent: 0.2200
+[ 60.3%] Iter 5885/9765 | SPS:  2381 | Reward:  0.072 | Value:  0.644 | VLoss: 0.0225 | PLoss: -0.0048 | Ent: 0.1718
+[ 60.3%] Iter 5886/9765 | SPS:  2381 | Reward:  0.074 | Value:  0.756 | VLoss: 0.0239 | PLoss: -0.0023 | Ent: 0.2023
+[ 60.3%] Iter 5887/9765 | SPS:  2381 | Reward:  0.065 | Value:  0.747 | VLoss: 0.0334 | PLoss: -0.0062 | Ent: 0.1944
+[ 60.3%] Iter 5888/9765 | SPS:  2381 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0230 | PLoss: -0.0155 | Ent: 0.2600
+[ 60.3%] Iter 5889/9765 | SPS:  2381 | Reward:  0.063 | Value:  0.718 | VLoss: 0.0193 | PLoss: -0.0101 | Ent: 0.2227
+[ 60.3%] Iter 5890/9765 | SPS:  2382 | Reward:  0.064 | Value:  0.764 | VLoss: 0.0181 | PLoss: -0.0068 | Ent: 0.2469
+[ 60.3%] Iter 5891/9765 | SPS:  2382 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0223 | PLoss: -0.0018 | Ent: 0.1994
+[ 60.3%] Iter 5892/9765 | SPS:  2382 | Reward:  0.059 | Value:  0.770 | VLoss: 0.0217 | PLoss: -0.0024 | Ent: 0.2012
+[ 60.3%] Iter 5893/9765 | SPS:  2382 | Reward:  0.054 | Value:  0.713 | VLoss: 0.0146 | PLoss: -0.0122 | Ent: 0.2342
+[ 60.4%] Iter 5894/9765 | SPS:  2382 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0198 | PLoss: -0.0135 | Ent: 0.2148
+[ 60.4%] Iter 5895/9765 | SPS:  2382 | Reward:  0.050 | Value:  0.751 | VLoss: 0.0209 | PLoss: -0.0036 | Ent: 0.1888
+[ 60.4%] Iter 5896/9765 | SPS:  2382 | Reward:  0.059 | Value:  0.726 | VLoss: 0.0197 | PLoss: -0.0067 | Ent: 0.2397
+[ 60.4%] Iter 5897/9765 | SPS:  2383 | Reward:  0.066 | Value:  0.725 | VLoss: 0.0203 | PLoss: -0.0189 | Ent: 0.2384
+[ 60.4%] Iter 5898/9765 | SPS:  2383 | Reward:  0.064 | Value:  0.781 | VLoss: 0.0203 | PLoss: -0.0062 | Ent: 0.1773
+[ 60.4%] Iter 5899/9765 | SPS:  2383 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0220 | PLoss: -0.0012 | Ent: 0.1960
+[ 60.4%] Iter 5900/9765 | SPS:  2383 | Reward:  0.071 | Value:  0.649 | VLoss: 0.0339 | PLoss: -0.0004 | Ent: 0.2098
+[ 60.4%] Iter 5901/9765 | SPS:  2383 | Reward:  0.065 | Value:  0.779 | VLoss: 0.0271 | PLoss: -0.0080 | Ent: 0.2261
+[ 60.4%] Iter 5902/9765 | SPS:  2383 | Reward:  0.071 | Value:  0.709 | VLoss: 0.0241 | PLoss: -0.0114 | Ent: 0.2022
+[ 60.5%] Iter 5903/9765 | SPS:  2383 | Reward:  0.051 | Value:  0.733 | VLoss: 0.0166 | PLoss: -0.0083 | Ent: 0.1344
+[ 60.5%] Iter 5904/9765 | SPS:  2384 | Reward:  0.060 | Value:  0.713 | VLoss: 0.0298 | PLoss: -0.0091 | Ent: 0.2401
+[ 60.5%] Iter 5905/9765 | SPS:  2384 | Reward:  0.056 | Value:  0.685 | VLoss: 0.0270 | PLoss: -0.0063 | Ent: 0.2556
+[ 60.5%] Iter 5906/9765 | SPS:  2384 | Reward:  0.057 | Value:  0.649 | VLoss: 0.0311 | PLoss: 0.0005 | Ent: 0.2062
+[ 60.5%] Iter 5907/9765 | SPS:  2384 | Reward:  0.058 | Value:  0.704 | VLoss: 0.0172 | PLoss: -0.0071 | Ent: 0.1749
+[ 60.5%] Iter 5908/9765 | SPS:  2384 | Reward:  0.070 | Value:  0.765 | VLoss: 0.0274 | PLoss: -0.0052 | Ent: 0.1982
+[ 60.5%] Iter 5909/9765 | SPS:  2384 | Reward:  0.064 | Value:  0.766 | VLoss: 0.0196 | PLoss: -0.0132 | Ent: 0.2470
+[ 60.5%] Iter 5910/9765 | SPS:  2384 | Reward:  0.051 | Value:  0.759 | VLoss: 0.0250 | PLoss: -0.0067 | Ent: 0.2090
+[ 60.5%] Iter 5911/9765 | SPS:  2385 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0238 | PLoss: -0.0031 | Ent: 0.2062
+[ 60.5%] Iter 5912/9765 | SPS:  2385 | Reward:  0.063 | Value:  0.789 | VLoss: 0.0290 | PLoss: -0.0006 | Ent: 0.2101
+[ 60.6%] Iter 5913/9765 | SPS:  2385 | Reward:  0.074 | Value:  0.662 | VLoss: 0.0273 | PLoss: -0.0117 | Ent: 0.2382
+[ 60.6%] Iter 5914/9765 | SPS:  2385 | Reward:  0.063 | Value:  0.701 | VLoss: 0.0221 | PLoss: -0.0091 | Ent: 0.2113
+[ 60.6%] Iter 5915/9765 | SPS:  2385 | Reward:  0.067 | Value:  0.770 | VLoss: 0.0236 | PLoss: -0.0065 | Ent: 0.2259
+[ 60.6%] Iter 5916/9765 | SPS:  2385 | Reward:  0.054 | Value:  0.763 | VLoss: 0.0253 | PLoss: -0.0019 | Ent: 0.2504
+[ 60.6%] Iter 5917/9765 | SPS:  2385 | Reward:  0.072 | Value:  0.771 | VLoss: 0.0262 | PLoss: -0.0101 | Ent: 0.2239
+[ 60.6%] Iter 5918/9765 | SPS:  2386 | Reward:  0.066 | Value:  0.796 | VLoss: 0.0192 | PLoss: -0.0038 | Ent: 0.2251
+[ 60.6%] Iter 5919/9765 | SPS:  2386 | Reward:  0.056 | Value:  0.759 | VLoss: 0.0270 | PLoss: -0.0031 | Ent: 0.1990
+[ 60.6%] Iter 5920/9765 | SPS:  2386 | Reward:  0.054 | Value:  0.665 | VLoss: 0.0223 | PLoss: -0.0085 | Ent: 0.2348
+[ 60.6%] Iter 5921/9765 | SPS:  2386 | Reward:  0.065 | Value:  0.676 | VLoss: 0.0233 | PLoss: -0.0149 | Ent: 0.2698
+[ 60.6%] Iter 5922/9765 | SPS:  2386 | Reward:  0.060 | Value:  0.790 | VLoss: 0.0205 | PLoss: -0.0053 | Ent: 0.2238
+[ 60.7%] Iter 5923/9765 | SPS:  2386 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0190 | PLoss: -0.0101 | Ent: 0.1529
+[ 60.7%] Iter 5924/9765 | SPS:  2386 | Reward:  0.054 | Value:  0.724 | VLoss: 0.0249 | PLoss: -0.0101 | Ent: 0.2308
+[ 60.7%] Iter 5925/9765 | SPS:  2387 | Reward:  0.063 | Value:  0.675 | VLoss: 0.0310 | PLoss: -0.0070 | Ent: 0.2024
+[ 60.7%] Iter 5926/9765 | SPS:  2387 | Reward:  0.054 | Value:  0.715 | VLoss: 0.0245 | PLoss: -0.0062 | Ent: 0.1885
+[ 60.7%] Iter 5927/9765 | SPS:  2387 | Reward:  0.062 | Value:  0.763 | VLoss: 0.0229 | PLoss: -0.0087 | Ent: 0.2164
+[ 60.7%] Iter 5928/9765 | SPS:  2387 | Reward:  0.071 | Value:  0.702 | VLoss: 0.0252 | PLoss: -0.0058 | Ent: 0.1818
+[ 60.7%] Iter 5929/9765 | SPS:  2387 | Reward:  0.066 | Value:  0.740 | VLoss: 0.0258 | PLoss: -0.0073 | Ent: 0.2439
+[ 60.7%] Iter 5930/9765 | SPS:  2387 | Reward:  0.069 | Value:  0.727 | VLoss: 0.0226 | PLoss: -0.0024 | Ent: 0.2524
+[ 60.7%] Iter 5931/9765 | SPS:  2387 | Reward:  0.069 | Value:  0.728 | VLoss: 0.0289 | PLoss: -0.0046 | Ent: 0.2383
+[ 60.7%] Iter 5932/9765 | SPS:  2388 | Reward:  0.064 | Value:  0.713 | VLoss: 0.0230 | PLoss: -0.0026 | Ent: 0.1828
+[ 60.8%] Iter 5933/9765 | SPS:  2388 | Reward:  0.069 | Value:  0.746 | VLoss: 0.0237 | PLoss: -0.0041 | Ent: 0.1732
+[ 60.8%] Iter 5934/9765 | SPS:  2388 | Reward:  0.077 | Value:  0.761 | VLoss: 0.0278 | PLoss: -0.0045 | Ent: 0.1683
+[ 60.8%] Iter 5935/9765 | SPS:  2388 | Reward:  0.066 | Value:  0.683 | VLoss: 0.0195 | PLoss: -0.0030 | Ent: 0.2177
+[ 60.8%] Iter 5936/9765 | SPS:  2388 | Reward:  0.055 | Value:  0.729 | VLoss: 0.0225 | PLoss: -0.0056 | Ent: 0.1569
+[ 60.8%] Iter 5937/9765 | SPS:  2388 | Reward:  0.069 | Value:  0.712 | VLoss: 0.0261 | PLoss: -0.0044 | Ent: 0.2361
+[ 60.8%] Iter 5938/9765 | SPS:  2388 | Reward:  0.070 | Value:  0.734 | VLoss: 0.0197 | PLoss: -0.0078 | Ent: 0.1813
+[ 60.8%] Iter 5939/9765 | SPS:  2388 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0334 | PLoss: -0.0089 | Ent: 0.2149
+[ 60.8%] Iter 5940/9765 | SPS:  2389 | Reward:  0.055 | Value:  0.676 | VLoss: 0.0204 | PLoss: -0.0051 | Ent: 0.1782
+[ 60.8%] Iter 5941/9765 | SPS:  2389 | Reward:  0.071 | Value:  0.717 | VLoss: 0.0224 | PLoss: -0.0074 | Ent: 0.2161
+[ 60.8%] Iter 5942/9765 | SPS:  2389 | Reward:  0.058 | Value:  0.796 | VLoss: 0.0209 | PLoss: -0.0091 | Ent: 0.2025
+[ 60.9%] Iter 5943/9765 | SPS:  2389 | Reward:  0.063 | Value:  0.691 | VLoss: 0.0240 | PLoss: -0.0046 | Ent: 0.1769
+[ 60.9%] Iter 5944/9765 | SPS:  2389 | Reward:  0.064 | Value:  0.763 | VLoss: 0.0250 | PLoss: -0.0063 | Ent: 0.1980
+[ 60.9%] Iter 5945/9765 | SPS:  2389 | Reward:  0.070 | Value:  0.702 | VLoss: 0.0218 | PLoss: -0.0037 | Ent: 0.1452
+[ 60.9%] Iter 5946/9765 | SPS:  2389 | Reward:  0.064 | Value:  0.709 | VLoss: 0.0185 | PLoss: -0.0060 | Ent: 0.1623
+[ 60.9%] Iter 5947/9765 | SPS:  2390 | Reward:  0.054 | Value:  0.652 | VLoss: 0.0211 | PLoss: -0.0036 | Ent: 0.2997
+[ 60.9%] Iter 5948/9765 | SPS:  2390 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0314 | PLoss: -0.0138 | Ent: 0.1981
+[ 60.9%] Iter 5949/9765 | SPS:  2390 | Reward:  0.064 | Value:  0.784 | VLoss: 0.0163 | PLoss: -0.0091 | Ent: 0.1836
+[ 60.9%] Iter 5950/9765 | SPS:  2390 | Reward:  0.055 | Value:  0.743 | VLoss: 0.0299 | PLoss: -0.0093 | Ent: 0.2286
+[ 60.9%] Iter 5951/9765 | SPS:  2390 | Reward:  0.069 | Value:  0.714 | VLoss: 0.0177 | PLoss: -0.0044 | Ent: 0.1658
+[ 61.0%] Iter 5952/9765 | SPS:  2390 | Reward:  0.064 | Value:  0.785 | VLoss: 0.0213 | PLoss: -0.0090 | Ent: 0.1680
+[ 61.0%] Iter 5953/9765 | SPS:  2390 | Reward:  0.055 | Value:  0.701 | VLoss: 0.0183 | PLoss: -0.0074 | Ent: 0.1729
+[ 61.0%] Iter 5954/9765 | SPS:  2391 | Reward:  0.065 | Value:  0.695 | VLoss: 0.0294 | PLoss: -0.0061 | Ent: 0.1646
+[ 61.0%] Iter 5955/9765 | SPS:  2391 | Reward:  0.057 | Value:  0.730 | VLoss: 0.0225 | PLoss: -0.0092 | Ent: 0.2409
+[ 61.0%] Iter 5956/9765 | SPS:  2391 | Reward:  0.070 | Value:  0.767 | VLoss: 0.0174 | PLoss: -0.0069 | Ent: 0.1961
+[ 61.0%] Iter 5957/9765 | SPS:  2391 | Reward:  0.068 | Value:  0.735 | VLoss: 0.0213 | PLoss: -0.0012 | Ent: 0.1564
+[ 61.0%] Iter 5958/9765 | SPS:  2391 | Reward:  0.067 | Value:  0.713 | VLoss: 0.0260 | PLoss: -0.0025 | Ent: 0.2222
+[ 61.0%] Iter 5959/9765 | SPS:  2391 | Reward:  0.066 | Value:  0.736 | VLoss: 0.0260 | PLoss: -0.0058 | Ent: 0.1650
+[ 61.0%] Iter 5960/9765 | SPS:  2391 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0260 | PLoss: -0.0056 | Ent: 0.1821
+[ 61.0%] Iter 5961/9765 | SPS:  2392 | Reward:  0.071 | Value:  0.643 | VLoss: 0.0343 | PLoss: -0.0051 | Ent: 0.2095
+[ 61.1%] Iter 5962/9765 | SPS:  2392 | Reward:  0.067 | Value:  0.752 | VLoss: 0.0243 | PLoss: -0.0099 | Ent: 0.1804
+[ 61.1%] Iter 5963/9765 | SPS:  2392 | Reward:  0.065 | Value:  0.763 | VLoss: 0.0179 | PLoss: -0.0047 | Ent: 0.2346
+[ 61.1%] Iter 5964/9765 | SPS:  2392 | Reward:  0.056 | Value:  0.681 | VLoss: 0.0159 | PLoss: -0.0097 | Ent: 0.1862
+[ 61.1%] Iter 5965/9765 | SPS:  2392 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0215 | PLoss: -0.0055 | Ent: 0.1988
+[ 61.1%] Iter 5966/9765 | SPS:  2392 | Reward:  0.066 | Value:  0.727 | VLoss: 0.0286 | PLoss: -0.0048 | Ent: 0.1773
+[ 61.1%] Iter 5967/9765 | SPS:  2392 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0198 | PLoss: -0.0022 | Ent: 0.1803
+[ 61.1%] Iter 5968/9765 | SPS:  2393 | Reward:  0.053 | Value:  0.663 | VLoss: 0.0270 | PLoss: -0.0138 | Ent: 0.1951
+[ 61.1%] Iter 5969/9765 | SPS:  2393 | Reward:  0.056 | Value:  0.671 | VLoss: 0.0188 | PLoss: -0.0052 | Ent: 0.2030
+[ 61.1%] Iter 5970/9765 | SPS:  2393 | Reward:  0.070 | Value:  0.727 | VLoss: 0.0224 | PLoss: 0.0029 | Ent: 0.1838
+[ 61.1%] Iter 5971/9765 | SPS:  2393 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0315 | PLoss: -0.0060 | Ent: 0.2113
+[ 61.2%] Iter 5972/9765 | SPS:  2393 | Reward:  0.069 | Value:  0.700 | VLoss: 0.0295 | PLoss: -0.0076 | Ent: 0.2291
+[ 61.2%] Iter 5973/9765 | SPS:  2393 | Reward:  0.056 | Value:  0.756 | VLoss: 0.0177 | PLoss: -0.0033 | Ent: 0.1941
+[ 61.2%] Iter 5974/9765 | SPS:  2393 | Reward:  0.065 | Value:  0.718 | VLoss: 0.0238 | PLoss: 0.0030 | Ent: 0.2189
+[ 61.2%] Iter 5975/9765 | SPS:  2394 | Reward:  0.070 | Value:  0.691 | VLoss: 0.0247 | PLoss: -0.0079 | Ent: 0.1779
+[ 61.2%] Iter 5976/9765 | SPS:  2394 | Reward:  0.064 | Value:  0.731 | VLoss: 0.0287 | PLoss: -0.0064 | Ent: 0.2498
+[ 61.2%] Iter 5977/9765 | SPS:  2394 | Reward:  0.048 | Value:  0.667 | VLoss: 0.0238 | PLoss: -0.0050 | Ent: 0.2135
+[ 61.2%] Iter 5978/9765 | SPS:  2394 | Reward:  0.049 | Value:  0.592 | VLoss: 0.0184 | PLoss: -0.0062 | Ent: 0.2031
+[ 61.2%] Iter 5979/9765 | SPS:  2394 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0262 | PLoss: -0.0041 | Ent: 0.1816
+[ 61.2%] Iter 5980/9765 | SPS:  2394 | Reward:  0.070 | Value:  0.697 | VLoss: 0.0143 | PLoss: -0.0128 | Ent: 0.2009
+[ 61.2%] Iter 5981/9765 | SPS:  2394 | Reward:  0.067 | Value:  0.793 | VLoss: 0.0262 | PLoss: -0.0073 | Ent: 0.1965
+[ 61.3%] Iter 5982/9765 | SPS:  2395 | Reward:  0.067 | Value:  0.784 | VLoss: 0.0217 | PLoss: -0.0049 | Ent: 0.1704
+[ 61.3%] Iter 5983/9765 | SPS:  2395 | Reward:  0.075 | Value:  0.679 | VLoss: 0.0292 | PLoss: -0.0050 | Ent: 0.1761
+[ 61.3%] Iter 5984/9765 | SPS:  2395 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0224 | PLoss: -0.0049 | Ent: 0.1673
+[ 61.3%] Iter 5985/9765 | SPS:  2395 | Reward:  0.052 | Value:  0.710 | VLoss: 0.0293 | PLoss: -0.0046 | Ent: 0.1429
+[ 61.3%] Iter 5986/9765 | SPS:  2395 | Reward:  0.062 | Value:  0.748 | VLoss: 0.0221 | PLoss: -0.0114 | Ent: 0.1737
+[ 61.3%] Iter 5987/9765 | SPS:  2395 | Reward:  0.060 | Value:  0.738 | VLoss: 0.0244 | PLoss: -0.0091 | Ent: 0.1986
+[ 61.3%] Iter 5988/9765 | SPS:  2395 | Reward:  0.049 | Value:  0.729 | VLoss: 0.0202 | PLoss: -0.0077 | Ent: 0.1761
+[ 61.3%] Iter 5989/9765 | SPS:  2396 | Reward:  0.070 | Value:  0.695 | VLoss: 0.0293 | PLoss: -0.0133 | Ent: 0.2052
+[ 61.3%] Iter 5990/9765 | SPS:  2396 | Reward:  0.052 | Value:  0.775 | VLoss: 0.0209 | PLoss: -0.0049 | Ent: 0.2013
+[ 61.4%] Iter 5991/9765 | SPS:  2396 | Reward:  0.069 | Value:  0.675 | VLoss: 0.0231 | PLoss: -0.0080 | Ent: 0.1690
+[ 61.4%] Iter 5992/9765 | SPS:  2396 | Reward:  0.061 | Value:  0.745 | VLoss: 0.0180 | PLoss: -0.0099 | Ent: 0.2012
+[ 61.4%] Iter 5993/9765 | SPS:  2396 | Reward:  0.071 | Value:  0.721 | VLoss: 0.0263 | PLoss: -0.0075 | Ent: 0.1825
+[ 61.4%] Iter 5994/9765 | SPS:  2396 | Reward:  0.069 | Value:  0.710 | VLoss: 0.0247 | PLoss: -0.0122 | Ent: 0.1795
+[ 61.4%] Iter 5995/9765 | SPS:  2396 | Reward:  0.058 | Value:  0.794 | VLoss: 0.0201 | PLoss: -0.0047 | Ent: 0.2015
+[ 61.4%] Iter 5996/9765 | SPS:  2397 | Reward:  0.058 | Value:  0.636 | VLoss: 0.0313 | PLoss: -0.0047 | Ent: 0.2575
+[ 61.4%] Iter 5997/9765 | SPS:  2397 | Reward:  0.058 | Value:  0.656 | VLoss: 0.0307 | PLoss: -0.0114 | Ent: 0.2038
+[ 61.4%] Iter 5998/9765 | SPS:  2397 | Reward:  0.046 | Value:  0.653 | VLoss: 0.0167 | PLoss: -0.0081 | Ent: 0.1879
+[ 61.4%] Iter 5999/9765 | SPS:  2397 | Reward:  0.060 | Value:  0.612 | VLoss: 0.0242 | PLoss: -0.0076 | Ent: 0.1782
+[ 61.4%] Iter 6000/9765 | SPS:  2397 | Reward:  0.061 | Value:  0.666 | VLoss: 0.0206 | PLoss: -0.0009 | Ent: 0.2074
+[ 61.5%] Iter 6001/9765 | SPS:  2397 | Reward:  0.058 | Value:  0.762 | VLoss: 0.0206 | PLoss: -0.0133 | Ent: 0.1488
+[ 61.5%] Iter 6002/9765 | SPS:  2397 | Reward:  0.056 | Value:  0.737 | VLoss: 0.0161 | PLoss: -0.0122 | Ent: 0.2086
+[ 61.5%] Iter 6003/9765 | SPS:  2398 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0197 | PLoss: -0.0020 | Ent: 0.2037
+[ 61.5%] Iter 6004/9765 | SPS:  2398 | Reward:  0.071 | Value:  0.680 | VLoss: 0.0262 | PLoss: -0.0078 | Ent: 0.2316
+[ 61.5%] Iter 6005/9765 | SPS:  2398 | Reward:  0.050 | Value:  0.708 | VLoss: 0.0193 | PLoss: -0.0046 | Ent: 0.2014
+[ 61.5%] Iter 6006/9765 | SPS:  2398 | Reward:  0.054 | Value:  0.671 | VLoss: 0.0235 | PLoss: -0.0092 | Ent: 0.2116
+[ 61.5%] Iter 6007/9765 | SPS:  2398 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0259 | PLoss: -0.0069 | Ent: 0.1717
+[ 61.5%] Iter 6008/9765 | SPS:  2398 | Reward:  0.050 | Value:  0.658 | VLoss: 0.0242 | PLoss: -0.0068 | Ent: 0.1905
+[ 61.5%] Iter 6009/9765 | SPS:  2398 | Reward:  0.060 | Value:  0.662 | VLoss: 0.0183 | PLoss: -0.0117 | Ent: 0.2538
+[ 61.5%] Iter 6010/9765 | SPS:  2399 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0245 | PLoss: -0.0065 | Ent: 0.1894
+[ 61.6%] Iter 6011/9765 | SPS:  2399 | Reward:  0.057 | Value:  0.783 | VLoss: 0.0261 | PLoss: 0.0000 | Ent: 0.1708
+[ 61.6%] Iter 6012/9765 | SPS:  2399 | Reward:  0.074 | Value:  0.734 | VLoss: 0.0284 | PLoss: -0.0038 | Ent: 0.1724
+[ 61.6%] Iter 6013/9765 | SPS:  2399 | Reward:  0.048 | Value:  0.663 | VLoss: 0.0181 | PLoss: -0.0045 | Ent: 0.1780
+[ 61.6%] Iter 6014/9765 | SPS:  2399 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0263 | PLoss: -0.0034 | Ent: 0.2025
+[ 61.6%] Iter 6015/9765 | SPS:  2399 | Reward:  0.070 | Value:  0.724 | VLoss: 0.0247 | PLoss: -0.0076 | Ent: 0.2224
+[ 61.6%] Iter 6016/9765 | SPS:  2400 | Reward:  0.058 | Value:  0.754 | VLoss: 0.0201 | PLoss: -0.0014 | Ent: 0.1866
+[ 61.6%] Iter 6017/9765 | SPS:  2400 | Reward:  0.071 | Value:  0.683 | VLoss: 0.0253 | PLoss: -0.0149 | Ent: 0.2405
+[ 61.6%] Iter 6018/9765 | SPS:  2400 | Reward:  0.066 | Value:  0.748 | VLoss: 0.0286 | PLoss: -0.0088 | Ent: 0.2347
+[ 61.6%] Iter 6019/9765 | SPS:  2400 | Reward:  0.059 | Value:  0.667 | VLoss: 0.0257 | PLoss: -0.0099 | Ent: 0.2499
+[ 61.6%] Iter 6020/9765 | SPS:  2400 | Reward:  0.075 | Value:  0.667 | VLoss: 0.0265 | PLoss: -0.0053 | Ent: 0.2109
+[ 61.7%] Iter 6021/9765 | SPS:  2400 | Reward:  0.041 | Value:  0.706 | VLoss: 0.0195 | PLoss: -0.0104 | Ent: 0.1696
+[ 61.7%] Iter 6022/9765 | SPS:  2400 | Reward:  0.064 | Value:  0.702 | VLoss: 0.0187 | PLoss: -0.0146 | Ent: 0.1631
+[ 61.7%] Iter 6023/9765 | SPS:  2400 | Reward:  0.062 | Value:  0.725 | VLoss: 0.0247 | PLoss: -0.0106 | Ent: 0.2252
+[ 61.7%] Iter 6024/9765 | SPS:  2401 | Reward:  0.076 | Value:  0.756 | VLoss: 0.0234 | PLoss: -0.0032 | Ent: 0.1821
+[ 61.7%] Iter 6025/9765 | SPS:  2401 | Reward:  0.053 | Value:  0.745 | VLoss: 0.0169 | PLoss: -0.0069 | Ent: 0.2192
+[ 61.7%] Iter 6026/9765 | SPS:  2401 | Reward:  0.048 | Value:  0.757 | VLoss: 0.0170 | PLoss: -0.0066 | Ent: 0.1929
+[ 61.7%] Iter 6027/9765 | SPS:  2401 | Reward:  0.059 | Value:  0.747 | VLoss: 0.0167 | PLoss: -0.0042 | Ent: 0.1724
+[ 61.7%] Iter 6028/9765 | SPS:  2401 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0257 | PLoss: -0.0043 | Ent: 0.1892
+[ 61.7%] Iter 6029/9765 | SPS:  2401 | Reward:  0.061 | Value:  0.674 | VLoss: 0.0203 | PLoss: -0.0050 | Ent: 0.1749
+[ 61.8%] Iter 6030/9765 | SPS:  2401 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0220 | PLoss: -0.0037 | Ent: 0.1655
+[ 61.8%] Iter 6031/9765 | SPS:  2402 | Reward:  0.060 | Value:  0.714 | VLoss: 0.0294 | PLoss: -0.0089 | Ent: 0.1803
+[ 61.8%] Iter 6032/9765 | SPS:  2402 | Reward:  0.050 | Value:  0.745 | VLoss: 0.0262 | PLoss: -0.0093 | Ent: 0.1938
+[ 61.8%] Iter 6033/9765 | SPS:  2402 | Reward:  0.073 | Value:  0.708 | VLoss: 0.0253 | PLoss: -0.0073 | Ent: 0.2033
+[ 61.8%] Iter 6034/9765 | SPS:  2402 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0269 | PLoss: -0.0069 | Ent: 0.2367
+[ 61.8%] Iter 6035/9765 | SPS:  2402 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0200 | PLoss: -0.0043 | Ent: 0.1728
+[ 61.8%] Iter 6036/9765 | SPS:  2402 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0292 | PLoss: -0.0085 | Ent: 0.1726
+[ 61.8%] Iter 6037/9765 | SPS:  2402 | Reward:  0.068 | Value:  0.746 | VLoss: 0.0216 | PLoss: -0.0179 | Ent: 0.1966
+[ 61.8%] Iter 6038/9765 | SPS:  2403 | Reward:  0.056 | Value:  0.818 | VLoss: 0.0165 | PLoss: -0.0011 | Ent: 0.1862
+[ 61.8%] Iter 6039/9765 | SPS:  2403 | Reward:  0.060 | Value:  0.776 | VLoss: 0.0167 | PLoss: -0.0025 | Ent: 0.1974
+[ 61.9%] Iter 6040/9765 | SPS:  2403 | Reward:  0.061 | Value:  0.712 | VLoss: 0.0220 | PLoss: -0.0028 | Ent: 0.2102
+[ 61.9%] Iter 6041/9765 | SPS:  2403 | Reward:  0.055 | Value:  0.789 | VLoss: 0.0149 | PLoss: -0.0038 | Ent: 0.2227
+[ 61.9%] Iter 6042/9765 | SPS:  2403 | Reward:  0.063 | Value:  0.784 | VLoss: 0.0164 | PLoss: -0.0020 | Ent: 0.1760
+[ 61.9%] Iter 6043/9765 | SPS:  2403 | Reward:  0.059 | Value:  0.730 | VLoss: 0.0224 | PLoss: -0.0038 | Ent: 0.1702
+[ 61.9%] Iter 6044/9765 | SPS:  2403 | Reward:  0.067 | Value:  0.655 | VLoss: 0.0349 | PLoss: -0.0098 | Ent: 0.1900
+[ 61.9%] Iter 6045/9765 | SPS:  2404 | Reward:  0.071 | Value:  0.698 | VLoss: 0.0241 | PLoss: -0.0031 | Ent: 0.1879
+[ 61.9%] Iter 6046/9765 | SPS:  2404 | Reward:  0.068 | Value:  0.811 | VLoss: 0.0248 | PLoss: -0.0111 | Ent: 0.2150
+[ 61.9%] Iter 6047/9765 | SPS:  2404 | Reward:  0.070 | Value:  0.770 | VLoss: 0.0236 | PLoss: -0.0078 | Ent: 0.1962
+[ 61.9%] Iter 6048/9765 | SPS:  2404 | Reward:  0.048 | Value:  0.695 | VLoss: 0.0217 | PLoss: -0.0099 | Ent: 0.2320
+[ 61.9%] Iter 6049/9765 | SPS:  2404 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0246 | PLoss: -0.0103 | Ent: 0.2002
+[ 62.0%] Iter 6050/9765 | SPS:  2404 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0263 | PLoss: -0.0057 | Ent: 0.2112
+[ 62.0%] Iter 6051/9765 | SPS:  2404 | Reward:  0.065 | Value:  0.699 | VLoss: 0.0273 | PLoss: -0.0061 | Ent: 0.2082
+[ 62.0%] Iter 6052/9765 | SPS:  2405 | Reward:  0.050 | Value:  0.664 | VLoss: 0.0220 | PLoss: -0.0112 | Ent: 0.2308
+[ 62.0%] Iter 6053/9765 | SPS:  2405 | Reward:  0.055 | Value:  0.646 | VLoss: 0.0272 | PLoss: -0.0088 | Ent: 0.1824
+[ 62.0%] Iter 6054/9765 | SPS:  2405 | Reward:  0.057 | Value:  0.745 | VLoss: 0.0296 | PLoss: -0.0046 | Ent: 0.1916
+[ 62.0%] Iter 6055/9765 | SPS:  2405 | Reward:  0.073 | Value:  0.673 | VLoss: 0.0242 | PLoss: -0.0047 | Ent: 0.1817
+[ 62.0%] Iter 6056/9765 | SPS:  2405 | Reward:  0.054 | Value:  0.746 | VLoss: 0.0236 | PLoss: -0.0068 | Ent: 0.1651
+[ 62.0%] Iter 6057/9765 | SPS:  2405 | Reward:  0.060 | Value:  0.721 | VLoss: 0.0285 | PLoss: -0.0053 | Ent: 0.2288
+[ 62.0%] Iter 6058/9765 | SPS:  2405 | Reward:  0.055 | Value:  0.681 | VLoss: 0.0126 | PLoss: -0.0028 | Ent: 0.1422
+[ 62.0%] Iter 6059/9765 | SPS:  2406 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0159 | PLoss: -0.0029 | Ent: 0.2108
+[ 62.1%] Iter 6060/9765 | SPS:  2406 | Reward:  0.067 | Value:  0.741 | VLoss: 0.0213 | PLoss: -0.0053 | Ent: 0.1993
+[ 62.1%] Iter 6061/9765 | SPS:  2406 | Reward:  0.072 | Value:  0.751 | VLoss: 0.0225 | PLoss: -0.0141 | Ent: 0.1784
+[ 62.1%] Iter 6062/9765 | SPS:  2406 | Reward:  0.074 | Value:  0.751 | VLoss: 0.0207 | PLoss: -0.0096 | Ent: 0.1825
+[ 62.1%] Iter 6063/9765 | SPS:  2406 | Reward:  0.075 | Value:  0.785 | VLoss: 0.0196 | PLoss: -0.0038 | Ent: 0.1823
+[ 62.1%] Iter 6064/9765 | SPS:  2406 | Reward:  0.058 | Value:  0.712 | VLoss: 0.0178 | PLoss: -0.0130 | Ent: 0.1748
+[ 62.1%] Iter 6065/9765 | SPS:  2406 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0277 | PLoss: -0.0099 | Ent: 0.1832
+[ 62.1%] Iter 6066/9765 | SPS:  2407 | Reward:  0.071 | Value:  0.720 | VLoss: 0.0304 | PLoss: -0.0058 | Ent: 0.1585
+[ 62.1%] Iter 6067/9765 | SPS:  2407 | Reward:  0.073 | Value:  0.729 | VLoss: 0.0266 | PLoss: -0.0062 | Ent: 0.2282
+[ 62.1%] Iter 6068/9765 | SPS:  2407 | Reward:  0.059 | Value:  0.784 | VLoss: 0.0269 | PLoss: -0.0042 | Ent: 0.1807
+[ 62.2%] Iter 6069/9765 | SPS:  2407 | Reward:  0.051 | Value:  0.681 | VLoss: 0.0183 | PLoss: -0.0080 | Ent: 0.2299
+[ 62.2%] Iter 6070/9765 | SPS:  2407 | Reward:  0.054 | Value:  0.723 | VLoss: 0.0158 | PLoss: -0.0076 | Ent: 0.2614
+[ 62.2%] Iter 6071/9765 | SPS:  2407 | Reward:  0.059 | Value:  0.640 | VLoss: 0.0201 | PLoss: -0.0023 | Ent: 0.1960
+[ 62.2%] Iter 6072/9765 | SPS:  2407 | Reward:  0.062 | Value:  0.709 | VLoss: 0.0205 | PLoss: -0.0043 | Ent: 0.1581
+[ 62.2%] Iter 6073/9765 | SPS:  2408 | Reward:  0.055 | Value:  0.769 | VLoss: 0.0218 | PLoss: -0.0070 | Ent: 0.2085
+[ 62.2%] Iter 6074/9765 | SPS:  2408 | Reward:  0.055 | Value:  0.678 | VLoss: 0.0244 | PLoss: -0.0045 | Ent: 0.2128
+[ 62.2%] Iter 6075/9765 | SPS:  2408 | Reward:  0.060 | Value:  0.737 | VLoss: 0.0188 | PLoss: -0.0036 | Ent: 0.1813
+[ 62.2%] Iter 6076/9765 | SPS:  2408 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0279 | PLoss: -0.0139 | Ent: 0.2174
+[ 62.2%] Iter 6077/9765 | SPS:  2408 | Reward:  0.054 | Value:  0.649 | VLoss: 0.0270 | PLoss: -0.0103 | Ent: 0.2484
+[ 62.2%] Iter 6078/9765 | SPS:  2408 | Reward:  0.060 | Value:  0.704 | VLoss: 0.0207 | PLoss: -0.0130 | Ent: 0.1903
+[ 62.3%] Iter 6079/9765 | SPS:  2408 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0286 | PLoss: -0.0074 | Ent: 0.1752
+[ 62.3%] Iter 6080/9765 | SPS:  2408 | Reward:  0.054 | Value:  0.716 | VLoss: 0.0219 | PLoss: -0.0084 | Ent: 0.1506
+[ 62.3%] Iter 6081/9765 | SPS:  2409 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0209 | PLoss: -0.0080 | Ent: 0.2297
+[ 62.3%] Iter 6082/9765 | SPS:  2409 | Reward:  0.067 | Value:  0.641 | VLoss: 0.0251 | PLoss: -0.0114 | Ent: 0.1901
+[ 62.3%] Iter 6083/9765 | SPS:  2409 | Reward:  0.060 | Value:  0.770 | VLoss: 0.0266 | PLoss: -0.0116 | Ent: 0.1849
+[ 62.3%] Iter 6084/9765 | SPS:  2409 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0252 | PLoss: -0.0048 | Ent: 0.2279
+[ 62.3%] Iter 6085/9765 | SPS:  2409 | Reward:  0.073 | Value:  0.753 | VLoss: 0.0280 | PLoss: -0.0067 | Ent: 0.1994
+[ 62.3%] Iter 6086/9765 | SPS:  2409 | Reward:  0.056 | Value:  0.737 | VLoss: 0.0229 | PLoss: -0.0044 | Ent: 0.1722
+[ 62.3%] Iter 6087/9765 | SPS:  2409 | Reward:  0.061 | Value:  0.696 | VLoss: 0.0229 | PLoss: -0.0042 | Ent: 0.1852
+[ 62.3%] Iter 6088/9765 | SPS:  2410 | Reward:  0.055 | Value:  0.727 | VLoss: 0.0167 | PLoss: 0.0006 | Ent: 0.1879
+[ 62.4%] Iter 6089/9765 | SPS:  2410 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0267 | PLoss: -0.0122 | Ent: 0.1574
+[ 62.4%] Iter 6090/9765 | SPS:  2410 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0235 | PLoss: -0.0104 | Ent: 0.1950
+[ 62.4%] Iter 6091/9765 | SPS:  2410 | Reward:  0.073 | Value:  0.772 | VLoss: 0.0253 | PLoss: -0.0066 | Ent: 0.1900
+[ 62.4%] Iter 6092/9765 | SPS:  2410 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0194 | PLoss: -0.0081 | Ent: 0.2777
+[ 62.4%] Iter 6093/9765 | SPS:  2410 | Reward:  0.062 | Value:  0.724 | VLoss: 0.0212 | PLoss: -0.0052 | Ent: 0.2578
+[ 62.4%] Iter 6094/9765 | SPS:  2410 | Reward:  0.053 | Value:  0.709 | VLoss: 0.0235 | PLoss: -0.0001 | Ent: 0.2106
+[ 62.4%] Iter 6095/9765 | SPS:  2411 | Reward:  0.055 | Value:  0.703 | VLoss: 0.0209 | PLoss: -0.0039 | Ent: 0.2107
+[ 62.4%] Iter 6096/9765 | SPS:  2411 | Reward:  0.051 | Value:  0.720 | VLoss: 0.0175 | PLoss: 0.0006 | Ent: 0.1919
+[ 62.4%] Iter 6097/9765 | SPS:  2411 | Reward:  0.069 | Value:  0.702 | VLoss: 0.0201 | PLoss: -0.0056 | Ent: 0.2279
+[ 62.4%] Iter 6098/9765 | SPS:  2411 | Reward:  0.045 | Value:  0.720 | VLoss: 0.0160 | PLoss: -0.0052 | Ent: 0.2001
+[ 62.5%] Iter 6099/9765 | SPS:  2411 | Reward:  0.054 | Value:  0.717 | VLoss: 0.0191 | PLoss: -0.0034 | Ent: 0.1954
+[ 62.5%] Iter 6100/9765 | SPS:  2411 | Reward:  0.049 | Value:  0.616 | VLoss: 0.0220 | PLoss: -0.0055 | Ent: 0.2019
+[ 62.5%] Iter 6101/9765 | SPS:  2411 | Reward:  0.071 | Value:  0.717 | VLoss: 0.0201 | PLoss: -0.0112 | Ent: 0.2249
+[ 62.5%] Iter 6102/9765 | SPS:  2412 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0289 | PLoss: -0.0091 | Ent: 0.2350
+[ 62.5%] Iter 6103/9765 | SPS:  2412 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0307 | PLoss: -0.0106 | Ent: 0.2148
+[ 62.5%] Iter 6104/9765 | SPS:  2412 | Reward:  0.066 | Value:  0.750 | VLoss: 0.0200 | PLoss: -0.0093 | Ent: 0.2240
+[ 62.5%] Iter 6105/9765 | SPS:  2412 | Reward:  0.060 | Value:  0.718 | VLoss: 0.0321 | PLoss: -0.0114 | Ent: 0.2038
+[ 62.5%] Iter 6106/9765 | SPS:  2412 | Reward:  0.061 | Value:  0.748 | VLoss: 0.0212 | PLoss: -0.0051 | Ent: 0.1816
+[ 62.5%] Iter 6107/9765 | SPS:  2412 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0278 | PLoss: -0.0057 | Ent: 0.1822
+[ 62.5%] Iter 6108/9765 | SPS:  2412 | Reward:  0.049 | Value:  0.632 | VLoss: 0.0201 | PLoss: -0.0151 | Ent: 0.2719
+[ 62.6%] Iter 6109/9765 | SPS:  2413 | Reward:  0.054 | Value:  0.689 | VLoss: 0.0227 | PLoss: -0.0087 | Ent: 0.1891
+[ 62.6%] Iter 6110/9765 | SPS:  2413 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0223 | PLoss: -0.0069 | Ent: 0.1988
+[ 62.6%] Iter 6111/9765 | SPS:  2413 | Reward:  0.058 | Value:  0.694 | VLoss: 0.0185 | PLoss: -0.0016 | Ent: 0.2045
+[ 62.6%] Iter 6112/9765 | SPS:  2413 | Reward:  0.063 | Value:  0.744 | VLoss: 0.0261 | PLoss: -0.0160 | Ent: 0.2068
+[ 62.6%] Iter 6113/9765 | SPS:  2413 | Reward:  0.066 | Value:  0.720 | VLoss: 0.0231 | PLoss: -0.0107 | Ent: 0.2084
+[ 62.6%] Iter 6114/9765 | SPS:  2413 | Reward:  0.068 | Value:  0.709 | VLoss: 0.0321 | PLoss: -0.0117 | Ent: 0.1868
+[ 62.6%] Iter 6115/9765 | SPS:  2413 | Reward:  0.063 | Value:  0.684 | VLoss: 0.0260 | PLoss: -0.0091 | Ent: 0.2066
+[ 62.6%] Iter 6116/9765 | SPS:  2414 | Reward:  0.066 | Value:  0.740 | VLoss: 0.0222 | PLoss: -0.0093 | Ent: 0.2095
+[ 62.6%] Iter 6117/9765 | SPS:  2414 | Reward:  0.068 | Value:  0.729 | VLoss: 0.0293 | PLoss: -0.0107 | Ent: 0.1841
+[ 62.7%] Iter 6118/9765 | SPS:  2414 | Reward:  0.054 | Value:  0.782 | VLoss: 0.0235 | PLoss: -0.0066 | Ent: 0.2255
+[ 62.7%] Iter 6119/9765 | SPS:  2414 | Reward:  0.051 | Value:  0.672 | VLoss: 0.0201 | PLoss: -0.0072 | Ent: 0.2077
+[ 62.7%] Iter 6120/9765 | SPS:  2414 | Reward:  0.048 | Value:  0.747 | VLoss: 0.0164 | PLoss: -0.0068 | Ent: 0.1765
+[ 62.7%] Iter 6121/9765 | SPS:  2414 | Reward:  0.074 | Value:  0.713 | VLoss: 0.0241 | PLoss: -0.0083 | Ent: 0.2119
+[ 62.7%] Iter 6122/9765 | SPS:  2414 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0240 | PLoss: -0.0073 | Ent: 0.2476
+[ 62.7%] Iter 6123/9765 | SPS:  2415 | Reward:  0.059 | Value:  0.719 | VLoss: 0.0204 | PLoss: -0.0084 | Ent: 0.2211
+[ 62.7%] Iter 6124/9765 | SPS:  2415 | Reward:  0.063 | Value:  0.681 | VLoss: 0.0237 | PLoss: -0.0074 | Ent: 0.1902
+[ 62.7%] Iter 6125/9765 | SPS:  2415 | Reward:  0.061 | Value:  0.735 | VLoss: 0.0166 | PLoss: -0.0032 | Ent: 0.1879
+[ 62.7%] Iter 6126/9765 | SPS:  2415 | Reward:  0.059 | Value:  0.730 | VLoss: 0.0241 | PLoss: -0.0077 | Ent: 0.2026
+[ 62.7%] Iter 6127/9765 | SPS:  2415 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0275 | PLoss: -0.0062 | Ent: 0.2305
+[ 62.8%] Iter 6128/9765 | SPS:  2415 | Reward:  0.064 | Value:  0.659 | VLoss: 0.0312 | PLoss: -0.0058 | Ent: 0.2276
+[ 62.8%] Iter 6129/9765 | SPS:  2415 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0240 | PLoss: -0.0048 | Ent: 0.2538
+[ 62.8%] Iter 6130/9765 | SPS:  2416 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0227 | PLoss: -0.0046 | Ent: 0.1975
+[ 62.8%] Iter 6131/9765 | SPS:  2416 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0339 | PLoss: -0.0098 | Ent: 0.2021
+[ 62.8%] Iter 6132/9765 | SPS:  2416 | Reward:  0.069 | Value:  0.689 | VLoss: 0.0279 | PLoss: -0.0073 | Ent: 0.2067
+[ 62.8%] Iter 6133/9765 | SPS:  2416 | Reward:  0.055 | Value:  0.721 | VLoss: 0.0231 | PLoss: -0.0119 | Ent: 0.2324
+[ 62.8%] Iter 6134/9765 | SPS:  2416 | Reward:  0.063 | Value:  0.681 | VLoss: 0.0290 | PLoss: -0.0039 | Ent: 0.2107
+[ 62.8%] Iter 6135/9765 | SPS:  2416 | Reward:  0.063 | Value:  0.627 | VLoss: 0.0252 | PLoss: -0.0043 | Ent: 0.1997
+[ 62.8%] Iter 6136/9765 | SPS:  2416 | Reward:  0.048 | Value:  0.712 | VLoss: 0.0194 | PLoss: -0.0080 | Ent: 0.2051
+[ 62.8%] Iter 6137/9765 | SPS:  2417 | Reward:  0.056 | Value:  0.642 | VLoss: 0.0195 | PLoss: -0.0077 | Ent: 0.1834
+[ 62.9%] Iter 6138/9765 | SPS:  2417 | Reward:  0.052 | Value:  0.717 | VLoss: 0.0255 | PLoss: -0.0107 | Ent: 0.1771
+[ 62.9%] Iter 6139/9765 | SPS:  2417 | Reward:  0.088 | Value:  0.695 | VLoss: 0.0327 | PLoss: -0.0075 | Ent: 0.1893
+[ 62.9%] Iter 6140/9765 | SPS:  2417 | Reward:  0.055 | Value:  0.708 | VLoss: 0.0240 | PLoss: -0.0075 | Ent: 0.2106
+[ 62.9%] Iter 6141/9765 | SPS:  2417 | Reward:  0.060 | Value:  0.730 | VLoss: 0.0208 | PLoss: -0.0095 | Ent: 0.1844
+[ 62.9%] Iter 6142/9765 | SPS:  2417 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0262 | PLoss: -0.0039 | Ent: 0.1972
+[ 62.9%] Iter 6143/9765 | SPS:  2417 | Reward:  0.055 | Value:  0.683 | VLoss: 0.0310 | PLoss: -0.0074 | Ent: 0.1873
+[ 62.9%] Iter 6144/9765 | SPS:  2418 | Reward:  0.065 | Value:  0.641 | VLoss: 0.0239 | PLoss: -0.0101 | Ent: 0.2056
+[ 62.9%] Iter 6145/9765 | SPS:  2418 | Reward:  0.056 | Value:  0.729 | VLoss: 0.0193 | PLoss: -0.0044 | Ent: 0.1933
+[ 62.9%] Iter 6146/9765 | SPS:  2418 | Reward:  0.060 | Value:  0.711 | VLoss: 0.0218 | PLoss: -0.0103 | Ent: 0.2003
+[ 62.9%] Iter 6147/9765 | SPS:  2418 | Reward:  0.045 | Value:  0.586 | VLoss: 0.0228 | PLoss: -0.0158 | Ent: 0.2123
+[ 63.0%] Iter 6148/9765 | SPS:  2418 | Reward:  0.057 | Value:  0.609 | VLoss: 0.0316 | PLoss: -0.0057 | Ent: 0.2371
+[ 63.0%] Iter 6149/9765 | SPS:  2418 | Reward:  0.065 | Value:  0.688 | VLoss: 0.0195 | PLoss: -0.0068 | Ent: 0.2194
+[ 63.0%] Iter 6150/9765 | SPS:  2418 | Reward:  0.045 | Value:  0.720 | VLoss: 0.0255 | PLoss: -0.0101 | Ent: 0.2448
+[ 63.0%] Iter 6151/9765 | SPS:  2419 | Reward:  0.066 | Value:  0.713 | VLoss: 0.0250 | PLoss: -0.0036 | Ent: 0.1710
+[ 63.0%] Iter 6152/9765 | SPS:  2419 | Reward:  0.048 | Value:  0.717 | VLoss: 0.0254 | PLoss: -0.0093 | Ent: 0.2448
+[ 63.0%] Iter 6153/9765 | SPS:  2419 | Reward:  0.064 | Value:  0.700 | VLoss: 0.0293 | PLoss: -0.0046 | Ent: 0.1420
+[ 63.0%] Iter 6154/9765 | SPS:  2419 | Reward:  0.039 | Value:  0.719 | VLoss: 0.0182 | PLoss: -0.0126 | Ent: 0.2022
+[ 63.0%] Iter 6155/9765 | SPS:  2419 | Reward:  0.060 | Value:  0.668 | VLoss: 0.0245 | PLoss: -0.0093 | Ent: 0.1510
+[ 63.0%] Iter 6156/9765 | SPS:  2419 | Reward:  0.042 | Value:  0.715 | VLoss: 0.0227 | PLoss: -0.0063 | Ent: 0.1999
+[ 63.1%] Iter 6157/9765 | SPS:  2419 | Reward:  0.050 | Value:  0.642 | VLoss: 0.0171 | PLoss: -0.0070 | Ent: 0.2236
+[ 63.1%] Iter 6158/9765 | SPS:  2420 | Reward:  0.052 | Value:  0.658 | VLoss: 0.0231 | PLoss: -0.0049 | Ent: 0.2229
+[ 63.1%] Iter 6159/9765 | SPS:  2420 | Reward:  0.064 | Value:  0.752 | VLoss: 0.0211 | PLoss: -0.0044 | Ent: 0.2137
+[ 63.1%] Iter 6160/9765 | SPS:  2420 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0290 | PLoss: -0.0029 | Ent: 0.1969
+[ 63.1%] Iter 6161/9765 | SPS:  2420 | Reward:  0.041 | Value:  0.760 | VLoss: 0.0217 | PLoss: -0.0051 | Ent: 0.1728
+[ 63.1%] Iter 6162/9765 | SPS:  2420 | Reward:  0.057 | Value:  0.696 | VLoss: 0.0315 | PLoss: -0.0142 | Ent: 0.2442
+[ 63.1%] Iter 6163/9765 | SPS:  2420 | Reward:  0.051 | Value:  0.660 | VLoss: 0.0235 | PLoss: -0.0143 | Ent: 0.2343
+[ 63.1%] Iter 6164/9765 | SPS:  2420 | Reward:  0.060 | Value:  0.693 | VLoss: 0.0264 | PLoss: -0.0121 | Ent: 0.2126
+[ 63.1%] Iter 6165/9765 | SPS:  2421 | Reward:  0.063 | Value:  0.674 | VLoss: 0.0214 | PLoss: -0.0075 | Ent: 0.1922
+[ 63.1%] Iter 6166/9765 | SPS:  2421 | Reward:  0.057 | Value:  0.658 | VLoss: 0.0218 | PLoss: -0.0005 | Ent: 0.1673
+[ 63.2%] Iter 6167/9765 | SPS:  2421 | Reward:  0.063 | Value:  0.682 | VLoss: 0.0258 | PLoss: -0.0099 | Ent: 0.2087
+[ 63.2%] Iter 6168/9765 | SPS:  2421 | Reward:  0.050 | Value:  0.666 | VLoss: 0.0208 | PLoss: -0.0069 | Ent: 0.2027
+[ 63.2%] Iter 6169/9765 | SPS:  2421 | Reward:  0.049 | Value:  0.639 | VLoss: 0.0198 | PLoss: -0.0024 | Ent: 0.1913
+[ 63.2%] Iter 6170/9765 | SPS:  2421 | Reward:  0.052 | Value:  0.603 | VLoss: 0.0270 | PLoss: -0.0041 | Ent: 0.2465
+[ 63.2%] Iter 6171/9765 | SPS:  2421 | Reward:  0.045 | Value:  0.720 | VLoss: 0.0277 | PLoss: -0.0080 | Ent: 0.2028
+[ 63.2%] Iter 6172/9765 | SPS:  2422 | Reward:  0.050 | Value:  0.671 | VLoss: 0.0216 | PLoss: -0.0125 | Ent: 0.1678
+[ 63.2%] Iter 6173/9765 | SPS:  2422 | Reward:  0.047 | Value:  0.583 | VLoss: 0.0141 | PLoss: -0.0121 | Ent: 0.2213
+[ 63.2%] Iter 6174/9765 | SPS:  2422 | Reward:  0.057 | Value:  0.616 | VLoss: 0.0273 | PLoss: -0.0048 | Ent: 0.2490
+[ 63.2%] Iter 6175/9765 | SPS:  2422 | Reward:  0.053 | Value:  0.742 | VLoss: 0.0189 | PLoss: -0.0033 | Ent: 0.2066
+[ 63.2%] Iter 6176/9765 | SPS:  2422 | Reward:  0.074 | Value:  0.720 | VLoss: 0.0279 | PLoss: -0.0067 | Ent: 0.1744
+[ 63.3%] Iter 6177/9765 | SPS:  2422 | Reward:  0.055 | Value:  0.770 | VLoss: 0.0218 | PLoss: -0.0099 | Ent: 0.2408
+[ 63.3%] Iter 6178/9765 | SPS:  2422 | Reward:  0.074 | Value:  0.675 | VLoss: 0.0222 | PLoss: -0.0097 | Ent: 0.1958
+[ 63.3%] Iter 6179/9765 | SPS:  2423 | Reward:  0.066 | Value:  0.711 | VLoss: 0.0307 | PLoss: -0.0072 | Ent: 0.1847
+[ 63.3%] Iter 6180/9765 | SPS:  2423 | Reward:  0.057 | Value:  0.669 | VLoss: 0.0149 | PLoss: -0.0034 | Ent: 0.2483
+[ 63.3%] Iter 6181/9765 | SPS:  2423 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0338 | PLoss: -0.0039 | Ent: 0.1949
+[ 63.3%] Iter 6182/9765 | SPS:  2423 | Reward:  0.062 | Value:  0.666 | VLoss: 0.0213 | PLoss: -0.0032 | Ent: 0.1886
+[ 63.3%] Iter 6183/9765 | SPS:  2423 | Reward:  0.069 | Value:  0.764 | VLoss: 0.0174 | PLoss: -0.0036 | Ent: 0.2345
+[ 63.3%] Iter 6184/9765 | SPS:  2423 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0188 | PLoss: -0.0038 | Ent: 0.2579
+[ 63.3%] Iter 6185/9765 | SPS:  2423 | Reward:  0.067 | Value:  0.786 | VLoss: 0.0214 | PLoss: -0.0087 | Ent: 0.2476
+[ 63.3%] Iter 6186/9765 | SPS:  2424 | Reward:  0.066 | Value:  0.796 | VLoss: 0.0205 | PLoss: -0.0039 | Ent: 0.2150
+[ 63.4%] Iter 6187/9765 | SPS:  2424 | Reward:  0.057 | Value:  0.628 | VLoss: 0.0201 | PLoss: -0.0102 | Ent: 0.2529
+[ 63.4%] Iter 6188/9765 | SPS:  2424 | Reward:  0.065 | Value:  0.637 | VLoss: 0.0268 | PLoss: -0.0077 | Ent: 0.2203
+[ 63.4%] Iter 6189/9765 | SPS:  2424 | Reward:  0.061 | Value:  0.765 | VLoss: 0.0301 | PLoss: -0.0034 | Ent: 0.1661
+[ 63.4%] Iter 6190/9765 | SPS:  2424 | Reward:  0.052 | Value:  0.690 | VLoss: 0.0200 | PLoss: -0.0101 | Ent: 0.1689
+[ 63.4%] Iter 6191/9765 | SPS:  2424 | Reward:  0.073 | Value:  0.706 | VLoss: 0.0246 | PLoss: -0.0051 | Ent: 0.2200
+[ 63.4%] Iter 6192/9765 | SPS:  2424 | Reward:  0.065 | Value:  0.785 | VLoss: 0.0213 | PLoss: -0.0044 | Ent: 0.2010
+[ 63.4%] Iter 6193/9765 | SPS:  2425 | Reward:  0.062 | Value:  0.773 | VLoss: 0.0258 | PLoss: -0.0058 | Ent: 0.2345
+[ 63.4%] Iter 6194/9765 | SPS:  2425 | Reward:  0.054 | Value:  0.726 | VLoss: 0.0160 | PLoss: -0.0062 | Ent: 0.1634
+[ 63.4%] Iter 6195/9765 | SPS:  2425 | Reward:  0.060 | Value:  0.701 | VLoss: 0.0172 | PLoss: -0.0072 | Ent: 0.1842
+[ 63.5%] Iter 6196/9765 | SPS:  2425 | Reward:  0.068 | Value:  0.708 | VLoss: 0.0206 | PLoss: -0.0013 | Ent: 0.1929
+[ 63.5%] Iter 6197/9765 | SPS:  2425 | Reward:  0.063 | Value:  0.771 | VLoss: 0.0219 | PLoss: -0.0039 | Ent: 0.1537
+[ 63.5%] Iter 6198/9765 | SPS:  2425 | Reward:  0.050 | Value:  0.742 | VLoss: 0.0227 | PLoss: -0.0047 | Ent: 0.2100
+[ 63.5%] Iter 6199/9765 | SPS:  2425 | Reward:  0.079 | Value:  0.667 | VLoss: 0.0319 | PLoss: -0.0063 | Ent: 0.2027
+[ 63.5%] Iter 6200/9765 | SPS:  2425 | Reward:  0.056 | Value:  0.746 | VLoss: 0.0296 | PLoss: -0.0115 | Ent: 0.2034
+[ 63.5%] Iter 6201/9765 | SPS:  2426 | Reward:  0.069 | Value:  0.648 | VLoss: 0.0203 | PLoss: -0.0047 | Ent: 0.1791
+[ 63.5%] Iter 6202/9765 | SPS:  2426 | Reward:  0.056 | Value:  0.682 | VLoss: 0.0228 | PLoss: -0.0071 | Ent: 0.2051
+[ 63.5%] Iter 6203/9765 | SPS:  2426 | Reward:  0.049 | Value:  0.659 | VLoss: 0.0319 | PLoss: -0.0011 | Ent: 0.1988
+[ 63.5%] Iter 6204/9765 | SPS:  2426 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0224 | PLoss: -0.0077 | Ent: 0.2197
+[ 63.5%] Iter 6205/9765 | SPS:  2426 | Reward:  0.063 | Value:  0.790 | VLoss: 0.0186 | PLoss: -0.0048 | Ent: 0.1853
+[ 63.6%] Iter 6206/9765 | SPS:  2426 | Reward:  0.063 | Value:  0.764 | VLoss: 0.0272 | PLoss: -0.0019 | Ent: 0.1870
+[ 63.6%] Iter 6207/9765 | SPS:  2426 | Reward:  0.062 | Value:  0.736 | VLoss: 0.0280 | PLoss: -0.0085 | Ent: 0.2182
+[ 63.6%] Iter 6208/9765 | SPS:  2427 | Reward:  0.049 | Value:  0.708 | VLoss: 0.0236 | PLoss: -0.0102 | Ent: 0.2473
+[ 63.6%] Iter 6209/9765 | SPS:  2427 | Reward:  0.070 | Value:  0.673 | VLoss: 0.0193 | PLoss: -0.0045 | Ent: 0.1836
+[ 63.6%] Iter 6210/9765 | SPS:  2427 | Reward:  0.066 | Value:  0.678 | VLoss: 0.0249 | PLoss: -0.0056 | Ent: 0.1877
+[ 63.6%] Iter 6211/9765 | SPS:  2427 | Reward:  0.058 | Value:  0.768 | VLoss: 0.0260 | PLoss: -0.0114 | Ent: 0.1908
+[ 63.6%] Iter 6212/9765 | SPS:  2427 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0250 | PLoss: -0.0026 | Ent: 0.2178
+[ 63.6%] Iter 6213/9765 | SPS:  2427 | Reward:  0.065 | Value:  0.699 | VLoss: 0.0234 | PLoss: -0.0130 | Ent: 0.2106
+[ 63.6%] Iter 6214/9765 | SPS:  2427 | Reward:  0.064 | Value:  0.774 | VLoss: 0.0264 | PLoss: -0.0068 | Ent: 0.1767
+[ 63.6%] Iter 6215/9765 | SPS:  2428 | Reward:  0.069 | Value:  0.722 | VLoss: 0.0235 | PLoss: -0.0036 | Ent: 0.2218
+[ 63.7%] Iter 6216/9765 | SPS:  2428 | Reward:  0.059 | Value:  0.796 | VLoss: 0.0149 | PLoss: 0.0002 | Ent: 0.2170
+[ 63.7%] Iter 6217/9765 | SPS:  2428 | Reward:  0.069 | Value:  0.691 | VLoss: 0.0253 | PLoss: -0.0120 | Ent: 0.2121
+[ 63.7%] Iter 6218/9765 | SPS:  2428 | Reward:  0.048 | Value:  0.706 | VLoss: 0.0205 | PLoss: -0.0069 | Ent: 0.2827
+[ 63.7%] Iter 6219/9765 | SPS:  2428 | Reward:  0.073 | Value:  0.721 | VLoss: 0.0198 | PLoss: -0.0040 | Ent: 0.1924
+[ 63.7%] Iter 6220/9765 | SPS:  2428 | Reward:  0.056 | Value:  0.726 | VLoss: 0.0189 | PLoss: -0.0079 | Ent: 0.1772
+[ 63.7%] Iter 6221/9765 | SPS:  2428 | Reward:  0.071 | Value:  0.684 | VLoss: 0.0283 | PLoss: -0.0038 | Ent: 0.1991
+[ 63.7%] Iter 6222/9765 | SPS:  2428 | Reward:  0.074 | Value:  0.667 | VLoss: 0.0203 | PLoss: -0.0044 | Ent: 0.2007
+[ 63.7%] Iter 6223/9765 | SPS:  2429 | Reward:  0.065 | Value:  0.811 | VLoss: 0.0163 | PLoss: -0.0071 | Ent: 0.1572
+[ 63.7%] Iter 6224/9765 | SPS:  2429 | Reward:  0.061 | Value:  0.756 | VLoss: 0.0248 | PLoss: -0.0133 | Ent: 0.1987
+[ 63.7%] Iter 6225/9765 | SPS:  2429 | Reward:  0.064 | Value:  0.747 | VLoss: 0.0262 | PLoss: -0.0144 | Ent: 0.1985
+[ 63.8%] Iter 6226/9765 | SPS:  2429 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0279 | PLoss: -0.0103 | Ent: 0.2575
+[ 63.8%] Iter 6227/9765 | SPS:  2429 | Reward:  0.068 | Value:  0.649 | VLoss: 0.0275 | PLoss: -0.0096 | Ent: 0.2213
+[ 63.8%] Iter 6228/9765 | SPS:  2429 | Reward:  0.051 | Value:  0.736 | VLoss: 0.0203 | PLoss: -0.0032 | Ent: 0.2027
+[ 63.8%] Iter 6229/9765 | SPS:  2429 | Reward:  0.066 | Value:  0.730 | VLoss: 0.0198 | PLoss: -0.0060 | Ent: 0.1782
+[ 63.8%] Iter 6230/9765 | SPS:  2430 | Reward:  0.049 | Value:  0.704 | VLoss: 0.0225 | PLoss: -0.0075 | Ent: 0.1888
+[ 63.8%] Iter 6231/9765 | SPS:  2430 | Reward:  0.062 | Value:  0.661 | VLoss: 0.0221 | PLoss: -0.0087 | Ent: 0.2307
+[ 63.8%] Iter 6232/9765 | SPS:  2430 | Reward:  0.058 | Value:  0.740 | VLoss: 0.0260 | PLoss: -0.0116 | Ent: 0.2244
+[ 63.8%] Iter 6233/9765 | SPS:  2430 | Reward:  0.062 | Value:  0.633 | VLoss: 0.0254 | PLoss: -0.0054 | Ent: 0.2078
+[ 63.8%] Iter 6234/9765 | SPS:  2430 | Reward:  0.049 | Value:  0.751 | VLoss: 0.0165 | PLoss: -0.0047 | Ent: 0.1940
+[ 63.9%] Iter 6235/9765 | SPS:  2430 | Reward:  0.068 | Value:  0.696 | VLoss: 0.0214 | PLoss: -0.0057 | Ent: 0.1753
+[ 63.9%] Iter 6236/9765 | SPS:  2430 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0307 | PLoss: -0.0062 | Ent: 0.1715
+[ 63.9%] Iter 6237/9765 | SPS:  2431 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0284 | PLoss: -0.0059 | Ent: 0.1779
+[ 63.9%] Iter 6238/9765 | SPS:  2431 | Reward:  0.050 | Value:  0.719 | VLoss: 0.0276 | PLoss: -0.0075 | Ent: 0.2176
+[ 63.9%] Iter 6239/9765 | SPS:  2431 | Reward:  0.062 | Value:  0.663 | VLoss: 0.0255 | PLoss: -0.0069 | Ent: 0.2231
+[ 63.9%] Iter 6240/9765 | SPS:  2431 | Reward:  0.066 | Value:  0.779 | VLoss: 0.0146 | PLoss: -0.0078 | Ent: 0.1815
+[ 63.9%] Iter 6241/9765 | SPS:  2431 | Reward:  0.053 | Value:  0.761 | VLoss: 0.0194 | PLoss: -0.0083 | Ent: 0.2082
+[ 63.9%] Iter 6242/9765 | SPS:  2431 | Reward:  0.058 | Value:  0.764 | VLoss: 0.0213 | PLoss: -0.0101 | Ent: 0.2007
+[ 63.9%] Iter 6243/9765 | SPS:  2431 | Reward:  0.057 | Value:  0.646 | VLoss: 0.0321 | PLoss: -0.0122 | Ent: 0.2387
+[ 63.9%] Iter 6244/9765 | SPS:  2431 | Reward:  0.069 | Value:  0.669 | VLoss: 0.0329 | PLoss: -0.0070 | Ent: 0.2084
+[ 64.0%] Iter 6245/9765 | SPS:  2432 | Reward:  0.053 | Value:  0.711 | VLoss: 0.0215 | PLoss: -0.0090 | Ent: 0.2526
+[ 64.0%] Iter 6246/9765 | SPS:  2432 | Reward:  0.067 | Value:  0.659 | VLoss: 0.0229 | PLoss: -0.0023 | Ent: 0.1759
+[ 64.0%] Iter 6247/9765 | SPS:  2432 | Reward:  0.063 | Value:  0.694 | VLoss: 0.0216 | PLoss: -0.0074 | Ent: 0.1754
+[ 64.0%] Iter 6248/9765 | SPS:  2432 | Reward:  0.054 | Value:  0.749 | VLoss: 0.0249 | PLoss: -0.0092 | Ent: 0.1832
+[ 64.0%] Iter 6249/9765 | SPS:  2432 | Reward:  0.067 | Value:  0.672 | VLoss: 0.0300 | PLoss: -0.0126 | Ent: 0.1825
+[ 64.0%] Iter 6250/9765 | SPS:  2432 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0284 | PLoss: -0.0122 | Ent: 0.1980
+[ 64.0%] Iter 6251/9765 | SPS:  2432 | Reward:  0.065 | Value:  0.722 | VLoss: 0.0270 | PLoss: -0.0128 | Ent: 0.2265
+[ 64.0%] Iter 6252/9765 | SPS:  2433 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0259 | PLoss: -0.0113 | Ent: 0.2042
+[ 64.0%] Iter 6253/9765 | SPS:  2433 | Reward:  0.065 | Value:  0.673 | VLoss: 0.0360 | PLoss: -0.0072 | Ent: 0.1993
+[ 64.0%] Iter 6254/9765 | SPS:  2433 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0242 | PLoss: -0.0050 | Ent: 0.1485
+[ 64.1%] Iter 6255/9765 | SPS:  2433 | Reward:  0.080 | Value:  0.752 | VLoss: 0.0278 | PLoss: 0.0037 | Ent: 0.2076
+[ 64.1%] Iter 6256/9765 | SPS:  2433 | Reward:  0.063 | Value:  0.782 | VLoss: 0.0269 | PLoss: -0.0065 | Ent: 0.2034
+[ 64.1%] Iter 6257/9765 | SPS:  2433 | Reward:  0.069 | Value:  0.705 | VLoss: 0.0263 | PLoss: -0.0105 | Ent: 0.1847
+[ 64.1%] Iter 6258/9765 | SPS:  2433 | Reward:  0.080 | Value:  0.765 | VLoss: 0.0298 | PLoss: -0.0125 | Ent: 0.1546
+[ 64.1%] Iter 6259/9765 | SPS:  2433 | Reward:  0.059 | Value:  0.761 | VLoss: 0.0213 | PLoss: -0.0065 | Ent: 0.1797
+[ 64.1%] Iter 6260/9765 | SPS:  2434 | Reward:  0.052 | Value:  0.636 | VLoss: 0.0306 | PLoss: -0.0044 | Ent: 0.2226
+[ 64.1%] Iter 6261/9765 | SPS:  2434 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0242 | PLoss: -0.0043 | Ent: 0.2054
+[ 64.1%] Iter 6262/9765 | SPS:  2434 | Reward:  0.067 | Value:  0.719 | VLoss: 0.0224 | PLoss: -0.0080 | Ent: 0.2231
+[ 64.1%] Iter 6263/9765 | SPS:  2434 | Reward:  0.067 | Value:  0.707 | VLoss: 0.0210 | PLoss: -0.0142 | Ent: 0.1591
+[ 64.1%] Iter 6264/9765 | SPS:  2434 | Reward:  0.045 | Value:  0.653 | VLoss: 0.0226 | PLoss: -0.0076 | Ent: 0.2368
+[ 64.2%] Iter 6265/9765 | SPS:  2434 | Reward:  0.065 | Value:  0.651 | VLoss: 0.0238 | PLoss: -0.0064 | Ent: 0.2050
+[ 64.2%] Iter 6266/9765 | SPS:  2434 | Reward:  0.063 | Value:  0.601 | VLoss: 0.0222 | PLoss: -0.0061 | Ent: 0.2405
+[ 64.2%] Iter 6267/9765 | SPS:  2435 | Reward:  0.061 | Value:  0.643 | VLoss: 0.0171 | PLoss: -0.0047 | Ent: 0.1970
+[ 64.2%] Iter 6268/9765 | SPS:  2435 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0264 | PLoss: -0.0036 | Ent: 0.1938
+[ 64.2%] Iter 6269/9765 | SPS:  2435 | Reward:  0.065 | Value:  0.680 | VLoss: 0.0280 | PLoss: -0.0036 | Ent: 0.1972
+[ 64.2%] Iter 6270/9765 | SPS:  2435 | Reward:  0.069 | Value:  0.732 | VLoss: 0.0253 | PLoss: -0.0006 | Ent: 0.2131
+[ 64.2%] Iter 6271/9765 | SPS:  2435 | Reward:  0.056 | Value:  0.658 | VLoss: 0.0173 | PLoss: -0.0069 | Ent: 0.2503
+[ 64.2%] Iter 6272/9765 | SPS:  2435 | Reward:  0.058 | Value:  0.688 | VLoss: 0.0337 | PLoss: -0.0061 | Ent: 0.2119
+[ 64.2%] Iter 6273/9765 | SPS:  2435 | Reward:  0.067 | Value:  0.648 | VLoss: 0.0271 | PLoss: -0.0048 | Ent: 0.2317
+[ 64.2%] Iter 6274/9765 | SPS:  2436 | Reward:  0.053 | Value:  0.726 | VLoss: 0.0240 | PLoss: -0.0045 | Ent: 0.2469
+[ 64.3%] Iter 6275/9765 | SPS:  2436 | Reward:  0.060 | Value:  0.712 | VLoss: 0.0224 | PLoss: -0.0150 | Ent: 0.1960
+[ 64.3%] Iter 6276/9765 | SPS:  2436 | Reward:  0.055 | Value:  0.696 | VLoss: 0.0284 | PLoss: -0.0041 | Ent: 0.2105
+[ 64.3%] Iter 6277/9765 | SPS:  2436 | Reward:  0.070 | Value:  0.722 | VLoss: 0.0186 | PLoss: -0.0080 | Ent: 0.1771
+[ 64.3%] Iter 6278/9765 | SPS:  2436 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0244 | PLoss: -0.0054 | Ent: 0.2519
+[ 64.3%] Iter 6279/9765 | SPS:  2436 | Reward:  0.062 | Value:  0.671 | VLoss: 0.0235 | PLoss: -0.0103 | Ent: 0.2201
+[ 64.3%] Iter 6280/9765 | SPS:  2436 | Reward:  0.064 | Value:  0.636 | VLoss: 0.0286 | PLoss: -0.0034 | Ent: 0.1668
+[ 64.3%] Iter 6281/9765 | SPS:  2436 | Reward:  0.056 | Value:  0.739 | VLoss: 0.0277 | PLoss: -0.0028 | Ent: 0.2024
+[ 64.3%] Iter 6282/9765 | SPS:  2437 | Reward:  0.059 | Value:  0.607 | VLoss: 0.0243 | PLoss: -0.0074 | Ent: 0.2267
+[ 64.3%] Iter 6283/9765 | SPS:  2437 | Reward:  0.077 | Value:  0.745 | VLoss: 0.0302 | PLoss: -0.0123 | Ent: 0.2273
+[ 64.4%] Iter 6284/9765 | SPS:  2437 | Reward:  0.058 | Value:  0.702 | VLoss: 0.0299 | PLoss: -0.0089 | Ent: 0.1597
+[ 64.4%] Iter 6285/9765 | SPS:  2437 | Reward:  0.065 | Value:  0.668 | VLoss: 0.0270 | PLoss: -0.0028 | Ent: 0.1754
+[ 64.4%] Iter 6286/9765 | SPS:  2437 | Reward:  0.077 | Value:  0.721 | VLoss: 0.0256 | PLoss: -0.0067 | Ent: 0.2182
+[ 64.4%] Iter 6287/9765 | SPS:  2437 | Reward:  0.072 | Value:  0.793 | VLoss: 0.0273 | PLoss: -0.0058 | Ent: 0.1935
+[ 64.4%] Iter 6288/9765 | SPS:  2437 | Reward:  0.061 | Value:  0.705 | VLoss: 0.0223 | PLoss: -0.0013 | Ent: 0.1698
+[ 64.4%] Iter 6289/9765 | SPS:  2438 | Reward:  0.058 | Value:  0.750 | VLoss: 0.0182 | PLoss: -0.0066 | Ent: 0.1622
+[ 64.4%] Iter 6290/9765 | SPS:  2438 | Reward:  0.057 | Value:  0.740 | VLoss: 0.0216 | PLoss: -0.0111 | Ent: 0.1842
+[ 64.4%] Iter 6291/9765 | SPS:  2438 | Reward:  0.065 | Value:  0.778 | VLoss: 0.0189 | PLoss: -0.0080 | Ent: 0.1908
+[ 64.4%] Iter 6292/9765 | SPS:  2438 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0234 | PLoss: -0.0069 | Ent: 0.2459
+[ 64.4%] Iter 6293/9765 | SPS:  2438 | Reward:  0.061 | Value:  0.704 | VLoss: 0.0231 | PLoss: -0.0039 | Ent: 0.2201
+[ 64.5%] Iter 6294/9765 | SPS:  2438 | Reward:  0.063 | Value:  0.689 | VLoss: 0.0317 | PLoss: -0.0078 | Ent: 0.2446
+[ 64.5%] Iter 6295/9765 | SPS:  2438 | Reward:  0.068 | Value:  0.726 | VLoss: 0.0176 | PLoss: -0.0043 | Ent: 0.1718
+[ 64.5%] Iter 6296/9765 | SPS:  2438 | Reward:  0.078 | Value:  0.720 | VLoss: 0.0266 | PLoss: -0.0033 | Ent: 0.1838
+[ 64.5%] Iter 6297/9765 | SPS:  2439 | Reward:  0.055 | Value:  0.777 | VLoss: 0.0184 | PLoss: -0.0095 | Ent: 0.2086
+[ 64.5%] Iter 6298/9765 | SPS:  2439 | Reward:  0.070 | Value:  0.705 | VLoss: 0.0298 | PLoss: -0.0006 | Ent: 0.1728
+[ 64.5%] Iter 6299/9765 | SPS:  2439 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0206 | PLoss: -0.0029 | Ent: 0.1902
+[ 64.5%] Iter 6300/9765 | SPS:  2439 | Reward:  0.057 | Value:  0.728 | VLoss: 0.0239 | PLoss: -0.0121 | Ent: 0.2219
+[ 64.5%] Iter 6301/9765 | SPS:  2439 | Reward:  0.053 | Value:  0.754 | VLoss: 0.0207 | PLoss: -0.0123 | Ent: 0.1650
+[ 64.5%] Iter 6302/9765 | SPS:  2439 | Reward:  0.069 | Value:  0.699 | VLoss: 0.0278 | PLoss: -0.0062 | Ent: 0.2025
+[ 64.5%] Iter 6303/9765 | SPS:  2439 | Reward:  0.069 | Value:  0.733 | VLoss: 0.0263 | PLoss: -0.0050 | Ent: 0.1863
+[ 64.6%] Iter 6304/9765 | SPS:  2440 | Reward:  0.066 | Value:  0.733 | VLoss: 0.0153 | PLoss: -0.0057 | Ent: 0.1809
+[ 64.6%] Iter 6305/9765 | SPS:  2440 | Reward:  0.069 | Value:  0.760 | VLoss: 0.0265 | PLoss: -0.0065 | Ent: 0.2149
+[ 64.6%] Iter 6306/9765 | SPS:  2440 | Reward:  0.065 | Value:  0.678 | VLoss: 0.0200 | PLoss: -0.0056 | Ent: 0.2046
+[ 64.6%] Iter 6307/9765 | SPS:  2440 | Reward:  0.071 | Value:  0.722 | VLoss: 0.0289 | PLoss: -0.0037 | Ent: 0.2162
+[ 64.6%] Iter 6308/9765 | SPS:  2440 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0265 | PLoss: -0.0129 | Ent: 0.2223
+[ 64.6%] Iter 6309/9765 | SPS:  2440 | Reward:  0.066 | Value:  0.723 | VLoss: 0.0173 | PLoss: 0.0004 | Ent: 0.2154
+[ 64.6%] Iter 6310/9765 | SPS:  2440 | Reward:  0.060 | Value:  0.678 | VLoss: 0.0257 | PLoss: -0.0038 | Ent: 0.2065
+[ 64.6%] Iter 6311/9765 | SPS:  2440 | Reward:  0.056 | Value:  0.692 | VLoss: 0.0311 | PLoss: -0.0112 | Ent: 0.2067
+[ 64.6%] Iter 6312/9765 | SPS:  2441 | Reward:  0.066 | Value:  0.669 | VLoss: 0.0232 | PLoss: -0.0042 | Ent: 0.1944
+[ 64.6%] Iter 6313/9765 | SPS:  2441 | Reward:  0.057 | Value:  0.741 | VLoss: 0.0194 | PLoss: -0.0108 | Ent: 0.2392
+[ 64.7%] Iter 6314/9765 | SPS:  2441 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0217 | PLoss: -0.0070 | Ent: 0.1847
+[ 64.7%] Iter 6315/9765 | SPS:  2441 | Reward:  0.072 | Value:  0.791 | VLoss: 0.0201 | PLoss: -0.0005 | Ent: 0.2082
+[ 64.7%] Iter 6316/9765 | SPS:  2441 | Reward:  0.074 | Value:  0.770 | VLoss: 0.0214 | PLoss: -0.0084 | Ent: 0.1875
+[ 64.7%] Iter 6317/9765 | SPS:  2441 | Reward:  0.073 | Value:  0.757 | VLoss: 0.0276 | PLoss: -0.0068 | Ent: 0.2620
+[ 64.7%] Iter 6318/9765 | SPS:  2441 | Reward:  0.061 | Value:  0.740 | VLoss: 0.0240 | PLoss: -0.0072 | Ent: 0.2043
+[ 64.7%] Iter 6319/9765 | SPS:  2442 | Reward:  0.061 | Value:  0.739 | VLoss: 0.0345 | PLoss: -0.0083 | Ent: 0.2129
+[ 64.7%] Iter 6320/9765 | SPS:  2442 | Reward:  0.077 | Value:  0.706 | VLoss: 0.0360 | PLoss: -0.0138 | Ent: 0.2430
+[ 64.7%] Iter 6321/9765 | SPS:  2442 | Reward:  0.057 | Value:  0.677 | VLoss: 0.0291 | PLoss: -0.0073 | Ent: 0.2224
+[ 64.7%] Iter 6322/9765 | SPS:  2442 | Reward:  0.064 | Value:  0.616 | VLoss: 0.0294 | PLoss: -0.0028 | Ent: 0.1968
+[ 64.8%] Iter 6323/9765 | SPS:  2442 | Reward:  0.057 | Value:  0.704 | VLoss: 0.0223 | PLoss: -0.0110 | Ent: 0.2191
+[ 64.8%] Iter 6324/9765 | SPS:  2442 | Reward:  0.072 | Value:  0.758 | VLoss: 0.0176 | PLoss: -0.0036 | Ent: 0.2205
+[ 64.8%] Iter 6325/9765 | SPS:  2442 | Reward:  0.057 | Value:  0.762 | VLoss: 0.0189 | PLoss: -0.0035 | Ent: 0.1871
+[ 64.8%] Iter 6326/9765 | SPS:  2443 | Reward:  0.060 | Value:  0.713 | VLoss: 0.0280 | PLoss: -0.0140 | Ent: 0.1936
+[ 64.8%] Iter 6327/9765 | SPS:  2443 | Reward:  0.073 | Value:  0.673 | VLoss: 0.0258 | PLoss: -0.0017 | Ent: 0.1899
+[ 64.8%] Iter 6328/9765 | SPS:  2443 | Reward:  0.056 | Value:  0.709 | VLoss: 0.0290 | PLoss: -0.0033 | Ent: 0.1686
+[ 64.8%] Iter 6329/9765 | SPS:  2443 | Reward:  0.072 | Value:  0.679 | VLoss: 0.0265 | PLoss: -0.0084 | Ent: 0.1814
+[ 64.8%] Iter 6330/9765 | SPS:  2443 | Reward:  0.059 | Value:  0.703 | VLoss: 0.0147 | PLoss: 0.0033 | Ent: 0.1610
+[ 64.8%] Iter 6331/9765 | SPS:  2443 | Reward:  0.068 | Value:  0.692 | VLoss: 0.0227 | PLoss: -0.0155 | Ent: 0.1915
+[ 64.8%] Iter 6332/9765 | SPS:  2443 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0259 | PLoss: -0.0099 | Ent: 0.1734
+[ 64.9%] Iter 6333/9765 | SPS:  2443 | Reward:  0.061 | Value:  0.696 | VLoss: 0.0287 | PLoss: -0.0046 | Ent: 0.2381
+[ 64.9%] Iter 6334/9765 | SPS:  2444 | Reward:  0.053 | Value:  0.701 | VLoss: 0.0198 | PLoss: -0.0062 | Ent: 0.2374
+[ 64.9%] Iter 6335/9765 | SPS:  2444 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0250 | PLoss: -0.0035 | Ent: 0.1913
+[ 64.9%] Iter 6336/9765 | SPS:  2444 | Reward:  0.079 | Value:  0.708 | VLoss: 0.0236 | PLoss: -0.0022 | Ent: 0.2192
+[ 64.9%] Iter 6337/9765 | SPS:  2444 | Reward:  0.073 | Value:  0.696 | VLoss: 0.0265 | PLoss: -0.0059 | Ent: 0.1984
+[ 64.9%] Iter 6338/9765 | SPS:  2444 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0271 | PLoss: -0.0146 | Ent: 0.2157
+[ 64.9%] Iter 6339/9765 | SPS:  2444 | Reward:  0.059 | Value:  0.667 | VLoss: 0.0230 | PLoss: -0.0086 | Ent: 0.2255
+[ 64.9%] Iter 6340/9765 | SPS:  2444 | Reward:  0.068 | Value:  0.678 | VLoss: 0.0221 | PLoss: -0.0080 | Ent: 0.2307
+[ 64.9%] Iter 6341/9765 | SPS:  2445 | Reward:  0.060 | Value:  0.676 | VLoss: 0.0290 | PLoss: -0.0077 | Ent: 0.1989
+[ 64.9%] Iter 6342/9765 | SPS:  2445 | Reward:  0.050 | Value:  0.710 | VLoss: 0.0126 | PLoss: -0.0061 | Ent: 0.1783
+[ 65.0%] Iter 6343/9765 | SPS:  2445 | Reward:  0.057 | Value:  0.716 | VLoss: 0.0214 | PLoss: -0.0149 | Ent: 0.2270
+[ 65.0%] Iter 6344/9765 | SPS:  2445 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0236 | PLoss: -0.0085 | Ent: 0.2323
+[ 65.0%] Iter 6345/9765 | SPS:  2445 | Reward:  0.066 | Value:  0.743 | VLoss: 0.0208 | PLoss: -0.0064 | Ent: 0.1789
+[ 65.0%] Iter 6346/9765 | SPS:  2445 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0200 | PLoss: -0.0056 | Ent: 0.2163
+[ 65.0%] Iter 6347/9765 | SPS:  2445 | Reward:  0.076 | Value:  0.669 | VLoss: 0.0242 | PLoss: -0.0079 | Ent: 0.1992
+[ 65.0%] Iter 6348/9765 | SPS:  2445 | Reward:  0.057 | Value:  0.785 | VLoss: 0.0244 | PLoss: -0.0055 | Ent: 0.1641
+[ 65.0%] Iter 6349/9765 | SPS:  2446 | Reward:  0.067 | Value:  0.719 | VLoss: 0.0209 | PLoss: -0.0078 | Ent: 0.1958
+[ 65.0%] Iter 6350/9765 | SPS:  2446 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0262 | PLoss: -0.0099 | Ent: 0.2095
+[ 65.0%] Iter 6351/9765 | SPS:  2446 | Reward:  0.061 | Value:  0.697 | VLoss: 0.0201 | PLoss: -0.0050 | Ent: 0.1850
+[ 65.0%] Iter 6352/9765 | SPS:  2446 | Reward:  0.067 | Value:  0.777 | VLoss: 0.0211 | PLoss: -0.0123 | Ent: 0.1789
+[ 65.1%] Iter 6353/9765 | SPS:  2446 | Reward:  0.070 | Value:  0.758 | VLoss: 0.0236 | PLoss: -0.0080 | Ent: 0.2632
+[ 65.1%] Iter 6354/9765 | SPS:  2446 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0243 | PLoss: -0.0078 | Ent: 0.2193
+[ 65.1%] Iter 6355/9765 | SPS:  2446 | Reward:  0.061 | Value:  0.709 | VLoss: 0.0255 | PLoss: -0.0102 | Ent: 0.1620
+[ 65.1%] Iter 6356/9765 | SPS:  2447 | Reward:  0.058 | Value:  0.658 | VLoss: 0.0200 | PLoss: -0.0004 | Ent: 0.2499
+[ 65.1%] Iter 6357/9765 | SPS:  2447 | Reward:  0.045 | Value:  0.620 | VLoss: 0.0195 | PLoss: -0.0078 | Ent: 0.2386
+[ 65.1%] Iter 6358/9765 | SPS:  2447 | Reward:  0.040 | Value:  0.614 | VLoss: 0.0214 | PLoss: -0.0029 | Ent: 0.2144
+[ 65.1%] Iter 6359/9765 | SPS:  2447 | Reward:  0.044 | Value:  0.642 | VLoss: 0.0176 | PLoss: -0.0120 | Ent: 0.2453
+[ 65.1%] Iter 6360/9765 | SPS:  2447 | Reward:  0.041 | Value:  0.508 | VLoss: 0.0147 | PLoss: -0.0125 | Ent: 0.2452
+[ 65.1%] Iter 6361/9765 | SPS:  2447 | Reward:  0.037 | Value:  0.534 | VLoss: 0.0121 | PLoss: -0.0051 | Ent: 0.3148
+[ 65.2%] Iter 6362/9765 | SPS:  2447 | Reward:  0.058 | Value:  0.625 | VLoss: 0.0217 | PLoss: -0.0149 | Ent: 0.2494
+[ 65.2%] Iter 6363/9765 | SPS:  2448 | Reward:  0.056 | Value:  0.644 | VLoss: 0.0315 | PLoss: -0.0106 | Ent: 0.2260
+[ 65.2%] Iter 6364/9765 | SPS:  2448 | Reward:  0.067 | Value:  0.692 | VLoss: 0.0345 | PLoss: -0.0075 | Ent: 0.1798
+[ 65.2%] Iter 6365/9765 | SPS:  2448 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0159 | PLoss: -0.0033 | Ent: 0.2036
+[ 65.2%] Iter 6366/9765 | SPS:  2448 | Reward:  0.065 | Value:  0.683 | VLoss: 0.0194 | PLoss: -0.0027 | Ent: 0.1902
+[ 65.2%] Iter 6367/9765 | SPS:  2448 | Reward:  0.060 | Value:  0.773 | VLoss: 0.0236 | PLoss: -0.0164 | Ent: 0.2132
+[ 65.2%] Iter 6368/9765 | SPS:  2448 | Reward:  0.073 | Value:  0.749 | VLoss: 0.0245 | PLoss: -0.0023 | Ent: 0.2061
+[ 65.2%] Iter 6369/9765 | SPS:  2448 | Reward:  0.058 | Value:  0.711 | VLoss: 0.0204 | PLoss: -0.0089 | Ent: 0.2081
+[ 65.2%] Iter 6370/9765 | SPS:  2448 | Reward:  0.062 | Value:  0.668 | VLoss: 0.0255 | PLoss: -0.0059 | Ent: 0.2449
+[ 65.2%] Iter 6371/9765 | SPS:  2449 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0170 | PLoss: -0.0035 | Ent: 0.2126
+[ 65.3%] Iter 6372/9765 | SPS:  2449 | Reward:  0.062 | Value:  0.709 | VLoss: 0.0253 | PLoss: -0.0100 | Ent: 0.1946
+[ 65.3%] Iter 6373/9765 | SPS:  2449 | Reward:  0.052 | Value:  0.698 | VLoss: 0.0185 | PLoss: -0.0000 | Ent: 0.1807
+[ 65.3%] Iter 6374/9765 | SPS:  2449 | Reward:  0.054 | Value:  0.648 | VLoss: 0.0170 | PLoss: -0.0050 | Ent: 0.2674
+[ 65.3%] Iter 6375/9765 | SPS:  2449 | Reward:  0.066 | Value:  0.688 | VLoss: 0.0210 | PLoss: 0.0004 | Ent: 0.2016
+[ 65.3%] Iter 6376/9765 | SPS:  2449 | Reward:  0.061 | Value:  0.740 | VLoss: 0.0275 | PLoss: -0.0063 | Ent: 0.1794
+[ 65.3%] Iter 6377/9765 | SPS:  2449 | Reward:  0.052 | Value:  0.736 | VLoss: 0.0231 | PLoss: -0.0080 | Ent: 0.2485
+[ 65.3%] Iter 6378/9765 | SPS:  2450 | Reward:  0.080 | Value:  0.697 | VLoss: 0.0203 | PLoss: -0.0036 | Ent: 0.2307
+[ 65.3%] Iter 6379/9765 | SPS:  2450 | Reward:  0.079 | Value:  0.796 | VLoss: 0.0253 | PLoss: -0.0017 | Ent: 0.1774
+[ 65.3%] Iter 6380/9765 | SPS:  2450 | Reward:  0.053 | Value:  0.753 | VLoss: 0.0214 | PLoss: -0.0069 | Ent: 0.2183
+[ 65.3%] Iter 6381/9765 | SPS:  2450 | Reward:  0.057 | Value:  0.697 | VLoss: 0.0245 | PLoss: -0.0112 | Ent: 0.2097
+[ 65.4%] Iter 6382/9765 | SPS:  2450 | Reward:  0.049 | Value:  0.634 | VLoss: 0.0236 | PLoss: -0.0038 | Ent: 0.1951
+[ 65.4%] Iter 6383/9765 | SPS:  2450 | Reward:  0.057 | Value:  0.711 | VLoss: 0.0268 | PLoss: -0.0040 | Ent: 0.1973
+[ 65.4%] Iter 6384/9765 | SPS:  2450 | Reward:  0.069 | Value:  0.723 | VLoss: 0.0260 | PLoss: -0.0051 | Ent: 0.2009
+[ 65.4%] Iter 6385/9765 | SPS:  2450 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0242 | PLoss: -0.0048 | Ent: 0.2007
+[ 65.4%] Iter 6386/9765 | SPS:  2451 | Reward:  0.072 | Value:  0.700 | VLoss: 0.0254 | PLoss: -0.0040 | Ent: 0.1960
+[ 65.4%] Iter 6387/9765 | SPS:  2451 | Reward:  0.062 | Value:  0.777 | VLoss: 0.0317 | PLoss: -0.0088 | Ent: 0.2281
+[ 65.4%] Iter 6388/9765 | SPS:  2451 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0185 | PLoss: -0.0104 | Ent: 0.2134
+[ 65.4%] Iter 6389/9765 | SPS:  2451 | Reward:  0.062 | Value:  0.724 | VLoss: 0.0244 | PLoss: -0.0040 | Ent: 0.2125
+[ 65.4%] Iter 6390/9765 | SPS:  2451 | Reward:  0.071 | Value:  0.700 | VLoss: 0.0174 | PLoss: -0.0066 | Ent: 0.2173
+[ 65.4%] Iter 6391/9765 | SPS:  2451 | Reward:  0.059 | Value:  0.823 | VLoss: 0.0221 | PLoss: -0.0073 | Ent: 0.1912
+[ 65.5%] Iter 6392/9765 | SPS:  2451 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0250 | PLoss: -0.0095 | Ent: 0.1995
+[ 65.5%] Iter 6393/9765 | SPS:  2452 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0238 | PLoss: -0.0066 | Ent: 0.2006
+[ 65.5%] Iter 6394/9765 | SPS:  2452 | Reward:  0.051 | Value:  0.794 | VLoss: 0.0224 | PLoss: -0.0123 | Ent: 0.2139
+[ 65.5%] Iter 6395/9765 | SPS:  2452 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0246 | PLoss: -0.0052 | Ent: 0.1598
+[ 65.5%] Iter 6396/9765 | SPS:  2452 | Reward:  0.056 | Value:  0.751 | VLoss: 0.0189 | PLoss: -0.0096 | Ent: 0.2383
+[ 65.5%] Iter 6397/9765 | SPS:  2452 | Reward:  0.065 | Value:  0.686 | VLoss: 0.0308 | PLoss: -0.0061 | Ent: 0.2145
+[ 65.5%] Iter 6398/9765 | SPS:  2452 | Reward:  0.060 | Value:  0.722 | VLoss: 0.0254 | PLoss: -0.0087 | Ent: 0.1891
+[ 65.5%] Iter 6399/9765 | SPS:  2452 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0198 | PLoss: -0.0048 | Ent: 0.1966
+[ 65.5%] Iter 6400/9765 | SPS:  2452 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0266 | PLoss: -0.0058 | Ent: 0.1966
+[ 65.6%] Iter 6401/9765 | SPS:  2453 | Reward:  0.071 | Value:  0.663 | VLoss: 0.0261 | PLoss: -0.0135 | Ent: 0.1796
+[ 65.6%] Iter 6402/9765 | SPS:  2453 | Reward:  0.069 | Value:  0.852 | VLoss: 0.0199 | PLoss: -0.0054 | Ent: 0.1572
+[ 65.6%] Iter 6403/9765 | SPS:  2453 | Reward:  0.077 | Value:  0.766 | VLoss: 0.0185 | PLoss: -0.0044 | Ent: 0.1739
+[ 65.6%] Iter 6404/9765 | SPS:  2453 | Reward:  0.091 | Value:  0.733 | VLoss: 0.0352 | PLoss: -0.0101 | Ent: 0.2211
+[ 65.6%] Iter 6405/9765 | SPS:  2453 | Reward:  0.065 | Value:  0.756 | VLoss: 0.0285 | PLoss: -0.0157 | Ent: 0.2458
+[ 65.6%] Iter 6406/9765 | SPS:  2453 | Reward:  0.075 | Value:  0.705 | VLoss: 0.0208 | PLoss: -0.0087 | Ent: 0.1802
+[ 65.6%] Iter 6407/9765 | SPS:  2453 | Reward:  0.073 | Value:  0.772 | VLoss: 0.0156 | PLoss: -0.0054 | Ent: 0.1985
+[ 65.6%] Iter 6408/9765 | SPS:  2453 | Reward:  0.058 | Value:  0.748 | VLoss: 0.0213 | PLoss: -0.0115 | Ent: 0.2034
+[ 65.6%] Iter 6409/9765 | SPS:  2454 | Reward:  0.072 | Value:  0.707 | VLoss: 0.0253 | PLoss: -0.0178 | Ent: 0.1651
+[ 65.6%] Iter 6410/9765 | SPS:  2454 | Reward:  0.069 | Value:  0.742 | VLoss: 0.0237 | PLoss: -0.0150 | Ent: 0.2097
+[ 65.7%] Iter 6411/9765 | SPS:  2454 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0225 | PLoss: -0.0012 | Ent: 0.2185
+[ 65.7%] Iter 6412/9765 | SPS:  2454 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0175 | PLoss: -0.0080 | Ent: 0.1626
+[ 65.7%] Iter 6413/9765 | SPS:  2454 | Reward:  0.061 | Value:  0.725 | VLoss: 0.0179 | PLoss: 0.0006 | Ent: 0.2285
+[ 65.7%] Iter 6414/9765 | SPS:  2454 | Reward:  0.048 | Value:  0.789 | VLoss: 0.0176 | PLoss: -0.0040 | Ent: 0.2356
+[ 65.7%] Iter 6415/9765 | SPS:  2454 | Reward:  0.060 | Value:  0.737 | VLoss: 0.0217 | PLoss: -0.0026 | Ent: 0.1876
+[ 65.7%] Iter 6416/9765 | SPS:  2455 | Reward:  0.057 | Value:  0.726 | VLoss: 0.0244 | PLoss: -0.0077 | Ent: 0.2193
+[ 65.7%] Iter 6417/9765 | SPS:  2455 | Reward:  0.061 | Value:  0.763 | VLoss: 0.0232 | PLoss: -0.0096 | Ent: 0.2279
+[ 65.7%] Iter 6418/9765 | SPS:  2455 | Reward:  0.055 | Value:  0.742 | VLoss: 0.0230 | PLoss: -0.0056 | Ent: 0.2190
+[ 65.7%] Iter 6419/9765 | SPS:  2455 | Reward:  0.080 | Value:  0.736 | VLoss: 0.0260 | PLoss: -0.0071 | Ent: 0.1890
+[ 65.7%] Iter 6420/9765 | SPS:  2455 | Reward:  0.085 | Value:  0.719 | VLoss: 0.0224 | PLoss: -0.0063 | Ent: 0.1859
+[ 65.8%] Iter 6421/9765 | SPS:  2455 | Reward:  0.065 | Value:  0.730 | VLoss: 0.0229 | PLoss: -0.0053 | Ent: 0.2442
+[ 65.8%] Iter 6422/9765 | SPS:  2455 | Reward:  0.076 | Value:  0.739 | VLoss: 0.0224 | PLoss: -0.0088 | Ent: 0.1979
+[ 65.8%] Iter 6423/9765 | SPS:  2455 | Reward:  0.073 | Value:  0.757 | VLoss: 0.0249 | PLoss: -0.0018 | Ent: 0.1725
+[ 65.8%] Iter 6424/9765 | SPS:  2456 | Reward:  0.059 | Value:  0.790 | VLoss: 0.0202 | PLoss: -0.0005 | Ent: 0.2286
+[ 65.8%] Iter 6425/9765 | SPS:  2456 | Reward:  0.056 | Value:  0.615 | VLoss: 0.0194 | PLoss: -0.0082 | Ent: 0.2630
+[ 65.8%] Iter 6426/9765 | SPS:  2456 | Reward:  0.066 | Value:  0.728 | VLoss: 0.0216 | PLoss: -0.0059 | Ent: 0.1559
+[ 65.8%] Iter 6427/9765 | SPS:  2456 | Reward:  0.071 | Value:  0.739 | VLoss: 0.0259 | PLoss: -0.0014 | Ent: 0.1764
+[ 65.8%] Iter 6428/9765 | SPS:  2456 | Reward:  0.071 | Value:  0.770 | VLoss: 0.0176 | PLoss: -0.0039 | Ent: 0.1608
+[ 65.8%] Iter 6429/9765 | SPS:  2456 | Reward:  0.055 | Value:  0.768 | VLoss: 0.0129 | PLoss: -0.0031 | Ent: 0.2060
+[ 65.8%] Iter 6430/9765 | SPS:  2456 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0307 | PLoss: -0.0111 | Ent: 0.1927
+[ 65.9%] Iter 6431/9765 | SPS:  2456 | Reward:  0.072 | Value:  0.630 | VLoss: 0.0277 | PLoss: -0.0099 | Ent: 0.2020
+[ 65.9%] Iter 6432/9765 | SPS:  2457 | Reward:  0.063 | Value:  0.734 | VLoss: 0.0159 | PLoss: -0.0066 | Ent: 0.2213
+[ 65.9%] Iter 6433/9765 | SPS:  2457 | Reward:  0.061 | Value:  0.755 | VLoss: 0.0211 | PLoss: -0.0021 | Ent: 0.1807
+[ 65.9%] Iter 6434/9765 | SPS:  2457 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0228 | PLoss: -0.0030 | Ent: 0.2314
+[ 65.9%] Iter 6435/9765 | SPS:  2457 | Reward:  0.058 | Value:  0.674 | VLoss: 0.0295 | PLoss: -0.0024 | Ent: 0.2353
+[ 65.9%] Iter 6436/9765 | SPS:  2457 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0226 | PLoss: -0.0123 | Ent: 0.1969
+[ 65.9%] Iter 6437/9765 | SPS:  2457 | Reward:  0.048 | Value:  0.677 | VLoss: 0.0213 | PLoss: -0.0094 | Ent: 0.2662
+[ 65.9%] Iter 6438/9765 | SPS:  2457 | Reward:  0.055 | Value:  0.685 | VLoss: 0.0219 | PLoss: -0.0081 | Ent: 0.2689
+[ 65.9%] Iter 6439/9765 | SPS:  2458 | Reward:  0.067 | Value:  0.688 | VLoss: 0.0216 | PLoss: -0.0038 | Ent: 0.1766
+[ 65.9%] Iter 6440/9765 | SPS:  2458 | Reward:  0.072 | Value:  0.779 | VLoss: 0.0217 | PLoss: -0.0016 | Ent: 0.1805
+[ 66.0%] Iter 6441/9765 | SPS:  2458 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0224 | PLoss: -0.0102 | Ent: 0.1527
+[ 66.0%] Iter 6442/9765 | SPS:  2458 | Reward:  0.054 | Value:  0.761 | VLoss: 0.0242 | PLoss: -0.0131 | Ent: 0.2044
+[ 66.0%] Iter 6443/9765 | SPS:  2458 | Reward:  0.061 | Value:  0.745 | VLoss: 0.0231 | PLoss: -0.0115 | Ent: 0.1835
+[ 66.0%] Iter 6444/9765 | SPS:  2458 | Reward:  0.062 | Value:  0.734 | VLoss: 0.0205 | PLoss: -0.0064 | Ent: 0.2165
+[ 66.0%] Iter 6445/9765 | SPS:  2458 | Reward:  0.056 | Value:  0.692 | VLoss: 0.0187 | PLoss: -0.0038 | Ent: 0.2398
+[ 66.0%] Iter 6446/9765 | SPS:  2458 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0229 | PLoss: -0.0061 | Ent: 0.2611
+[ 66.0%] Iter 6447/9765 | SPS:  2459 | Reward:  0.059 | Value:  0.647 | VLoss: 0.0224 | PLoss: -0.0115 | Ent: 0.1508
+[ 66.0%] Iter 6448/9765 | SPS:  2459 | Reward:  0.071 | Value:  0.806 | VLoss: 0.0203 | PLoss: -0.0082 | Ent: 0.2223
+[ 66.0%] Iter 6449/9765 | SPS:  2459 | Reward:  0.068 | Value:  0.761 | VLoss: 0.0241 | PLoss: -0.0062 | Ent: 0.1832
+[ 66.1%] Iter 6450/9765 | SPS:  2459 | Reward:  0.054 | Value:  0.765 | VLoss: 0.0190 | PLoss: -0.0071 | Ent: 0.1557
+[ 66.1%] Iter 6451/9765 | SPS:  2459 | Reward:  0.058 | Value:  0.713 | VLoss: 0.0252 | PLoss: -0.0036 | Ent: 0.1948
+[ 66.1%] Iter 6452/9765 | SPS:  2459 | Reward:  0.049 | Value:  0.590 | VLoss: 0.0179 | PLoss: -0.0003 | Ent: 0.2178
+[ 66.1%] Iter 6453/9765 | SPS:  2459 | Reward:  0.054 | Value:  0.739 | VLoss: 0.0205 | PLoss: -0.0077 | Ent: 0.2167
+[ 66.1%] Iter 6454/9765 | SPS:  2460 | Reward:  0.054 | Value:  0.684 | VLoss: 0.0156 | PLoss: -0.0047 | Ent: 0.1597
+[ 66.1%] Iter 6455/9765 | SPS:  2460 | Reward:  0.070 | Value:  0.752 | VLoss: 0.0198 | PLoss: -0.0152 | Ent: 0.1793
+[ 66.1%] Iter 6456/9765 | SPS:  2460 | Reward:  0.072 | Value:  0.819 | VLoss: 0.0172 | PLoss: -0.0076 | Ent: 0.1718
+[ 66.1%] Iter 6457/9765 | SPS:  2460 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0231 | PLoss: -0.0030 | Ent: 0.2309
+[ 66.1%] Iter 6458/9765 | SPS:  2460 | Reward:  0.061 | Value:  0.726 | VLoss: 0.0209 | PLoss: -0.0113 | Ent: 0.1916
+[ 66.1%] Iter 6459/9765 | SPS:  2460 | Reward:  0.061 | Value:  0.722 | VLoss: 0.0268 | PLoss: -0.0121 | Ent: 0.2292
+[ 66.2%] Iter 6460/9765 | SPS:  2460 | Reward:  0.054 | Value:  0.706 | VLoss: 0.0183 | PLoss: -0.0036 | Ent: 0.2267
+[ 66.2%] Iter 6461/9765 | SPS:  2460 | Reward:  0.053 | Value:  0.699 | VLoss: 0.0196 | PLoss: -0.0043 | Ent: 0.1882
+[ 66.2%] Iter 6462/9765 | SPS:  2461 | Reward:  0.066 | Value:  0.673 | VLoss: 0.0272 | PLoss: -0.0094 | Ent: 0.1816
+[ 66.2%] Iter 6463/9765 | SPS:  2461 | Reward:  0.066 | Value:  0.750 | VLoss: 0.0198 | PLoss: -0.0069 | Ent: 0.2289
+[ 66.2%] Iter 6464/9765 | SPS:  2461 | Reward:  0.060 | Value:  0.732 | VLoss: 0.0212 | PLoss: -0.0056 | Ent: 0.1977
+[ 66.2%] Iter 6465/9765 | SPS:  2461 | Reward:  0.068 | Value:  0.699 | VLoss: 0.0176 | PLoss: -0.0043 | Ent: 0.2360
+[ 66.2%] Iter 6466/9765 | SPS:  2461 | Reward:  0.063 | Value:  0.787 | VLoss: 0.0198 | PLoss: -0.0011 | Ent: 0.1797
+[ 66.2%] Iter 6467/9765 | SPS:  2461 | Reward:  0.077 | Value:  0.761 | VLoss: 0.0299 | PLoss: -0.0080 | Ent: 0.1883
+[ 66.2%] Iter 6468/9765 | SPS:  2461 | Reward:  0.081 | Value:  0.735 | VLoss: 0.0277 | PLoss: -0.0032 | Ent: 0.2043
+[ 66.2%] Iter 6469/9765 | SPS:  2461 | Reward:  0.061 | Value:  0.782 | VLoss: 0.0256 | PLoss: -0.0104 | Ent: 0.1827
+[ 66.3%] Iter 6470/9765 | SPS:  2462 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0237 | PLoss: -0.0132 | Ent: 0.2750
+[ 66.3%] Iter 6471/9765 | SPS:  2462 | Reward:  0.055 | Value:  0.712 | VLoss: 0.0210 | PLoss: -0.0120 | Ent: 0.2195
+[ 66.3%] Iter 6472/9765 | SPS:  2462 | Reward:  0.075 | Value:  0.685 | VLoss: 0.0311 | PLoss: -0.0080 | Ent: 0.2315
+[ 66.3%] Iter 6473/9765 | SPS:  2462 | Reward:  0.068 | Value:  0.764 | VLoss: 0.0213 | PLoss: -0.0163 | Ent: 0.2112
+[ 66.3%] Iter 6474/9765 | SPS:  2462 | Reward:  0.068 | Value:  0.753 | VLoss: 0.0199 | PLoss: -0.0026 | Ent: 0.2001
+[ 66.3%] Iter 6475/9765 | SPS:  2462 | Reward:  0.073 | Value:  0.763 | VLoss: 0.0168 | PLoss: -0.0013 | Ent: 0.2621
+[ 66.3%] Iter 6476/9765 | SPS:  2462 | Reward:  0.068 | Value:  0.764 | VLoss: 0.0220 | PLoss: 0.0028 | Ent: 0.1755
+[ 66.3%] Iter 6477/9765 | SPS:  2463 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0257 | PLoss: -0.0098 | Ent: 0.1865
+[ 66.3%] Iter 6478/9765 | SPS:  2463 | Reward:  0.052 | Value:  0.700 | VLoss: 0.0197 | PLoss: -0.0056 | Ent: 0.1428
+[ 66.3%] Iter 6479/9765 | SPS:  2463 | Reward:  0.070 | Value:  0.730 | VLoss: 0.0193 | PLoss: -0.0095 | Ent: 0.2164
+[ 66.4%] Iter 6480/9765 | SPS:  2463 | Reward:  0.064 | Value:  0.806 | VLoss: 0.0231 | PLoss: -0.0046 | Ent: 0.1843
+[ 66.4%] Iter 6481/9765 | SPS:  2463 | Reward:  0.058 | Value:  0.771 | VLoss: 0.0236 | PLoss: -0.0039 | Ent: 0.1953
+[ 66.4%] Iter 6482/9765 | SPS:  2463 | Reward:  0.063 | Value:  0.699 | VLoss: 0.0251 | PLoss: -0.0127 | Ent: 0.2486
+[ 66.4%] Iter 6483/9765 | SPS:  2463 | Reward:  0.056 | Value:  0.724 | VLoss: 0.0219 | PLoss: -0.0069 | Ent: 0.2291
+[ 66.4%] Iter 6484/9765 | SPS:  2463 | Reward:  0.053 | Value:  0.697 | VLoss: 0.0213 | PLoss: -0.0110 | Ent: 0.1826
+[ 66.4%] Iter 6485/9765 | SPS:  2464 | Reward:  0.060 | Value:  0.690 | VLoss: 0.0303 | PLoss: -0.0050 | Ent: 0.2058
+[ 66.4%] Iter 6486/9765 | SPS:  2464 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0262 | PLoss: -0.0064 | Ent: 0.2042
+[ 66.4%] Iter 6487/9765 | SPS:  2464 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0289 | PLoss: -0.0082 | Ent: 0.1967
+[ 66.4%] Iter 6488/9765 | SPS:  2464 | Reward:  0.062 | Value:  0.623 | VLoss: 0.0299 | PLoss: -0.0057 | Ent: 0.1943
+[ 66.5%] Iter 6489/9765 | SPS:  2464 | Reward:  0.063 | Value:  0.691 | VLoss: 0.0255 | PLoss: -0.0070 | Ent: 0.1572
+[ 66.5%] Iter 6490/9765 | SPS:  2464 | Reward:  0.064 | Value:  0.742 | VLoss: 0.0262 | PLoss: -0.0062 | Ent: 0.1845
+[ 66.5%] Iter 6491/9765 | SPS:  2464 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0252 | PLoss: -0.0096 | Ent: 0.2138
+[ 66.5%] Iter 6492/9765 | SPS:  2464 | Reward:  0.059 | Value:  0.639 | VLoss: 0.0259 | PLoss: -0.0114 | Ent: 0.2081
+[ 66.5%] Iter 6493/9765 | SPS:  2465 | Reward:  0.043 | Value:  0.627 | VLoss: 0.0198 | PLoss: -0.0081 | Ent: 0.2081
+[ 66.5%] Iter 6494/9765 | SPS:  2465 | Reward:  0.066 | Value:  0.678 | VLoss: 0.0272 | PLoss: -0.0063 | Ent: 0.2215
+[ 66.5%] Iter 6495/9765 | SPS:  2465 | Reward:  0.065 | Value:  0.726 | VLoss: 0.0231 | PLoss: -0.0053 | Ent: 0.1919
+[ 66.5%] Iter 6496/9765 | SPS:  2465 | Reward:  0.050 | Value:  0.727 | VLoss: 0.0208 | PLoss: -0.0103 | Ent: 0.1671
+[ 66.5%] Iter 6497/9765 | SPS:  2465 | Reward:  0.061 | Value:  0.702 | VLoss: 0.0147 | PLoss: -0.0121 | Ent: 0.2387
+[ 66.5%] Iter 6498/9765 | SPS:  2465 | Reward:  0.064 | Value:  0.755 | VLoss: 0.0238 | PLoss: -0.0131 | Ent: 0.2428
+[ 66.6%] Iter 6499/9765 | SPS:  2465 | Reward:  0.057 | Value:  0.692 | VLoss: 0.0179 | PLoss: -0.0067 | Ent: 0.2527
+[ 66.6%] Iter 6500/9765 | SPS:  2465 | Reward:  0.069 | Value:  0.729 | VLoss: 0.0285 | PLoss: -0.0080 | Ent: 0.2320
+[ 66.6%] Iter 6501/9765 | SPS:  2466 | Reward:  0.057 | Value:  0.664 | VLoss: 0.0314 | PLoss: -0.0115 | Ent: 0.2138
+[ 66.6%] Iter 6502/9765 | SPS:  2466 | Reward:  0.067 | Value:  0.638 | VLoss: 0.0269 | PLoss: -0.0078 | Ent: 0.1929
+[ 66.6%] Iter 6503/9765 | SPS:  2466 | Reward:  0.049 | Value:  0.750 | VLoss: 0.0227 | PLoss: -0.0057 | Ent: 0.1853
+[ 66.6%] Iter 6504/9765 | SPS:  2466 | Reward:  0.054 | Value:  0.654 | VLoss: 0.0216 | PLoss: -0.0038 | Ent: 0.1927
+[ 66.6%] Iter 6505/9765 | SPS:  2466 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0218 | PLoss: -0.0013 | Ent: 0.1753
+[ 66.6%] Iter 6506/9765 | SPS:  2466 | Reward:  0.077 | Value:  0.748 | VLoss: 0.0257 | PLoss: -0.0085 | Ent: 0.2106
+[ 66.6%] Iter 6507/9765 | SPS:  2466 | Reward:  0.067 | Value:  0.771 | VLoss: 0.0199 | PLoss: -0.0037 | Ent: 0.2186
+[ 66.6%] Iter 6508/9765 | SPS:  2466 | Reward:  0.069 | Value:  0.789 | VLoss: 0.0316 | PLoss: -0.0145 | Ent: 0.2727
+[ 66.7%] Iter 6509/9765 | SPS:  2467 | Reward:  0.070 | Value:  0.722 | VLoss: 0.0213 | PLoss: -0.0039 | Ent: 0.1977
+[ 66.7%] Iter 6510/9765 | SPS:  2467 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0231 | PLoss: -0.0062 | Ent: 0.1892
+[ 66.7%] Iter 6511/9765 | SPS:  2467 | Reward:  0.067 | Value:  0.740 | VLoss: 0.0312 | PLoss: -0.0069 | Ent: 0.1937
+[ 66.7%] Iter 6512/9765 | SPS:  2467 | Reward:  0.059 | Value:  0.722 | VLoss: 0.0207 | PLoss: -0.0122 | Ent: 0.1883
+[ 66.7%] Iter 6513/9765 | SPS:  2467 | Reward:  0.074 | Value:  0.748 | VLoss: 0.0194 | PLoss: -0.0137 | Ent: 0.1970
+[ 66.7%] Iter 6514/9765 | SPS:  2467 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0271 | PLoss: -0.0045 | Ent: 0.1990
+[ 66.7%] Iter 6515/9765 | SPS:  2467 | Reward:  0.059 | Value:  0.677 | VLoss: 0.0249 | PLoss: -0.0067 | Ent: 0.2056
+[ 66.7%] Iter 6516/9765 | SPS:  2467 | Reward:  0.055 | Value:  0.704 | VLoss: 0.0185 | PLoss: -0.0071 | Ent: 0.2095
+[ 66.7%] Iter 6517/9765 | SPS:  2468 | Reward:  0.055 | Value:  0.728 | VLoss: 0.0199 | PLoss: -0.0158 | Ent: 0.2403
+[ 66.7%] Iter 6518/9765 | SPS:  2468 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0326 | PLoss: -0.0035 | Ent: 0.1938
+[ 66.8%] Iter 6519/9765 | SPS:  2468 | Reward:  0.069 | Value:  0.726 | VLoss: 0.0191 | PLoss: -0.0085 | Ent: 0.2107
+[ 66.8%] Iter 6520/9765 | SPS:  2468 | Reward:  0.069 | Value:  0.741 | VLoss: 0.0264 | PLoss: -0.0083 | Ent: 0.2075
+[ 66.8%] Iter 6521/9765 | SPS:  2468 | Reward:  0.053 | Value:  0.746 | VLoss: 0.0227 | PLoss: -0.0103 | Ent: 0.2463
+[ 66.8%] Iter 6522/9765 | SPS:  2468 | Reward:  0.046 | Value:  0.698 | VLoss: 0.0231 | PLoss: -0.0075 | Ent: 0.2072
+[ 66.8%] Iter 6523/9765 | SPS:  2468 | Reward:  0.065 | Value:  0.664 | VLoss: 0.0232 | PLoss: -0.0086 | Ent: 0.2410
+[ 66.8%] Iter 6524/9765 | SPS:  2469 | Reward:  0.073 | Value:  0.814 | VLoss: 0.0296 | PLoss: -0.0035 | Ent: 0.2080
+[ 66.8%] Iter 6525/9765 | SPS:  2469 | Reward:  0.056 | Value:  0.757 | VLoss: 0.0173 | PLoss: -0.0008 | Ent: 0.1950
+[ 66.8%] Iter 6526/9765 | SPS:  2469 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0322 | PLoss: -0.0079 | Ent: 0.2337
+[ 66.8%] Iter 6527/9765 | SPS:  2469 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0295 | PLoss: -0.0072 | Ent: 0.2255
+[ 66.9%] Iter 6528/9765 | SPS:  2469 | Reward:  0.063 | Value:  0.761 | VLoss: 0.0207 | PLoss: -0.0105 | Ent: 0.2254
+[ 66.9%] Iter 6529/9765 | SPS:  2469 | Reward:  0.060 | Value:  0.789 | VLoss: 0.0250 | PLoss: -0.0063 | Ent: 0.1927
+[ 66.9%] Iter 6530/9765 | SPS:  2469 | Reward:  0.064 | Value:  0.704 | VLoss: 0.0243 | PLoss: -0.0089 | Ent: 0.2406
+[ 66.9%] Iter 6531/9765 | SPS:  2469 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0191 | PLoss: -0.0068 | Ent: 0.2421
+[ 66.9%] Iter 6532/9765 | SPS:  2470 | Reward:  0.061 | Value:  0.769 | VLoss: 0.0164 | PLoss: -0.0069 | Ent: 0.1899
+[ 66.9%] Iter 6533/9765 | SPS:  2470 | Reward:  0.066 | Value:  0.707 | VLoss: 0.0159 | PLoss: -0.0035 | Ent: 0.2232
+[ 66.9%] Iter 6534/9765 | SPS:  2470 | Reward:  0.068 | Value:  0.759 | VLoss: 0.0298 | PLoss: -0.0113 | Ent: 0.1945
+[ 66.9%] Iter 6535/9765 | SPS:  2470 | Reward:  0.056 | Value:  0.671 | VLoss: 0.0222 | PLoss: -0.0049 | Ent: 0.2085
+[ 66.9%] Iter 6536/9765 | SPS:  2470 | Reward:  0.052 | Value:  0.685 | VLoss: 0.0206 | PLoss: -0.0073 | Ent: 0.2209
+[ 66.9%] Iter 6537/9765 | SPS:  2470 | Reward:  0.075 | Value:  0.708 | VLoss: 0.0307 | PLoss: -0.0054 | Ent: 0.2112
+[ 67.0%] Iter 6538/9765 | SPS:  2470 | Reward:  0.073 | Value:  0.781 | VLoss: 0.0173 | PLoss: -0.0029 | Ent: 0.1676
+[ 67.0%] Iter 6539/9765 | SPS:  2470 | Reward:  0.062 | Value:  0.805 | VLoss: 0.0139 | PLoss: -0.0113 | Ent: 0.1760
+[ 67.0%] Iter 6540/9765 | SPS:  2471 | Reward:  0.065 | Value:  0.714 | VLoss: 0.0274 | PLoss: -0.0040 | Ent: 0.1715
+[ 67.0%] Iter 6541/9765 | SPS:  2471 | Reward:  0.074 | Value:  0.714 | VLoss: 0.0198 | PLoss: -0.0059 | Ent: 0.2530
+[ 67.0%] Iter 6542/9765 | SPS:  2471 | Reward:  0.062 | Value:  0.757 | VLoss: 0.0196 | PLoss: -0.0058 | Ent: 0.1653
+[ 67.0%] Iter 6543/9765 | SPS:  2471 | Reward:  0.062 | Value:  0.741 | VLoss: 0.0272 | PLoss: -0.0057 | Ent: 0.2124
+[ 67.0%] Iter 6544/9765 | SPS:  2471 | Reward:  0.057 | Value:  0.725 | VLoss: 0.0160 | PLoss: -0.0077 | Ent: 0.2084
+[ 67.0%] Iter 6545/9765 | SPS:  2471 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0244 | PLoss: -0.0184 | Ent: 0.2253
+[ 67.0%] Iter 6546/9765 | SPS:  2471 | Reward:  0.076 | Value:  0.684 | VLoss: 0.0233 | PLoss: -0.0058 | Ent: 0.2381
+[ 67.0%] Iter 6547/9765 | SPS:  2471 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0258 | PLoss: -0.0071 | Ent: 0.2084
+[ 67.1%] Iter 6548/9765 | SPS:  2472 | Reward:  0.056 | Value:  0.719 | VLoss: 0.0298 | PLoss: -0.0020 | Ent: 0.2268
+[ 67.1%] Iter 6549/9765 | SPS:  2472 | Reward:  0.072 | Value:  0.726 | VLoss: 0.0244 | PLoss: -0.0031 | Ent: 0.1611
+[ 67.1%] Iter 6550/9765 | SPS:  2472 | Reward:  0.064 | Value:  0.739 | VLoss: 0.0255 | PLoss: -0.0137 | Ent: 0.1991
+[ 67.1%] Iter 6551/9765 | SPS:  2472 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0236 | PLoss: -0.0081 | Ent: 0.2076
+[ 67.1%] Iter 6552/9765 | SPS:  2472 | Reward:  0.063 | Value:  0.675 | VLoss: 0.0209 | PLoss: -0.0079 | Ent: 0.1859
+[ 67.1%] Iter 6553/9765 | SPS:  2472 | Reward:  0.062 | Value:  0.676 | VLoss: 0.0226 | PLoss: -0.0058 | Ent: 0.2151
+[ 67.1%] Iter 6554/9765 | SPS:  2472 | Reward:  0.063 | Value:  0.630 | VLoss: 0.0266 | PLoss: -0.0081 | Ent: 0.2112
+[ 67.1%] Iter 6555/9765 | SPS:  2472 | Reward:  0.056 | Value:  0.720 | VLoss: 0.0248 | PLoss: -0.0034 | Ent: 0.2001
+[ 67.1%] Iter 6556/9765 | SPS:  2473 | Reward:  0.057 | Value:  0.734 | VLoss: 0.0169 | PLoss: -0.0150 | Ent: 0.1705
+[ 67.1%] Iter 6557/9765 | SPS:  2473 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0248 | PLoss: -0.0086 | Ent: 0.2054
+[ 67.2%] Iter 6558/9765 | SPS:  2473 | Reward:  0.068 | Value:  0.752 | VLoss: 0.0293 | PLoss: -0.0057 | Ent: 0.2083
+[ 67.2%] Iter 6559/9765 | SPS:  2473 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0228 | PLoss: -0.0058 | Ent: 0.2186
+[ 67.2%] Iter 6560/9765 | SPS:  2473 | Reward:  0.075 | Value:  0.743 | VLoss: 0.0162 | PLoss: -0.0107 | Ent: 0.1951
+[ 67.2%] Iter 6561/9765 | SPS:  2473 | Reward:  0.072 | Value:  0.782 | VLoss: 0.0345 | PLoss: -0.0111 | Ent: 0.2277
+[ 67.2%] Iter 6562/9765 | SPS:  2473 | Reward:  0.053 | Value:  0.777 | VLoss: 0.0185 | PLoss: -0.0062 | Ent: 0.1711
+[ 67.2%] Iter 6563/9765 | SPS:  2473 | Reward:  0.055 | Value:  0.670 | VLoss: 0.0308 | PLoss: -0.0133 | Ent: 0.2385
+[ 67.2%] Iter 6564/9765 | SPS:  2474 | Reward:  0.047 | Value:  0.686 | VLoss: 0.0279 | PLoss: -0.0057 | Ent: 0.2002
+[ 67.2%] Iter 6565/9765 | SPS:  2474 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0198 | PLoss: -0.0123 | Ent: 0.2520
+[ 67.2%] Iter 6566/9765 | SPS:  2474 | Reward:  0.051 | Value:  0.686 | VLoss: 0.0292 | PLoss: -0.0041 | Ent: 0.2308
+[ 67.3%] Iter 6567/9765 | SPS:  2474 | Reward:  0.053 | Value:  0.706 | VLoss: 0.0299 | PLoss: -0.0036 | Ent: 0.2640
+[ 67.3%] Iter 6568/9765 | SPS:  2474 | Reward:  0.061 | Value:  0.726 | VLoss: 0.0188 | PLoss: -0.0110 | Ent: 0.2046
+[ 67.3%] Iter 6569/9765 | SPS:  2474 | Reward:  0.064 | Value:  0.748 | VLoss: 0.0189 | PLoss: -0.0040 | Ent: 0.2453
+[ 67.3%] Iter 6570/9765 | SPS:  2474 | Reward:  0.039 | Value:  0.734 | VLoss: 0.0147 | PLoss: -0.0137 | Ent: 0.2519
+[ 67.3%] Iter 6571/9765 | SPS:  2474 | Reward:  0.074 | Value:  0.695 | VLoss: 0.0284 | PLoss: -0.0024 | Ent: 0.2028
+[ 67.3%] Iter 6572/9765 | SPS:  2475 | Reward:  0.057 | Value:  0.727 | VLoss: 0.0220 | PLoss: -0.0027 | Ent: 0.1603
+[ 67.3%] Iter 6573/9765 | SPS:  2475 | Reward:  0.061 | Value:  0.696 | VLoss: 0.0228 | PLoss: -0.0123 | Ent: 0.2045
+[ 67.3%] Iter 6574/9765 | SPS:  2475 | Reward:  0.063 | Value:  0.671 | VLoss: 0.0400 | PLoss: -0.0055 | Ent: 0.2353
+[ 67.3%] Iter 6575/9765 | SPS:  2475 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0278 | PLoss: -0.0034 | Ent: 0.2215
+[ 67.3%] Iter 6576/9765 | SPS:  2475 | Reward:  0.063 | Value:  0.794 | VLoss: 0.0220 | PLoss: -0.0115 | Ent: 0.2170
+[ 67.4%] Iter 6577/9765 | SPS:  2475 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0303 | PLoss: -0.0126 | Ent: 0.1845
+[ 67.4%] Iter 6578/9765 | SPS:  2475 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0211 | PLoss: -0.0072 | Ent: 0.2079
+[ 67.4%] Iter 6579/9765 | SPS:  2475 | Reward:  0.054 | Value:  0.717 | VLoss: 0.0136 | PLoss: -0.0070 | Ent: 0.2083
+[ 67.4%] Iter 6580/9765 | SPS:  2476 | Reward:  0.063 | Value:  0.638 | VLoss: 0.0243 | PLoss: -0.0117 | Ent: 0.1736
+[ 67.4%] Iter 6581/9765 | SPS:  2476 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0208 | PLoss: -0.0071 | Ent: 0.2294
+[ 67.4%] Iter 6582/9765 | SPS:  2476 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0227 | PLoss: -0.0024 | Ent: 0.1765
+[ 67.4%] Iter 6583/9765 | SPS:  2476 | Reward:  0.054 | Value:  0.673 | VLoss: 0.0220 | PLoss: -0.0023 | Ent: 0.2309
+[ 67.4%] Iter 6584/9765 | SPS:  2476 | Reward:  0.065 | Value:  0.613 | VLoss: 0.0226 | PLoss: -0.0019 | Ent: 0.2009
+[ 67.4%] Iter 6585/9765 | SPS:  2476 | Reward:  0.066 | Value:  0.647 | VLoss: 0.0312 | PLoss: -0.0070 | Ent: 0.2175
+[ 67.4%] Iter 6586/9765 | SPS:  2476 | Reward:  0.051 | Value:  0.711 | VLoss: 0.0213 | PLoss: -0.0097 | Ent: 0.2153
+[ 67.5%] Iter 6587/9765 | SPS:  2476 | Reward:  0.069 | Value:  0.600 | VLoss: 0.0326 | PLoss: -0.0069 | Ent: 0.2106
+[ 67.5%] Iter 6588/9765 | SPS:  2476 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0222 | PLoss: -0.0075 | Ent: 0.1920
+[ 67.5%] Iter 6589/9765 | SPS:  2477 | Reward:  0.073 | Value:  0.745 | VLoss: 0.0236 | PLoss: -0.0157 | Ent: 0.1890
+[ 67.5%] Iter 6590/9765 | SPS:  2477 | Reward:  0.061 | Value:  0.803 | VLoss: 0.0248 | PLoss: -0.0036 | Ent: 0.2091
+[ 67.5%] Iter 6591/9765 | SPS:  2477 | Reward:  0.056 | Value:  0.658 | VLoss: 0.0281 | PLoss: -0.0045 | Ent: 0.2235
+[ 67.5%] Iter 6592/9765 | SPS:  2477 | Reward:  0.066 | Value:  0.655 | VLoss: 0.0332 | PLoss: -0.0113 | Ent: 0.2266
+[ 67.5%] Iter 6593/9765 | SPS:  2477 | Reward:  0.068 | Value:  0.711 | VLoss: 0.0311 | PLoss: -0.0080 | Ent: 0.2160
+[ 67.5%] Iter 6594/9765 | SPS:  2477 | Reward:  0.053 | Value:  0.743 | VLoss: 0.0285 | PLoss: -0.0065 | Ent: 0.2042
+[ 67.5%] Iter 6595/9765 | SPS:  2477 | Reward:  0.069 | Value:  0.671 | VLoss: 0.0187 | PLoss: -0.0097 | Ent: 0.2533
+[ 67.5%] Iter 6596/9765 | SPS:  2477 | Reward:  0.059 | Value:  0.673 | VLoss: 0.0226 | PLoss: -0.0061 | Ent: 0.2211
+[ 67.6%] Iter 6597/9765 | SPS:  2478 | Reward:  0.065 | Value:  0.711 | VLoss: 0.0193 | PLoss: -0.0077 | Ent: 0.2722
+[ 67.6%] Iter 6598/9765 | SPS:  2478 | Reward:  0.056 | Value:  0.754 | VLoss: 0.0163 | PLoss: -0.0083 | Ent: 0.2107
+[ 67.6%] Iter 6599/9765 | SPS:  2478 | Reward:  0.056 | Value:  0.732 | VLoss: 0.0187 | PLoss: -0.0103 | Ent: 0.2106
+[ 67.6%] Iter 6600/9765 | SPS:  2478 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0249 | PLoss: -0.0013 | Ent: 0.2613
+[ 67.6%] Iter 6601/9765 | SPS:  2478 | Reward:  0.058 | Value:  0.823 | VLoss: 0.0230 | PLoss: -0.0150 | Ent: 0.2368
+[ 67.6%] Iter 6602/9765 | SPS:  2478 | Reward:  0.070 | Value:  0.674 | VLoss: 0.0228 | PLoss: -0.0039 | Ent: 0.2316
+[ 67.6%] Iter 6603/9765 | SPS:  2478 | Reward:  0.070 | Value:  0.649 | VLoss: 0.0364 | PLoss: -0.0067 | Ent: 0.2205
+[ 67.6%] Iter 6604/9765 | SPS:  2478 | Reward:  0.068 | Value:  0.768 | VLoss: 0.0157 | PLoss: -0.0043 | Ent: 0.2112
+[ 67.6%] Iter 6605/9765 | SPS:  2478 | Reward:  0.072 | Value:  0.790 | VLoss: 0.0297 | PLoss: -0.0064 | Ent: 0.2138
+[ 67.6%] Iter 6606/9765 | SPS:  2479 | Reward:  0.069 | Value:  0.786 | VLoss: 0.0220 | PLoss: -0.0037 | Ent: 0.2143
+[ 67.7%] Iter 6607/9765 | SPS:  2479 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0251 | PLoss: -0.0077 | Ent: 0.2047
+[ 67.7%] Iter 6608/9765 | SPS:  2479 | Reward:  0.062 | Value:  0.750 | VLoss: 0.0207 | PLoss: -0.0045 | Ent: 0.1821
+[ 67.7%] Iter 6609/9765 | SPS:  2479 | Reward:  0.068 | Value:  0.714 | VLoss: 0.0225 | PLoss: -0.0023 | Ent: 0.2223
+[ 67.7%] Iter 6610/9765 | SPS:  2479 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0196 | PLoss: -0.0162 | Ent: 0.2027
+[ 67.7%] Iter 6611/9765 | SPS:  2479 | Reward:  0.048 | Value:  0.715 | VLoss: 0.0210 | PLoss: -0.0078 | Ent: 0.2173
+[ 67.7%] Iter 6612/9765 | SPS:  2479 | Reward:  0.056 | Value:  0.674 | VLoss: 0.0306 | PLoss: -0.0037 | Ent: 0.2283
+[ 67.7%] Iter 6613/9765 | SPS:  2479 | Reward:  0.061 | Value:  0.744 | VLoss: 0.0237 | PLoss: -0.0121 | Ent: 0.2252
+[ 67.7%] Iter 6614/9765 | SPS:  2480 | Reward:  0.062 | Value:  0.658 | VLoss: 0.0206 | PLoss: -0.0041 | Ent: 0.1913
+[ 67.7%] Iter 6615/9765 | SPS:  2480 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0255 | PLoss: -0.0037 | Ent: 0.1902
+[ 67.8%] Iter 6616/9765 | SPS:  2480 | Reward:  0.072 | Value:  0.697 | VLoss: 0.0242 | PLoss: -0.0063 | Ent: 0.2068
+[ 67.8%] Iter 6617/9765 | SPS:  2480 | Reward:  0.052 | Value:  0.723 | VLoss: 0.0283 | PLoss: -0.0124 | Ent: 0.2327
+[ 67.8%] Iter 6618/9765 | SPS:  2480 | Reward:  0.050 | Value:  0.670 | VLoss: 0.0186 | PLoss: -0.0022 | Ent: 0.2158
+[ 67.8%] Iter 6619/9765 | SPS:  2480 | Reward:  0.059 | Value:  0.680 | VLoss: 0.0216 | PLoss: -0.0072 | Ent: 0.1646
+[ 67.8%] Iter 6620/9765 | SPS:  2480 | Reward:  0.062 | Value:  0.755 | VLoss: 0.0239 | PLoss: -0.0076 | Ent: 0.1749
+[ 67.8%] Iter 6621/9765 | SPS:  2480 | Reward:  0.060 | Value:  0.718 | VLoss: 0.0192 | PLoss: -0.0009 | Ent: 0.1801
+[ 67.8%] Iter 6622/9765 | SPS:  2481 | Reward:  0.058 | Value:  0.721 | VLoss: 0.0209 | PLoss: -0.0066 | Ent: 0.1358
+[ 67.8%] Iter 6623/9765 | SPS:  2481 | Reward:  0.059 | Value:  0.745 | VLoss: 0.0238 | PLoss: -0.0090 | Ent: 0.1974
+[ 67.8%] Iter 6624/9765 | SPS:  2481 | Reward:  0.063 | Value:  0.733 | VLoss: 0.0193 | PLoss: -0.0120 | Ent: 0.2413
+[ 67.8%] Iter 6625/9765 | SPS:  2481 | Reward:  0.051 | Value:  0.764 | VLoss: 0.0192 | PLoss: -0.0051 | Ent: 0.1823
+[ 67.9%] Iter 6626/9765 | SPS:  2481 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0150 | PLoss: -0.0069 | Ent: 0.1560
+[ 67.9%] Iter 6627/9765 | SPS:  2481 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0210 | PLoss: -0.0060 | Ent: 0.1902
+[ 67.9%] Iter 6628/9765 | SPS:  2481 | Reward:  0.055 | Value:  0.711 | VLoss: 0.0162 | PLoss: -0.0131 | Ent: 0.1929
+[ 67.9%] Iter 6629/9765 | SPS:  2482 | Reward:  0.044 | Value:  0.674 | VLoss: 0.0239 | PLoss: -0.0091 | Ent: 0.2120
+[ 67.9%] Iter 6630/9765 | SPS:  2482 | Reward:  0.064 | Value:  0.610 | VLoss: 0.0304 | PLoss: -0.0048 | Ent: 0.2220
+[ 67.9%] Iter 6631/9765 | SPS:  2482 | Reward:  0.047 | Value:  0.573 | VLoss: 0.0210 | PLoss: -0.0082 | Ent: 0.2438
+[ 67.9%] Iter 6632/9765 | SPS:  2482 | Reward:  0.054 | Value:  0.621 | VLoss: 0.0246 | PLoss: -0.0117 | Ent: 0.2526
+[ 67.9%] Iter 6633/9765 | SPS:  2482 | Reward:  0.068 | Value:  0.682 | VLoss: 0.0349 | PLoss: -0.0072 | Ent: 0.2538
+[ 67.9%] Iter 6634/9765 | SPS:  2482 | Reward:  0.064 | Value:  0.642 | VLoss: 0.0271 | PLoss: -0.0060 | Ent: 0.2313
+[ 67.9%] Iter 6635/9765 | SPS:  2482 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0198 | PLoss: -0.0057 | Ent: 0.2403
+[ 68.0%] Iter 6636/9765 | SPS:  2482 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0244 | PLoss: -0.0064 | Ent: 0.2610
+[ 68.0%] Iter 6637/9765 | SPS:  2482 | Reward:  0.075 | Value:  0.691 | VLoss: 0.0316 | PLoss: -0.0046 | Ent: 0.2215
+[ 68.0%] Iter 6638/9765 | SPS:  2483 | Reward:  0.057 | Value:  0.750 | VLoss: 0.0213 | PLoss: -0.0127 | Ent: 0.1856
+[ 68.0%] Iter 6639/9765 | SPS:  2483 | Reward:  0.064 | Value:  0.661 | VLoss: 0.0194 | PLoss: -0.0120 | Ent: 0.2042
+[ 68.0%] Iter 6640/9765 | SPS:  2483 | Reward:  0.058 | Value:  0.656 | VLoss: 0.0258 | PLoss: -0.0082 | Ent: 0.2210
+[ 68.0%] Iter 6641/9765 | SPS:  2483 | Reward:  0.057 | Value:  0.735 | VLoss: 0.0245 | PLoss: -0.0046 | Ent: 0.1641
+[ 68.0%] Iter 6642/9765 | SPS:  2483 | Reward:  0.059 | Value:  0.695 | VLoss: 0.0192 | PLoss: -0.0013 | Ent: 0.1700
+[ 68.0%] Iter 6643/9765 | SPS:  2483 | Reward:  0.041 | Value:  0.757 | VLoss: 0.0203 | PLoss: -0.0048 | Ent: 0.1878
+[ 68.0%] Iter 6644/9765 | SPS:  2483 | Reward:  0.055 | Value:  0.665 | VLoss: 0.0185 | PLoss: -0.0069 | Ent: 0.2199
+[ 68.0%] Iter 6645/9765 | SPS:  2484 | Reward:  0.056 | Value:  0.726 | VLoss: 0.0206 | PLoss: -0.0085 | Ent: 0.1957
+[ 68.1%] Iter 6646/9765 | SPS:  2484 | Reward:  0.073 | Value:  0.812 | VLoss: 0.0209 | PLoss: -0.0051 | Ent: 0.2222
+[ 68.1%] Iter 6647/9765 | SPS:  2484 | Reward:  0.049 | Value:  0.684 | VLoss: 0.0212 | PLoss: -0.0204 | Ent: 0.1992
+[ 68.1%] Iter 6648/9765 | SPS:  2484 | Reward:  0.077 | Value:  0.731 | VLoss: 0.0284 | PLoss: -0.0061 | Ent: 0.1942
+[ 68.1%] Iter 6649/9765 | SPS:  2484 | Reward:  0.051 | Value:  0.721 | VLoss: 0.0235 | PLoss: -0.0014 | Ent: 0.1880
+[ 68.1%] Iter 6650/9765 | SPS:  2484 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0164 | PLoss: -0.0055 | Ent: 0.1793
+[ 68.1%] Iter 6651/9765 | SPS:  2484 | Reward:  0.057 | Value:  0.722 | VLoss: 0.0262 | PLoss: -0.0108 | Ent: 0.2557
+[ 68.1%] Iter 6652/9765 | SPS:  2484 | Reward:  0.065 | Value:  0.678 | VLoss: 0.0294 | PLoss: -0.0076 | Ent: 0.1776
+[ 68.1%] Iter 6653/9765 | SPS:  2485 | Reward:  0.062 | Value:  0.768 | VLoss: 0.0212 | PLoss: -0.0073 | Ent: 0.2154
+[ 68.1%] Iter 6654/9765 | SPS:  2485 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0205 | PLoss: -0.0088 | Ent: 0.2258
+[ 68.2%] Iter 6655/9765 | SPS:  2485 | Reward:  0.057 | Value:  0.734 | VLoss: 0.0217 | PLoss: -0.0078 | Ent: 0.2120
+[ 68.2%] Iter 6656/9765 | SPS:  2485 | Reward:  0.048 | Value:  0.714 | VLoss: 0.0173 | PLoss: -0.0058 | Ent: 0.2306
+[ 68.2%] Iter 6657/9765 | SPS:  2485 | Reward:  0.062 | Value:  0.640 | VLoss: 0.0306 | PLoss: -0.0039 | Ent: 0.2378
+[ 68.2%] Iter 6658/9765 | SPS:  2485 | Reward:  0.048 | Value:  0.686 | VLoss: 0.0216 | PLoss: -0.0025 | Ent: 0.1986
+[ 68.2%] Iter 6659/9765 | SPS:  2485 | Reward:  0.063 | Value:  0.675 | VLoss: 0.0259 | PLoss: -0.0093 | Ent: 0.1936
+[ 68.2%] Iter 6660/9765 | SPS:  2485 | Reward:  0.056 | Value:  0.649 | VLoss: 0.0232 | PLoss: -0.0124 | Ent: 0.2187
+[ 68.2%] Iter 6661/9765 | SPS:  2486 | Reward:  0.069 | Value:  0.700 | VLoss: 0.0216 | PLoss: -0.0067 | Ent: 0.2757
+[ 68.2%] Iter 6662/9765 | SPS:  2486 | Reward:  0.067 | Value:  0.751 | VLoss: 0.0170 | PLoss: -0.0049 | Ent: 0.1736
+[ 68.2%] Iter 6663/9765 | SPS:  2486 | Reward:  0.058 | Value:  0.780 | VLoss: 0.0250 | PLoss: -0.0126 | Ent: 0.2422
+[ 68.2%] Iter 6664/9765 | SPS:  2486 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0238 | PLoss: -0.0040 | Ent: 0.1923
+[ 68.3%] Iter 6665/9765 | SPS:  2486 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0219 | PLoss: -0.0102 | Ent: 0.2122
+[ 68.3%] Iter 6666/9765 | SPS:  2486 | Reward:  0.076 | Value:  0.746 | VLoss: 0.0174 | PLoss: -0.0046 | Ent: 0.1954
+[ 68.3%] Iter 6667/9765 | SPS:  2486 | Reward:  0.060 | Value:  0.779 | VLoss: 0.0245 | PLoss: -0.0036 | Ent: 0.2060
+[ 68.3%] Iter 6668/9765 | SPS:  2486 | Reward:  0.066 | Value:  0.705 | VLoss: 0.0230 | PLoss: -0.0052 | Ent: 0.1924
+[ 68.3%] Iter 6669/9765 | SPS:  2487 | Reward:  0.066 | Value:  0.659 | VLoss: 0.0239 | PLoss: -0.0071 | Ent: 0.2534
+[ 68.3%] Iter 6670/9765 | SPS:  2487 | Reward:  0.066 | Value:  0.767 | VLoss: 0.0198 | PLoss: -0.0068 | Ent: 0.1925
+[ 68.3%] Iter 6671/9765 | SPS:  2487 | Reward:  0.059 | Value:  0.762 | VLoss: 0.0210 | PLoss: -0.0019 | Ent: 0.1587
+[ 68.3%] Iter 6672/9765 | SPS:  2487 | Reward:  0.051 | Value:  0.765 | VLoss: 0.0164 | PLoss: -0.0052 | Ent: 0.1640
+[ 68.3%] Iter 6673/9765 | SPS:  2487 | Reward:  0.072 | Value:  0.679 | VLoss: 0.0256 | PLoss: -0.0018 | Ent: 0.2312
+[ 68.3%] Iter 6674/9765 | SPS:  2487 | Reward:  0.067 | Value:  0.781 | VLoss: 0.0284 | PLoss: -0.0062 | Ent: 0.1698
+[ 68.4%] Iter 6675/9765 | SPS:  2487 | Reward:  0.071 | Value:  0.671 | VLoss: 0.0251 | PLoss: -0.0046 | Ent: 0.2438
+[ 68.4%] Iter 6676/9765 | SPS:  2487 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0179 | PLoss: -0.0094 | Ent: 0.2265
+[ 68.4%] Iter 6677/9765 | SPS:  2488 | Reward:  0.048 | Value:  0.719 | VLoss: 0.0187 | PLoss: -0.0080 | Ent: 0.2308
+[ 68.4%] Iter 6678/9765 | SPS:  2488 | Reward:  0.053 | Value:  0.699 | VLoss: 0.0213 | PLoss: -0.0025 | Ent: 0.2031
+[ 68.4%] Iter 6679/9765 | SPS:  2488 | Reward:  0.059 | Value:  0.747 | VLoss: 0.0190 | PLoss: -0.0060 | Ent: 0.2009
+[ 68.4%] Iter 6680/9765 | SPS:  2488 | Reward:  0.061 | Value:  0.672 | VLoss: 0.0202 | PLoss: -0.0048 | Ent: 0.2442
+[ 68.4%] Iter 6681/9765 | SPS:  2488 | Reward:  0.058 | Value:  0.713 | VLoss: 0.0261 | PLoss: -0.0074 | Ent: 0.2376
+[ 68.4%] Iter 6682/9765 | SPS:  2488 | Reward:  0.057 | Value:  0.662 | VLoss: 0.0241 | PLoss: -0.0064 | Ent: 0.2105
+[ 68.4%] Iter 6683/9765 | SPS:  2488 | Reward:  0.074 | Value:  0.749 | VLoss: 0.0209 | PLoss: -0.0079 | Ent: 0.1772
+[ 68.4%] Iter 6684/9765 | SPS:  2488 | Reward:  0.060 | Value:  0.750 | VLoss: 0.0278 | PLoss: -0.0128 | Ent: 0.2275
+[ 68.5%] Iter 6685/9765 | SPS:  2489 | Reward:  0.070 | Value:  0.746 | VLoss: 0.0240 | PLoss: -0.0073 | Ent: 0.2141
+[ 68.5%] Iter 6686/9765 | SPS:  2489 | Reward:  0.061 | Value:  0.622 | VLoss: 0.0184 | PLoss: -0.0097 | Ent: 0.2528
+[ 68.5%] Iter 6687/9765 | SPS:  2489 | Reward:  0.069 | Value:  0.721 | VLoss: 0.0310 | PLoss: -0.0092 | Ent: 0.2313
+[ 68.5%] Iter 6688/9765 | SPS:  2489 | Reward:  0.055 | Value:  0.667 | VLoss: 0.0213 | PLoss: -0.0025 | Ent: 0.2322
+[ 68.5%] Iter 6689/9765 | SPS:  2489 | Reward:  0.059 | Value:  0.675 | VLoss: 0.0159 | PLoss: -0.0088 | Ent: 0.1858
+[ 68.5%] Iter 6690/9765 | SPS:  2489 | Reward:  0.045 | Value:  0.694 | VLoss: 0.0137 | PLoss: 0.0058 | Ent: 0.2293
+[ 68.5%] Iter 6691/9765 | SPS:  2489 | Reward:  0.065 | Value:  0.618 | VLoss: 0.0307 | PLoss: -0.0100 | Ent: 0.2197
+[ 68.5%] Iter 6692/9765 | SPS:  2489 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0208 | PLoss: -0.0047 | Ent: 0.2301
+[ 68.5%] Iter 6693/9765 | SPS:  2490 | Reward:  0.062 | Value:  0.783 | VLoss: 0.0216 | PLoss: -0.0053 | Ent: 0.1956
+[ 68.6%] Iter 6694/9765 | SPS:  2490 | Reward:  0.053 | Value:  0.708 | VLoss: 0.0247 | PLoss: -0.0081 | Ent: 0.2208
+[ 68.6%] Iter 6695/9765 | SPS:  2490 | Reward:  0.063 | Value:  0.642 | VLoss: 0.0308 | PLoss: -0.0092 | Ent: 0.1912
+[ 68.6%] Iter 6696/9765 | SPS:  2490 | Reward:  0.051 | Value:  0.641 | VLoss: 0.0212 | PLoss: -0.0081 | Ent: 0.1971
+[ 68.6%] Iter 6697/9765 | SPS:  2490 | Reward:  0.054 | Value:  0.693 | VLoss: 0.0289 | PLoss: -0.0073 | Ent: 0.1815
+[ 68.6%] Iter 6698/9765 | SPS:  2490 | Reward:  0.055 | Value:  0.705 | VLoss: 0.0252 | PLoss: -0.0132 | Ent: 0.1951
+[ 68.6%] Iter 6699/9765 | SPS:  2490 | Reward:  0.056 | Value:  0.660 | VLoss: 0.0294 | PLoss: -0.0117 | Ent: 0.2070
+[ 68.6%] Iter 6700/9765 | SPS:  2490 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0298 | PLoss: -0.0056 | Ent: 0.1911
+[ 68.6%] Iter 6701/9765 | SPS:  2491 | Reward:  0.058 | Value:  0.732 | VLoss: 0.0148 | PLoss: -0.0048 | Ent: 0.2248
+[ 68.6%] Iter 6702/9765 | SPS:  2491 | Reward:  0.063 | Value:  0.763 | VLoss: 0.0220 | PLoss: -0.0066 | Ent: 0.1770
+[ 68.6%] Iter 6703/9765 | SPS:  2491 | Reward:  0.071 | Value:  0.674 | VLoss: 0.0285 | PLoss: -0.0085 | Ent: 0.2583
+[ 68.7%] Iter 6704/9765 | SPS:  2491 | Reward:  0.055 | Value:  0.705 | VLoss: 0.0226 | PLoss: -0.0042 | Ent: 0.2312
+[ 68.7%] Iter 6705/9765 | SPS:  2491 | Reward:  0.066 | Value:  0.692 | VLoss: 0.0318 | PLoss: -0.0111 | Ent: 0.2062
+[ 68.7%] Iter 6706/9765 | SPS:  2491 | Reward:  0.065 | Value:  0.770 | VLoss: 0.0200 | PLoss: -0.0120 | Ent: 0.2021
+[ 68.7%] Iter 6707/9765 | SPS:  2491 | Reward:  0.056 | Value:  0.743 | VLoss: 0.0182 | PLoss: -0.0123 | Ent: 0.1750
+[ 68.7%] Iter 6708/9765 | SPS:  2491 | Reward:  0.067 | Value:  0.686 | VLoss: 0.0227 | PLoss: -0.0056 | Ent: 0.1763
+[ 68.7%] Iter 6709/9765 | SPS:  2492 | Reward:  0.077 | Value:  0.800 | VLoss: 0.0260 | PLoss: -0.0059 | Ent: 0.2085
+[ 68.7%] Iter 6710/9765 | SPS:  2492 | Reward:  0.053 | Value:  0.723 | VLoss: 0.0229 | PLoss: -0.0064 | Ent: 0.2118
+[ 68.7%] Iter 6711/9765 | SPS:  2492 | Reward:  0.059 | Value:  0.722 | VLoss: 0.0215 | PLoss: -0.0081 | Ent: 0.1905
+[ 68.7%] Iter 6712/9765 | SPS:  2492 | Reward:  0.057 | Value:  0.731 | VLoss: 0.0161 | PLoss: -0.0011 | Ent: 0.1771
+[ 68.7%] Iter 6713/9765 | SPS:  2492 | Reward:  0.059 | Value:  0.745 | VLoss: 0.0170 | PLoss: -0.0052 | Ent: 0.1695
+[ 68.8%] Iter 6714/9765 | SPS:  2492 | Reward:  0.077 | Value:  0.656 | VLoss: 0.0344 | PLoss: -0.0115 | Ent: 0.2074
+[ 68.8%] Iter 6715/9765 | SPS:  2492 | Reward:  0.059 | Value:  0.797 | VLoss: 0.0198 | PLoss: -0.0162 | Ent: 0.2050
+[ 68.8%] Iter 6716/9765 | SPS:  2492 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0257 | PLoss: -0.0043 | Ent: 0.2283
+[ 68.8%] Iter 6717/9765 | SPS:  2493 | Reward:  0.062 | Value:  0.709 | VLoss: 0.0309 | PLoss: -0.0051 | Ent: 0.1697
+[ 68.8%] Iter 6718/9765 | SPS:  2493 | Reward:  0.069 | Value:  0.727 | VLoss: 0.0229 | PLoss: -0.0058 | Ent: 0.1728
+[ 68.8%] Iter 6719/9765 | SPS:  2493 | Reward:  0.045 | Value:  0.749 | VLoss: 0.0203 | PLoss: -0.0053 | Ent: 0.1672
+[ 68.8%] Iter 6720/9765 | SPS:  2493 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0181 | PLoss: -0.0087 | Ent: 0.1653
+[ 68.8%] Iter 6721/9765 | SPS:  2493 | Reward:  0.062 | Value:  0.656 | VLoss: 0.0290 | PLoss: -0.0069 | Ent: 0.2414
+[ 68.8%] Iter 6722/9765 | SPS:  2493 | Reward:  0.073 | Value:  0.788 | VLoss: 0.0285 | PLoss: -0.0049 | Ent: 0.2139
+[ 68.8%] Iter 6723/9765 | SPS:  2493 | Reward:  0.072 | Value:  0.688 | VLoss: 0.0270 | PLoss: -0.0071 | Ent: 0.2288
+[ 68.9%] Iter 6724/9765 | SPS:  2493 | Reward:  0.063 | Value:  0.770 | VLoss: 0.0286 | PLoss: -0.0023 | Ent: 0.2074
+[ 68.9%] Iter 6725/9765 | SPS:  2494 | Reward:  0.068 | Value:  0.670 | VLoss: 0.0258 | PLoss: -0.0061 | Ent: 0.2081
+[ 68.9%] Iter 6726/9765 | SPS:  2494 | Reward:  0.050 | Value:  0.726 | VLoss: 0.0235 | PLoss: -0.0085 | Ent: 0.2246
+[ 68.9%] Iter 6727/9765 | SPS:  2494 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0220 | PLoss: 0.0012 | Ent: 0.2062
+[ 68.9%] Iter 6728/9765 | SPS:  2494 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0255 | PLoss: -0.0082 | Ent: 0.2027
+[ 68.9%] Iter 6729/9765 | SPS:  2494 | Reward:  0.071 | Value:  0.675 | VLoss: 0.0295 | PLoss: -0.0103 | Ent: 0.1995
+[ 68.9%] Iter 6730/9765 | SPS:  2494 | Reward:  0.055 | Value:  0.721 | VLoss: 0.0238 | PLoss: -0.0080 | Ent: 0.2099
+[ 68.9%] Iter 6731/9765 | SPS:  2494 | Reward:  0.043 | Value:  0.687 | VLoss: 0.0206 | PLoss: -0.0049 | Ent: 0.2040
+[ 68.9%] Iter 6732/9765 | SPS:  2494 | Reward:  0.046 | Value:  0.683 | VLoss: 0.0166 | PLoss: -0.0071 | Ent: 0.2112
+[ 69.0%] Iter 6733/9765 | SPS:  2495 | Reward:  0.054 | Value:  0.671 | VLoss: 0.0267 | PLoss: -0.0062 | Ent: 0.1850
+[ 69.0%] Iter 6734/9765 | SPS:  2495 | Reward:  0.062 | Value:  0.637 | VLoss: 0.0253 | PLoss: -0.0065 | Ent: 0.2108
+[ 69.0%] Iter 6735/9765 | SPS:  2495 | Reward:  0.060 | Value:  0.672 | VLoss: 0.0245 | PLoss: -0.0101 | Ent: 0.1894
+[ 69.0%] Iter 6736/9765 | SPS:  2495 | Reward:  0.061 | Value:  0.698 | VLoss: 0.0199 | PLoss: -0.0048 | Ent: 0.2013
+[ 69.0%] Iter 6737/9765 | SPS:  2495 | Reward:  0.065 | Value:  0.715 | VLoss: 0.0244 | PLoss: -0.0101 | Ent: 0.2589
+[ 69.0%] Iter 6738/9765 | SPS:  2495 | Reward:  0.055 | Value:  0.750 | VLoss: 0.0220 | PLoss: -0.0075 | Ent: 0.1529
+[ 69.0%] Iter 6739/9765 | SPS:  2495 | Reward:  0.055 | Value:  0.725 | VLoss: 0.0231 | PLoss: -0.0091 | Ent: 0.2363
+[ 69.0%] Iter 6740/9765 | SPS:  2495 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0180 | PLoss: -0.0131 | Ent: 0.2054
+[ 69.0%] Iter 6741/9765 | SPS:  2496 | Reward:  0.069 | Value:  0.698 | VLoss: 0.0245 | PLoss: -0.0061 | Ent: 0.1751
+[ 69.0%] Iter 6742/9765 | SPS:  2496 | Reward:  0.061 | Value:  0.803 | VLoss: 0.0275 | PLoss: -0.0107 | Ent: 0.2371
+[ 69.1%] Iter 6743/9765 | SPS:  2496 | Reward:  0.066 | Value:  0.733 | VLoss: 0.0154 | PLoss: -0.0107 | Ent: 0.2493
+[ 69.1%] Iter 6744/9765 | SPS:  2496 | Reward:  0.045 | Value:  0.732 | VLoss: 0.0220 | PLoss: -0.0077 | Ent: 0.2311
+[ 69.1%] Iter 6745/9765 | SPS:  2496 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0187 | PLoss: -0.0011 | Ent: 0.2160
+[ 69.1%] Iter 6746/9765 | SPS:  2496 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0257 | PLoss: -0.0039 | Ent: 0.2116
+[ 69.1%] Iter 6747/9765 | SPS:  2496 | Reward:  0.058 | Value:  0.677 | VLoss: 0.0177 | PLoss: -0.0094 | Ent: 0.2274
+[ 69.1%] Iter 6748/9765 | SPS:  2496 | Reward:  0.036 | Value:  0.644 | VLoss: 0.0213 | PLoss: -0.0141 | Ent: 0.1911
+[ 69.1%] Iter 6749/9765 | SPS:  2497 | Reward:  0.058 | Value:  0.610 | VLoss: 0.0223 | PLoss: -0.0070 | Ent: 0.1864
+[ 69.1%] Iter 6750/9765 | SPS:  2497 | Reward:  0.057 | Value:  0.738 | VLoss: 0.0142 | PLoss: -0.0046 | Ent: 0.1716
+[ 69.1%] Iter 6751/9765 | SPS:  2497 | Reward:  0.060 | Value:  0.707 | VLoss: 0.0194 | PLoss: -0.0104 | Ent: 0.1994
+[ 69.1%] Iter 6752/9765 | SPS:  2497 | Reward:  0.072 | Value:  0.685 | VLoss: 0.0266 | PLoss: -0.0097 | Ent: 0.1545
+[ 69.2%] Iter 6753/9765 | SPS:  2497 | Reward:  0.059 | Value:  0.731 | VLoss: 0.0291 | PLoss: -0.0055 | Ent: 0.1547
+[ 69.2%] Iter 6754/9765 | SPS:  2497 | Reward:  0.056 | Value:  0.654 | VLoss: 0.0229 | PLoss: -0.0053 | Ent: 0.2266
+[ 69.2%] Iter 6755/9765 | SPS:  2497 | Reward:  0.055 | Value:  0.689 | VLoss: 0.0218 | PLoss: -0.0034 | Ent: 0.1903
+[ 69.2%] Iter 6756/9765 | SPS:  2497 | Reward:  0.057 | Value:  0.686 | VLoss: 0.0129 | PLoss: -0.0055 | Ent: 0.1701
+[ 69.2%] Iter 6757/9765 | SPS:  2498 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0211 | PLoss: -0.0069 | Ent: 0.2343
+[ 69.2%] Iter 6758/9765 | SPS:  2498 | Reward:  0.068 | Value:  0.726 | VLoss: 0.0331 | PLoss: -0.0022 | Ent: 0.2636
+[ 69.2%] Iter 6759/9765 | SPS:  2498 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0262 | PLoss: -0.0052 | Ent: 0.2349
+[ 69.2%] Iter 6760/9765 | SPS:  2498 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0200 | PLoss: -0.0052 | Ent: 0.2472
+[ 69.2%] Iter 6761/9765 | SPS:  2498 | Reward:  0.067 | Value:  0.752 | VLoss: 0.0241 | PLoss: -0.0060 | Ent: 0.2120
+[ 69.2%] Iter 6762/9765 | SPS:  2498 | Reward:  0.058 | Value:  0.744 | VLoss: 0.0211 | PLoss: -0.0126 | Ent: 0.2015
+[ 69.3%] Iter 6763/9765 | SPS:  2498 | Reward:  0.061 | Value:  0.771 | VLoss: 0.0191 | PLoss: -0.0089 | Ent: 0.2066
+[ 69.3%] Iter 6764/9765 | SPS:  2498 | Reward:  0.048 | Value:  0.741 | VLoss: 0.0178 | PLoss: -0.0120 | Ent: 0.2018
+[ 69.3%] Iter 6765/9765 | SPS:  2499 | Reward:  0.063 | Value:  0.634 | VLoss: 0.0246 | PLoss: -0.0079 | Ent: 0.2017
+[ 69.3%] Iter 6766/9765 | SPS:  2499 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0162 | PLoss: -0.0041 | Ent: 0.1573
+[ 69.3%] Iter 6767/9765 | SPS:  2499 | Reward:  0.049 | Value:  0.707 | VLoss: 0.0154 | PLoss: -0.0095 | Ent: 0.2128
+[ 69.3%] Iter 6768/9765 | SPS:  2499 | Reward:  0.045 | Value:  0.691 | VLoss: 0.0196 | PLoss: -0.0017 | Ent: 0.2428
+[ 69.3%] Iter 6769/9765 | SPS:  2499 | Reward:  0.061 | Value:  0.688 | VLoss: 0.0166 | PLoss: -0.0110 | Ent: 0.1615
+[ 69.3%] Iter 6770/9765 | SPS:  2499 | Reward:  0.067 | Value:  0.728 | VLoss: 0.0195 | PLoss: -0.0038 | Ent: 0.2178
+[ 69.3%] Iter 6771/9765 | SPS:  2499 | Reward:  0.063 | Value:  0.762 | VLoss: 0.0207 | PLoss: -0.0019 | Ent: 0.2083
+[ 69.3%] Iter 6772/9765 | SPS:  2499 | Reward:  0.074 | Value:  0.712 | VLoss: 0.0219 | PLoss: -0.0066 | Ent: 0.2270
+[ 69.4%] Iter 6773/9765 | SPS:  2500 | Reward:  0.056 | Value:  0.807 | VLoss: 0.0315 | PLoss: -0.0058 | Ent: 0.1590
+[ 69.4%] Iter 6774/9765 | SPS:  2500 | Reward:  0.061 | Value:  0.690 | VLoss: 0.0195 | PLoss: -0.0051 | Ent: 0.1586
+[ 69.4%] Iter 6775/9765 | SPS:  2500 | Reward:  0.067 | Value:  0.676 | VLoss: 0.0185 | PLoss: -0.0001 | Ent: 0.2093
+[ 69.4%] Iter 6776/9765 | SPS:  2500 | Reward:  0.060 | Value:  0.770 | VLoss: 0.0182 | PLoss: -0.0051 | Ent: 0.2197
+[ 69.4%] Iter 6777/9765 | SPS:  2500 | Reward:  0.054 | Value:  0.793 | VLoss: 0.0170 | PLoss: -0.0066 | Ent: 0.2246
+[ 69.4%] Iter 6778/9765 | SPS:  2500 | Reward:  0.070 | Value:  0.677 | VLoss: 0.0200 | PLoss: -0.0085 | Ent: 0.2011
+[ 69.4%] Iter 6779/9765 | SPS:  2500 | Reward:  0.065 | Value:  0.850 | VLoss: 0.0200 | PLoss: -0.0064 | Ent: 0.2182
+[ 69.4%] Iter 6780/9765 | SPS:  2500 | Reward:  0.058 | Value:  0.747 | VLoss: 0.0153 | PLoss: -0.0052 | Ent: 0.2071
+[ 69.4%] Iter 6781/9765 | SPS:  2501 | Reward:  0.067 | Value:  0.757 | VLoss: 0.0240 | PLoss: -0.0022 | Ent: 0.1458
+[ 69.5%] Iter 6782/9765 | SPS:  2501 | Reward:  0.056 | Value:  0.714 | VLoss: 0.0251 | PLoss: -0.0090 | Ent: 0.2504
+[ 69.5%] Iter 6783/9765 | SPS:  2501 | Reward:  0.064 | Value:  0.675 | VLoss: 0.0271 | PLoss: -0.0014 | Ent: 0.2019
+[ 69.5%] Iter 6784/9765 | SPS:  2501 | Reward:  0.054 | Value:  0.669 | VLoss: 0.0191 | PLoss: -0.0110 | Ent: 0.2459
+[ 69.5%] Iter 6785/9765 | SPS:  2501 | Reward:  0.053 | Value:  0.657 | VLoss: 0.0200 | PLoss: -0.0022 | Ent: 0.1884
+[ 69.5%] Iter 6786/9765 | SPS:  2501 | Reward:  0.068 | Value:  0.724 | VLoss: 0.0253 | PLoss: -0.0040 | Ent: 0.1593
+[ 69.5%] Iter 6787/9765 | SPS:  2501 | Reward:  0.048 | Value:  0.649 | VLoss: 0.0289 | PLoss: -0.0032 | Ent: 0.2552
+[ 69.5%] Iter 6788/9765 | SPS:  2501 | Reward:  0.063 | Value:  0.730 | VLoss: 0.0177 | PLoss: -0.0047 | Ent: 0.1646
+[ 69.5%] Iter 6789/9765 | SPS:  2502 | Reward:  0.057 | Value:  0.733 | VLoss: 0.0208 | PLoss: -0.0026 | Ent: 0.1638
+[ 69.5%] Iter 6790/9765 | SPS:  2502 | Reward:  0.058 | Value:  0.666 | VLoss: 0.0214 | PLoss: -0.0078 | Ent: 0.2267
+[ 69.5%] Iter 6791/9765 | SPS:  2502 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0181 | PLoss: -0.0045 | Ent: 0.2068
+[ 69.6%] Iter 6792/9765 | SPS:  2502 | Reward:  0.076 | Value:  0.734 | VLoss: 0.0265 | PLoss: -0.0060 | Ent: 0.2170
+[ 69.6%] Iter 6793/9765 | SPS:  2502 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0258 | PLoss: -0.0055 | Ent: 0.1506
+[ 69.6%] Iter 6794/9765 | SPS:  2502 | Reward:  0.061 | Value:  0.725 | VLoss: 0.0182 | PLoss: -0.0099 | Ent: 0.2311
+[ 69.6%] Iter 6795/9765 | SPS:  2502 | Reward:  0.056 | Value:  0.647 | VLoss: 0.0249 | PLoss: -0.0009 | Ent: 0.2641
+[ 69.6%] Iter 6796/9765 | SPS:  2502 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0208 | PLoss: -0.0101 | Ent: 0.2177
+[ 69.6%] Iter 6797/9765 | SPS:  2503 | Reward:  0.061 | Value:  0.680 | VLoss: 0.0242 | PLoss: -0.0042 | Ent: 0.2202
+[ 69.6%] Iter 6798/9765 | SPS:  2503 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0209 | PLoss: -0.0055 | Ent: 0.2089
+[ 69.6%] Iter 6799/9765 | SPS:  2503 | Reward:  0.065 | Value:  0.783 | VLoss: 0.0204 | PLoss: -0.0058 | Ent: 0.2004
+[ 69.6%] Iter 6800/9765 | SPS:  2503 | Reward:  0.053 | Value:  0.772 | VLoss: 0.0223 | PLoss: -0.0066 | Ent: 0.2001
+[ 69.6%] Iter 6801/9765 | SPS:  2503 | Reward:  0.061 | Value:  0.734 | VLoss: 0.0262 | PLoss: -0.0025 | Ent: 0.2000
+[ 69.7%] Iter 6802/9765 | SPS:  2503 | Reward:  0.064 | Value:  0.745 | VLoss: 0.0213 | PLoss: -0.0036 | Ent: 0.1911
+[ 69.7%] Iter 6803/9765 | SPS:  2503 | Reward:  0.066 | Value:  0.793 | VLoss: 0.0218 | PLoss: -0.0053 | Ent: 0.1959
+[ 69.7%] Iter 6804/9765 | SPS:  2503 | Reward:  0.070 | Value:  0.660 | VLoss: 0.0257 | PLoss: -0.0094 | Ent: 0.2891
+[ 69.7%] Iter 6805/9765 | SPS:  2504 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0205 | PLoss: -0.0020 | Ent: 0.2451
+[ 69.7%] Iter 6806/9765 | SPS:  2504 | Reward:  0.062 | Value:  0.757 | VLoss: 0.0200 | PLoss: -0.0077 | Ent: 0.2492
+[ 69.7%] Iter 6807/9765 | SPS:  2504 | Reward:  0.062 | Value:  0.733 | VLoss: 0.0241 | PLoss: -0.0094 | Ent: 0.2444
+[ 69.7%] Iter 6808/9765 | SPS:  2504 | Reward:  0.060 | Value:  0.756 | VLoss: 0.0276 | PLoss: -0.0022 | Ent: 0.2695
+[ 69.7%] Iter 6809/9765 | SPS:  2504 | Reward:  0.063 | Value:  0.642 | VLoss: 0.0207 | PLoss: -0.0067 | Ent: 0.2754
+[ 69.7%] Iter 6810/9765 | SPS:  2504 | Reward:  0.053 | Value:  0.685 | VLoss: 0.0185 | PLoss: -0.0081 | Ent: 0.2214
+[ 69.7%] Iter 6811/9765 | SPS:  2504 | Reward:  0.056 | Value:  0.729 | VLoss: 0.0203 | PLoss: -0.0044 | Ent: 0.1808
+[ 69.8%] Iter 6812/9765 | SPS:  2504 | Reward:  0.060 | Value:  0.662 | VLoss: 0.0257 | PLoss: -0.0069 | Ent: 0.2472
+[ 69.8%] Iter 6813/9765 | SPS:  2504 | Reward:  0.060 | Value:  0.727 | VLoss: 0.0209 | PLoss: -0.0070 | Ent: 0.2170
+[ 69.8%] Iter 6814/9765 | SPS:  2505 | Reward:  0.070 | Value:  0.741 | VLoss: 0.0202 | PLoss: -0.0113 | Ent: 0.2134
+[ 69.8%] Iter 6815/9765 | SPS:  2505 | Reward:  0.059 | Value:  0.723 | VLoss: 0.0239 | PLoss: -0.0022 | Ent: 0.2447
+[ 69.8%] Iter 6816/9765 | SPS:  2505 | Reward:  0.062 | Value:  0.752 | VLoss: 0.0188 | PLoss: -0.0039 | Ent: 0.2046
+[ 69.8%] Iter 6817/9765 | SPS:  2505 | Reward:  0.074 | Value:  0.730 | VLoss: 0.0260 | PLoss: -0.0059 | Ent: 0.2137
+[ 69.8%] Iter 6818/9765 | SPS:  2505 | Reward:  0.071 | Value:  0.725 | VLoss: 0.0262 | PLoss: -0.0008 | Ent: 0.2141
+[ 69.8%] Iter 6819/9765 | SPS:  2505 | Reward:  0.061 | Value:  0.713 | VLoss: 0.0263 | PLoss: -0.0050 | Ent: 0.2100
+[ 69.8%] Iter 6820/9765 | SPS:  2505 | Reward:  0.048 | Value:  0.747 | VLoss: 0.0173 | PLoss: -0.0017 | Ent: 0.1766
+[ 69.9%] Iter 6821/9765 | SPS:  2505 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0167 | PLoss: -0.0114 | Ent: 0.1874
+[ 69.9%] Iter 6822/9765 | SPS:  2506 | Reward:  0.057 | Value:  0.736 | VLoss: 0.0253 | PLoss: -0.0039 | Ent: 0.1847
+[ 69.9%] Iter 6823/9765 | SPS:  2506 | Reward:  0.067 | Value:  0.692 | VLoss: 0.0295 | PLoss: -0.0045 | Ent: 0.2170
+[ 69.9%] Iter 6824/9765 | SPS:  2506 | Reward:  0.055 | Value:  0.762 | VLoss: 0.0187 | PLoss: -0.0059 | Ent: 0.2371
+[ 69.9%] Iter 6825/9765 | SPS:  2506 | Reward:  0.068 | Value:  0.715 | VLoss: 0.0194 | PLoss: -0.0135 | Ent: 0.1975
+[ 69.9%] Iter 6826/9765 | SPS:  2506 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0178 | PLoss: -0.0075 | Ent: 0.1944
+[ 69.9%] Iter 6827/9765 | SPS:  2506 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0257 | PLoss: -0.0027 | Ent: 0.2267
+[ 69.9%] Iter 6828/9765 | SPS:  2506 | Reward:  0.061 | Value:  0.715 | VLoss: 0.0242 | PLoss: -0.0053 | Ent: 0.2146
+[ 69.9%] Iter 6829/9765 | SPS:  2506 | Reward:  0.057 | Value:  0.697 | VLoss: 0.0201 | PLoss: -0.0088 | Ent: 0.2169
+[ 69.9%] Iter 6830/9765 | SPS:  2507 | Reward:  0.059 | Value:  0.729 | VLoss: 0.0158 | PLoss: -0.0058 | Ent: 0.2201
+[ 70.0%] Iter 6831/9765 | SPS:  2507 | Reward:  0.072 | Value:  0.672 | VLoss: 0.0287 | PLoss: -0.0033 | Ent: 0.2058
+[ 70.0%] Iter 6832/9765 | SPS:  2507 | Reward:  0.053 | Value:  0.791 | VLoss: 0.0170 | PLoss: -0.0043 | Ent: 0.1772
+[ 70.0%] Iter 6833/9765 | SPS:  2507 | Reward:  0.065 | Value:  0.753 | VLoss: 0.0197 | PLoss: -0.0141 | Ent: 0.2156
+[ 70.0%] Iter 6834/9765 | SPS:  2507 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0226 | PLoss: -0.0097 | Ent: 0.1825
+[ 70.0%] Iter 6835/9765 | SPS:  2507 | Reward:  0.057 | Value:  0.653 | VLoss: 0.0256 | PLoss: -0.0087 | Ent: 0.2453
+[ 70.0%] Iter 6836/9765 | SPS:  2507 | Reward:  0.051 | Value:  0.650 | VLoss: 0.0197 | PLoss: -0.0039 | Ent: 0.1954
+[ 70.0%] Iter 6837/9765 | SPS:  2507 | Reward:  0.078 | Value:  0.674 | VLoss: 0.0263 | PLoss: -0.0076 | Ent: 0.2059
+[ 70.0%] Iter 6838/9765 | SPS:  2507 | Reward:  0.064 | Value:  0.819 | VLoss: 0.0173 | PLoss: -0.0108 | Ent: 0.2390
+[ 70.0%] Iter 6839/9765 | SPS:  2508 | Reward:  0.069 | Value:  0.777 | VLoss: 0.0239 | PLoss: -0.0052 | Ent: 0.2100
+[ 70.0%] Iter 6840/9765 | SPS:  2508 | Reward:  0.058 | Value:  0.827 | VLoss: 0.0190 | PLoss: -0.0020 | Ent: 0.1886
+[ 70.1%] Iter 6841/9765 | SPS:  2508 | Reward:  0.064 | Value:  0.755 | VLoss: 0.0198 | PLoss: -0.0064 | Ent: 0.1925
+[ 70.1%] Iter 6842/9765 | SPS:  2508 | Reward:  0.070 | Value:  0.724 | VLoss: 0.0273 | PLoss: -0.0074 | Ent: 0.1922
+[ 70.1%] Iter 6843/9765 | SPS:  2508 | Reward:  0.062 | Value:  0.750 | VLoss: 0.0238 | PLoss: -0.0045 | Ent: 0.1997
+[ 70.1%] Iter 6844/9765 | SPS:  2508 | Reward:  0.064 | Value:  0.788 | VLoss: 0.0190 | PLoss: -0.0026 | Ent: 0.1616
+[ 70.1%] Iter 6845/9765 | SPS:  2508 | Reward:  0.073 | Value:  0.721 | VLoss: 0.0276 | PLoss: -0.0073 | Ent: 0.2293
+[ 70.1%] Iter 6846/9765 | SPS:  2508 | Reward:  0.034 | Value:  0.699 | VLoss: 0.0155 | PLoss: -0.0082 | Ent: 0.2467
+[ 70.1%] Iter 6847/9765 | SPS:  2508 | Reward:  0.062 | Value:  0.606 | VLoss: 0.0249 | PLoss: -0.0079 | Ent: 0.2424
+[ 70.1%] Iter 6848/9765 | SPS:  2509 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0190 | PLoss: -0.0066 | Ent: 0.2123
+[ 70.1%] Iter 6849/9765 | SPS:  2509 | Reward:  0.069 | Value:  0.723 | VLoss: 0.0325 | PLoss: -0.0054 | Ent: 0.2005
+[ 70.1%] Iter 6850/9765 | SPS:  2509 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0212 | PLoss: -0.0006 | Ent: 0.1986
+[ 70.2%] Iter 6851/9765 | SPS:  2509 | Reward:  0.055 | Value:  0.743 | VLoss: 0.0221 | PLoss: -0.0011 | Ent: 0.2117
+[ 70.2%] Iter 6852/9765 | SPS:  2509 | Reward:  0.061 | Value:  0.738 | VLoss: 0.0248 | PLoss: -0.0066 | Ent: 0.2518
+[ 70.2%] Iter 6853/9765 | SPS:  2509 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0254 | PLoss: -0.0092 | Ent: 0.2142
+[ 70.2%] Iter 6854/9765 | SPS:  2509 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0336 | PLoss: -0.0026 | Ent: 0.2201
+[ 70.2%] Iter 6855/9765 | SPS:  2509 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0147 | PLoss: -0.0057 | Ent: 0.1913
+[ 70.2%] Iter 6856/9765 | SPS:  2510 | Reward:  0.058 | Value:  0.706 | VLoss: 0.0250 | PLoss: -0.0030 | Ent: 0.2041
+[ 70.2%] Iter 6857/9765 | SPS:  2510 | Reward:  0.066 | Value:  0.688 | VLoss: 0.0257 | PLoss: -0.0055 | Ent: 0.2293
+[ 70.2%] Iter 6858/9765 | SPS:  2510 | Reward:  0.054 | Value:  0.769 | VLoss: 0.0260 | PLoss: -0.0065 | Ent: 0.2396
+[ 70.2%] Iter 6859/9765 | SPS:  2510 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0219 | PLoss: -0.0046 | Ent: 0.2082
+[ 70.3%] Iter 6860/9765 | SPS:  2510 | Reward:  0.052 | Value:  0.757 | VLoss: 0.0251 | PLoss: -0.0059 | Ent: 0.2093
+[ 70.3%] Iter 6861/9765 | SPS:  2510 | Reward:  0.059 | Value:  0.762 | VLoss: 0.0196 | PLoss: -0.0089 | Ent: 0.2128
+[ 70.3%] Iter 6862/9765 | SPS:  2510 | Reward:  0.066 | Value:  0.736 | VLoss: 0.0208 | PLoss: -0.0071 | Ent: 0.2344
+[ 70.3%] Iter 6863/9765 | SPS:  2510 | Reward:  0.043 | Value:  0.715 | VLoss: 0.0222 | PLoss: -0.0043 | Ent: 0.2627
+[ 70.3%] Iter 6864/9765 | SPS:  2511 | Reward:  0.064 | Value:  0.605 | VLoss: 0.0225 | PLoss: -0.0095 | Ent: 0.2166
+[ 70.3%] Iter 6865/9765 | SPS:  2511 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0295 | PLoss: -0.0066 | Ent: 0.2313
+[ 70.3%] Iter 6866/9765 | SPS:  2511 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0217 | PLoss: -0.0113 | Ent: 0.1915
+[ 70.3%] Iter 6867/9765 | SPS:  2511 | Reward:  0.074 | Value:  0.750 | VLoss: 0.0269 | PLoss: -0.0148 | Ent: 0.2133
+[ 70.3%] Iter 6868/9765 | SPS:  2511 | Reward:  0.070 | Value:  0.675 | VLoss: 0.0195 | PLoss: -0.0007 | Ent: 0.2031
+[ 70.3%] Iter 6869/9765 | SPS:  2511 | Reward:  0.068 | Value:  0.748 | VLoss: 0.0216 | PLoss: -0.0055 | Ent: 0.2043
+[ 70.4%] Iter 6870/9765 | SPS:  2511 | Reward:  0.066 | Value:  0.752 | VLoss: 0.0203 | PLoss: -0.0047 | Ent: 0.2261
+[ 70.4%] Iter 6871/9765 | SPS:  2511 | Reward:  0.062 | Value:  0.776 | VLoss: 0.0137 | PLoss: -0.0048 | Ent: 0.2359
+[ 70.4%] Iter 6872/9765 | SPS:  2511 | Reward:  0.051 | Value:  0.742 | VLoss: 0.0238 | PLoss: -0.0051 | Ent: 0.1806
+[ 70.4%] Iter 6873/9765 | SPS:  2512 | Reward:  0.059 | Value:  0.740 | VLoss: 0.0176 | PLoss: -0.0021 | Ent: 0.1813
+[ 70.4%] Iter 6874/9765 | SPS:  2512 | Reward:  0.064 | Value:  0.750 | VLoss: 0.0207 | PLoss: -0.0047 | Ent: 0.2238
+[ 70.4%] Iter 6875/9765 | SPS:  2512 | Reward:  0.059 | Value:  0.741 | VLoss: 0.0161 | PLoss: -0.0015 | Ent: 0.2284
+[ 70.4%] Iter 6876/9765 | SPS:  2512 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0238 | PLoss: -0.0126 | Ent: 0.2710
+[ 70.4%] Iter 6877/9765 | SPS:  2512 | Reward:  0.079 | Value:  0.743 | VLoss: 0.0269 | PLoss: -0.0070 | Ent: 0.2145
+[ 70.4%] Iter 6878/9765 | SPS:  2512 | Reward:  0.052 | Value:  0.747 | VLoss: 0.0279 | PLoss: -0.0098 | Ent: 0.2002
+[ 70.4%] Iter 6879/9765 | SPS:  2512 | Reward:  0.059 | Value:  0.644 | VLoss: 0.0212 | PLoss: -0.0124 | Ent: 0.2418
+[ 70.5%] Iter 6880/9765 | SPS:  2512 | Reward:  0.066 | Value:  0.658 | VLoss: 0.0203 | PLoss: -0.0008 | Ent: 0.1703
+[ 70.5%] Iter 6881/9765 | SPS:  2513 | Reward:  0.058 | Value:  0.741 | VLoss: 0.0265 | PLoss: -0.0020 | Ent: 0.2174
+[ 70.5%] Iter 6882/9765 | SPS:  2513 | Reward:  0.065 | Value:  0.750 | VLoss: 0.0154 | PLoss: -0.0080 | Ent: 0.2107
+[ 70.5%] Iter 6883/9765 | SPS:  2513 | Reward:  0.062 | Value:  0.677 | VLoss: 0.0286 | PLoss: -0.0095 | Ent: 0.2065
+[ 70.5%] Iter 6884/9765 | SPS:  2513 | Reward:  0.052 | Value:  0.686 | VLoss: 0.0235 | PLoss: -0.0027 | Ent: 0.2038
+[ 70.5%] Iter 6885/9765 | SPS:  2513 | Reward:  0.058 | Value:  0.762 | VLoss: 0.0138 | PLoss: -0.0091 | Ent: 0.2158
+[ 70.5%] Iter 6886/9765 | SPS:  2513 | Reward:  0.059 | Value:  0.720 | VLoss: 0.0203 | PLoss: -0.0017 | Ent: 0.2507
+[ 70.5%] Iter 6887/9765 | SPS:  2513 | Reward:  0.068 | Value:  0.745 | VLoss: 0.0277 | PLoss: -0.0028 | Ent: 0.2098
+[ 70.5%] Iter 6888/9765 | SPS:  2513 | Reward:  0.071 | Value:  0.778 | VLoss: 0.0222 | PLoss: -0.0004 | Ent: 0.2594
+[ 70.5%] Iter 6889/9765 | SPS:  2513 | Reward:  0.058 | Value:  0.698 | VLoss: 0.0267 | PLoss: -0.0050 | Ent: 0.1686
+[ 70.6%] Iter 6890/9765 | SPS:  2514 | Reward:  0.058 | Value:  0.744 | VLoss: 0.0210 | PLoss: -0.0053 | Ent: 0.2354
+[ 70.6%] Iter 6891/9765 | SPS:  2514 | Reward:  0.073 | Value:  0.694 | VLoss: 0.0253 | PLoss: -0.0088 | Ent: 0.2059
+[ 70.6%] Iter 6892/9765 | SPS:  2514 | Reward:  0.058 | Value:  0.737 | VLoss: 0.0170 | PLoss: -0.0017 | Ent: 0.2114
+[ 70.6%] Iter 6893/9765 | SPS:  2514 | Reward:  0.060 | Value:  0.716 | VLoss: 0.0239 | PLoss: -0.0035 | Ent: 0.2316
+[ 70.6%] Iter 6894/9765 | SPS:  2514 | Reward:  0.063 | Value:  0.752 | VLoss: 0.0180 | PLoss: -0.0093 | Ent: 0.2076
+[ 70.6%] Iter 6895/9765 | SPS:  2514 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0224 | PLoss: -0.0058 | Ent: 0.2505
+[ 70.6%] Iter 6896/9765 | SPS:  2514 | Reward:  0.066 | Value:  0.743 | VLoss: 0.0189 | PLoss: -0.0084 | Ent: 0.2081
+[ 70.6%] Iter 6897/9765 | SPS:  2514 | Reward:  0.065 | Value:  0.755 | VLoss: 0.0170 | PLoss: -0.0038 | Ent: 0.2182
+[ 70.6%] Iter 6898/9765 | SPS:  2515 | Reward:  0.048 | Value:  0.690 | VLoss: 0.0234 | PLoss: -0.0026 | Ent: 0.2643
+[ 70.7%] Iter 6899/9765 | SPS:  2515 | Reward:  0.063 | Value:  0.621 | VLoss: 0.0248 | PLoss: -0.0064 | Ent: 0.2262
+[ 70.7%] Iter 6900/9765 | SPS:  2515 | Reward:  0.049 | Value:  0.737 | VLoss: 0.0201 | PLoss: -0.0020 | Ent: 0.1833
+[ 70.7%] Iter 6901/9765 | SPS:  2515 | Reward:  0.058 | Value:  0.735 | VLoss: 0.0146 | PLoss: -0.0041 | Ent: 0.1817
+[ 70.7%] Iter 6902/9765 | SPS:  2515 | Reward:  0.072 | Value:  0.733 | VLoss: 0.0206 | PLoss: -0.0019 | Ent: 0.2181
+[ 70.7%] Iter 6903/9765 | SPS:  2515 | Reward:  0.073 | Value:  0.778 | VLoss: 0.0220 | PLoss: -0.0067 | Ent: 0.2016
+[ 70.7%] Iter 6904/9765 | SPS:  2515 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0177 | PLoss: -0.0122 | Ent: 0.2586
+[ 70.7%] Iter 6905/9765 | SPS:  2515 | Reward:  0.072 | Value:  0.700 | VLoss: 0.0282 | PLoss: -0.0077 | Ent: 0.2247
+[ 70.7%] Iter 6906/9765 | SPS:  2515 | Reward:  0.061 | Value:  0.800 | VLoss: 0.0311 | PLoss: -0.0010 | Ent: 0.2020
+[ 70.7%] Iter 6907/9765 | SPS:  2516 | Reward:  0.071 | Value:  0.754 | VLoss: 0.0174 | PLoss: -0.0093 | Ent: 0.2355
+[ 70.7%] Iter 6908/9765 | SPS:  2516 | Reward:  0.067 | Value:  0.737 | VLoss: 0.0242 | PLoss: -0.0044 | Ent: 0.1754
+[ 70.8%] Iter 6909/9765 | SPS:  2516 | Reward:  0.070 | Value:  0.792 | VLoss: 0.0205 | PLoss: -0.0074 | Ent: 0.1803
+[ 70.8%] Iter 6910/9765 | SPS:  2516 | Reward:  0.068 | Value:  0.742 | VLoss: 0.0291 | PLoss: -0.0057 | Ent: 0.2378
+[ 70.8%] Iter 6911/9765 | SPS:  2516 | Reward:  0.074 | Value:  0.749 | VLoss: 0.0266 | PLoss: -0.0051 | Ent: 0.2281
+[ 70.8%] Iter 6912/9765 | SPS:  2516 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0245 | PLoss: -0.0092 | Ent: 0.1781
+[ 70.8%] Iter 6913/9765 | SPS:  2516 | Reward:  0.049 | Value:  0.703 | VLoss: 0.0245 | PLoss: -0.0148 | Ent: 0.2803
+[ 70.8%] Iter 6914/9765 | SPS:  2516 | Reward:  0.054 | Value:  0.664 | VLoss: 0.0252 | PLoss: -0.0102 | Ent: 0.2262
+[ 70.8%] Iter 6915/9765 | SPS:  2517 | Reward:  0.055 | Value:  0.731 | VLoss: 0.0182 | PLoss: -0.0042 | Ent: 0.2465
+[ 70.8%] Iter 6916/9765 | SPS:  2517 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0231 | PLoss: -0.0102 | Ent: 0.2315
+[ 70.8%] Iter 6917/9765 | SPS:  2517 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0183 | PLoss: -0.0096 | Ent: 0.2315
+[ 70.8%] Iter 6918/9765 | SPS:  2517 | Reward:  0.068 | Value:  0.745 | VLoss: 0.0312 | PLoss: -0.0032 | Ent: 0.2065
+[ 70.9%] Iter 6919/9765 | SPS:  2517 | Reward:  0.074 | Value:  0.800 | VLoss: 0.0165 | PLoss: -0.0058 | Ent: 0.1991
+[ 70.9%] Iter 6920/9765 | SPS:  2517 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0233 | PLoss: -0.0064 | Ent: 0.1814
+[ 70.9%] Iter 6921/9765 | SPS:  2517 | Reward:  0.061 | Value:  0.779 | VLoss: 0.0192 | PLoss: -0.0063 | Ent: 0.2113
+[ 70.9%] Iter 6922/9765 | SPS:  2517 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0228 | PLoss: -0.0081 | Ent: 0.2276
+[ 70.9%] Iter 6923/9765 | SPS:  2518 | Reward:  0.049 | Value:  0.669 | VLoss: 0.0248 | PLoss: -0.0024 | Ent: 0.2458
+[ 70.9%] Iter 6924/9765 | SPS:  2518 | Reward:  0.055 | Value:  0.636 | VLoss: 0.0319 | PLoss: -0.0075 | Ent: 0.2671
+[ 70.9%] Iter 6925/9765 | SPS:  2518 | Reward:  0.051 | Value:  0.613 | VLoss: 0.0317 | PLoss: -0.0078 | Ent: 0.1960
+[ 70.9%] Iter 6926/9765 | SPS:  2518 | Reward:  0.051 | Value:  0.625 | VLoss: 0.0226 | PLoss: -0.0083 | Ent: 0.2461
+[ 70.9%] Iter 6927/9765 | SPS:  2518 | Reward:  0.053 | Value:  0.641 | VLoss: 0.0242 | PLoss: -0.0053 | Ent: 0.1959
+[ 70.9%] Iter 6928/9765 | SPS:  2518 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0259 | PLoss: -0.0045 | Ent: 0.2358
+[ 71.0%] Iter 6929/9765 | SPS:  2518 | Reward:  0.047 | Value:  0.711 | VLoss: 0.0209 | PLoss: -0.0063 | Ent: 0.2276
+[ 71.0%] Iter 6930/9765 | SPS:  2518 | Reward:  0.068 | Value:  0.717 | VLoss: 0.0240 | PLoss: -0.0033 | Ent: 0.2026
+[ 71.0%] Iter 6931/9765 | SPS:  2518 | Reward:  0.069 | Value:  0.752 | VLoss: 0.0198 | PLoss: -0.0112 | Ent: 0.2222
+[ 71.0%] Iter 6932/9765 | SPS:  2519 | Reward:  0.072 | Value:  0.740 | VLoss: 0.0233 | PLoss: -0.0064 | Ent: 0.1909
+[ 71.0%] Iter 6933/9765 | SPS:  2519 | Reward:  0.069 | Value:  0.775 | VLoss: 0.0245 | PLoss: -0.0100 | Ent: 0.2155
+[ 71.0%] Iter 6934/9765 | SPS:  2519 | Reward:  0.057 | Value:  0.690 | VLoss: 0.0193 | PLoss: -0.0042 | Ent: 0.1963
+[ 71.0%] Iter 6935/9765 | SPS:  2519 | Reward:  0.058 | Value:  0.746 | VLoss: 0.0291 | PLoss: -0.0027 | Ent: 0.1877
+[ 71.0%] Iter 6936/9765 | SPS:  2519 | Reward:  0.059 | Value:  0.696 | VLoss: 0.0254 | PLoss: -0.0062 | Ent: 0.1932
+[ 71.0%] Iter 6937/9765 | SPS:  2519 | Reward:  0.055 | Value:  0.740 | VLoss: 0.0182 | PLoss: -0.0020 | Ent: 0.2185
+[ 71.0%] Iter 6938/9765 | SPS:  2519 | Reward:  0.057 | Value:  0.731 | VLoss: 0.0234 | PLoss: -0.0063 | Ent: 0.2529
+[ 71.1%] Iter 6939/9765 | SPS:  2519 | Reward:  0.063 | Value:  0.623 | VLoss: 0.0233 | PLoss: -0.0027 | Ent: 0.2500
+[ 71.1%] Iter 6940/9765 | SPS:  2519 | Reward:  0.068 | Value:  0.742 | VLoss: 0.0241 | PLoss: -0.0103 | Ent: 0.2238
+[ 71.1%] Iter 6941/9765 | SPS:  2520 | Reward:  0.059 | Value:  0.694 | VLoss: 0.0210 | PLoss: -0.0109 | Ent: 0.2209
+[ 71.1%] Iter 6942/9765 | SPS:  2520 | Reward:  0.055 | Value:  0.671 | VLoss: 0.0148 | PLoss: -0.0048 | Ent: 0.1875
+[ 71.1%] Iter 6943/9765 | SPS:  2520 | Reward:  0.054 | Value:  0.728 | VLoss: 0.0276 | PLoss: -0.0089 | Ent: 0.2140
+[ 71.1%] Iter 6944/9765 | SPS:  2520 | Reward:  0.060 | Value:  0.770 | VLoss: 0.0188 | PLoss: -0.0076 | Ent: 0.1999
+[ 71.1%] Iter 6945/9765 | SPS:  2520 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0202 | PLoss: -0.0047 | Ent: 0.1947
+[ 71.1%] Iter 6946/9765 | SPS:  2520 | Reward:  0.062 | Value:  0.734 | VLoss: 0.0183 | PLoss: -0.0107 | Ent: 0.1852
+[ 71.1%] Iter 6947/9765 | SPS:  2520 | Reward:  0.063 | Value:  0.779 | VLoss: 0.0267 | PLoss: -0.0081 | Ent: 0.2338
+[ 71.2%] Iter 6948/9765 | SPS:  2520 | Reward:  0.044 | Value:  0.685 | VLoss: 0.0236 | PLoss: -0.0087 | Ent: 0.1669
+[ 71.2%] Iter 6949/9765 | SPS:  2521 | Reward:  0.068 | Value:  0.695 | VLoss: 0.0232 | PLoss: -0.0036 | Ent: 0.2413
+[ 71.2%] Iter 6950/9765 | SPS:  2521 | Reward:  0.043 | Value:  0.748 | VLoss: 0.0193 | PLoss: -0.0081 | Ent: 0.1569
+[ 71.2%] Iter 6951/9765 | SPS:  2521 | Reward:  0.060 | Value:  0.742 | VLoss: 0.0236 | PLoss: -0.0093 | Ent: 0.2223
+[ 71.2%] Iter 6952/9765 | SPS:  2521 | Reward:  0.058 | Value:  0.756 | VLoss: 0.0259 | PLoss: -0.0071 | Ent: 0.1616
+[ 71.2%] Iter 6953/9765 | SPS:  2521 | Reward:  0.061 | Value:  0.698 | VLoss: 0.0194 | PLoss: -0.0048 | Ent: 0.2116
+[ 71.2%] Iter 6954/9765 | SPS:  2521 | Reward:  0.058 | Value:  0.736 | VLoss: 0.0168 | PLoss: -0.0052 | Ent: 0.1962
+[ 71.2%] Iter 6955/9765 | SPS:  2521 | Reward:  0.064 | Value:  0.721 | VLoss: 0.0250 | PLoss: -0.0070 | Ent: 0.1981
+[ 71.2%] Iter 6956/9765 | SPS:  2521 | Reward:  0.056 | Value:  0.757 | VLoss: 0.0272 | PLoss: -0.0104 | Ent: 0.2153
+[ 71.2%] Iter 6957/9765 | SPS:  2522 | Reward:  0.057 | Value:  0.691 | VLoss: 0.0234 | PLoss: -0.0070 | Ent: 0.2039
+[ 71.3%] Iter 6958/9765 | SPS:  2522 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0222 | PLoss: 0.0032 | Ent: 0.2244
+[ 71.3%] Iter 6959/9765 | SPS:  2522 | Reward:  0.066 | Value:  0.721 | VLoss: 0.0260 | PLoss: -0.0050 | Ent: 0.2111
+[ 71.3%] Iter 6960/9765 | SPS:  2522 | Reward:  0.052 | Value:  0.763 | VLoss: 0.0153 | PLoss: -0.0015 | Ent: 0.2200
+[ 71.3%] Iter 6961/9765 | SPS:  2522 | Reward:  0.079 | Value:  0.781 | VLoss: 0.0196 | PLoss: -0.0067 | Ent: 0.2043
+[ 71.3%] Iter 6962/9765 | SPS:  2522 | Reward:  0.078 | Value:  0.781 | VLoss: 0.0326 | PLoss: -0.0125 | Ent: 0.1785
+[ 71.3%] Iter 6963/9765 | SPS:  2522 | Reward:  0.052 | Value:  0.728 | VLoss: 0.0228 | PLoss: -0.0033 | Ent: 0.2199
+[ 71.3%] Iter 6964/9765 | SPS:  2522 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0233 | PLoss: -0.0058 | Ent: 0.2387
+[ 71.3%] Iter 6965/9765 | SPS:  2522 | Reward:  0.062 | Value:  0.765 | VLoss: 0.0229 | PLoss: -0.0055 | Ent: 0.2055
+[ 71.3%] Iter 6966/9765 | SPS:  2523 | Reward:  0.049 | Value:  0.730 | VLoss: 0.0144 | PLoss: -0.0062 | Ent: 0.2049
+[ 71.3%] Iter 6967/9765 | SPS:  2523 | Reward:  0.063 | Value:  0.728 | VLoss: 0.0229 | PLoss: -0.0098 | Ent: 0.2013
+[ 71.4%] Iter 6968/9765 | SPS:  2523 | Reward:  0.060 | Value:  0.739 | VLoss: 0.0193 | PLoss: -0.0079 | Ent: 0.2707
+[ 71.4%] Iter 6969/9765 | SPS:  2523 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0277 | PLoss: 0.0010 | Ent: 0.2140
+[ 71.4%] Iter 6970/9765 | SPS:  2523 | Reward:  0.068 | Value:  0.702 | VLoss: 0.0204 | PLoss: -0.0081 | Ent: 0.2145
+[ 71.4%] Iter 6971/9765 | SPS:  2523 | Reward:  0.063 | Value:  0.718 | VLoss: 0.0208 | PLoss: -0.0091 | Ent: 0.2192
+[ 71.4%] Iter 6972/9765 | SPS:  2523 | Reward:  0.072 | Value:  0.741 | VLoss: 0.0253 | PLoss: -0.0031 | Ent: 0.2251
+[ 71.4%] Iter 6973/9765 | SPS:  2523 | Reward:  0.063 | Value:  0.738 | VLoss: 0.0193 | PLoss: -0.0043 | Ent: 0.2066
+[ 71.4%] Iter 6974/9765 | SPS:  2523 | Reward:  0.062 | Value:  0.636 | VLoss: 0.0181 | PLoss: -0.0065 | Ent: 0.2098
+[ 71.4%] Iter 6975/9765 | SPS:  2524 | Reward:  0.061 | Value:  0.720 | VLoss: 0.0235 | PLoss: -0.0067 | Ent: 0.2058
+[ 71.4%] Iter 6976/9765 | SPS:  2524 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0223 | PLoss: -0.0091 | Ent: 0.2693
+[ 71.4%] Iter 6977/9765 | SPS:  2524 | Reward:  0.066 | Value:  0.797 | VLoss: 0.0190 | PLoss: -0.0058 | Ent: 0.1909
+[ 71.5%] Iter 6978/9765 | SPS:  2524 | Reward:  0.048 | Value:  0.718 | VLoss: 0.0216 | PLoss: -0.0079 | Ent: 0.2050
+[ 71.5%] Iter 6979/9765 | SPS:  2524 | Reward:  0.061 | Value:  0.664 | VLoss: 0.0161 | PLoss: 0.0008 | Ent: 0.2270
+[ 71.5%] Iter 6980/9765 | SPS:  2524 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0150 | PLoss: -0.0051 | Ent: 0.1627
+[ 71.5%] Iter 6981/9765 | SPS:  2524 | Reward:  0.049 | Value:  0.695 | VLoss: 0.0206 | PLoss: -0.0058 | Ent: 0.2241
+[ 71.5%] Iter 6982/9765 | SPS:  2524 | Reward:  0.062 | Value:  0.672 | VLoss: 0.0293 | PLoss: -0.0155 | Ent: 0.2423
+[ 71.5%] Iter 6983/9765 | SPS:  2524 | Reward:  0.049 | Value:  0.680 | VLoss: 0.0262 | PLoss: -0.0074 | Ent: 0.2187
+[ 71.5%] Iter 6984/9765 | SPS:  2525 | Reward:  0.059 | Value:  0.744 | VLoss: 0.0148 | PLoss: -0.0079 | Ent: 0.1930
+[ 71.5%] Iter 6985/9765 | SPS:  2525 | Reward:  0.043 | Value:  0.713 | VLoss: 0.0118 | PLoss: -0.0030 | Ent: 0.1927
+[ 71.5%] Iter 6986/9765 | SPS:  2525 | Reward:  0.041 | Value:  0.709 | VLoss: 0.0178 | PLoss: -0.0072 | Ent: 0.1404
+[ 71.6%] Iter 6987/9765 | SPS:  2525 | Reward:  0.052 | Value:  0.609 | VLoss: 0.0216 | PLoss: -0.0080 | Ent: 0.1939
+[ 71.6%] Iter 6988/9765 | SPS:  2525 | Reward:  0.044 | Value:  0.629 | VLoss: 0.0248 | PLoss: -0.0059 | Ent: 0.2704
+[ 71.6%] Iter 6989/9765 | SPS:  2525 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0235 | PLoss: -0.0103 | Ent: 0.2160
+[ 71.6%] Iter 6990/9765 | SPS:  2525 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0236 | PLoss: -0.0141 | Ent: 0.2019
+[ 71.6%] Iter 6991/9765 | SPS:  2525 | Reward:  0.060 | Value:  0.714 | VLoss: 0.0220 | PLoss: -0.0066 | Ent: 0.1852
+[ 71.6%] Iter 6992/9765 | SPS:  2525 | Reward:  0.058 | Value:  0.739 | VLoss: 0.0201 | PLoss: -0.0012 | Ent: 0.1621
+[ 71.6%] Iter 6993/9765 | SPS:  2526 | Reward:  0.044 | Value:  0.712 | VLoss: 0.0144 | PLoss: -0.0100 | Ent: 0.1939
+[ 71.6%] Iter 6994/9765 | SPS:  2526 | Reward:  0.058 | Value:  0.683 | VLoss: 0.0192 | PLoss: -0.0081 | Ent: 0.2306
+[ 71.6%] Iter 6995/9765 | SPS:  2526 | Reward:  0.070 | Value:  0.761 | VLoss: 0.0211 | PLoss: -0.0038 | Ent: 0.2007
+[ 71.6%] Iter 6996/9765 | SPS:  2526 | Reward:  0.054 | Value:  0.756 | VLoss: 0.0192 | PLoss: -0.0031 | Ent: 0.1731
+[ 71.7%] Iter 6997/9765 | SPS:  2526 | Reward:  0.070 | Value:  0.736 | VLoss: 0.0238 | PLoss: -0.0083 | Ent: 0.2328
+[ 71.7%] Iter 6998/9765 | SPS:  2526 | Reward:  0.062 | Value:  0.748 | VLoss: 0.0209 | PLoss: -0.0131 | Ent: 0.2196
+[ 71.7%] Iter 6999/9765 | SPS:  2526 | Reward:  0.072 | Value:  0.746 | VLoss: 0.0220 | PLoss: -0.0073 | Ent: 0.1848
+[ 71.7%] Iter 7000/9765 | SPS:  2526 | Reward:  0.052 | Value:  0.656 | VLoss: 0.0281 | PLoss: -0.0044 | Ent: 0.1648
+[ 71.7%] Iter 7001/9765 | SPS:  2527 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0191 | PLoss: 0.0014 | Ent: 0.1781
+[ 71.7%] Iter 7002/9765 | SPS:  2527 | Reward:  0.057 | Value:  0.716 | VLoss: 0.0158 | PLoss: -0.0036 | Ent: 0.2303
+[ 71.7%] Iter 7003/9765 | SPS:  2527 | Reward:  0.057 | Value:  0.746 | VLoss: 0.0203 | PLoss: -0.0055 | Ent: 0.1591
+[ 71.7%] Iter 7004/9765 | SPS:  2527 | Reward:  0.059 | Value:  0.720 | VLoss: 0.0263 | PLoss: -0.0038 | Ent: 0.2413
+[ 71.7%] Iter 7005/9765 | SPS:  2527 | Reward:  0.065 | Value:  0.673 | VLoss: 0.0261 | PLoss: -0.0084 | Ent: 0.2325
+[ 71.7%] Iter 7006/9765 | SPS:  2527 | Reward:  0.045 | Value:  0.790 | VLoss: 0.0183 | PLoss: -0.0092 | Ent: 0.2024
+[ 71.8%] Iter 7007/9765 | SPS:  2527 | Reward:  0.059 | Value:  0.706 | VLoss: 0.0240 | PLoss: -0.0101 | Ent: 0.2376
+[ 71.8%] Iter 7008/9765 | SPS:  2527 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0186 | PLoss: -0.0104 | Ent: 0.1958
+[ 71.8%] Iter 7009/9765 | SPS:  2527 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0220 | PLoss: -0.0040 | Ent: 0.2195
+[ 71.8%] Iter 7010/9765 | SPS:  2528 | Reward:  0.065 | Value:  0.742 | VLoss: 0.0197 | PLoss: -0.0049 | Ent: 0.2188
+[ 71.8%] Iter 7011/9765 | SPS:  2528 | Reward:  0.064 | Value:  0.728 | VLoss: 0.0387 | PLoss: -0.0004 | Ent: 0.1817
+[ 71.8%] Iter 7012/9765 | SPS:  2528 | Reward:  0.049 | Value:  0.692 | VLoss: 0.0293 | PLoss: -0.0101 | Ent: 0.2533
+[ 71.8%] Iter 7013/9765 | SPS:  2528 | Reward:  0.058 | Value:  0.657 | VLoss: 0.0257 | PLoss: -0.0054 | Ent: 0.2023
+[ 71.8%] Iter 7014/9765 | SPS:  2528 | Reward:  0.049 | Value:  0.679 | VLoss: 0.0305 | PLoss: -0.0062 | Ent: 0.2485
+[ 71.8%] Iter 7015/9765 | SPS:  2528 | Reward:  0.076 | Value:  0.699 | VLoss: 0.0216 | PLoss: -0.0019 | Ent: 0.1660
+[ 71.8%] Iter 7016/9765 | SPS:  2528 | Reward:  0.071 | Value:  0.706 | VLoss: 0.0314 | PLoss: -0.0090 | Ent: 0.2193
+[ 71.9%] Iter 7017/9765 | SPS:  2528 | Reward:  0.067 | Value:  0.750 | VLoss: 0.0216 | PLoss: -0.0084 | Ent: 0.1762
+[ 71.9%] Iter 7018/9765 | SPS:  2529 | Reward:  0.062 | Value:  0.757 | VLoss: 0.0195 | PLoss: -0.0057 | Ent: 0.1463
+[ 71.9%] Iter 7019/9765 | SPS:  2529 | Reward:  0.062 | Value:  0.655 | VLoss: 0.0203 | PLoss: -0.0018 | Ent: 0.2284
+[ 71.9%] Iter 7020/9765 | SPS:  2529 | Reward:  0.057 | Value:  0.639 | VLoss: 0.0182 | PLoss: -0.0151 | Ent: 0.2104
+[ 71.9%] Iter 7021/9765 | SPS:  2529 | Reward:  0.065 | Value:  0.649 | VLoss: 0.0240 | PLoss: -0.0078 | Ent: 0.1942
+[ 71.9%] Iter 7022/9765 | SPS:  2529 | Reward:  0.064 | Value:  0.642 | VLoss: 0.0258 | PLoss: -0.0061 | Ent: 0.1875
+[ 71.9%] Iter 7023/9765 | SPS:  2529 | Reward:  0.057 | Value:  0.725 | VLoss: 0.0188 | PLoss: -0.0052 | Ent: 0.2180
+[ 71.9%] Iter 7024/9765 | SPS:  2529 | Reward:  0.054 | Value:  0.794 | VLoss: 0.0201 | PLoss: -0.0127 | Ent: 0.2527
+[ 71.9%] Iter 7025/9765 | SPS:  2529 | Reward:  0.061 | Value:  0.776 | VLoss: 0.0214 | PLoss: -0.0040 | Ent: 0.2152
+[ 72.0%] Iter 7026/9765 | SPS:  2529 | Reward:  0.061 | Value:  0.740 | VLoss: 0.0284 | PLoss: -0.0076 | Ent: 0.2026
+[ 72.0%] Iter 7027/9765 | SPS:  2530 | Reward:  0.068 | Value:  0.743 | VLoss: 0.0271 | PLoss: -0.0078 | Ent: 0.2139
+[ 72.0%] Iter 7028/9765 | SPS:  2530 | Reward:  0.054 | Value:  0.688 | VLoss: 0.0187 | PLoss: -0.0051 | Ent: 0.2267
+[ 72.0%] Iter 7029/9765 | SPS:  2530 | Reward:  0.069 | Value:  0.666 | VLoss: 0.0179 | PLoss: -0.0040 | Ent: 0.1820
+[ 72.0%] Iter 7030/9765 | SPS:  2530 | Reward:  0.050 | Value:  0.728 | VLoss: 0.0176 | PLoss: -0.0025 | Ent: 0.1963
+[ 72.0%] Iter 7031/9765 | SPS:  2530 | Reward:  0.041 | Value:  0.693 | VLoss: 0.0167 | PLoss: -0.0043 | Ent: 0.2660
+[ 72.0%] Iter 7032/9765 | SPS:  2530 | Reward:  0.049 | Value:  0.642 | VLoss: 0.0222 | PLoss: -0.0077 | Ent: 0.2051
+[ 72.0%] Iter 7033/9765 | SPS:  2530 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0186 | PLoss: -0.0121 | Ent: 0.2167
+[ 72.0%] Iter 7034/9765 | SPS:  2530 | Reward:  0.070 | Value:  0.691 | VLoss: 0.0284 | PLoss: -0.0073 | Ent: 0.2246
+[ 72.0%] Iter 7035/9765 | SPS:  2531 | Reward:  0.047 | Value:  0.731 | VLoss: 0.0218 | PLoss: -0.0076 | Ent: 0.2267
+[ 72.1%] Iter 7036/9765 | SPS:  2531 | Reward:  0.046 | Value:  0.655 | VLoss: 0.0231 | PLoss: -0.0114 | Ent: 0.2215
+[ 72.1%] Iter 7037/9765 | SPS:  2531 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0309 | PLoss: -0.0181 | Ent: 0.2222
+[ 72.1%] Iter 7038/9765 | SPS:  2531 | Reward:  0.076 | Value:  0.608 | VLoss: 0.0258 | PLoss: -0.0034 | Ent: 0.2480
+[ 72.1%] Iter 7039/9765 | SPS:  2531 | Reward:  0.066 | Value:  0.747 | VLoss: 0.0195 | PLoss: -0.0060 | Ent: 0.1870
+[ 72.1%] Iter 7040/9765 | SPS:  2531 | Reward:  0.053 | Value:  0.700 | VLoss: 0.0205 | PLoss: -0.0149 | Ent: 0.2000
+[ 72.1%] Iter 7041/9765 | SPS:  2531 | Reward:  0.059 | Value:  0.716 | VLoss: 0.0218 | PLoss: -0.0081 | Ent: 0.2201
+[ 72.1%] Iter 7042/9765 | SPS:  2531 | Reward:  0.052 | Value:  0.735 | VLoss: 0.0233 | PLoss: -0.0068 | Ent: 0.2163
+[ 72.1%] Iter 7043/9765 | SPS:  2531 | Reward:  0.060 | Value:  0.766 | VLoss: 0.0174 | PLoss: -0.0156 | Ent: 0.2210
+[ 72.1%] Iter 7044/9765 | SPS:  2532 | Reward:  0.043 | Value:  0.756 | VLoss: 0.0184 | PLoss: -0.0059 | Ent: 0.1939
+[ 72.1%] Iter 7045/9765 | SPS:  2532 | Reward:  0.053 | Value:  0.639 | VLoss: 0.0207 | PLoss: -0.0090 | Ent: 0.2173
+[ 72.2%] Iter 7046/9765 | SPS:  2532 | Reward:  0.048 | Value:  0.736 | VLoss: 0.0132 | PLoss: -0.0008 | Ent: 0.1873
+[ 72.2%] Iter 7047/9765 | SPS:  2532 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0237 | PLoss: -0.0090 | Ent: 0.2729
+[ 72.2%] Iter 7048/9765 | SPS:  2532 | Reward:  0.062 | Value:  0.728 | VLoss: 0.0201 | PLoss: 0.0002 | Ent: 0.2491
+[ 72.2%] Iter 7049/9765 | SPS:  2532 | Reward:  0.051 | Value:  0.711 | VLoss: 0.0201 | PLoss: -0.0057 | Ent: 0.2362
+[ 72.2%] Iter 7050/9765 | SPS:  2532 | Reward:  0.065 | Value:  0.702 | VLoss: 0.0250 | PLoss: -0.0089 | Ent: 0.2464
+[ 72.2%] Iter 7051/9765 | SPS:  2532 | Reward:  0.061 | Value:  0.751 | VLoss: 0.0197 | PLoss: 0.0026 | Ent: 0.2287
+[ 72.2%] Iter 7052/9765 | SPS:  2533 | Reward:  0.055 | Value:  0.742 | VLoss: 0.0216 | PLoss: -0.0081 | Ent: 0.2600
+[ 72.2%] Iter 7053/9765 | SPS:  2533 | Reward:  0.061 | Value:  0.702 | VLoss: 0.0189 | PLoss: 0.0004 | Ent: 0.2038
+[ 72.2%] Iter 7054/9765 | SPS:  2533 | Reward:  0.058 | Value:  0.758 | VLoss: 0.0144 | PLoss: -0.0052 | Ent: 0.1888
+[ 72.2%] Iter 7055/9765 | SPS:  2533 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0268 | PLoss: -0.0054 | Ent: 0.1924
+[ 72.3%] Iter 7056/9765 | SPS:  2533 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0264 | PLoss: -0.0085 | Ent: 0.2654
+[ 72.3%] Iter 7057/9765 | SPS:  2533 | Reward:  0.062 | Value:  0.769 | VLoss: 0.0266 | PLoss: -0.0017 | Ent: 0.2496
+[ 72.3%] Iter 7058/9765 | SPS:  2533 | Reward:  0.065 | Value:  0.739 | VLoss: 0.0188 | PLoss: -0.0089 | Ent: 0.2103
+[ 72.3%] Iter 7059/9765 | SPS:  2533 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0167 | PLoss: -0.0062 | Ent: 0.2153
+[ 72.3%] Iter 7060/9765 | SPS:  2533 | Reward:  0.068 | Value:  0.756 | VLoss: 0.0188 | PLoss: -0.0091 | Ent: 0.2300
+[ 72.3%] Iter 7061/9765 | SPS:  2534 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0254 | PLoss: -0.0087 | Ent: 0.2485
+[ 72.3%] Iter 7062/9765 | SPS:  2534 | Reward:  0.069 | Value:  0.652 | VLoss: 0.0348 | PLoss: -0.0027 | Ent: 0.2609
+[ 72.3%] Iter 7063/9765 | SPS:  2534 | Reward:  0.056 | Value:  0.643 | VLoss: 0.0229 | PLoss: -0.0112 | Ent: 0.3014
+[ 72.3%] Iter 7064/9765 | SPS:  2534 | Reward:  0.065 | Value:  0.769 | VLoss: 0.0207 | PLoss: -0.0002 | Ent: 0.2351
+[ 72.4%] Iter 7065/9765 | SPS:  2534 | Reward:  0.053 | Value:  0.707 | VLoss: 0.0190 | PLoss: -0.0055 | Ent: 0.2111
+[ 72.4%] Iter 7066/9765 | SPS:  2534 | Reward:  0.060 | Value:  0.724 | VLoss: 0.0265 | PLoss: -0.0106 | Ent: 0.2739
+[ 72.4%] Iter 7067/9765 | SPS:  2534 | Reward:  0.072 | Value:  0.691 | VLoss: 0.0222 | PLoss: -0.0109 | Ent: 0.1701
+[ 72.4%] Iter 7068/9765 | SPS:  2534 | Reward:  0.064 | Value:  0.721 | VLoss: 0.0187 | PLoss: -0.0018 | Ent: 0.2002
+[ 72.4%] Iter 7069/9765 | SPS:  2534 | Reward:  0.069 | Value:  0.761 | VLoss: 0.0151 | PLoss: -0.0024 | Ent: 0.2141
+[ 72.4%] Iter 7070/9765 | SPS:  2535 | Reward:  0.058 | Value:  0.769 | VLoss: 0.0202 | PLoss: -0.0017 | Ent: 0.2267
+[ 72.4%] Iter 7071/9765 | SPS:  2535 | Reward:  0.057 | Value:  0.737 | VLoss: 0.0146 | PLoss: -0.0106 | Ent: 0.2580
+[ 72.4%] Iter 7072/9765 | SPS:  2535 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0280 | PLoss: -0.0053 | Ent: 0.2148
+[ 72.4%] Iter 7073/9765 | SPS:  2535 | Reward:  0.063 | Value:  0.762 | VLoss: 0.0268 | PLoss: -0.0093 | Ent: 0.2361
+[ 72.4%] Iter 7074/9765 | SPS:  2535 | Reward:  0.068 | Value:  0.711 | VLoss: 0.0229 | PLoss: -0.0058 | Ent: 0.2065
+[ 72.5%] Iter 7075/9765 | SPS:  2535 | Reward:  0.065 | Value:  0.770 | VLoss: 0.0190 | PLoss: -0.0014 | Ent: 0.2028
+[ 72.5%] Iter 7076/9765 | SPS:  2535 | Reward:  0.047 | Value:  0.764 | VLoss: 0.0168 | PLoss: -0.0094 | Ent: 0.2142
+[ 72.5%] Iter 7077/9765 | SPS:  2535 | Reward:  0.057 | Value:  0.722 | VLoss: 0.0180 | PLoss: -0.0053 | Ent: 0.2097
+[ 72.5%] Iter 7078/9765 | SPS:  2536 | Reward:  0.055 | Value:  0.663 | VLoss: 0.0156 | PLoss: -0.0015 | Ent: 0.1826
+[ 72.5%] Iter 7079/9765 | SPS:  2536 | Reward:  0.057 | Value:  0.682 | VLoss: 0.0198 | PLoss: -0.0106 | Ent: 0.1990
+[ 72.5%] Iter 7080/9765 | SPS:  2536 | Reward:  0.068 | Value:  0.750 | VLoss: 0.0242 | PLoss: -0.0040 | Ent: 0.2203
+[ 72.5%] Iter 7081/9765 | SPS:  2536 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0227 | PLoss: -0.0036 | Ent: 0.2212
+[ 72.5%] Iter 7082/9765 | SPS:  2536 | Reward:  0.070 | Value:  0.708 | VLoss: 0.0233 | PLoss: -0.0090 | Ent: 0.2078
+[ 72.5%] Iter 7083/9765 | SPS:  2536 | Reward:  0.053 | Value:  0.748 | VLoss: 0.0155 | PLoss: -0.0096 | Ent: 0.2138
+[ 72.5%] Iter 7084/9765 | SPS:  2536 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0270 | PLoss: -0.0053 | Ent: 0.2254
+[ 72.6%] Iter 7085/9765 | SPS:  2536 | Reward:  0.057 | Value:  0.721 | VLoss: 0.0123 | PLoss: -0.0081 | Ent: 0.1548
+[ 72.6%] Iter 7086/9765 | SPS:  2536 | Reward:  0.057 | Value:  0.753 | VLoss: 0.0195 | PLoss: -0.0061 | Ent: 0.1838
+[ 72.6%] Iter 7087/9765 | SPS:  2537 | Reward:  0.057 | Value:  0.673 | VLoss: 0.0268 | PLoss: -0.0041 | Ent: 0.2368
+[ 72.6%] Iter 7088/9765 | SPS:  2537 | Reward:  0.056 | Value:  0.699 | VLoss: 0.0242 | PLoss: -0.0036 | Ent: 0.2390
+[ 72.6%] Iter 7089/9765 | SPS:  2537 | Reward:  0.048 | Value:  0.598 | VLoss: 0.0131 | PLoss: 0.0010 | Ent: 0.1769
+[ 72.6%] Iter 7090/9765 | SPS:  2537 | Reward:  0.053 | Value:  0.693 | VLoss: 0.0213 | PLoss: -0.0010 | Ent: 0.2473
+[ 72.6%] Iter 7091/9765 | SPS:  2537 | Reward:  0.047 | Value:  0.696 | VLoss: 0.0251 | PLoss: -0.0034 | Ent: 0.1808
+[ 72.6%] Iter 7092/9765 | SPS:  2537 | Reward:  0.061 | Value:  0.619 | VLoss: 0.0219 | PLoss: -0.0155 | Ent: 0.1993
+[ 72.6%] Iter 7093/9765 | SPS:  2537 | Reward:  0.069 | Value:  0.676 | VLoss: 0.0260 | PLoss: -0.0112 | Ent: 0.2188
+[ 72.6%] Iter 7094/9765 | SPS:  2537 | Reward:  0.043 | Value:  0.741 | VLoss: 0.0210 | PLoss: -0.0014 | Ent: 0.2369
+[ 72.7%] Iter 7095/9765 | SPS:  2538 | Reward:  0.062 | Value:  0.742 | VLoss: 0.0162 | PLoss: -0.0048 | Ent: 0.1662
+[ 72.7%] Iter 7096/9765 | SPS:  2538 | Reward:  0.061 | Value:  0.739 | VLoss: 0.0256 | PLoss: -0.0149 | Ent: 0.2613
+[ 72.7%] Iter 7097/9765 | SPS:  2538 | Reward:  0.073 | Value:  0.739 | VLoss: 0.0268 | PLoss: -0.0063 | Ent: 0.2748
+[ 72.7%] Iter 7098/9765 | SPS:  2538 | Reward:  0.050 | Value:  0.640 | VLoss: 0.0235 | PLoss: -0.0034 | Ent: 0.1943
+[ 72.7%] Iter 7099/9765 | SPS:  2538 | Reward:  0.054 | Value:  0.713 | VLoss: 0.0178 | PLoss: -0.0023 | Ent: 0.2246
+[ 72.7%] Iter 7100/9765 | SPS:  2538 | Reward:  0.063 | Value:  0.674 | VLoss: 0.0192 | PLoss: -0.0054 | Ent: 0.1662
+[ 72.7%] Iter 7101/9765 | SPS:  2538 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0268 | PLoss: -0.0031 | Ent: 0.1927
+[ 72.7%] Iter 7102/9765 | SPS:  2538 | Reward:  0.062 | Value:  0.628 | VLoss: 0.0191 | PLoss: -0.0081 | Ent: 0.2157
+[ 72.7%] Iter 7103/9765 | SPS:  2538 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0170 | PLoss: -0.0088 | Ent: 0.2145
+[ 72.7%] Iter 7104/9765 | SPS:  2539 | Reward:  0.065 | Value:  0.746 | VLoss: 0.0245 | PLoss: -0.0030 | Ent: 0.2277
+[ 72.8%] Iter 7105/9765 | SPS:  2539 | Reward:  0.059 | Value:  0.779 | VLoss: 0.0270 | PLoss: -0.0092 | Ent: 0.2110
+[ 72.8%] Iter 7106/9765 | SPS:  2539 | Reward:  0.052 | Value:  0.628 | VLoss: 0.0268 | PLoss: -0.0084 | Ent: 0.2374
+[ 72.8%] Iter 7107/9765 | SPS:  2539 | Reward:  0.055 | Value:  0.704 | VLoss: 0.0168 | PLoss: -0.0009 | Ent: 0.1802
+[ 72.8%] Iter 7108/9765 | SPS:  2539 | Reward:  0.063 | Value:  0.769 | VLoss: 0.0308 | PLoss: -0.0087 | Ent: 0.1861
+[ 72.8%] Iter 7109/9765 | SPS:  2539 | Reward:  0.063 | Value:  0.736 | VLoss: 0.0136 | PLoss: -0.0096 | Ent: 0.1475
+[ 72.8%] Iter 7110/9765 | SPS:  2539 | Reward:  0.069 | Value:  0.760 | VLoss: 0.0193 | PLoss: -0.0099 | Ent: 0.2149
+[ 72.8%] Iter 7111/9765 | SPS:  2539 | Reward:  0.048 | Value:  0.745 | VLoss: 0.0210 | PLoss: -0.0065 | Ent: 0.2023
+[ 72.8%] Iter 7112/9765 | SPS:  2540 | Reward:  0.048 | Value:  0.572 | VLoss: 0.0259 | PLoss: -0.0043 | Ent: 0.1953
+[ 72.8%] Iter 7113/9765 | SPS:  2540 | Reward:  0.058 | Value:  0.718 | VLoss: 0.0227 | PLoss: -0.0124 | Ent: 0.2228
+[ 72.9%] Iter 7114/9765 | SPS:  2540 | Reward:  0.051 | Value:  0.692 | VLoss: 0.0167 | PLoss: -0.0005 | Ent: 0.1897
+[ 72.9%] Iter 7115/9765 | SPS:  2540 | Reward:  0.063 | Value:  0.643 | VLoss: 0.0188 | PLoss: -0.0047 | Ent: 0.1774
+[ 72.9%] Iter 7116/9765 | SPS:  2540 | Reward:  0.053 | Value:  0.641 | VLoss: 0.0195 | PLoss: -0.0022 | Ent: 0.2133
+[ 72.9%] Iter 7117/9765 | SPS:  2540 | Reward:  0.061 | Value:  0.788 | VLoss: 0.0143 | PLoss: -0.0028 | Ent: 0.1913
+[ 72.9%] Iter 7118/9765 | SPS:  2540 | Reward:  0.059 | Value:  0.790 | VLoss: 0.0213 | PLoss: -0.0081 | Ent: 0.2141
+[ 72.9%] Iter 7119/9765 | SPS:  2540 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0242 | PLoss: -0.0056 | Ent: 0.1894
+[ 72.9%] Iter 7120/9765 | SPS:  2540 | Reward:  0.053 | Value:  0.682 | VLoss: 0.0176 | PLoss: -0.0058 | Ent: 0.2280
+[ 72.9%] Iter 7121/9765 | SPS:  2541 | Reward:  0.055 | Value:  0.750 | VLoss: 0.0234 | PLoss: -0.0123 | Ent: 0.1839
+[ 72.9%] Iter 7122/9765 | SPS:  2541 | Reward:  0.064 | Value:  0.767 | VLoss: 0.0178 | PLoss: -0.0075 | Ent: 0.1613
+[ 72.9%] Iter 7123/9765 | SPS:  2541 | Reward:  0.061 | Value:  0.789 | VLoss: 0.0190 | PLoss: -0.0076 | Ent: 0.1691
+[ 73.0%] Iter 7124/9765 | SPS:  2541 | Reward:  0.060 | Value:  0.769 | VLoss: 0.0217 | PLoss: -0.0080 | Ent: 0.1945
+[ 73.0%] Iter 7125/9765 | SPS:  2541 | Reward:  0.056 | Value:  0.709 | VLoss: 0.0227 | PLoss: -0.0043 | Ent: 0.1913
+[ 73.0%] Iter 7126/9765 | SPS:  2541 | Reward:  0.055 | Value:  0.729 | VLoss: 0.0185 | PLoss: -0.0084 | Ent: 0.1695
+[ 73.0%] Iter 7127/9765 | SPS:  2541 | Reward:  0.055 | Value:  0.717 | VLoss: 0.0141 | PLoss: -0.0012 | Ent: 0.2046
+[ 73.0%] Iter 7128/9765 | SPS:  2541 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0218 | PLoss: -0.0052 | Ent: 0.2219
+[ 73.0%] Iter 7129/9765 | SPS:  2542 | Reward:  0.061 | Value:  0.676 | VLoss: 0.0256 | PLoss: -0.0045 | Ent: 0.2344
+[ 73.0%] Iter 7130/9765 | SPS:  2542 | Reward:  0.054 | Value:  0.747 | VLoss: 0.0203 | PLoss: -0.0040 | Ent: 0.2224
+[ 73.0%] Iter 7131/9765 | SPS:  2542 | Reward:  0.059 | Value:  0.712 | VLoss: 0.0223 | PLoss: -0.0113 | Ent: 0.2474
+[ 73.0%] Iter 7132/9765 | SPS:  2542 | Reward:  0.060 | Value:  0.668 | VLoss: 0.0278 | PLoss: -0.0123 | Ent: 0.2141
+[ 73.0%] Iter 7133/9765 | SPS:  2542 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0188 | PLoss: -0.0033 | Ent: 0.2228
+[ 73.1%] Iter 7134/9765 | SPS:  2542 | Reward:  0.069 | Value:  0.776 | VLoss: 0.0201 | PLoss: -0.0038 | Ent: 0.1842
+[ 73.1%] Iter 7135/9765 | SPS:  2542 | Reward:  0.078 | Value:  0.759 | VLoss: 0.0318 | PLoss: -0.0078 | Ent: 0.1827
+[ 73.1%] Iter 7136/9765 | SPS:  2542 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0253 | PLoss: -0.0109 | Ent: 0.1968
+[ 73.1%] Iter 7137/9765 | SPS:  2542 | Reward:  0.066 | Value:  0.781 | VLoss: 0.0251 | PLoss: -0.0111 | Ent: 0.2414
+[ 73.1%] Iter 7138/9765 | SPS:  2543 | Reward:  0.063 | Value:  0.738 | VLoss: 0.0203 | PLoss: -0.0008 | Ent: 0.2050
+[ 73.1%] Iter 7139/9765 | SPS:  2543 | Reward:  0.067 | Value:  0.704 | VLoss: 0.0216 | PLoss: -0.0042 | Ent: 0.1908
+[ 73.1%] Iter 7140/9765 | SPS:  2543 | Reward:  0.076 | Value:  0.739 | VLoss: 0.0217 | PLoss: -0.0042 | Ent: 0.1958
+[ 73.1%] Iter 7141/9765 | SPS:  2543 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0204 | PLoss: -0.0051 | Ent: 0.2037
+[ 73.1%] Iter 7142/9765 | SPS:  2543 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0263 | PLoss: -0.0025 | Ent: 0.2293
+[ 73.1%] Iter 7143/9765 | SPS:  2543 | Reward:  0.068 | Value:  0.703 | VLoss: 0.0228 | PLoss: -0.0084 | Ent: 0.1954
+[ 73.2%] Iter 7144/9765 | SPS:  2543 | Reward:  0.063 | Value:  0.741 | VLoss: 0.0248 | PLoss: -0.0029 | Ent: 0.1679
+[ 73.2%] Iter 7145/9765 | SPS:  2543 | Reward:  0.058 | Value:  0.700 | VLoss: 0.0252 | PLoss: -0.0064 | Ent: 0.1864
+[ 73.2%] Iter 7146/9765 | SPS:  2544 | Reward:  0.053 | Value:  0.734 | VLoss: 0.0312 | PLoss: -0.0058 | Ent: 0.2375
+[ 73.2%] Iter 7147/9765 | SPS:  2544 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0218 | PLoss: -0.0063 | Ent: 0.1879
+[ 73.2%] Iter 7148/9765 | SPS:  2544 | Reward:  0.059 | Value:  0.791 | VLoss: 0.0220 | PLoss: -0.0128 | Ent: 0.2219
+[ 73.2%] Iter 7149/9765 | SPS:  2544 | Reward:  0.066 | Value:  0.728 | VLoss: 0.0203 | PLoss: -0.0025 | Ent: 0.1956
+[ 73.2%] Iter 7150/9765 | SPS:  2544 | Reward:  0.061 | Value:  0.755 | VLoss: 0.0190 | PLoss: 0.0053 | Ent: 0.1939
+[ 73.2%] Iter 7151/9765 | SPS:  2544 | Reward:  0.061 | Value:  0.729 | VLoss: 0.0205 | PLoss: -0.0072 | Ent: 0.1672
+[ 73.2%] Iter 7152/9765 | SPS:  2544 | Reward:  0.058 | Value:  0.664 | VLoss: 0.0124 | PLoss: -0.0081 | Ent: 0.2015
+[ 73.3%] Iter 7153/9765 | SPS:  2544 | Reward:  0.052 | Value:  0.754 | VLoss: 0.0170 | PLoss: -0.0075 | Ent: 0.1710
+[ 73.3%] Iter 7154/9765 | SPS:  2544 | Reward:  0.065 | Value:  0.769 | VLoss: 0.0234 | PLoss: -0.0086 | Ent: 0.1710
+[ 73.3%] Iter 7155/9765 | SPS:  2545 | Reward:  0.059 | Value:  0.750 | VLoss: 0.0188 | PLoss: -0.0015 | Ent: 0.2175
+[ 73.3%] Iter 7156/9765 | SPS:  2545 | Reward:  0.046 | Value:  0.725 | VLoss: 0.0117 | PLoss: -0.0072 | Ent: 0.2561
+[ 73.3%] Iter 7157/9765 | SPS:  2545 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0283 | PLoss: -0.0062 | Ent: 0.2329
+[ 73.3%] Iter 7158/9765 | SPS:  2545 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0151 | PLoss: -0.0031 | Ent: 0.2325
+[ 73.3%] Iter 7159/9765 | SPS:  2545 | Reward:  0.060 | Value:  0.753 | VLoss: 0.0243 | PLoss: -0.0070 | Ent: 0.2238
+[ 73.3%] Iter 7160/9765 | SPS:  2545 | Reward:  0.076 | Value:  0.731 | VLoss: 0.0210 | PLoss: -0.0086 | Ent: 0.2463
+[ 73.3%] Iter 7161/9765 | SPS:  2545 | Reward:  0.053 | Value:  0.813 | VLoss: 0.0175 | PLoss: -0.0017 | Ent: 0.1354
+[ 73.3%] Iter 7162/9765 | SPS:  2545 | Reward:  0.054 | Value:  0.754 | VLoss: 0.0181 | PLoss: -0.0034 | Ent: 0.1835
+[ 73.4%] Iter 7163/9765 | SPS:  2545 | Reward:  0.068 | Value:  0.690 | VLoss: 0.0179 | PLoss: -0.0055 | Ent: 0.2089
+[ 73.4%] Iter 7164/9765 | SPS:  2546 | Reward:  0.071 | Value:  0.724 | VLoss: 0.0255 | PLoss: -0.0017 | Ent: 0.2345
+[ 73.4%] Iter 7165/9765 | SPS:  2546 | Reward:  0.058 | Value:  0.752 | VLoss: 0.0309 | PLoss: -0.0018 | Ent: 0.2180
+[ 73.4%] Iter 7166/9765 | SPS:  2546 | Reward:  0.054 | Value:  0.680 | VLoss: 0.0277 | PLoss: -0.0076 | Ent: 0.1757
+[ 73.4%] Iter 7167/9765 | SPS:  2546 | Reward:  0.062 | Value:  0.755 | VLoss: 0.0211 | PLoss: -0.0113 | Ent: 0.1955
+[ 73.4%] Iter 7168/9765 | SPS:  2546 | Reward:  0.069 | Value:  0.751 | VLoss: 0.0301 | PLoss: -0.0007 | Ent: 0.2328
+[ 73.4%] Iter 7169/9765 | SPS:  2546 | Reward:  0.052 | Value:  0.746 | VLoss: 0.0235 | PLoss: -0.0085 | Ent: 0.1920
+[ 73.4%] Iter 7170/9765 | SPS:  2546 | Reward:  0.060 | Value:  0.654 | VLoss: 0.0290 | PLoss: -0.0127 | Ent: 0.2278
+[ 73.4%] Iter 7171/9765 | SPS:  2546 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0335 | PLoss: -0.0082 | Ent: 0.2091
+[ 73.4%] Iter 7172/9765 | SPS:  2547 | Reward:  0.052 | Value:  0.754 | VLoss: 0.0247 | PLoss: -0.0087 | Ent: 0.1788
+[ 73.5%] Iter 7173/9765 | SPS:  2547 | Reward:  0.062 | Value:  0.714 | VLoss: 0.0216 | PLoss: -0.0073 | Ent: 0.1779
+[ 73.5%] Iter 7174/9765 | SPS:  2547 | Reward:  0.061 | Value:  0.737 | VLoss: 0.0229 | PLoss: -0.0009 | Ent: 0.2219
+[ 73.5%] Iter 7175/9765 | SPS:  2547 | Reward:  0.061 | Value:  0.673 | VLoss: 0.0195 | PLoss: -0.0024 | Ent: 0.2286
+[ 73.5%] Iter 7176/9765 | SPS:  2547 | Reward:  0.054 | Value:  0.742 | VLoss: 0.0161 | PLoss: -0.0023 | Ent: 0.1983
+[ 73.5%] Iter 7177/9765 | SPS:  2547 | Reward:  0.063 | Value:  0.663 | VLoss: 0.0253 | PLoss: -0.0100 | Ent: 0.2093
+[ 73.5%] Iter 7178/9765 | SPS:  2547 | Reward:  0.067 | Value:  0.710 | VLoss: 0.0296 | PLoss: -0.0072 | Ent: 0.2151
+[ 73.5%] Iter 7179/9765 | SPS:  2547 | Reward:  0.060 | Value:  0.724 | VLoss: 0.0172 | PLoss: -0.0088 | Ent: 0.1958
+[ 73.5%] Iter 7180/9765 | SPS:  2547 | Reward:  0.056 | Value:  0.751 | VLoss: 0.0163 | PLoss: -0.0081 | Ent: 0.2234
+[ 73.5%] Iter 7181/9765 | SPS:  2548 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0167 | PLoss: -0.0042 | Ent: 0.2594
+[ 73.5%] Iter 7182/9765 | SPS:  2548 | Reward:  0.056 | Value:  0.781 | VLoss: 0.0215 | PLoss: -0.0065 | Ent: 0.2849
+[ 73.6%] Iter 7183/9765 | SPS:  2548 | Reward:  0.057 | Value:  0.681 | VLoss: 0.0299 | PLoss: -0.0023 | Ent: 0.2287
+[ 73.6%] Iter 7184/9765 | SPS:  2548 | Reward:  0.058 | Value:  0.646 | VLoss: 0.0217 | PLoss: -0.0110 | Ent: 0.2307
+[ 73.6%] Iter 7185/9765 | SPS:  2548 | Reward:  0.055 | Value:  0.690 | VLoss: 0.0257 | PLoss: -0.0070 | Ent: 0.2529
+[ 73.6%] Iter 7186/9765 | SPS:  2548 | Reward:  0.065 | Value:  0.708 | VLoss: 0.0231 | PLoss: -0.0078 | Ent: 0.2398
+[ 73.6%] Iter 7187/9765 | SPS:  2548 | Reward:  0.070 | Value:  0.775 | VLoss: 0.0305 | PLoss: -0.0139 | Ent: 0.2574
+[ 73.6%] Iter 7188/9765 | SPS:  2548 | Reward:  0.058 | Value:  0.730 | VLoss: 0.0247 | PLoss: -0.0039 | Ent: 0.2011
+[ 73.6%] Iter 7189/9765 | SPS:  2548 | Reward:  0.049 | Value:  0.711 | VLoss: 0.0161 | PLoss: -0.0066 | Ent: 0.2191
+[ 73.6%] Iter 7190/9765 | SPS:  2549 | Reward:  0.050 | Value:  0.690 | VLoss: 0.0161 | PLoss: -0.0032 | Ent: 0.2847
+[ 73.6%] Iter 7191/9765 | SPS:  2549 | Reward:  0.057 | Value:  0.667 | VLoss: 0.0194 | PLoss: -0.0030 | Ent: 0.2003
+[ 73.7%] Iter 7192/9765 | SPS:  2549 | Reward:  0.046 | Value:  0.656 | VLoss: 0.0249 | PLoss: -0.0064 | Ent: 0.2099
+[ 73.7%] Iter 7193/9765 | SPS:  2549 | Reward:  0.067 | Value:  0.691 | VLoss: 0.0280 | PLoss: -0.0078 | Ent: 0.2276
+[ 73.7%] Iter 7194/9765 | SPS:  2549 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0240 | PLoss: -0.0071 | Ent: 0.2275
+[ 73.7%] Iter 7195/9765 | SPS:  2549 | Reward:  0.058 | Value:  0.743 | VLoss: 0.0198 | PLoss: -0.0112 | Ent: 0.1957
+[ 73.7%] Iter 7196/9765 | SPS:  2549 | Reward:  0.060 | Value:  0.757 | VLoss: 0.0211 | PLoss: -0.0055 | Ent: 0.2291
+[ 73.7%] Iter 7197/9765 | SPS:  2549 | Reward:  0.049 | Value:  0.709 | VLoss: 0.0188 | PLoss: -0.0027 | Ent: 0.2404
+[ 73.7%] Iter 7198/9765 | SPS:  2550 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0180 | PLoss: -0.0049 | Ent: 0.2370
+[ 73.7%] Iter 7199/9765 | SPS:  2550 | Reward:  0.047 | Value:  0.628 | VLoss: 0.0183 | PLoss: -0.0073 | Ent: 0.2470
+[ 73.7%] Iter 7200/9765 | SPS:  2550 | Reward:  0.059 | Value:  0.723 | VLoss: 0.0182 | PLoss: -0.0056 | Ent: 0.2117
+[ 73.7%] Iter 7201/9765 | SPS:  2550 | Reward:  0.069 | Value:  0.745 | VLoss: 0.0282 | PLoss: -0.0094 | Ent: 0.1843
+[ 73.8%] Iter 7202/9765 | SPS:  2550 | Reward:  0.067 | Value:  0.700 | VLoss: 0.0222 | PLoss: -0.0100 | Ent: 0.2315
+[ 73.8%] Iter 7203/9765 | SPS:  2550 | Reward:  0.053 | Value:  0.741 | VLoss: 0.0220 | PLoss: -0.0077 | Ent: 0.1697
+[ 73.8%] Iter 7204/9765 | SPS:  2550 | Reward:  0.066 | Value:  0.740 | VLoss: 0.0170 | PLoss: -0.0040 | Ent: 0.2113
+[ 73.8%] Iter 7205/9765 | SPS:  2550 | Reward:  0.061 | Value:  0.751 | VLoss: 0.0197 | PLoss: -0.0109 | Ent: 0.2054
+[ 73.8%] Iter 7206/9765 | SPS:  2550 | Reward:  0.059 | Value:  0.709 | VLoss: 0.0110 | PLoss: -0.0076 | Ent: 0.2365
+[ 73.8%] Iter 7207/9765 | SPS:  2551 | Reward:  0.065 | Value:  0.703 | VLoss: 0.0252 | PLoss: -0.0069 | Ent: 0.2563
+[ 73.8%] Iter 7208/9765 | SPS:  2551 | Reward:  0.065 | Value:  0.746 | VLoss: 0.0276 | PLoss: -0.0098 | Ent: 0.2491
+[ 73.8%] Iter 7209/9765 | SPS:  2551 | Reward:  0.056 | Value:  0.727 | VLoss: 0.0168 | PLoss: -0.0036 | Ent: 0.2312
+[ 73.8%] Iter 7210/9765 | SPS:  2551 | Reward:  0.051 | Value:  0.724 | VLoss: 0.0173 | PLoss: -0.0049 | Ent: 0.2247
+[ 73.8%] Iter 7211/9765 | SPS:  2551 | Reward:  0.065 | Value:  0.724 | VLoss: 0.0255 | PLoss: -0.0018 | Ent: 0.1824
+[ 73.9%] Iter 7212/9765 | SPS:  2551 | Reward:  0.052 | Value:  0.663 | VLoss: 0.0248 | PLoss: -0.0124 | Ent: 0.2004
+[ 73.9%] Iter 7213/9765 | SPS:  2551 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0183 | PLoss: -0.0036 | Ent: 0.1857
+[ 73.9%] Iter 7214/9765 | SPS:  2551 | Reward:  0.057 | Value:  0.606 | VLoss: 0.0340 | PLoss: -0.0009 | Ent: 0.2446
+[ 73.9%] Iter 7215/9765 | SPS:  2551 | Reward:  0.066 | Value:  0.751 | VLoss: 0.0227 | PLoss: -0.0067 | Ent: 0.2261
+[ 73.9%] Iter 7216/9765 | SPS:  2552 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0200 | PLoss: -0.0094 | Ent: 0.1935
+[ 73.9%] Iter 7217/9765 | SPS:  2552 | Reward:  0.072 | Value:  0.782 | VLoss: 0.0209 | PLoss: -0.0058 | Ent: 0.2391
+[ 73.9%] Iter 7218/9765 | SPS:  2552 | Reward:  0.062 | Value:  0.809 | VLoss: 0.0200 | PLoss: -0.0037 | Ent: 0.2175
+[ 73.9%] Iter 7219/9765 | SPS:  2552 | Reward:  0.070 | Value:  0.777 | VLoss: 0.0226 | PLoss: -0.0036 | Ent: 0.1965
+[ 73.9%] Iter 7220/9765 | SPS:  2552 | Reward:  0.061 | Value:  0.691 | VLoss: 0.0188 | PLoss: -0.0039 | Ent: 0.1973
+[ 73.9%] Iter 7221/9765 | SPS:  2552 | Reward:  0.047 | Value:  0.678 | VLoss: 0.0305 | PLoss: -0.0079 | Ent: 0.2285
+[ 74.0%] Iter 7222/9765 | SPS:  2552 | Reward:  0.069 | Value:  0.619 | VLoss: 0.0238 | PLoss: -0.0044 | Ent: 0.2344
+[ 74.0%] Iter 7223/9765 | SPS:  2552 | Reward:  0.060 | Value:  0.721 | VLoss: 0.0293 | PLoss: -0.0094 | Ent: 0.2404
+[ 74.0%] Iter 7224/9765 | SPS:  2553 | Reward:  0.053 | Value:  0.710 | VLoss: 0.0220 | PLoss: -0.0088 | Ent: 0.2463
+[ 74.0%] Iter 7225/9765 | SPS:  2553 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0305 | PLoss: -0.0063 | Ent: 0.2443
+[ 74.0%] Iter 7226/9765 | SPS:  2553 | Reward:  0.056 | Value:  0.746 | VLoss: 0.0244 | PLoss: -0.0025 | Ent: 0.2245
+[ 74.0%] Iter 7227/9765 | SPS:  2553 | Reward:  0.055 | Value:  0.702 | VLoss: 0.0151 | PLoss: -0.0030 | Ent: 0.2173
+[ 74.0%] Iter 7228/9765 | SPS:  2553 | Reward:  0.068 | Value:  0.738 | VLoss: 0.0185 | PLoss: -0.0034 | Ent: 0.2257
+[ 74.0%] Iter 7229/9765 | SPS:  2553 | Reward:  0.065 | Value:  0.685 | VLoss: 0.0259 | PLoss: -0.0071 | Ent: 0.1918
+[ 74.0%] Iter 7230/9765 | SPS:  2553 | Reward:  0.054 | Value:  0.682 | VLoss: 0.0258 | PLoss: -0.0057 | Ent: 0.2306
+[ 74.1%] Iter 7231/9765 | SPS:  2553 | Reward:  0.052 | Value:  0.700 | VLoss: 0.0186 | PLoss: -0.0065 | Ent: 0.2459
+[ 74.1%] Iter 7232/9765 | SPS:  2553 | Reward:  0.063 | Value:  0.776 | VLoss: 0.0243 | PLoss: -0.0095 | Ent: 0.1889
+[ 74.1%] Iter 7233/9765 | SPS:  2554 | Reward:  0.079 | Value:  0.743 | VLoss: 0.0234 | PLoss: -0.0112 | Ent: 0.2181
+[ 74.1%] Iter 7234/9765 | SPS:  2554 | Reward:  0.057 | Value:  0.751 | VLoss: 0.0168 | PLoss: -0.0007 | Ent: 0.1906
+[ 74.1%] Iter 7235/9765 | SPS:  2554 | Reward:  0.054 | Value:  0.680 | VLoss: 0.0245 | PLoss: -0.0082 | Ent: 0.2455
+[ 74.1%] Iter 7236/9765 | SPS:  2554 | Reward:  0.054 | Value:  0.691 | VLoss: 0.0187 | PLoss: -0.0076 | Ent: 0.2491
+[ 74.1%] Iter 7237/9765 | SPS:  2554 | Reward:  0.059 | Value:  0.749 | VLoss: 0.0237 | PLoss: -0.0056 | Ent: 0.2365
+[ 74.1%] Iter 7238/9765 | SPS:  2554 | Reward:  0.059 | Value:  0.764 | VLoss: 0.0255 | PLoss: -0.0051 | Ent: 0.2131
+[ 74.1%] Iter 7239/9765 | SPS:  2554 | Reward:  0.064 | Value:  0.746 | VLoss: 0.0172 | PLoss: -0.0083 | Ent: 0.1807
+[ 74.1%] Iter 7240/9765 | SPS:  2554 | Reward:  0.057 | Value:  0.740 | VLoss: 0.0233 | PLoss: -0.0169 | Ent: 0.1950
+[ 74.2%] Iter 7241/9765 | SPS:  2555 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0258 | PLoss: -0.0062 | Ent: 0.2377
+[ 74.2%] Iter 7242/9765 | SPS:  2555 | Reward:  0.073 | Value:  0.673 | VLoss: 0.0257 | PLoss: -0.0041 | Ent: 0.2125
+[ 74.2%] Iter 7243/9765 | SPS:  2555 | Reward:  0.079 | Value:  0.756 | VLoss: 0.0259 | PLoss: -0.0051 | Ent: 0.2566
+[ 74.2%] Iter 7244/9765 | SPS:  2555 | Reward:  0.071 | Value:  0.842 | VLoss: 0.0155 | PLoss: -0.0062 | Ent: 0.2008
+[ 74.2%] Iter 7245/9765 | SPS:  2555 | Reward:  0.057 | Value:  0.760 | VLoss: 0.0245 | PLoss: -0.0047 | Ent: 0.2034
+[ 74.2%] Iter 7246/9765 | SPS:  2555 | Reward:  0.067 | Value:  0.680 | VLoss: 0.0135 | PLoss: -0.0057 | Ent: 0.1690
+[ 74.2%] Iter 7247/9765 | SPS:  2555 | Reward:  0.071 | Value:  0.754 | VLoss: 0.0252 | PLoss: -0.0015 | Ent: 0.2035
+[ 74.2%] Iter 7248/9765 | SPS:  2555 | Reward:  0.044 | Value:  0.688 | VLoss: 0.0324 | PLoss: -0.0109 | Ent: 0.1981
+[ 74.2%] Iter 7249/9765 | SPS:  2555 | Reward:  0.071 | Value:  0.621 | VLoss: 0.0216 | PLoss: -0.0072 | Ent: 0.2146
+[ 74.2%] Iter 7250/9765 | SPS:  2556 | Reward:  0.069 | Value:  0.717 | VLoss: 0.0249 | PLoss: -0.0126 | Ent: 0.2048
+[ 74.3%] Iter 7251/9765 | SPS:  2556 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0221 | PLoss: -0.0062 | Ent: 0.1920
+[ 74.3%] Iter 7252/9765 | SPS:  2556 | Reward:  0.060 | Value:  0.721 | VLoss: 0.0239 | PLoss: -0.0102 | Ent: 0.1851
+[ 74.3%] Iter 7253/9765 | SPS:  2556 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0219 | PLoss: -0.0063 | Ent: 0.2652
+[ 74.3%] Iter 7254/9765 | SPS:  2556 | Reward:  0.070 | Value:  0.804 | VLoss: 0.0154 | PLoss: -0.0060 | Ent: 0.1865
+[ 74.3%] Iter 7255/9765 | SPS:  2556 | Reward:  0.061 | Value:  0.714 | VLoss: 0.0172 | PLoss: -0.0039 | Ent: 0.2131
+[ 74.3%] Iter 7256/9765 | SPS:  2556 | Reward:  0.059 | Value:  0.762 | VLoss: 0.0135 | PLoss: -0.0133 | Ent: 0.2198
+[ 74.3%] Iter 7257/9765 | SPS:  2556 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0254 | PLoss: -0.0066 | Ent: 0.2356
+[ 74.3%] Iter 7258/9765 | SPS:  2556 | Reward:  0.045 | Value:  0.688 | VLoss: 0.0200 | PLoss: -0.0127 | Ent: 0.2267
+[ 74.3%] Iter 7259/9765 | SPS:  2557 | Reward:  0.073 | Value:  0.633 | VLoss: 0.0315 | PLoss: -0.0118 | Ent: 0.2293
+[ 74.3%] Iter 7260/9765 | SPS:  2557 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0172 | PLoss: -0.0114 | Ent: 0.2285
+[ 74.4%] Iter 7261/9765 | SPS:  2557 | Reward:  0.056 | Value:  0.740 | VLoss: 0.0179 | PLoss: -0.0033 | Ent: 0.1905
+[ 74.4%] Iter 7262/9765 | SPS:  2557 | Reward:  0.072 | Value:  0.756 | VLoss: 0.0179 | PLoss: -0.0092 | Ent: 0.1770
+[ 74.4%] Iter 7263/9765 | SPS:  2557 | Reward:  0.051 | Value:  0.703 | VLoss: 0.0201 | PLoss: -0.0100 | Ent: 0.2229
+[ 74.4%] Iter 7264/9765 | SPS:  2557 | Reward:  0.074 | Value:  0.717 | VLoss: 0.0251 | PLoss: -0.0023 | Ent: 0.2089
+[ 74.4%] Iter 7265/9765 | SPS:  2557 | Reward:  0.064 | Value:  0.699 | VLoss: 0.0196 | PLoss: -0.0077 | Ent: 0.2728
+[ 74.4%] Iter 7266/9765 | SPS:  2557 | Reward:  0.066 | Value:  0.706 | VLoss: 0.0223 | PLoss: -0.0045 | Ent: 0.2103
+[ 74.4%] Iter 7267/9765 | SPS:  2557 | Reward:  0.056 | Value:  0.717 | VLoss: 0.0221 | PLoss: -0.0075 | Ent: 0.1743
+[ 74.4%] Iter 7268/9765 | SPS:  2558 | Reward:  0.056 | Value:  0.748 | VLoss: 0.0141 | PLoss: -0.0036 | Ent: 0.1933
+[ 74.4%] Iter 7269/9765 | SPS:  2558 | Reward:  0.066 | Value:  0.740 | VLoss: 0.0220 | PLoss: -0.0031 | Ent: 0.2247
+[ 74.4%] Iter 7270/9765 | SPS:  2558 | Reward:  0.066 | Value:  0.744 | VLoss: 0.0174 | PLoss: -0.0065 | Ent: 0.2290
+[ 74.5%] Iter 7271/9765 | SPS:  2558 | Reward:  0.085 | Value:  0.802 | VLoss: 0.0162 | PLoss: -0.0027 | Ent: 0.1982
+[ 74.5%] Iter 7272/9765 | SPS:  2558 | Reward:  0.061 | Value:  0.765 | VLoss: 0.0187 | PLoss: -0.0031 | Ent: 0.1967
+[ 74.5%] Iter 7273/9765 | SPS:  2558 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0155 | PLoss: -0.0051 | Ent: 0.2513
+[ 74.5%] Iter 7274/9765 | SPS:  2558 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0219 | PLoss: -0.0015 | Ent: 0.1939
+[ 74.5%] Iter 7275/9765 | SPS:  2558 | Reward:  0.051 | Value:  0.701 | VLoss: 0.0250 | PLoss: -0.0097 | Ent: 0.2202
+[ 74.5%] Iter 7276/9765 | SPS:  2558 | Reward:  0.060 | Value:  0.621 | VLoss: 0.0257 | PLoss: -0.0062 | Ent: 0.2380
+[ 74.5%] Iter 7277/9765 | SPS:  2559 | Reward:  0.048 | Value:  0.740 | VLoss: 0.0219 | PLoss: -0.0090 | Ent: 0.1891
+[ 74.5%] Iter 7278/9765 | SPS:  2559 | Reward:  0.049 | Value:  0.654 | VLoss: 0.0192 | PLoss: -0.0034 | Ent: 0.2325
+[ 74.5%] Iter 7279/9765 | SPS:  2559 | Reward:  0.059 | Value:  0.631 | VLoss: 0.0273 | PLoss: -0.0045 | Ent: 0.2338
+[ 74.6%] Iter 7280/9765 | SPS:  2559 | Reward:  0.060 | Value:  0.746 | VLoss: 0.0194 | PLoss: -0.0054 | Ent: 0.2603
+[ 74.6%] Iter 7281/9765 | SPS:  2559 | Reward:  0.078 | Value:  0.774 | VLoss: 0.0314 | PLoss: -0.0071 | Ent: 0.2262
+[ 74.6%] Iter 7282/9765 | SPS:  2559 | Reward:  0.053 | Value:  0.714 | VLoss: 0.0227 | PLoss: -0.0086 | Ent: 0.2429
+[ 74.6%] Iter 7283/9765 | SPS:  2559 | Reward:  0.064 | Value:  0.762 | VLoss: 0.0280 | PLoss: -0.0115 | Ent: 0.2597
+[ 74.6%] Iter 7284/9765 | SPS:  2559 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0234 | PLoss: -0.0057 | Ent: 0.1997
+[ 74.6%] Iter 7285/9765 | SPS:  2559 | Reward:  0.059 | Value:  0.718 | VLoss: 0.0244 | PLoss: -0.0079 | Ent: 0.2130
+[ 74.6%] Iter 7286/9765 | SPS:  2560 | Reward:  0.071 | Value:  0.697 | VLoss: 0.0184 | PLoss: -0.0073 | Ent: 0.1800
+[ 74.6%] Iter 7287/9765 | SPS:  2560 | Reward:  0.060 | Value:  0.732 | VLoss: 0.0243 | PLoss: -0.0074 | Ent: 0.2253
+[ 74.6%] Iter 7288/9765 | SPS:  2560 | Reward:  0.056 | Value:  0.704 | VLoss: 0.0198 | PLoss: -0.0073 | Ent: 0.2600
+[ 74.6%] Iter 7289/9765 | SPS:  2560 | Reward:  0.057 | Value:  0.683 | VLoss: 0.0222 | PLoss: -0.0045 | Ent: 0.2096
+[ 74.7%] Iter 7290/9765 | SPS:  2560 | Reward:  0.055 | Value:  0.764 | VLoss: 0.0193 | PLoss: -0.0057 | Ent: 0.1765
+[ 74.7%] Iter 7291/9765 | SPS:  2560 | Reward:  0.066 | Value:  0.744 | VLoss: 0.0228 | PLoss: -0.0065 | Ent: 0.2255
+[ 74.7%] Iter 7292/9765 | SPS:  2560 | Reward:  0.064 | Value:  0.696 | VLoss: 0.0200 | PLoss: -0.0040 | Ent: 0.1889
+[ 74.7%] Iter 7293/9765 | SPS:  2560 | Reward:  0.068 | Value:  0.768 | VLoss: 0.0198 | PLoss: -0.0052 | Ent: 0.2475
+[ 74.7%] Iter 7294/9765 | SPS:  2560 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0270 | PLoss: -0.0068 | Ent: 0.1868
+[ 74.7%] Iter 7295/9765 | SPS:  2561 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0190 | PLoss: -0.0025 | Ent: 0.2204
+[ 74.7%] Iter 7296/9765 | SPS:  2561 | Reward:  0.051 | Value:  0.717 | VLoss: 0.0249 | PLoss: -0.0020 | Ent: 0.2491
+[ 74.7%] Iter 7297/9765 | SPS:  2561 | Reward:  0.065 | Value:  0.788 | VLoss: 0.0290 | PLoss: -0.0090 | Ent: 0.1994
+[ 74.7%] Iter 7298/9765 | SPS:  2561 | Reward:  0.056 | Value:  0.720 | VLoss: 0.0229 | PLoss: -0.0028 | Ent: 0.1680
+[ 74.7%] Iter 7299/9765 | SPS:  2561 | Reward:  0.050 | Value:  0.705 | VLoss: 0.0234 | PLoss: -0.0085 | Ent: 0.2387
+[ 74.8%] Iter 7300/9765 | SPS:  2561 | Reward:  0.063 | Value:  0.762 | VLoss: 0.0171 | PLoss: -0.0176 | Ent: 0.2165
+[ 74.8%] Iter 7301/9765 | SPS:  2561 | Reward:  0.052 | Value:  0.760 | VLoss: 0.0198 | PLoss: -0.0049 | Ent: 0.2126
+[ 74.8%] Iter 7302/9765 | SPS:  2561 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0229 | PLoss: -0.0078 | Ent: 0.1861
+[ 74.8%] Iter 7303/9765 | SPS:  2561 | Reward:  0.060 | Value:  0.690 | VLoss: 0.0221 | PLoss: -0.0047 | Ent: 0.2265
+[ 74.8%] Iter 7304/9765 | SPS:  2562 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0163 | PLoss: -0.0047 | Ent: 0.2478
+[ 74.8%] Iter 7305/9765 | SPS:  2562 | Reward:  0.069 | Value:  0.733 | VLoss: 0.0186 | PLoss: -0.0015 | Ent: 0.2101
+[ 74.8%] Iter 7306/9765 | SPS:  2562 | Reward:  0.053 | Value:  0.666 | VLoss: 0.0324 | PLoss: -0.0063 | Ent: 0.2556
+[ 74.8%] Iter 7307/9765 | SPS:  2562 | Reward:  0.061 | Value:  0.722 | VLoss: 0.0237 | PLoss: -0.0078 | Ent: 0.2217
+[ 74.8%] Iter 7308/9765 | SPS:  2562 | Reward:  0.073 | Value:  0.733 | VLoss: 0.0204 | PLoss: -0.0130 | Ent: 0.2067
+[ 74.8%] Iter 7309/9765 | SPS:  2562 | Reward:  0.072 | Value:  0.739 | VLoss: 0.0257 | PLoss: -0.0116 | Ent: 0.1817
+[ 74.9%] Iter 7310/9765 | SPS:  2562 | Reward:  0.079 | Value:  0.726 | VLoss: 0.0260 | PLoss: -0.0083 | Ent: 0.1628
+[ 74.9%] Iter 7311/9765 | SPS:  2562 | Reward:  0.061 | Value:  0.692 | VLoss: 0.0272 | PLoss: -0.0083 | Ent: 0.1942
+[ 74.9%] Iter 7312/9765 | SPS:  2562 | Reward:  0.062 | Value:  0.688 | VLoss: 0.0174 | PLoss: -0.0078 | Ent: 0.2210
+[ 74.9%] Iter 7313/9765 | SPS:  2563 | Reward:  0.055 | Value:  0.714 | VLoss: 0.0208 | PLoss: -0.0099 | Ent: 0.2294
+[ 74.9%] Iter 7314/9765 | SPS:  2563 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0282 | PLoss: -0.0074 | Ent: 0.2145
+[ 74.9%] Iter 7315/9765 | SPS:  2563 | Reward:  0.068 | Value:  0.680 | VLoss: 0.0276 | PLoss: -0.0007 | Ent: 0.2648
+[ 74.9%] Iter 7316/9765 | SPS:  2563 | Reward:  0.050 | Value:  0.714 | VLoss: 0.0152 | PLoss: -0.0089 | Ent: 0.2386
+[ 74.9%] Iter 7317/9765 | SPS:  2563 | Reward:  0.078 | Value:  0.753 | VLoss: 0.0242 | PLoss: 0.0013 | Ent: 0.1867
+[ 74.9%] Iter 7318/9765 | SPS:  2563 | Reward:  0.074 | Value:  0.718 | VLoss: 0.0248 | PLoss: -0.0134 | Ent: 0.2342
+[ 75.0%] Iter 7319/9765 | SPS:  2563 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0316 | PLoss: -0.0030 | Ent: 0.2079
+[ 75.0%] Iter 7320/9765 | SPS:  2563 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0289 | PLoss: -0.0118 | Ent: 0.2168
+[ 75.0%] Iter 7321/9765 | SPS:  2563 | Reward:  0.058 | Value:  0.661 | VLoss: 0.0199 | PLoss: -0.0025 | Ent: 0.1906
+[ 75.0%] Iter 7322/9765 | SPS:  2564 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0184 | PLoss: -0.0046 | Ent: 0.2120
+[ 75.0%] Iter 7323/9765 | SPS:  2564 | Reward:  0.071 | Value:  0.722 | VLoss: 0.0229 | PLoss: -0.0081 | Ent: 0.2017
+[ 75.0%] Iter 7324/9765 | SPS:  2564 | Reward:  0.050 | Value:  0.709 | VLoss: 0.0183 | PLoss: -0.0106 | Ent: 0.2167
+[ 75.0%] Iter 7325/9765 | SPS:  2564 | Reward:  0.066 | Value:  0.735 | VLoss: 0.0222 | PLoss: -0.0059 | Ent: 0.2306
+[ 75.0%] Iter 7326/9765 | SPS:  2564 | Reward:  0.059 | Value:  0.752 | VLoss: 0.0209 | PLoss: -0.0082 | Ent: 0.2030
+[ 75.0%] Iter 7327/9765 | SPS:  2564 | Reward:  0.064 | Value:  0.701 | VLoss: 0.0238 | PLoss: -0.0003 | Ent: 0.1868
+[ 75.0%] Iter 7328/9765 | SPS:  2564 | Reward:  0.057 | Value:  0.673 | VLoss: 0.0253 | PLoss: -0.0035 | Ent: 0.2040
+[ 75.1%] Iter 7329/9765 | SPS:  2564 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0197 | PLoss: 0.0012 | Ent: 0.2468
+[ 75.1%] Iter 7330/9765 | SPS:  2564 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0335 | PLoss: -0.0063 | Ent: 0.2084
+[ 75.1%] Iter 7331/9765 | SPS:  2565 | Reward:  0.058 | Value:  0.643 | VLoss: 0.0234 | PLoss: -0.0039 | Ent: 0.1907
+[ 75.1%] Iter 7332/9765 | SPS:  2565 | Reward:  0.057 | Value:  0.694 | VLoss: 0.0234 | PLoss: -0.0134 | Ent: 0.2803
+[ 75.1%] Iter 7333/9765 | SPS:  2565 | Reward:  0.066 | Value:  0.711 | VLoss: 0.0188 | PLoss: -0.0044 | Ent: 0.2353
+[ 75.1%] Iter 7334/9765 | SPS:  2565 | Reward:  0.062 | Value:  0.689 | VLoss: 0.0213 | PLoss: -0.0087 | Ent: 0.1922
+[ 75.1%] Iter 7335/9765 | SPS:  2565 | Reward:  0.074 | Value:  0.781 | VLoss: 0.0168 | PLoss: -0.0101 | Ent: 0.2376
+[ 75.1%] Iter 7336/9765 | SPS:  2565 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0290 | PLoss: -0.0034 | Ent: 0.2237
+[ 75.1%] Iter 7337/9765 | SPS:  2565 | Reward:  0.065 | Value:  0.768 | VLoss: 0.0248 | PLoss: -0.0093 | Ent: 0.2317
+[ 75.1%] Iter 7338/9765 | SPS:  2565 | Reward:  0.071 | Value:  0.672 | VLoss: 0.0202 | PLoss: -0.0059 | Ent: 0.2108
+[ 75.2%] Iter 7339/9765 | SPS:  2565 | Reward:  0.053 | Value:  0.734 | VLoss: 0.0219 | PLoss: -0.0058 | Ent: 0.1832
+[ 75.2%] Iter 7340/9765 | SPS:  2566 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0275 | PLoss: -0.0082 | Ent: 0.2280
+[ 75.2%] Iter 7341/9765 | SPS:  2566 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0155 | PLoss: -0.0038 | Ent: 0.2145
+[ 75.2%] Iter 7342/9765 | SPS:  2566 | Reward:  0.051 | Value:  0.765 | VLoss: 0.0233 | PLoss: -0.0091 | Ent: 0.2460
+[ 75.2%] Iter 7343/9765 | SPS:  2566 | Reward:  0.063 | Value:  0.684 | VLoss: 0.0347 | PLoss: -0.0042 | Ent: 0.1887
+[ 75.2%] Iter 7344/9765 | SPS:  2566 | Reward:  0.066 | Value:  0.666 | VLoss: 0.0228 | PLoss: -0.0054 | Ent: 0.2240
+[ 75.2%] Iter 7345/9765 | SPS:  2566 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0272 | PLoss: -0.0058 | Ent: 0.2065
+[ 75.2%] Iter 7346/9765 | SPS:  2566 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0242 | PLoss: -0.0111 | Ent: 0.2702
+[ 75.2%] Iter 7347/9765 | SPS:  2566 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0249 | PLoss: -0.0063 | Ent: 0.2424
+[ 75.2%] Iter 7348/9765 | SPS:  2566 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0147 | PLoss: -0.0071 | Ent: 0.1974
+[ 75.3%] Iter 7349/9765 | SPS:  2567 | Reward:  0.061 | Value:  0.705 | VLoss: 0.0244 | PLoss: -0.0047 | Ent: 0.2023
+[ 75.3%] Iter 7350/9765 | SPS:  2567 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0224 | PLoss: -0.0054 | Ent: 0.2453
+[ 75.3%] Iter 7351/9765 | SPS:  2567 | Reward:  0.040 | Value:  0.643 | VLoss: 0.0213 | PLoss: -0.0104 | Ent: 0.2751
+[ 75.3%] Iter 7352/9765 | SPS:  2567 | Reward:  0.060 | Value:  0.674 | VLoss: 0.0212 | PLoss: -0.0021 | Ent: 0.1913
+[ 75.3%] Iter 7353/9765 | SPS:  2567 | Reward:  0.062 | Value:  0.733 | VLoss: 0.0195 | PLoss: -0.0083 | Ent: 0.2308
+[ 75.3%] Iter 7354/9765 | SPS:  2567 | Reward:  0.057 | Value:  0.709 | VLoss: 0.0302 | PLoss: -0.0049 | Ent: 0.2143
+[ 75.3%] Iter 7355/9765 | SPS:  2567 | Reward:  0.060 | Value:  0.661 | VLoss: 0.0214 | PLoss: -0.0041 | Ent: 0.1810
+[ 75.3%] Iter 7356/9765 | SPS:  2567 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0244 | PLoss: -0.0063 | Ent: 0.2071
+[ 75.3%] Iter 7357/9765 | SPS:  2567 | Reward:  0.062 | Value:  0.659 | VLoss: 0.0202 | PLoss: -0.0094 | Ent: 0.2420
+[ 75.4%] Iter 7358/9765 | SPS:  2568 | Reward:  0.054 | Value:  0.814 | VLoss: 0.0138 | PLoss: -0.0037 | Ent: 0.1836
+[ 75.4%] Iter 7359/9765 | SPS:  2568 | Reward:  0.055 | Value:  0.793 | VLoss: 0.0169 | PLoss: -0.0051 | Ent: 0.1947
+[ 75.4%] Iter 7360/9765 | SPS:  2568 | Reward:  0.056 | Value:  0.777 | VLoss: 0.0297 | PLoss: 0.0005 | Ent: 0.2522
+[ 75.4%] Iter 7361/9765 | SPS:  2568 | Reward:  0.077 | Value:  0.699 | VLoss: 0.0241 | PLoss: -0.0056 | Ent: 0.2504
+[ 75.4%] Iter 7362/9765 | SPS:  2568 | Reward:  0.068 | Value:  0.802 | VLoss: 0.0169 | PLoss: -0.0043 | Ent: 0.1987
+[ 75.4%] Iter 7363/9765 | SPS:  2568 | Reward:  0.056 | Value:  0.730 | VLoss: 0.0202 | PLoss: -0.0087 | Ent: 0.2498
+[ 75.4%] Iter 7364/9765 | SPS:  2568 | Reward:  0.071 | Value:  0.674 | VLoss: 0.0338 | PLoss: -0.0092 | Ent: 0.1472
+[ 75.4%] Iter 7365/9765 | SPS:  2568 | Reward:  0.060 | Value:  0.777 | VLoss: 0.0145 | PLoss: -0.0073 | Ent: 0.1724
+[ 75.4%] Iter 7366/9765 | SPS:  2568 | Reward:  0.067 | Value:  0.772 | VLoss: 0.0249 | PLoss: -0.0026 | Ent: 0.2414
+[ 75.4%] Iter 7367/9765 | SPS:  2569 | Reward:  0.063 | Value:  0.679 | VLoss: 0.0284 | PLoss: -0.0026 | Ent: 0.2390
+[ 75.5%] Iter 7368/9765 | SPS:  2569 | Reward:  0.063 | Value:  0.729 | VLoss: 0.0227 | PLoss: -0.0084 | Ent: 0.2072
+[ 75.5%] Iter 7369/9765 | SPS:  2569 | Reward:  0.062 | Value:  0.772 | VLoss: 0.0252 | PLoss: -0.0107 | Ent: 0.1950
+[ 75.5%] Iter 7370/9765 | SPS:  2569 | Reward:  0.074 | Value:  0.703 | VLoss: 0.0271 | PLoss: -0.0046 | Ent: 0.2208
+[ 75.5%] Iter 7371/9765 | SPS:  2569 | Reward:  0.066 | Value:  0.727 | VLoss: 0.0289 | PLoss: -0.0053 | Ent: 0.2231
+[ 75.5%] Iter 7372/9765 | SPS:  2569 | Reward:  0.074 | Value:  0.788 | VLoss: 0.0250 | PLoss: -0.0041 | Ent: 0.2271
+[ 75.5%] Iter 7373/9765 | SPS:  2569 | Reward:  0.048 | Value:  0.773 | VLoss: 0.0187 | PLoss: -0.0053 | Ent: 0.1953
+[ 75.5%] Iter 7374/9765 | SPS:  2569 | Reward:  0.066 | Value:  0.670 | VLoss: 0.0205 | PLoss: -0.0103 | Ent: 0.2046
+[ 75.5%] Iter 7375/9765 | SPS:  2569 | Reward:  0.070 | Value:  0.722 | VLoss: 0.0219 | PLoss: -0.0069 | Ent: 0.2144
+[ 75.5%] Iter 7376/9765 | SPS:  2570 | Reward:  0.054 | Value:  0.658 | VLoss: 0.0233 | PLoss: -0.0081 | Ent: 0.2109
+[ 75.5%] Iter 7377/9765 | SPS:  2570 | Reward:  0.059 | Value:  0.716 | VLoss: 0.0210 | PLoss: -0.0082 | Ent: 0.1885
+[ 75.6%] Iter 7378/9765 | SPS:  2570 | Reward:  0.053 | Value:  0.737 | VLoss: 0.0203 | PLoss: -0.0023 | Ent: 0.1837
+[ 75.6%] Iter 7379/9765 | SPS:  2570 | Reward:  0.068 | Value:  0.708 | VLoss: 0.0265 | PLoss: -0.0049 | Ent: 0.1976
+[ 75.6%] Iter 7380/9765 | SPS:  2570 | Reward:  0.063 | Value:  0.740 | VLoss: 0.0209 | PLoss: -0.0094 | Ent: 0.2200
+[ 75.6%] Iter 7381/9765 | SPS:  2570 | Reward:  0.070 | Value:  0.700 | VLoss: 0.0230 | PLoss: -0.0099 | Ent: 0.2523
+[ 75.6%] Iter 7382/9765 | SPS:  2570 | Reward:  0.059 | Value:  0.771 | VLoss: 0.0148 | PLoss: -0.0037 | Ent: 0.2163
+[ 75.6%] Iter 7383/9765 | SPS:  2570 | Reward:  0.060 | Value:  0.717 | VLoss: 0.0130 | PLoss: -0.0051 | Ent: 0.2088
+[ 75.6%] Iter 7384/9765 | SPS:  2570 | Reward:  0.056 | Value:  0.732 | VLoss: 0.0205 | PLoss: -0.0073 | Ent: 0.2267
+[ 75.6%] Iter 7385/9765 | SPS:  2571 | Reward:  0.058 | Value:  0.728 | VLoss: 0.0269 | PLoss: -0.0074 | Ent: 0.2380
+[ 75.6%] Iter 7386/9765 | SPS:  2571 | Reward:  0.065 | Value:  0.693 | VLoss: 0.0207 | PLoss: -0.0084 | Ent: 0.2432
+[ 75.6%] Iter 7387/9765 | SPS:  2571 | Reward:  0.049 | Value:  0.683 | VLoss: 0.0190 | PLoss: -0.0032 | Ent: 0.1933
+[ 75.7%] Iter 7388/9765 | SPS:  2571 | Reward:  0.069 | Value:  0.748 | VLoss: 0.0320 | PLoss: -0.0070 | Ent: 0.2084
+[ 75.7%] Iter 7389/9765 | SPS:  2571 | Reward:  0.069 | Value:  0.622 | VLoss: 0.0331 | PLoss: -0.0146 | Ent: 0.2700
+[ 75.7%] Iter 7390/9765 | SPS:  2571 | Reward:  0.060 | Value:  0.728 | VLoss: 0.0157 | PLoss: -0.0054 | Ent: 0.1935
+[ 75.7%] Iter 7391/9765 | SPS:  2571 | Reward:  0.073 | Value:  0.653 | VLoss: 0.0279 | PLoss: -0.0134 | Ent: 0.2237
+[ 75.7%] Iter 7392/9765 | SPS:  2571 | Reward:  0.061 | Value:  0.696 | VLoss: 0.0237 | PLoss: -0.0046 | Ent: 0.2123
+[ 75.7%] Iter 7393/9765 | SPS:  2571 | Reward:  0.054 | Value:  0.744 | VLoss: 0.0216 | PLoss: -0.0075 | Ent: 0.2071
+[ 75.7%] Iter 7394/9765 | SPS:  2571 | Reward:  0.064 | Value:  0.642 | VLoss: 0.0254 | PLoss: -0.0028 | Ent: 0.2366
+[ 75.7%] Iter 7395/9765 | SPS:  2572 | Reward:  0.057 | Value:  0.698 | VLoss: 0.0221 | PLoss: -0.0053 | Ent: 0.2357
+[ 75.7%] Iter 7396/9765 | SPS:  2572 | Reward:  0.067 | Value:  0.764 | VLoss: 0.0204 | PLoss: -0.0102 | Ent: 0.2271
+[ 75.8%] Iter 7397/9765 | SPS:  2572 | Reward:  0.059 | Value:  0.753 | VLoss: 0.0164 | PLoss: -0.0072 | Ent: 0.2249
+[ 75.8%] Iter 7398/9765 | SPS:  2572 | Reward:  0.071 | Value:  0.755 | VLoss: 0.0241 | PLoss: -0.0056 | Ent: 0.2531
+[ 75.8%] Iter 7399/9765 | SPS:  2572 | Reward:  0.061 | Value:  0.671 | VLoss: 0.0195 | PLoss: -0.0050 | Ent: 0.2123
+[ 75.8%] Iter 7400/9765 | SPS:  2572 | Reward:  0.075 | Value:  0.725 | VLoss: 0.0161 | PLoss: -0.0109 | Ent: 0.2285
+[ 75.8%] Iter 7401/9765 | SPS:  2572 | Reward:  0.058 | Value:  0.769 | VLoss: 0.0160 | PLoss: 0.0002 | Ent: 0.2833
+[ 75.8%] Iter 7402/9765 | SPS:  2572 | Reward:  0.060 | Value:  0.739 | VLoss: 0.0215 | PLoss: -0.0063 | Ent: 0.2376
+[ 75.8%] Iter 7403/9765 | SPS:  2573 | Reward:  0.054 | Value:  0.747 | VLoss: 0.0176 | PLoss: -0.0071 | Ent: 0.2203
+[ 75.8%] Iter 7404/9765 | SPS:  2573 | Reward:  0.064 | Value:  0.711 | VLoss: 0.0221 | PLoss: -0.0060 | Ent: 0.1693
+[ 75.8%] Iter 7405/9765 | SPS:  2573 | Reward:  0.058 | Value:  0.739 | VLoss: 0.0307 | PLoss: -0.0073 | Ent: 0.2156
+[ 75.8%] Iter 7406/9765 | SPS:  2573 | Reward:  0.053 | Value:  0.658 | VLoss: 0.0208 | PLoss: -0.0016 | Ent: 0.1820
+[ 75.9%] Iter 7407/9765 | SPS:  2573 | Reward:  0.043 | Value:  0.682 | VLoss: 0.0161 | PLoss: -0.0069 | Ent: 0.2193
+[ 75.9%] Iter 7408/9765 | SPS:  2573 | Reward:  0.056 | Value:  0.641 | VLoss: 0.0185 | PLoss: -0.0046 | Ent: 0.1939
+[ 75.9%] Iter 7409/9765 | SPS:  2573 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0271 | PLoss: -0.0036 | Ent: 0.1860
+[ 75.9%] Iter 7410/9765 | SPS:  2573 | Reward:  0.072 | Value:  0.683 | VLoss: 0.0284 | PLoss: -0.0031 | Ent: 0.2282
+[ 75.9%] Iter 7411/9765 | SPS:  2573 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0192 | PLoss: -0.0021 | Ent: 0.1869
+[ 75.9%] Iter 7412/9765 | SPS:  2573 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0213 | PLoss: -0.0026 | Ent: 0.2431
+[ 75.9%] Iter 7413/9765 | SPS:  2574 | Reward:  0.064 | Value:  0.678 | VLoss: 0.0233 | PLoss: -0.0062 | Ent: 0.2143
+[ 75.9%] Iter 7414/9765 | SPS:  2574 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0240 | PLoss: -0.0098 | Ent: 0.2633
+[ 75.9%] Iter 7415/9765 | SPS:  2574 | Reward:  0.067 | Value:  0.750 | VLoss: 0.0212 | PLoss: -0.0036 | Ent: 0.2218
+[ 75.9%] Iter 7416/9765 | SPS:  2574 | Reward:  0.059 | Value:  0.661 | VLoss: 0.0300 | PLoss: -0.0106 | Ent: 0.2184
+[ 76.0%] Iter 7417/9765 | SPS:  2574 | Reward:  0.072 | Value:  0.713 | VLoss: 0.0248 | PLoss: -0.0063 | Ent: 0.1907
+[ 76.0%] Iter 7418/9765 | SPS:  2574 | Reward:  0.067 | Value:  0.810 | VLoss: 0.0244 | PLoss: -0.0070 | Ent: 0.1967
+[ 76.0%] Iter 7419/9765 | SPS:  2574 | Reward:  0.065 | Value:  0.743 | VLoss: 0.0192 | PLoss: -0.0030 | Ent: 0.2012
+[ 76.0%] Iter 7420/9765 | SPS:  2574 | Reward:  0.078 | Value:  0.740 | VLoss: 0.0219 | PLoss: -0.0038 | Ent: 0.2207
+[ 76.0%] Iter 7421/9765 | SPS:  2574 | Reward:  0.068 | Value:  0.765 | VLoss: 0.0247 | PLoss: -0.0101 | Ent: 0.1940
+[ 76.0%] Iter 7422/9765 | SPS:  2575 | Reward:  0.065 | Value:  0.706 | VLoss: 0.0267 | PLoss: -0.0038 | Ent: 0.2050
+[ 76.0%] Iter 7423/9765 | SPS:  2575 | Reward:  0.062 | Value:  0.760 | VLoss: 0.0192 | PLoss: -0.0054 | Ent: 0.1934
+[ 76.0%] Iter 7424/9765 | SPS:  2575 | Reward:  0.060 | Value:  0.680 | VLoss: 0.0202 | PLoss: -0.0007 | Ent: 0.2558
+[ 76.0%] Iter 7425/9765 | SPS:  2575 | Reward:  0.066 | Value:  0.750 | VLoss: 0.0264 | PLoss: -0.0079 | Ent: 0.2491
+[ 76.0%] Iter 7426/9765 | SPS:  2575 | Reward:  0.070 | Value:  0.711 | VLoss: 0.0289 | PLoss: -0.0036 | Ent: 0.2518
+[ 76.1%] Iter 7427/9765 | SPS:  2575 | Reward:  0.061 | Value:  0.728 | VLoss: 0.0114 | PLoss: -0.0056 | Ent: 0.1954
+[ 76.1%] Iter 7428/9765 | SPS:  2575 | Reward:  0.070 | Value:  0.791 | VLoss: 0.0282 | PLoss: -0.0057 | Ent: 0.2079
+[ 76.1%] Iter 7429/9765 | SPS:  2575 | Reward:  0.062 | Value:  0.734 | VLoss: 0.0260 | PLoss: -0.0066 | Ent: 0.1872
+[ 76.1%] Iter 7430/9765 | SPS:  2575 | Reward:  0.068 | Value:  0.767 | VLoss: 0.0166 | PLoss: -0.0080 | Ent: 0.2298
+[ 76.1%] Iter 7431/9765 | SPS:  2576 | Reward:  0.054 | Value:  0.748 | VLoss: 0.0252 | PLoss: -0.0050 | Ent: 0.1929
+[ 76.1%] Iter 7432/9765 | SPS:  2576 | Reward:  0.056 | Value:  0.730 | VLoss: 0.0265 | PLoss: -0.0074 | Ent: 0.2299
+[ 76.1%] Iter 7433/9765 | SPS:  2576 | Reward:  0.059 | Value:  0.669 | VLoss: 0.0313 | PLoss: -0.0105 | Ent: 0.1477
+[ 76.1%] Iter 7434/9765 | SPS:  2576 | Reward:  0.057 | Value:  0.751 | VLoss: 0.0254 | PLoss: -0.0106 | Ent: 0.2708
+[ 76.1%] Iter 7435/9765 | SPS:  2576 | Reward:  0.060 | Value:  0.689 | VLoss: 0.0253 | PLoss: -0.0041 | Ent: 0.2292
+[ 76.1%] Iter 7436/9765 | SPS:  2576 | Reward:  0.074 | Value:  0.692 | VLoss: 0.0182 | PLoss: -0.0005 | Ent: 0.2030
+[ 76.2%] Iter 7437/9765 | SPS:  2576 | Reward:  0.060 | Value:  0.831 | VLoss: 0.0169 | PLoss: -0.0017 | Ent: 0.2224
+[ 76.2%] Iter 7438/9765 | SPS:  2576 | Reward:  0.065 | Value:  0.781 | VLoss: 0.0256 | PLoss: -0.0082 | Ent: 0.1608
+[ 76.2%] Iter 7439/9765 | SPS:  2576 | Reward:  0.062 | Value:  0.742 | VLoss: 0.0166 | PLoss: -0.0045 | Ent: 0.1607
+[ 76.2%] Iter 7440/9765 | SPS:  2577 | Reward:  0.057 | Value:  0.623 | VLoss: 0.0305 | PLoss: -0.0087 | Ent: 0.2305
+[ 76.2%] Iter 7441/9765 | SPS:  2577 | Reward:  0.046 | Value:  0.711 | VLoss: 0.0212 | PLoss: -0.0015 | Ent: 0.2249
+[ 76.2%] Iter 7442/9765 | SPS:  2577 | Reward:  0.068 | Value:  0.675 | VLoss: 0.0261 | PLoss: -0.0007 | Ent: 0.2205
+[ 76.2%] Iter 7443/9765 | SPS:  2577 | Reward:  0.066 | Value:  0.690 | VLoss: 0.0237 | PLoss: -0.0107 | Ent: 0.2323
+[ 76.2%] Iter 7444/9765 | SPS:  2577 | Reward:  0.061 | Value:  0.807 | VLoss: 0.0170 | PLoss: -0.0039 | Ent: 0.1752
+[ 76.2%] Iter 7445/9765 | SPS:  2577 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0228 | PLoss: -0.0018 | Ent: 0.1797
+[ 76.3%] Iter 7446/9765 | SPS:  2577 | Reward:  0.053 | Value:  0.738 | VLoss: 0.0166 | PLoss: -0.0054 | Ent: 0.2263
+[ 76.3%] Iter 7447/9765 | SPS:  2577 | Reward:  0.056 | Value:  0.720 | VLoss: 0.0153 | PLoss: -0.0045 | Ent: 0.1911
+[ 76.3%] Iter 7448/9765 | SPS:  2577 | Reward:  0.057 | Value:  0.664 | VLoss: 0.0227 | PLoss: -0.0089 | Ent: 0.2383
+[ 76.3%] Iter 7449/9765 | SPS:  2577 | Reward:  0.066 | Value:  0.710 | VLoss: 0.0270 | PLoss: -0.0047 | Ent: 0.2587
+[ 76.3%] Iter 7450/9765 | SPS:  2578 | Reward:  0.065 | Value:  0.721 | VLoss: 0.0197 | PLoss: -0.0088 | Ent: 0.1920
+[ 76.3%] Iter 7451/9765 | SPS:  2578 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0214 | PLoss: -0.0044 | Ent: 0.1944
+[ 76.3%] Iter 7452/9765 | SPS:  2578 | Reward:  0.061 | Value:  0.756 | VLoss: 0.0212 | PLoss: -0.0107 | Ent: 0.1987
+[ 76.3%] Iter 7453/9765 | SPS:  2578 | Reward:  0.063 | Value:  0.678 | VLoss: 0.0249 | PLoss: -0.0040 | Ent: 0.2020
+[ 76.3%] Iter 7454/9765 | SPS:  2578 | Reward:  0.042 | Value:  0.688 | VLoss: 0.0194 | PLoss: -0.0068 | Ent: 0.2739
+[ 76.3%] Iter 7455/9765 | SPS:  2578 | Reward:  0.058 | Value:  0.622 | VLoss: 0.0263 | PLoss: -0.0081 | Ent: 0.2530
+[ 76.4%] Iter 7456/9765 | SPS:  2578 | Reward:  0.067 | Value:  0.708 | VLoss: 0.0227 | PLoss: -0.0040 | Ent: 0.2587
+[ 76.4%] Iter 7457/9765 | SPS:  2578 | Reward:  0.055 | Value:  0.724 | VLoss: 0.0213 | PLoss: -0.0073 | Ent: 0.2157
+[ 76.4%] Iter 7458/9765 | SPS:  2578 | Reward:  0.056 | Value:  0.689 | VLoss: 0.0260 | PLoss: -0.0055 | Ent: 0.2313
+[ 76.4%] Iter 7459/9765 | SPS:  2579 | Reward:  0.072 | Value:  0.715 | VLoss: 0.0148 | PLoss: -0.0055 | Ent: 0.1994
+[ 76.4%] Iter 7460/9765 | SPS:  2579 | Reward:  0.059 | Value:  0.727 | VLoss: 0.0194 | PLoss: -0.0012 | Ent: 0.2316
+[ 76.4%] Iter 7461/9765 | SPS:  2579 | Reward:  0.046 | Value:  0.713 | VLoss: 0.0194 | PLoss: -0.0102 | Ent: 0.2626
+[ 76.4%] Iter 7462/9765 | SPS:  2579 | Reward:  0.067 | Value:  0.709 | VLoss: 0.0223 | PLoss: -0.0031 | Ent: 0.1566
+[ 76.4%] Iter 7463/9765 | SPS:  2579 | Reward:  0.055 | Value:  0.743 | VLoss: 0.0189 | PLoss: -0.0095 | Ent: 0.1685
+[ 76.4%] Iter 7464/9765 | SPS:  2579 | Reward:  0.056 | Value:  0.668 | VLoss: 0.0235 | PLoss: -0.0047 | Ent: 0.2271
+[ 76.4%] Iter 7465/9765 | SPS:  2579 | Reward:  0.061 | Value:  0.748 | VLoss: 0.0175 | PLoss: -0.0092 | Ent: 0.2210
+[ 76.5%] Iter 7466/9765 | SPS:  2579 | Reward:  0.075 | Value:  0.716 | VLoss: 0.0183 | PLoss: -0.0057 | Ent: 0.1951
+[ 76.5%] Iter 7467/9765 | SPS:  2579 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0340 | PLoss: -0.0069 | Ent: 0.2291
+[ 76.5%] Iter 7468/9765 | SPS:  2580 | Reward:  0.055 | Value:  0.748 | VLoss: 0.0192 | PLoss: -0.0036 | Ent: 0.2138
+[ 76.5%] Iter 7469/9765 | SPS:  2580 | Reward:  0.080 | Value:  0.759 | VLoss: 0.0290 | PLoss: -0.0035 | Ent: 0.2158
+[ 76.5%] Iter 7470/9765 | SPS:  2580 | Reward:  0.051 | Value:  0.786 | VLoss: 0.0229 | PLoss: -0.0105 | Ent: 0.1842
+[ 76.5%] Iter 7471/9765 | SPS:  2580 | Reward:  0.069 | Value:  0.718 | VLoss: 0.0297 | PLoss: -0.0085 | Ent: 0.1966
+[ 76.5%] Iter 7472/9765 | SPS:  2580 | Reward:  0.059 | Value:  0.662 | VLoss: 0.0224 | PLoss: -0.0059 | Ent: 0.2794
+[ 76.5%] Iter 7473/9765 | SPS:  2580 | Reward:  0.064 | Value:  0.666 | VLoss: 0.0329 | PLoss: -0.0053 | Ent: 0.2223
+[ 76.5%] Iter 7474/9765 | SPS:  2580 | Reward:  0.054 | Value:  0.797 | VLoss: 0.0229 | PLoss: -0.0058 | Ent: 0.1982
+[ 76.5%] Iter 7475/9765 | SPS:  2580 | Reward:  0.072 | Value:  0.723 | VLoss: 0.0230 | PLoss: -0.0052 | Ent: 0.2619
+[ 76.6%] Iter 7476/9765 | SPS:  2580 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0205 | PLoss: -0.0035 | Ent: 0.2061
+[ 76.6%] Iter 7477/9765 | SPS:  2581 | Reward:  0.066 | Value:  0.773 | VLoss: 0.0237 | PLoss: -0.0123 | Ent: 0.2296
+[ 76.6%] Iter 7478/9765 | SPS:  2581 | Reward:  0.064 | Value:  0.782 | VLoss: 0.0280 | PLoss: -0.0039 | Ent: 0.1990
+[ 76.6%] Iter 7479/9765 | SPS:  2581 | Reward:  0.074 | Value:  0.761 | VLoss: 0.0219 | PLoss: -0.0058 | Ent: 0.1819
+[ 76.6%] Iter 7480/9765 | SPS:  2581 | Reward:  0.056 | Value:  0.693 | VLoss: 0.0205 | PLoss: -0.0041 | Ent: 0.2251
+[ 76.6%] Iter 7481/9765 | SPS:  2581 | Reward:  0.075 | Value:  0.763 | VLoss: 0.0169 | PLoss: -0.0039 | Ent: 0.2091
+[ 76.6%] Iter 7482/9765 | SPS:  2581 | Reward:  0.081 | Value:  0.722 | VLoss: 0.0250 | PLoss: -0.0061 | Ent: 0.2012
+[ 76.6%] Iter 7483/9765 | SPS:  2581 | Reward:  0.059 | Value:  0.698 | VLoss: 0.0209 | PLoss: -0.0062 | Ent: 0.2735
+[ 76.6%] Iter 7484/9765 | SPS:  2581 | Reward:  0.071 | Value:  0.711 | VLoss: 0.0199 | PLoss: -0.0055 | Ent: 0.1803
+[ 76.7%] Iter 7485/9765 | SPS:  2581 | Reward:  0.055 | Value:  0.801 | VLoss: 0.0184 | PLoss: -0.0012 | Ent: 0.2317
+[ 76.7%] Iter 7486/9765 | SPS:  2581 | Reward:  0.047 | Value:  0.676 | VLoss: 0.0147 | PLoss: -0.0047 | Ent: 0.2103
+[ 76.7%] Iter 7487/9765 | SPS:  2582 | Reward:  0.067 | Value:  0.737 | VLoss: 0.0326 | PLoss: -0.0086 | Ent: 0.2549
+[ 76.7%] Iter 7488/9765 | SPS:  2582 | Reward:  0.050 | Value:  0.695 | VLoss: 0.0211 | PLoss: -0.0051 | Ent: 0.2074
+[ 76.7%] Iter 7489/9765 | SPS:  2582 | Reward:  0.069 | Value:  0.751 | VLoss: 0.0208 | PLoss: -0.0106 | Ent: 0.2079
+[ 76.7%] Iter 7490/9765 | SPS:  2582 | Reward:  0.051 | Value:  0.727 | VLoss: 0.0268 | PLoss: -0.0092 | Ent: 0.2418
+[ 76.7%] Iter 7491/9765 | SPS:  2582 | Reward:  0.066 | Value:  0.649 | VLoss: 0.0196 | PLoss: -0.0026 | Ent: 0.1717
+[ 76.7%] Iter 7492/9765 | SPS:  2582 | Reward:  0.072 | Value:  0.762 | VLoss: 0.0207 | PLoss: -0.0028 | Ent: 0.1607
+[ 76.7%] Iter 7493/9765 | SPS:  2582 | Reward:  0.063 | Value:  0.811 | VLoss: 0.0206 | PLoss: -0.0060 | Ent: 0.1660
+[ 76.7%] Iter 7494/9765 | SPS:  2582 | Reward:  0.062 | Value:  0.778 | VLoss: 0.0206 | PLoss: -0.0130 | Ent: 0.2558
+[ 76.8%] Iter 7495/9765 | SPS:  2582 | Reward:  0.060 | Value:  0.724 | VLoss: 0.0180 | PLoss: -0.0087 | Ent: 0.2709
+[ 76.8%] Iter 7496/9765 | SPS:  2583 | Reward:  0.072 | Value:  0.786 | VLoss: 0.0289 | PLoss: -0.0095 | Ent: 0.2512
+[ 76.8%] Iter 7497/9765 | SPS:  2583 | Reward:  0.064 | Value:  0.721 | VLoss: 0.0239 | PLoss: -0.0062 | Ent: 0.2355
+[ 76.8%] Iter 7498/9765 | SPS:  2583 | Reward:  0.060 | Value:  0.743 | VLoss: 0.0176 | PLoss: -0.0028 | Ent: 0.2234
+[ 76.8%] Iter 7499/9765 | SPS:  2583 | Reward:  0.067 | Value:  0.713 | VLoss: 0.0272 | PLoss: -0.0073 | Ent: 0.2103
+[ 76.8%] Iter 7500/9765 | SPS:  2583 | Reward:  0.073 | Value:  0.704 | VLoss: 0.0311 | PLoss: -0.0050 | Ent: 0.2120
+[ 76.8%] Iter 7501/9765 | SPS:  2583 | Reward:  0.053 | Value:  0.770 | VLoss: 0.0181 | PLoss: -0.0056 | Ent: 0.2140
+[ 76.8%] Iter 7502/9765 | SPS:  2583 | Reward:  0.063 | Value:  0.716 | VLoss: 0.0280 | PLoss: -0.0105 | Ent: 0.2101
+[ 76.8%] Iter 7503/9765 | SPS:  2583 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0273 | PLoss: -0.0054 | Ent: 0.2614
+[ 76.8%] Iter 7504/9765 | SPS:  2583 | Reward:  0.062 | Value:  0.756 | VLoss: 0.0221 | PLoss: -0.0109 | Ent: 0.2523
+[ 76.9%] Iter 7505/9765 | SPS:  2583 | Reward:  0.065 | Value:  0.746 | VLoss: 0.0168 | PLoss: -0.0047 | Ent: 0.2057
+[ 76.9%] Iter 7506/9765 | SPS:  2584 | Reward:  0.056 | Value:  0.735 | VLoss: 0.0167 | PLoss: -0.0092 | Ent: 0.2297
+[ 76.9%] Iter 7507/9765 | SPS:  2584 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0158 | PLoss: -0.0097 | Ent: 0.1798
+[ 76.9%] Iter 7508/9765 | SPS:  2584 | Reward:  0.059 | Value:  0.782 | VLoss: 0.0183 | PLoss: -0.0079 | Ent: 0.1811
+[ 76.9%] Iter 7509/9765 | SPS:  2584 | Reward:  0.056 | Value:  0.673 | VLoss: 0.0261 | PLoss: -0.0055 | Ent: 0.2504
+[ 76.9%] Iter 7510/9765 | SPS:  2584 | Reward:  0.059 | Value:  0.715 | VLoss: 0.0218 | PLoss: -0.0022 | Ent: 0.2335
+[ 76.9%] Iter 7511/9765 | SPS:  2584 | Reward:  0.050 | Value:  0.730 | VLoss: 0.0231 | PLoss: -0.0060 | Ent: 0.1940
+[ 76.9%] Iter 7512/9765 | SPS:  2584 | Reward:  0.056 | Value:  0.660 | VLoss: 0.0197 | PLoss: -0.0044 | Ent: 0.2189
+[ 76.9%] Iter 7513/9765 | SPS:  2584 | Reward:  0.067 | Value:  0.728 | VLoss: 0.0260 | PLoss: -0.0044 | Ent: 0.2593
+[ 76.9%] Iter 7514/9765 | SPS:  2584 | Reward:  0.069 | Value:  0.748 | VLoss: 0.0241 | PLoss: -0.0047 | Ent: 0.2432
+[ 77.0%] Iter 7515/9765 | SPS:  2585 | Reward:  0.055 | Value:  0.820 | VLoss: 0.0088 | PLoss: 0.0010 | Ent: 0.1932
+[ 77.0%] Iter 7516/9765 | SPS:  2585 | Reward:  0.058 | Value:  0.744 | VLoss: 0.0196 | PLoss: -0.0058 | Ent: 0.2060
+[ 77.0%] Iter 7517/9765 | SPS:  2585 | Reward:  0.061 | Value:  0.661 | VLoss: 0.0280 | PLoss: -0.0096 | Ent: 0.2667
+[ 77.0%] Iter 7518/9765 | SPS:  2585 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0211 | PLoss: -0.0049 | Ent: 0.2105
+[ 77.0%] Iter 7519/9765 | SPS:  2585 | Reward:  0.070 | Value:  0.741 | VLoss: 0.0176 | PLoss: 0.0001 | Ent: 0.2024
+[ 77.0%] Iter 7520/9765 | SPS:  2585 | Reward:  0.060 | Value:  0.605 | VLoss: 0.0236 | PLoss: -0.0039 | Ent: 0.2061
+[ 77.0%] Iter 7521/9765 | SPS:  2585 | Reward:  0.064 | Value:  0.783 | VLoss: 0.0192 | PLoss: -0.0070 | Ent: 0.2068
+[ 77.0%] Iter 7522/9765 | SPS:  2585 | Reward:  0.067 | Value:  0.761 | VLoss: 0.0167 | PLoss: -0.0058 | Ent: 0.2209
+[ 77.0%] Iter 7523/9765 | SPS:  2585 | Reward:  0.055 | Value:  0.823 | VLoss: 0.0187 | PLoss: -0.0051 | Ent: 0.1987
+[ 77.1%] Iter 7524/9765 | SPS:  2586 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0271 | PLoss: -0.0049 | Ent: 0.2190
+[ 77.1%] Iter 7525/9765 | SPS:  2586 | Reward:  0.063 | Value:  0.675 | VLoss: 0.0223 | PLoss: -0.0016 | Ent: 0.2043
+[ 77.1%] Iter 7526/9765 | SPS:  2586 | Reward:  0.057 | Value:  0.726 | VLoss: 0.0178 | PLoss: -0.0047 | Ent: 0.2175
+[ 77.1%] Iter 7527/9765 | SPS:  2586 | Reward:  0.069 | Value:  0.729 | VLoss: 0.0197 | PLoss: -0.0062 | Ent: 0.1831
+[ 77.1%] Iter 7528/9765 | SPS:  2586 | Reward:  0.071 | Value:  0.724 | VLoss: 0.0255 | PLoss: -0.0084 | Ent: 0.1757
+[ 77.1%] Iter 7529/9765 | SPS:  2586 | Reward:  0.059 | Value:  0.739 | VLoss: 0.0265 | PLoss: -0.0077 | Ent: 0.2551
+[ 77.1%] Iter 7530/9765 | SPS:  2586 | Reward:  0.057 | Value:  0.742 | VLoss: 0.0169 | PLoss: 0.0013 | Ent: 0.1811
+[ 77.1%] Iter 7531/9765 | SPS:  2586 | Reward:  0.060 | Value:  0.784 | VLoss: 0.0151 | PLoss: -0.0013 | Ent: 0.1901
+[ 77.1%] Iter 7532/9765 | SPS:  2586 | Reward:  0.070 | Value:  0.761 | VLoss: 0.0178 | PLoss: -0.0055 | Ent: 0.2139
+[ 77.1%] Iter 7533/9765 | SPS:  2587 | Reward:  0.059 | Value:  0.730 | VLoss: 0.0246 | PLoss: -0.0072 | Ent: 0.2320
+[ 77.2%] Iter 7534/9765 | SPS:  2587 | Reward:  0.063 | Value:  0.713 | VLoss: 0.0215 | PLoss: -0.0040 | Ent: 0.2084
+[ 77.2%] Iter 7535/9765 | SPS:  2587 | Reward:  0.066 | Value:  0.756 | VLoss: 0.0241 | PLoss: -0.0040 | Ent: 0.1792
+[ 77.2%] Iter 7536/9765 | SPS:  2587 | Reward:  0.067 | Value:  0.745 | VLoss: 0.0281 | PLoss: -0.0048 | Ent: 0.1871
+[ 77.2%] Iter 7537/9765 | SPS:  2587 | Reward:  0.070 | Value:  0.712 | VLoss: 0.0239 | PLoss: -0.0017 | Ent: 0.2163
+[ 77.2%] Iter 7538/9765 | SPS:  2587 | Reward:  0.070 | Value:  0.764 | VLoss: 0.0195 | PLoss: -0.0053 | Ent: 0.2355
+[ 77.2%] Iter 7539/9765 | SPS:  2587 | Reward:  0.069 | Value:  0.708 | VLoss: 0.0269 | PLoss: 0.0009 | Ent: 0.2024
+[ 77.2%] Iter 7540/9765 | SPS:  2587 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0228 | PLoss: -0.0051 | Ent: 0.2150
+[ 77.2%] Iter 7541/9765 | SPS:  2587 | Reward:  0.063 | Value:  0.764 | VLoss: 0.0236 | PLoss: -0.0062 | Ent: 0.2217
+[ 77.2%] Iter 7542/9765 | SPS:  2587 | Reward:  0.059 | Value:  0.740 | VLoss: 0.0240 | PLoss: -0.0028 | Ent: 0.2084
+[ 77.2%] Iter 7543/9765 | SPS:  2588 | Reward:  0.058 | Value:  0.673 | VLoss: 0.0191 | PLoss: -0.0027 | Ent: 0.1955
+[ 77.3%] Iter 7544/9765 | SPS:  2588 | Reward:  0.047 | Value:  0.713 | VLoss: 0.0183 | PLoss: -0.0040 | Ent: 0.2327
+[ 77.3%] Iter 7545/9765 | SPS:  2588 | Reward:  0.069 | Value:  0.751 | VLoss: 0.0214 | PLoss: -0.0076 | Ent: 0.1559
+[ 77.3%] Iter 7546/9765 | SPS:  2588 | Reward:  0.078 | Value:  0.706 | VLoss: 0.0291 | PLoss: -0.0063 | Ent: 0.2371
+[ 77.3%] Iter 7547/9765 | SPS:  2588 | Reward:  0.068 | Value:  0.766 | VLoss: 0.0275 | PLoss: -0.0047 | Ent: 0.2348
+[ 77.3%] Iter 7548/9765 | SPS:  2588 | Reward:  0.065 | Value:  0.723 | VLoss: 0.0212 | PLoss: -0.0048 | Ent: 0.1727
+[ 77.3%] Iter 7549/9765 | SPS:  2588 | Reward:  0.047 | Value:  0.754 | VLoss: 0.0173 | PLoss: -0.0047 | Ent: 0.2074
+[ 77.3%] Iter 7550/9765 | SPS:  2588 | Reward:  0.058 | Value:  0.700 | VLoss: 0.0172 | PLoss: -0.0052 | Ent: 0.1693
+[ 77.3%] Iter 7551/9765 | SPS:  2588 | Reward:  0.041 | Value:  0.730 | VLoss: 0.0088 | PLoss: -0.0035 | Ent: 0.2396
+[ 77.3%] Iter 7552/9765 | SPS:  2589 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0212 | PLoss: -0.0100 | Ent: 0.2255
+[ 77.3%] Iter 7553/9765 | SPS:  2589 | Reward:  0.058 | Value:  0.698 | VLoss: 0.0196 | PLoss: -0.0069 | Ent: 0.2460
+[ 77.4%] Iter 7554/9765 | SPS:  2589 | Reward:  0.067 | Value:  0.719 | VLoss: 0.0261 | PLoss: -0.0044 | Ent: 0.1999
+[ 77.4%] Iter 7555/9765 | SPS:  2589 | Reward:  0.068 | Value:  0.649 | VLoss: 0.0228 | PLoss: -0.0062 | Ent: 0.2278
+[ 77.4%] Iter 7556/9765 | SPS:  2589 | Reward:  0.056 | Value:  0.773 | VLoss: 0.0180 | PLoss: -0.0095 | Ent: 0.2281
+[ 77.4%] Iter 7557/9765 | SPS:  2589 | Reward:  0.066 | Value:  0.765 | VLoss: 0.0237 | PLoss: -0.0058 | Ent: 0.2308
+[ 77.4%] Iter 7558/9765 | SPS:  2589 | Reward:  0.067 | Value:  0.714 | VLoss: 0.0163 | PLoss: -0.0052 | Ent: 0.2126
+[ 77.4%] Iter 7559/9765 | SPS:  2589 | Reward:  0.064 | Value:  0.816 | VLoss: 0.0150 | PLoss: -0.0061 | Ent: 0.2117
+[ 77.4%] Iter 7560/9765 | SPS:  2589 | Reward:  0.068 | Value:  0.768 | VLoss: 0.0222 | PLoss: -0.0030 | Ent: 0.1620
+[ 77.4%] Iter 7561/9765 | SPS:  2589 | Reward:  0.063 | Value:  0.742 | VLoss: 0.0200 | PLoss: -0.0017 | Ent: 0.2098
+[ 77.4%] Iter 7562/9765 | SPS:  2590 | Reward:  0.063 | Value:  0.804 | VLoss: 0.0183 | PLoss: -0.0077 | Ent: 0.2445
+[ 77.5%] Iter 7563/9765 | SPS:  2590 | Reward:  0.064 | Value:  0.771 | VLoss: 0.0143 | PLoss: -0.0065 | Ent: 0.2309
+[ 77.5%] Iter 7564/9765 | SPS:  2590 | Reward:  0.054 | Value:  0.683 | VLoss: 0.0188 | PLoss: -0.0090 | Ent: 0.2370
+[ 77.5%] Iter 7565/9765 | SPS:  2590 | Reward:  0.041 | Value:  0.654 | VLoss: 0.0244 | PLoss: -0.0027 | Ent: 0.2458
+[ 77.5%] Iter 7566/9765 | SPS:  2590 | Reward:  0.073 | Value:  0.685 | VLoss: 0.0219 | PLoss: -0.0085 | Ent: 0.1679
+[ 77.5%] Iter 7567/9765 | SPS:  2590 | Reward:  0.061 | Value:  0.755 | VLoss: 0.0189 | PLoss: 0.0034 | Ent: 0.1552
+[ 77.5%] Iter 7568/9765 | SPS:  2590 | Reward:  0.063 | Value:  0.761 | VLoss: 0.0188 | PLoss: -0.0041 | Ent: 0.2269
+[ 77.5%] Iter 7569/9765 | SPS:  2590 | Reward:  0.046 | Value:  0.724 | VLoss: 0.0151 | PLoss: -0.0044 | Ent: 0.2410
+[ 77.5%] Iter 7570/9765 | SPS:  2590 | Reward:  0.069 | Value:  0.670 | VLoss: 0.0184 | PLoss: -0.0054 | Ent: 0.1887
+[ 77.5%] Iter 7571/9765 | SPS:  2591 | Reward:  0.072 | Value:  0.760 | VLoss: 0.0215 | PLoss: -0.0079 | Ent: 0.2442
+[ 77.5%] Iter 7572/9765 | SPS:  2591 | Reward:  0.070 | Value:  0.759 | VLoss: 0.0133 | PLoss: -0.0113 | Ent: 0.2133
+[ 77.6%] Iter 7573/9765 | SPS:  2591 | Reward:  0.058 | Value:  0.749 | VLoss: 0.0276 | PLoss: -0.0040 | Ent: 0.2411
+[ 77.6%] Iter 7574/9765 | SPS:  2591 | Reward:  0.051 | Value:  0.593 | VLoss: 0.0230 | PLoss: -0.0040 | Ent: 0.2346
+[ 77.6%] Iter 7575/9765 | SPS:  2591 | Reward:  0.059 | Value:  0.634 | VLoss: 0.0200 | PLoss: -0.0060 | Ent: 0.1985
+[ 77.6%] Iter 7576/9765 | SPS:  2591 | Reward:  0.054 | Value:  0.695 | VLoss: 0.0187 | PLoss: -0.0009 | Ent: 0.2314
+[ 77.6%] Iter 7577/9765 | SPS:  2591 | Reward:  0.063 | Value:  0.717 | VLoss: 0.0188 | PLoss: -0.0055 | Ent: 0.1760
+[ 77.6%] Iter 7578/9765 | SPS:  2591 | Reward:  0.069 | Value:  0.751 | VLoss: 0.0270 | PLoss: -0.0008 | Ent: 0.2190
+[ 77.6%] Iter 7579/9765 | SPS:  2591 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0337 | PLoss: -0.0052 | Ent: 0.1944
+[ 77.6%] Iter 7580/9765 | SPS:  2592 | Reward:  0.055 | Value:  0.708 | VLoss: 0.0279 | PLoss: -0.0057 | Ent: 0.2118
+[ 77.6%] Iter 7581/9765 | SPS:  2592 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0206 | PLoss: -0.0058 | Ent: 0.2077
+[ 77.6%] Iter 7582/9765 | SPS:  2592 | Reward:  0.065 | Value:  0.788 | VLoss: 0.0199 | PLoss: -0.0042 | Ent: 0.2642
+[ 77.7%] Iter 7583/9765 | SPS:  2592 | Reward:  0.052 | Value:  0.721 | VLoss: 0.0155 | PLoss: -0.0070 | Ent: 0.2379
+[ 77.7%] Iter 7584/9765 | SPS:  2592 | Reward:  0.070 | Value:  0.659 | VLoss: 0.0251 | PLoss: -0.0066 | Ent: 0.2135
+[ 77.7%] Iter 7585/9765 | SPS:  2592 | Reward:  0.062 | Value:  0.596 | VLoss: 0.0278 | PLoss: -0.0039 | Ent: 0.2660
+[ 77.7%] Iter 7586/9765 | SPS:  2592 | Reward:  0.060 | Value:  0.743 | VLoss: 0.0213 | PLoss: -0.0110 | Ent: 0.2481
+[ 77.7%] Iter 7587/9765 | SPS:  2592 | Reward:  0.058 | Value:  0.630 | VLoss: 0.0291 | PLoss: 0.0009 | Ent: 0.2338
+[ 77.7%] Iter 7588/9765 | SPS:  2592 | Reward:  0.057 | Value:  0.687 | VLoss: 0.0183 | PLoss: -0.0021 | Ent: 0.2249
+[ 77.7%] Iter 7589/9765 | SPS:  2592 | Reward:  0.066 | Value:  0.795 | VLoss: 0.0195 | PLoss: -0.0029 | Ent: 0.2114
+[ 77.7%] Iter 7590/9765 | SPS:  2593 | Reward:  0.071 | Value:  0.748 | VLoss: 0.0182 | PLoss: -0.0081 | Ent: 0.2051
+[ 77.7%] Iter 7591/9765 | SPS:  2593 | Reward:  0.075 | Value:  0.753 | VLoss: 0.0227 | PLoss: -0.0073 | Ent: 0.2170
+[ 77.7%] Iter 7592/9765 | SPS:  2593 | Reward:  0.061 | Value:  0.769 | VLoss: 0.0134 | PLoss: -0.0038 | Ent: 0.2195
+[ 77.8%] Iter 7593/9765 | SPS:  2593 | Reward:  0.067 | Value:  0.733 | VLoss: 0.0197 | PLoss: -0.0100 | Ent: 0.2454
+[ 77.8%] Iter 7594/9765 | SPS:  2593 | Reward:  0.055 | Value:  0.826 | VLoss: 0.0200 | PLoss: -0.0083 | Ent: 0.2246
+[ 77.8%] Iter 7595/9765 | SPS:  2593 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0208 | PLoss: -0.0008 | Ent: 0.1741
+[ 77.8%] Iter 7596/9765 | SPS:  2593 | Reward:  0.049 | Value:  0.673 | VLoss: 0.0264 | PLoss: -0.0038 | Ent: 0.2271
+[ 77.8%] Iter 7597/9765 | SPS:  2593 | Reward:  0.052 | Value:  0.684 | VLoss: 0.0196 | PLoss: -0.0070 | Ent: 0.2189
+[ 77.8%] Iter 7598/9765 | SPS:  2593 | Reward:  0.056 | Value:  0.620 | VLoss: 0.0211 | PLoss: -0.0087 | Ent: 0.1825
+[ 77.8%] Iter 7599/9765 | SPS:  2594 | Reward:  0.056 | Value:  0.709 | VLoss: 0.0282 | PLoss: -0.0014 | Ent: 0.1917
+[ 77.8%] Iter 7600/9765 | SPS:  2594 | Reward:  0.066 | Value:  0.706 | VLoss: 0.0234 | PLoss: -0.0111 | Ent: 0.2022
+[ 77.8%] Iter 7601/9765 | SPS:  2594 | Reward:  0.046 | Value:  0.666 | VLoss: 0.0245 | PLoss: -0.0043 | Ent: 0.2071
+[ 77.8%] Iter 7602/9765 | SPS:  2594 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0244 | PLoss: -0.0067 | Ent: 0.2062
+[ 77.9%] Iter 7603/9765 | SPS:  2594 | Reward:  0.059 | Value:  0.640 | VLoss: 0.0212 | PLoss: -0.0054 | Ent: 0.2616
+[ 77.9%] Iter 7604/9765 | SPS:  2594 | Reward:  0.070 | Value:  0.714 | VLoss: 0.0188 | PLoss: -0.0037 | Ent: 0.2168
+[ 77.9%] Iter 7605/9765 | SPS:  2594 | Reward:  0.050 | Value:  0.746 | VLoss: 0.0194 | PLoss: -0.0033 | Ent: 0.1942
+[ 77.9%] Iter 7606/9765 | SPS:  2594 | Reward:  0.057 | Value:  0.765 | VLoss: 0.0147 | PLoss: -0.0042 | Ent: 0.2019
+[ 77.9%] Iter 7607/9765 | SPS:  2594 | Reward:  0.066 | Value:  0.754 | VLoss: 0.0267 | PLoss: -0.0079 | Ent: 0.2330
+[ 77.9%] Iter 7608/9765 | SPS:  2595 | Reward:  0.049 | Value:  0.679 | VLoss: 0.0193 | PLoss: -0.0100 | Ent: 0.2388
+[ 77.9%] Iter 7609/9765 | SPS:  2595 | Reward:  0.058 | Value:  0.687 | VLoss: 0.0293 | PLoss: -0.0039 | Ent: 0.1786
+[ 77.9%] Iter 7610/9765 | SPS:  2595 | Reward:  0.059 | Value:  0.648 | VLoss: 0.0186 | PLoss: -0.0061 | Ent: 0.1915
+[ 77.9%] Iter 7611/9765 | SPS:  2595 | Reward:  0.063 | Value:  0.690 | VLoss: 0.0169 | PLoss: -0.0042 | Ent: 0.2183
+[ 78.0%] Iter 7612/9765 | SPS:  2595 | Reward:  0.054 | Value:  0.728 | VLoss: 0.0286 | PLoss: -0.0071 | Ent: 0.2315
+[ 78.0%] Iter 7613/9765 | SPS:  2595 | Reward:  0.081 | Value:  0.709 | VLoss: 0.0175 | PLoss: -0.0041 | Ent: 0.2168
+[ 78.0%] Iter 7614/9765 | SPS:  2595 | Reward:  0.067 | Value:  0.754 | VLoss: 0.0197 | PLoss: -0.0069 | Ent: 0.2198
+[ 78.0%] Iter 7615/9765 | SPS:  2595 | Reward:  0.054 | Value:  0.706 | VLoss: 0.0251 | PLoss: -0.0059 | Ent: 0.2650
+[ 78.0%] Iter 7616/9765 | SPS:  2595 | Reward:  0.058 | Value:  0.727 | VLoss: 0.0239 | PLoss: -0.0052 | Ent: 0.2112
+[ 78.0%] Iter 7617/9765 | SPS:  2596 | Reward:  0.068 | Value:  0.755 | VLoss: 0.0210 | PLoss: -0.0058 | Ent: 0.2207
+[ 78.0%] Iter 7618/9765 | SPS:  2596 | Reward:  0.066 | Value:  0.740 | VLoss: 0.0229 | PLoss: -0.0073 | Ent: 0.2656
+[ 78.0%] Iter 7619/9765 | SPS:  2596 | Reward:  0.066 | Value:  0.759 | VLoss: 0.0184 | PLoss: -0.0045 | Ent: 0.2105
+[ 78.0%] Iter 7620/9765 | SPS:  2596 | Reward:  0.047 | Value:  0.692 | VLoss: 0.0237 | PLoss: -0.0067 | Ent: 0.1924
+[ 78.0%] Iter 7621/9765 | SPS:  2596 | Reward:  0.056 | Value:  0.774 | VLoss: 0.0144 | PLoss: -0.0054 | Ent: 0.1948
+[ 78.1%] Iter 7622/9765 | SPS:  2596 | Reward:  0.061 | Value:  0.729 | VLoss: 0.0302 | PLoss: -0.0063 | Ent: 0.2335
+[ 78.1%] Iter 7623/9765 | SPS:  2596 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0168 | PLoss: -0.0028 | Ent: 0.1785
+[ 78.1%] Iter 7624/9765 | SPS:  2596 | Reward:  0.067 | Value:  0.765 | VLoss: 0.0204 | PLoss: -0.0024 | Ent: 0.2195
+[ 78.1%] Iter 7625/9765 | SPS:  2596 | Reward:  0.068 | Value:  0.747 | VLoss: 0.0230 | PLoss: -0.0030 | Ent: 0.2040
+[ 78.1%] Iter 7626/9765 | SPS:  2596 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0253 | PLoss: -0.0041 | Ent: 0.2577
+[ 78.1%] Iter 7627/9765 | SPS:  2597 | Reward:  0.065 | Value:  0.712 | VLoss: 0.0194 | PLoss: -0.0076 | Ent: 0.1970
+[ 78.1%] Iter 7628/9765 | SPS:  2597 | Reward:  0.058 | Value:  0.675 | VLoss: 0.0192 | PLoss: -0.0038 | Ent: 0.2314
+[ 78.1%] Iter 7629/9765 | SPS:  2597 | Reward:  0.068 | Value:  0.683 | VLoss: 0.0226 | PLoss: -0.0058 | Ent: 0.2279
+[ 78.1%] Iter 7630/9765 | SPS:  2597 | Reward:  0.042 | Value:  0.735 | VLoss: 0.0244 | PLoss: -0.0013 | Ent: 0.2028
+[ 78.1%] Iter 7631/9765 | SPS:  2597 | Reward:  0.050 | Value:  0.594 | VLoss: 0.0296 | PLoss: -0.0033 | Ent: 0.2146
+[ 78.2%] Iter 7632/9765 | SPS:  2597 | Reward:  0.062 | Value:  0.631 | VLoss: 0.0249 | PLoss: -0.0072 | Ent: 0.2028
+[ 78.2%] Iter 7633/9765 | SPS:  2597 | Reward:  0.042 | Value:  0.690 | VLoss: 0.0218 | PLoss: -0.0021 | Ent: 0.1471
+[ 78.2%] Iter 7634/9765 | SPS:  2597 | Reward:  0.049 | Value:  0.665 | VLoss: 0.0151 | PLoss: -0.0021 | Ent: 0.2266
+[ 78.2%] Iter 7635/9765 | SPS:  2597 | Reward:  0.059 | Value:  0.641 | VLoss: 0.0164 | PLoss: -0.0079 | Ent: 0.2204
+[ 78.2%] Iter 7636/9765 | SPS:  2598 | Reward:  0.062 | Value:  0.744 | VLoss: 0.0225 | PLoss: -0.0030 | Ent: 0.2256
+[ 78.2%] Iter 7637/9765 | SPS:  2598 | Reward:  0.064 | Value:  0.791 | VLoss: 0.0164 | PLoss: -0.0091 | Ent: 0.2467
+[ 78.2%] Iter 7638/9765 | SPS:  2598 | Reward:  0.054 | Value:  0.744 | VLoss: 0.0197 | PLoss: -0.0060 | Ent: 0.2374
+[ 78.2%] Iter 7639/9765 | SPS:  2598 | Reward:  0.056 | Value:  0.715 | VLoss: 0.0202 | PLoss: -0.0151 | Ent: 0.1909
+[ 78.2%] Iter 7640/9765 | SPS:  2598 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0263 | PLoss: -0.0042 | Ent: 0.2081
+[ 78.2%] Iter 7641/9765 | SPS:  2598 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0214 | PLoss: -0.0044 | Ent: 0.2333
+[ 78.3%] Iter 7642/9765 | SPS:  2598 | Reward:  0.050 | Value:  0.696 | VLoss: 0.0198 | PLoss: -0.0026 | Ent: 0.1764
+[ 78.3%] Iter 7643/9765 | SPS:  2598 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0258 | PLoss: -0.0032 | Ent: 0.1936
+[ 78.3%] Iter 7644/9765 | SPS:  2598 | Reward:  0.054 | Value:  0.665 | VLoss: 0.0204 | PLoss: -0.0082 | Ent: 0.1987
+[ 78.3%] Iter 7645/9765 | SPS:  2599 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0241 | PLoss: -0.0060 | Ent: 0.2686
+[ 78.3%] Iter 7646/9765 | SPS:  2599 | Reward:  0.069 | Value:  0.709 | VLoss: 0.0202 | PLoss: -0.0086 | Ent: 0.1802
+[ 78.3%] Iter 7647/9765 | SPS:  2599 | Reward:  0.060 | Value:  0.804 | VLoss: 0.0317 | PLoss: -0.0058 | Ent: 0.2438
+[ 78.3%] Iter 7648/9765 | SPS:  2599 | Reward:  0.058 | Value:  0.695 | VLoss: 0.0213 | PLoss: -0.0074 | Ent: 0.1832
+[ 78.3%] Iter 7649/9765 | SPS:  2599 | Reward:  0.045 | Value:  0.658 | VLoss: 0.0235 | PLoss: -0.0076 | Ent: 0.2412
+[ 78.3%] Iter 7650/9765 | SPS:  2599 | Reward:  0.051 | Value:  0.673 | VLoss: 0.0232 | PLoss: -0.0008 | Ent: 0.1651
+[ 78.4%] Iter 7651/9765 | SPS:  2599 | Reward:  0.050 | Value:  0.683 | VLoss: 0.0175 | PLoss: -0.0046 | Ent: 0.1882
+[ 78.4%] Iter 7652/9765 | SPS:  2599 | Reward:  0.056 | Value:  0.696 | VLoss: 0.0288 | PLoss: -0.0048 | Ent: 0.2231
+[ 78.4%] Iter 7653/9765 | SPS:  2599 | Reward:  0.062 | Value:  0.674 | VLoss: 0.0297 | PLoss: -0.0080 | Ent: 0.2912
+[ 78.4%] Iter 7654/9765 | SPS:  2600 | Reward:  0.054 | Value:  0.709 | VLoss: 0.0208 | PLoss: -0.0069 | Ent: 0.2292
+[ 78.4%] Iter 7655/9765 | SPS:  2600 | Reward:  0.081 | Value:  0.759 | VLoss: 0.0307 | PLoss: -0.0079 | Ent: 0.1998
+[ 78.4%] Iter 7656/9765 | SPS:  2600 | Reward:  0.053 | Value:  0.775 | VLoss: 0.0173 | PLoss: -0.0094 | Ent: 0.2233
+[ 78.4%] Iter 7657/9765 | SPS:  2600 | Reward:  0.050 | Value:  0.738 | VLoss: 0.0152 | PLoss: -0.0059 | Ent: 0.2677
+[ 78.4%] Iter 7658/9765 | SPS:  2600 | Reward:  0.061 | Value:  0.771 | VLoss: 0.0256 | PLoss: -0.0085 | Ent: 0.2599
+[ 78.4%] Iter 7659/9765 | SPS:  2600 | Reward:  0.066 | Value:  0.657 | VLoss: 0.0256 | PLoss: -0.0121 | Ent: 0.2472
+[ 78.4%] Iter 7660/9765 | SPS:  2600 | Reward:  0.045 | Value:  0.730 | VLoss: 0.0207 | PLoss: -0.0074 | Ent: 0.2278
+[ 78.5%] Iter 7661/9765 | SPS:  2600 | Reward:  0.056 | Value:  0.729 | VLoss: 0.0194 | PLoss: -0.0082 | Ent: 0.1845
+[ 78.5%] Iter 7662/9765 | SPS:  2600 | Reward:  0.061 | Value:  0.728 | VLoss: 0.0257 | PLoss: -0.0009 | Ent: 0.2096
+[ 78.5%] Iter 7663/9765 | SPS:  2601 | Reward:  0.063 | Value:  0.736 | VLoss: 0.0198 | PLoss: -0.0065 | Ent: 0.2059
+[ 78.5%] Iter 7664/9765 | SPS:  2601 | Reward:  0.050 | Value:  0.660 | VLoss: 0.0212 | PLoss: -0.0043 | Ent: 0.2125
+[ 78.5%] Iter 7665/9765 | SPS:  2601 | Reward:  0.052 | Value:  0.677 | VLoss: 0.0238 | PLoss: -0.0068 | Ent: 0.2159
+[ 78.5%] Iter 7666/9765 | SPS:  2601 | Reward:  0.049 | Value:  0.696 | VLoss: 0.0177 | PLoss: -0.0068 | Ent: 0.2061
+[ 78.5%] Iter 7667/9765 | SPS:  2601 | Reward:  0.060 | Value:  0.726 | VLoss: 0.0183 | PLoss: -0.0033 | Ent: 0.2499
+[ 78.5%] Iter 7668/9765 | SPS:  2601 | Reward:  0.055 | Value:  0.678 | VLoss: 0.0265 | PLoss: -0.0055 | Ent: 0.2515
+[ 78.5%] Iter 7669/9765 | SPS:  2601 | Reward:  0.063 | Value:  0.768 | VLoss: 0.0207 | PLoss: -0.0041 | Ent: 0.2278
+[ 78.5%] Iter 7670/9765 | SPS:  2601 | Reward:  0.058 | Value:  0.716 | VLoss: 0.0204 | PLoss: -0.0052 | Ent: 0.2051
+[ 78.6%] Iter 7671/9765 | SPS:  2601 | Reward:  0.054 | Value:  0.711 | VLoss: 0.0209 | PLoss: -0.0053 | Ent: 0.2431
+[ 78.6%] Iter 7672/9765 | SPS:  2601 | Reward:  0.058 | Value:  0.688 | VLoss: 0.0200 | PLoss: -0.0080 | Ent: 0.1740
+[ 78.6%] Iter 7673/9765 | SPS:  2602 | Reward:  0.061 | Value:  0.716 | VLoss: 0.0268 | PLoss: -0.0053 | Ent: 0.2473
+[ 78.6%] Iter 7674/9765 | SPS:  2602 | Reward:  0.046 | Value:  0.679 | VLoss: 0.0269 | PLoss: -0.0054 | Ent: 0.2369
+[ 78.6%] Iter 7675/9765 | SPS:  2602 | Reward:  0.060 | Value:  0.612 | VLoss: 0.0295 | PLoss: -0.0068 | Ent: 0.2187
+[ 78.6%] Iter 7676/9765 | SPS:  2602 | Reward:  0.065 | Value:  0.767 | VLoss: 0.0164 | PLoss: -0.0028 | Ent: 0.1725
+[ 78.6%] Iter 7677/9765 | SPS:  2602 | Reward:  0.063 | Value:  0.800 | VLoss: 0.0215 | PLoss: -0.0017 | Ent: 0.2230
+[ 78.6%] Iter 7678/9765 | SPS:  2602 | Reward:  0.063 | Value:  0.755 | VLoss: 0.0165 | PLoss: -0.0039 | Ent: 0.2141
+[ 78.6%] Iter 7679/9765 | SPS:  2602 | Reward:  0.048 | Value:  0.729 | VLoss: 0.0148 | PLoss: -0.0052 | Ent: 0.1971
+[ 78.6%] Iter 7680/9765 | SPS:  2602 | Reward:  0.045 | Value:  0.699 | VLoss: 0.0177 | PLoss: -0.0072 | Ent: 0.2036
+[ 78.7%] Iter 7681/9765 | SPS:  2602 | Reward:  0.042 | Value:  0.623 | VLoss: 0.0171 | PLoss: -0.0061 | Ent: 0.1804
+[ 78.7%] Iter 7682/9765 | SPS:  2603 | Reward:  0.055 | Value:  0.636 | VLoss: 0.0130 | PLoss: -0.0029 | Ent: 0.1903
+[ 78.7%] Iter 7683/9765 | SPS:  2603 | Reward:  0.054 | Value:  0.694 | VLoss: 0.0228 | PLoss: -0.0036 | Ent: 0.2211
+[ 78.7%] Iter 7684/9765 | SPS:  2603 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0283 | PLoss: -0.0042 | Ent: 0.2442
+[ 78.7%] Iter 7685/9765 | SPS:  2603 | Reward:  0.073 | Value:  0.719 | VLoss: 0.0218 | PLoss: -0.0050 | Ent: 0.2271
+[ 78.7%] Iter 7686/9765 | SPS:  2603 | Reward:  0.068 | Value:  0.740 | VLoss: 0.0278 | PLoss: -0.0114 | Ent: 0.1915
+[ 78.7%] Iter 7687/9765 | SPS:  2603 | Reward:  0.053 | Value:  0.707 | VLoss: 0.0221 | PLoss: -0.0081 | Ent: 0.2208
+[ 78.7%] Iter 7688/9765 | SPS:  2603 | Reward:  0.065 | Value:  0.717 | VLoss: 0.0222 | PLoss: -0.0033 | Ent: 0.1944
+[ 78.7%] Iter 7689/9765 | SPS:  2603 | Reward:  0.065 | Value:  0.754 | VLoss: 0.0181 | PLoss: -0.0096 | Ent: 0.2602
+[ 78.8%] Iter 7690/9765 | SPS:  2603 | Reward:  0.051 | Value:  0.650 | VLoss: 0.0212 | PLoss: -0.0060 | Ent: 0.2406
+[ 78.8%] Iter 7691/9765 | SPS:  2603 | Reward:  0.052 | Value:  0.681 | VLoss: 0.0201 | PLoss: -0.0054 | Ent: 0.2084
+[ 78.8%] Iter 7692/9765 | SPS:  2604 | Reward:  0.075 | Value:  0.735 | VLoss: 0.0217 | PLoss: -0.0081 | Ent: 0.2377
+[ 78.8%] Iter 7693/9765 | SPS:  2604 | Reward:  0.058 | Value:  0.759 | VLoss: 0.0227 | PLoss: -0.0044 | Ent: 0.2067
+[ 78.8%] Iter 7694/9765 | SPS:  2604 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0264 | PLoss: -0.0067 | Ent: 0.2284
+[ 78.8%] Iter 7695/9765 | SPS:  2604 | Reward:  0.063 | Value:  0.705 | VLoss: 0.0291 | PLoss: -0.0011 | Ent: 0.1958
+[ 78.8%] Iter 7696/9765 | SPS:  2604 | Reward:  0.049 | Value:  0.719 | VLoss: 0.0177 | PLoss: -0.0102 | Ent: 0.2102
+[ 78.8%] Iter 7697/9765 | SPS:  2604 | Reward:  0.059 | Value:  0.722 | VLoss: 0.0211 | PLoss: -0.0018 | Ent: 0.2222
+[ 78.8%] Iter 7698/9765 | SPS:  2604 | Reward:  0.073 | Value:  0.694 | VLoss: 0.0290 | PLoss: -0.0012 | Ent: 0.2354
+[ 78.8%] Iter 7699/9765 | SPS:  2604 | Reward:  0.061 | Value:  0.682 | VLoss: 0.0201 | PLoss: -0.0097 | Ent: 0.2333
+[ 78.9%] Iter 7700/9765 | SPS:  2604 | Reward:  0.059 | Value:  0.752 | VLoss: 0.0178 | PLoss: -0.0028 | Ent: 0.2214
+[ 78.9%] Iter 7701/9765 | SPS:  2605 | Reward:  0.059 | Value:  0.693 | VLoss: 0.0209 | PLoss: -0.0051 | Ent: 0.2179
+[ 78.9%] Iter 7702/9765 | SPS:  2605 | Reward:  0.062 | Value:  0.692 | VLoss: 0.0222 | PLoss: -0.0085 | Ent: 0.2643
+[ 78.9%] Iter 7703/9765 | SPS:  2605 | Reward:  0.063 | Value:  0.770 | VLoss: 0.0217 | PLoss: -0.0063 | Ent: 0.2212
+[ 78.9%] Iter 7704/9765 | SPS:  2605 | Reward:  0.054 | Value:  0.760 | VLoss: 0.0127 | PLoss: -0.0083 | Ent: 0.2485
+[ 78.9%] Iter 7705/9765 | SPS:  2605 | Reward:  0.062 | Value:  0.770 | VLoss: 0.0293 | PLoss: -0.0051 | Ent: 0.2363
+[ 78.9%] Iter 7706/9765 | SPS:  2605 | Reward:  0.046 | Value:  0.714 | VLoss: 0.0329 | PLoss: -0.0056 | Ent: 0.1951
+[ 78.9%] Iter 7707/9765 | SPS:  2605 | Reward:  0.070 | Value:  0.631 | VLoss: 0.0205 | PLoss: -0.0005 | Ent: 0.2394
+[ 78.9%] Iter 7708/9765 | SPS:  2605 | Reward:  0.066 | Value:  0.662 | VLoss: 0.0211 | PLoss: -0.0074 | Ent: 0.2170
+[ 78.9%] Iter 7709/9765 | SPS:  2605 | Reward:  0.061 | Value:  0.771 | VLoss: 0.0316 | PLoss: -0.0039 | Ent: 0.2368
+[ 79.0%] Iter 7710/9765 | SPS:  2605 | Reward:  0.061 | Value:  0.744 | VLoss: 0.0269 | PLoss: -0.0049 | Ent: 0.1961
+[ 79.0%] Iter 7711/9765 | SPS:  2606 | Reward:  0.061 | Value:  0.691 | VLoss: 0.0273 | PLoss: -0.0098 | Ent: 0.2130
+[ 79.0%] Iter 7712/9765 | SPS:  2606 | Reward:  0.064 | Value:  0.693 | VLoss: 0.0307 | PLoss: -0.0025 | Ent: 0.2325
+[ 79.0%] Iter 7713/9765 | SPS:  2606 | Reward:  0.047 | Value:  0.610 | VLoss: 0.0218 | PLoss: -0.0086 | Ent: 0.2645
+[ 79.0%] Iter 7714/9765 | SPS:  2606 | Reward:  0.050 | Value:  0.731 | VLoss: 0.0220 | PLoss: -0.0013 | Ent: 0.2241
+[ 79.0%] Iter 7715/9765 | SPS:  2606 | Reward:  0.075 | Value:  0.712 | VLoss: 0.0228 | PLoss: -0.0037 | Ent: 0.2180
+[ 79.0%] Iter 7716/9765 | SPS:  2606 | Reward:  0.065 | Value:  0.756 | VLoss: 0.0214 | PLoss: -0.0021 | Ent: 0.2141
+[ 79.0%] Iter 7717/9765 | SPS:  2606 | Reward:  0.063 | Value:  0.704 | VLoss: 0.0253 | PLoss: -0.0083 | Ent: 0.2046
+[ 79.0%] Iter 7718/9765 | SPS:  2606 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0203 | PLoss: -0.0030 | Ent: 0.2186
+[ 79.0%] Iter 7719/9765 | SPS:  2606 | Reward:  0.062 | Value:  0.750 | VLoss: 0.0246 | PLoss: -0.0021 | Ent: 0.2499
+[ 79.1%] Iter 7720/9765 | SPS:  2606 | Reward:  0.062 | Value:  0.699 | VLoss: 0.0260 | PLoss: -0.0018 | Ent: 0.2702
+[ 79.1%] Iter 7721/9765 | SPS:  2607 | Reward:  0.059 | Value:  0.766 | VLoss: 0.0225 | PLoss: -0.0066 | Ent: 0.2296
+[ 79.1%] Iter 7722/9765 | SPS:  2607 | Reward:  0.067 | Value:  0.752 | VLoss: 0.0158 | PLoss: -0.0021 | Ent: 0.2527
+[ 79.1%] Iter 7723/9765 | SPS:  2607 | Reward:  0.063 | Value:  0.757 | VLoss: 0.0177 | PLoss: -0.0095 | Ent: 0.2237
+[ 79.1%] Iter 7724/9765 | SPS:  2607 | Reward:  0.060 | Value:  0.732 | VLoss: 0.0132 | PLoss: -0.0063 | Ent: 0.2169
+[ 79.1%] Iter 7725/9765 | SPS:  2607 | Reward:  0.058 | Value:  0.740 | VLoss: 0.0277 | PLoss: -0.0040 | Ent: 0.2450
+[ 79.1%] Iter 7726/9765 | SPS:  2607 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0241 | PLoss: -0.0046 | Ent: 0.2386
+[ 79.1%] Iter 7727/9765 | SPS:  2607 | Reward:  0.057 | Value:  0.708 | VLoss: 0.0236 | PLoss: -0.0084 | Ent: 0.2618
+[ 79.1%] Iter 7728/9765 | SPS:  2607 | Reward:  0.060 | Value:  0.714 | VLoss: 0.0247 | PLoss: -0.0043 | Ent: 0.2591
+[ 79.2%] Iter 7729/9765 | SPS:  2607 | Reward:  0.062 | Value:  0.756 | VLoss: 0.0173 | PLoss: -0.0014 | Ent: 0.2026
+[ 79.2%] Iter 7730/9765 | SPS:  2608 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0181 | PLoss: -0.0101 | Ent: 0.2450
+[ 79.2%] Iter 7731/9765 | SPS:  2608 | Reward:  0.056 | Value:  0.765 | VLoss: 0.0167 | PLoss: -0.0091 | Ent: 0.1941
+[ 79.2%] Iter 7732/9765 | SPS:  2608 | Reward:  0.065 | Value:  0.710 | VLoss: 0.0254 | PLoss: -0.0028 | Ent: 0.2358
+[ 79.2%] Iter 7733/9765 | SPS:  2608 | Reward:  0.052 | Value:  0.632 | VLoss: 0.0231 | PLoss: -0.0022 | Ent: 0.2606
+[ 79.2%] Iter 7734/9765 | SPS:  2608 | Reward:  0.059 | Value:  0.745 | VLoss: 0.0201 | PLoss: -0.0068 | Ent: 0.2551
+[ 79.2%] Iter 7735/9765 | SPS:  2608 | Reward:  0.058 | Value:  0.669 | VLoss: 0.0252 | PLoss: -0.0076 | Ent: 0.2212
+[ 79.2%] Iter 7736/9765 | SPS:  2608 | Reward:  0.055 | Value:  0.714 | VLoss: 0.0224 | PLoss: -0.0069 | Ent: 0.2368
+[ 79.2%] Iter 7737/9765 | SPS:  2608 | Reward:  0.060 | Value:  0.707 | VLoss: 0.0223 | PLoss: -0.0018 | Ent: 0.2179
+[ 79.2%] Iter 7738/9765 | SPS:  2608 | Reward:  0.059 | Value:  0.686 | VLoss: 0.0234 | PLoss: -0.0029 | Ent: 0.2435
+[ 79.3%] Iter 7739/9765 | SPS:  2608 | Reward:  0.059 | Value:  0.727 | VLoss: 0.0245 | PLoss: -0.0057 | Ent: 0.2072
+[ 79.3%] Iter 7740/9765 | SPS:  2609 | Reward:  0.062 | Value:  0.740 | VLoss: 0.0158 | PLoss: -0.0056 | Ent: 0.2507
+[ 79.3%] Iter 7741/9765 | SPS:  2609 | Reward:  0.063 | Value:  0.729 | VLoss: 0.0263 | PLoss: -0.0025 | Ent: 0.2428
+[ 79.3%] Iter 7742/9765 | SPS:  2609 | Reward:  0.066 | Value:  0.711 | VLoss: 0.0229 | PLoss: -0.0045 | Ent: 0.2305
+[ 79.3%] Iter 7743/9765 | SPS:  2609 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0213 | PLoss: -0.0084 | Ent: 0.2190
+[ 79.3%] Iter 7744/9765 | SPS:  2609 | Reward:  0.056 | Value:  0.691 | VLoss: 0.0251 | PLoss: 0.0001 | Ent: 0.2112
+[ 79.3%] Iter 7745/9765 | SPS:  2609 | Reward:  0.066 | Value:  0.750 | VLoss: 0.0192 | PLoss: -0.0035 | Ent: 0.1909
+[ 79.3%] Iter 7746/9765 | SPS:  2609 | Reward:  0.060 | Value:  0.676 | VLoss: 0.0225 | PLoss: -0.0089 | Ent: 0.2734
+[ 79.3%] Iter 7747/9765 | SPS:  2609 | Reward:  0.056 | Value:  0.690 | VLoss: 0.0181 | PLoss: -0.0056 | Ent: 0.1947
+[ 79.3%] Iter 7748/9765 | SPS:  2609 | Reward:  0.058 | Value:  0.647 | VLoss: 0.0263 | PLoss: -0.0048 | Ent: 0.2172
+[ 79.4%] Iter 7749/9765 | SPS:  2610 | Reward:  0.060 | Value:  0.686 | VLoss: 0.0232 | PLoss: -0.0119 | Ent: 0.2741
+[ 79.4%] Iter 7750/9765 | SPS:  2610 | Reward:  0.047 | Value:  0.733 | VLoss: 0.0209 | PLoss: -0.0079 | Ent: 0.2398
+[ 79.4%] Iter 7751/9765 | SPS:  2610 | Reward:  0.069 | Value:  0.700 | VLoss: 0.0195 | PLoss: -0.0082 | Ent: 0.2060
+[ 79.4%] Iter 7752/9765 | SPS:  2610 | Reward:  0.056 | Value:  0.775 | VLoss: 0.0245 | PLoss: -0.0028 | Ent: 0.2135
+[ 79.4%] Iter 7753/9765 | SPS:  2610 | Reward:  0.064 | Value:  0.709 | VLoss: 0.0199 | PLoss: -0.0057 | Ent: 0.2006
+[ 79.4%] Iter 7754/9765 | SPS:  2610 | Reward:  0.059 | Value:  0.684 | VLoss: 0.0184 | PLoss: -0.0007 | Ent: 0.2273
+[ 79.4%] Iter 7755/9765 | SPS:  2610 | Reward:  0.068 | Value:  0.733 | VLoss: 0.0257 | PLoss: -0.0045 | Ent: 0.2705
+[ 79.4%] Iter 7756/9765 | SPS:  2610 | Reward:  0.065 | Value:  0.760 | VLoss: 0.0248 | PLoss: -0.0083 | Ent: 0.2086
+[ 79.4%] Iter 7757/9765 | SPS:  2610 | Reward:  0.052 | Value:  0.626 | VLoss: 0.0273 | PLoss: -0.0113 | Ent: 0.2686
+[ 79.4%] Iter 7758/9765 | SPS:  2610 | Reward:  0.074 | Value:  0.697 | VLoss: 0.0218 | PLoss: -0.0026 | Ent: 0.2213
+[ 79.5%] Iter 7759/9765 | SPS:  2611 | Reward:  0.078 | Value:  0.723 | VLoss: 0.0221 | PLoss: -0.0038 | Ent: 0.2273
+[ 79.5%] Iter 7760/9765 | SPS:  2611 | Reward:  0.062 | Value:  0.789 | VLoss: 0.0206 | PLoss: -0.0016 | Ent: 0.1844
+[ 79.5%] Iter 7761/9765 | SPS:  2611 | Reward:  0.063 | Value:  0.736 | VLoss: 0.0227 | PLoss: -0.0063 | Ent: 0.2093
+[ 79.5%] Iter 7762/9765 | SPS:  2611 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0311 | PLoss: -0.0037 | Ent: 0.2050
+[ 79.5%] Iter 7763/9765 | SPS:  2611 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0259 | PLoss: -0.0048 | Ent: 0.2299
+[ 79.5%] Iter 7764/9765 | SPS:  2611 | Reward:  0.065 | Value:  0.680 | VLoss: 0.0291 | PLoss: -0.0056 | Ent: 0.1953
+[ 79.5%] Iter 7765/9765 | SPS:  2611 | Reward:  0.061 | Value:  0.651 | VLoss: 0.0225 | PLoss: -0.0045 | Ent: 0.2762
+[ 79.5%] Iter 7766/9765 | SPS:  2611 | Reward:  0.057 | Value:  0.693 | VLoss: 0.0218 | PLoss: -0.0076 | Ent: 0.2202
+[ 79.5%] Iter 7767/9765 | SPS:  2611 | Reward:  0.059 | Value:  0.734 | VLoss: 0.0196 | PLoss: -0.0115 | Ent: 0.2639
+[ 79.5%] Iter 7768/9765 | SPS:  2611 | Reward:  0.068 | Value:  0.676 | VLoss: 0.0182 | PLoss: -0.0079 | Ent: 0.2054
+[ 79.6%] Iter 7769/9765 | SPS:  2612 | Reward:  0.074 | Value:  0.793 | VLoss: 0.0254 | PLoss: -0.0093 | Ent: 0.2254
+[ 79.6%] Iter 7770/9765 | SPS:  2612 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0280 | PLoss: -0.0066 | Ent: 0.2527
+[ 79.6%] Iter 7771/9765 | SPS:  2612 | Reward:  0.066 | Value:  0.762 | VLoss: 0.0220 | PLoss: -0.0028 | Ent: 0.2235
+[ 79.6%] Iter 7772/9765 | SPS:  2612 | Reward:  0.068 | Value:  0.703 | VLoss: 0.0211 | PLoss: -0.0050 | Ent: 0.2147
+[ 79.6%] Iter 7773/9765 | SPS:  2612 | Reward:  0.062 | Value:  0.765 | VLoss: 0.0131 | PLoss: -0.0060 | Ent: 0.2144
+[ 79.6%] Iter 7774/9765 | SPS:  2612 | Reward:  0.059 | Value:  0.727 | VLoss: 0.0162 | PLoss: -0.0043 | Ent: 0.2640
+[ 79.6%] Iter 7775/9765 | SPS:  2612 | Reward:  0.045 | Value:  0.674 | VLoss: 0.0277 | PLoss: -0.0119 | Ent: 0.2518
+[ 79.6%] Iter 7776/9765 | SPS:  2612 | Reward:  0.060 | Value:  0.688 | VLoss: 0.0199 | PLoss: -0.0075 | Ent: 0.2475
+[ 79.6%] Iter 7777/9765 | SPS:  2612 | Reward:  0.067 | Value:  0.694 | VLoss: 0.0293 | PLoss: -0.0047 | Ent: 0.2020
+[ 79.7%] Iter 7778/9765 | SPS:  2612 | Reward:  0.054 | Value:  0.709 | VLoss: 0.0219 | PLoss: -0.0069 | Ent: 0.2489
+[ 79.7%] Iter 7779/9765 | SPS:  2613 | Reward:  0.051 | Value:  0.765 | VLoss: 0.0214 | PLoss: -0.0109 | Ent: 0.2442
+[ 79.7%] Iter 7780/9765 | SPS:  2613 | Reward:  0.072 | Value:  0.666 | VLoss: 0.0224 | PLoss: -0.0044 | Ent: 0.2261
+[ 79.7%] Iter 7781/9765 | SPS:  2613 | Reward:  0.066 | Value:  0.739 | VLoss: 0.0252 | PLoss: -0.0047 | Ent: 0.2465
+[ 79.7%] Iter 7782/9765 | SPS:  2613 | Reward:  0.061 | Value:  0.752 | VLoss: 0.0276 | PLoss: -0.0092 | Ent: 0.2676
+[ 79.7%] Iter 7783/9765 | SPS:  2613 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0229 | PLoss: -0.0017 | Ent: 0.2243
+[ 79.7%] Iter 7784/9765 | SPS:  2613 | Reward:  0.057 | Value:  0.712 | VLoss: 0.0161 | PLoss: -0.0053 | Ent: 0.1946
+[ 79.7%] Iter 7785/9765 | SPS:  2613 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0228 | PLoss: -0.0061 | Ent: 0.2243
+[ 79.7%] Iter 7786/9765 | SPS:  2613 | Reward:  0.052 | Value:  0.754 | VLoss: 0.0263 | PLoss: -0.0082 | Ent: 0.2558
+[ 79.7%] Iter 7787/9765 | SPS:  2613 | Reward:  0.069 | Value:  0.693 | VLoss: 0.0308 | PLoss: -0.0015 | Ent: 0.2331
+[ 79.8%] Iter 7788/9765 | SPS:  2613 | Reward:  0.067 | Value:  0.651 | VLoss: 0.0218 | PLoss: -0.0035 | Ent: 0.1806
+[ 79.8%] Iter 7789/9765 | SPS:  2614 | Reward:  0.055 | Value:  0.705 | VLoss: 0.0188 | PLoss: -0.0037 | Ent: 0.1977
+[ 79.8%] Iter 7790/9765 | SPS:  2614 | Reward:  0.058 | Value:  0.739 | VLoss: 0.0270 | PLoss: -0.0059 | Ent: 0.1761
+[ 79.8%] Iter 7791/9765 | SPS:  2614 | Reward:  0.056 | Value:  0.672 | VLoss: 0.0225 | PLoss: -0.0080 | Ent: 0.2165
+[ 79.8%] Iter 7792/9765 | SPS:  2614 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0251 | PLoss: -0.0053 | Ent: 0.2083
+[ 79.8%] Iter 7793/9765 | SPS:  2614 | Reward:  0.070 | Value:  0.707 | VLoss: 0.0234 | PLoss: -0.0025 | Ent: 0.2253
+[ 79.8%] Iter 7794/9765 | SPS:  2614 | Reward:  0.055 | Value:  0.763 | VLoss: 0.0162 | PLoss: -0.0042 | Ent: 0.2359
+[ 79.8%] Iter 7795/9765 | SPS:  2614 | Reward:  0.059 | Value:  0.680 | VLoss: 0.0226 | PLoss: -0.0069 | Ent: 0.2326
+[ 79.8%] Iter 7796/9765 | SPS:  2614 | Reward:  0.065 | Value:  0.721 | VLoss: 0.0343 | PLoss: -0.0027 | Ent: 0.2159
+[ 79.8%] Iter 7797/9765 | SPS:  2614 | Reward:  0.055 | Value:  0.649 | VLoss: 0.0171 | PLoss: -0.0002 | Ent: 0.2255
+[ 79.9%] Iter 7798/9765 | SPS:  2615 | Reward:  0.071 | Value:  0.682 | VLoss: 0.0187 | PLoss: -0.0073 | Ent: 0.2537
+[ 79.9%] Iter 7799/9765 | SPS:  2615 | Reward:  0.058 | Value:  0.712 | VLoss: 0.0132 | PLoss: -0.0019 | Ent: 0.2777
+[ 79.9%] Iter 7800/9765 | SPS:  2615 | Reward:  0.048 | Value:  0.687 | VLoss: 0.0213 | PLoss: -0.0055 | Ent: 0.2089
+[ 79.9%] Iter 7801/9765 | SPS:  2615 | Reward:  0.070 | Value:  0.773 | VLoss: 0.0246 | PLoss: -0.0033 | Ent: 0.1646
+[ 79.9%] Iter 7802/9765 | SPS:  2615 | Reward:  0.056 | Value:  0.788 | VLoss: 0.0228 | PLoss: -0.0021 | Ent: 0.2248
+[ 79.9%] Iter 7803/9765 | SPS:  2615 | Reward:  0.071 | Value:  0.767 | VLoss: 0.0210 | PLoss: -0.0042 | Ent: 0.2402
+[ 79.9%] Iter 7804/9765 | SPS:  2615 | Reward:  0.063 | Value:  0.794 | VLoss: 0.0184 | PLoss: -0.0053 | Ent: 0.1925
+[ 79.9%] Iter 7805/9765 | SPS:  2615 | Reward:  0.055 | Value:  0.710 | VLoss: 0.0207 | PLoss: -0.0064 | Ent: 0.2459
+[ 79.9%] Iter 7806/9765 | SPS:  2615 | Reward:  0.056 | Value:  0.669 | VLoss: 0.0236 | PLoss: -0.0056 | Ent: 0.2577
+[ 79.9%] Iter 7807/9765 | SPS:  2615 | Reward:  0.061 | Value:  0.689 | VLoss: 0.0242 | PLoss: -0.0060 | Ent: 0.2279
+[ 80.0%] Iter 7808/9765 | SPS:  2616 | Reward:  0.058 | Value:  0.745 | VLoss: 0.0125 | PLoss: -0.0057 | Ent: 0.2063
+[ 80.0%] Iter 7809/9765 | SPS:  2616 | Reward:  0.059 | Value:  0.770 | VLoss: 0.0181 | PLoss: -0.0122 | Ent: 0.2836
+[ 80.0%] Iter 7810/9765 | SPS:  2616 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0244 | PLoss: -0.0035 | Ent: 0.2459
+[ 80.0%] Iter 7811/9765 | SPS:  2616 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0164 | PLoss: -0.0040 | Ent: 0.2500
+[ 80.0%] Iter 7812/9765 | SPS:  2616 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0241 | PLoss: -0.0020 | Ent: 0.2054
+[ 80.0%] Iter 7813/9765 | SPS:  2616 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0210 | PLoss: -0.0076 | Ent: 0.2298
+[ 80.0%] Iter 7814/9765 | SPS:  2616 | Reward:  0.060 | Value:  0.754 | VLoss: 0.0169 | PLoss: -0.0052 | Ent: 0.1842
+[ 80.0%] Iter 7815/9765 | SPS:  2616 | Reward:  0.071 | Value:  0.717 | VLoss: 0.0256 | PLoss: -0.0017 | Ent: 0.1739
+[ 80.0%] Iter 7816/9765 | SPS:  2616 | Reward:  0.059 | Value:  0.783 | VLoss: 0.0201 | PLoss: -0.0074 | Ent: 0.1894
+[ 80.1%] Iter 7817/9765 | SPS:  2616 | Reward:  0.065 | Value:  0.661 | VLoss: 0.0205 | PLoss: 0.0016 | Ent: 0.2221
+[ 80.1%] Iter 7818/9765 | SPS:  2617 | Reward:  0.055 | Value:  0.720 | VLoss: 0.0139 | PLoss: -0.0069 | Ent: 0.2549
+[ 80.1%] Iter 7819/9765 | SPS:  2617 | Reward:  0.057 | Value:  0.741 | VLoss: 0.0246 | PLoss: -0.0072 | Ent: 0.2165
+[ 80.1%] Iter 7820/9765 | SPS:  2617 | Reward:  0.042 | Value:  0.732 | VLoss: 0.0136 | PLoss: -0.0003 | Ent: 0.1769
+[ 80.1%] Iter 7821/9765 | SPS:  2617 | Reward:  0.069 | Value:  0.689 | VLoss: 0.0267 | PLoss: -0.0078 | Ent: 0.2079
+[ 80.1%] Iter 7822/9765 | SPS:  2617 | Reward:  0.059 | Value:  0.719 | VLoss: 0.0215 | PLoss: -0.0067 | Ent: 0.2377
+[ 80.1%] Iter 7823/9765 | SPS:  2617 | Reward:  0.066 | Value:  0.750 | VLoss: 0.0213 | PLoss: -0.0028 | Ent: 0.1887
+[ 80.1%] Iter 7824/9765 | SPS:  2617 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0208 | PLoss: -0.0049 | Ent: 0.2587
+[ 80.1%] Iter 7825/9765 | SPS:  2617 | Reward:  0.057 | Value:  0.649 | VLoss: 0.0209 | PLoss: 0.0010 | Ent: 0.2129
+[ 80.1%] Iter 7826/9765 | SPS:  2617 | Reward:  0.063 | Value:  0.669 | VLoss: 0.0213 | PLoss: -0.0085 | Ent: 0.1940
+[ 80.2%] Iter 7827/9765 | SPS:  2617 | Reward:  0.070 | Value:  0.752 | VLoss: 0.0159 | PLoss: -0.0050 | Ent: 0.1661
+[ 80.2%] Iter 7828/9765 | SPS:  2618 | Reward:  0.067 | Value:  0.801 | VLoss: 0.0172 | PLoss: -0.0049 | Ent: 0.1989
+[ 80.2%] Iter 7829/9765 | SPS:  2618 | Reward:  0.077 | Value:  0.731 | VLoss: 0.0287 | PLoss: -0.0073 | Ent: 0.2076
+[ 80.2%] Iter 7830/9765 | SPS:  2618 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0205 | PLoss: -0.0014 | Ent: 0.2060
+[ 80.2%] Iter 7831/9765 | SPS:  2618 | Reward:  0.072 | Value:  0.741 | VLoss: 0.0297 | PLoss: -0.0041 | Ent: 0.2333
+[ 80.2%] Iter 7832/9765 | SPS:  2618 | Reward:  0.065 | Value:  0.684 | VLoss: 0.0261 | PLoss: -0.0059 | Ent: 0.2588
+[ 80.2%] Iter 7833/9765 | SPS:  2618 | Reward:  0.060 | Value:  0.725 | VLoss: 0.0229 | PLoss: -0.0071 | Ent: 0.1734
+[ 80.2%] Iter 7834/9765 | SPS:  2618 | Reward:  0.061 | Value:  0.720 | VLoss: 0.0226 | PLoss: -0.0050 | Ent: 0.2268
+[ 80.2%] Iter 7835/9765 | SPS:  2618 | Reward:  0.054 | Value:  0.727 | VLoss: 0.0163 | PLoss: -0.0102 | Ent: 0.1834
+[ 80.2%] Iter 7836/9765 | SPS:  2618 | Reward:  0.061 | Value:  0.706 | VLoss: 0.0237 | PLoss: -0.0059 | Ent: 0.1966
+[ 80.3%] Iter 7837/9765 | SPS:  2618 | Reward:  0.057 | Value:  0.669 | VLoss: 0.0200 | PLoss: -0.0050 | Ent: 0.2076
+[ 80.3%] Iter 7838/9765 | SPS:  2619 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0238 | PLoss: -0.0095 | Ent: 0.1709
+[ 80.3%] Iter 7839/9765 | SPS:  2619 | Reward:  0.067 | Value:  0.732 | VLoss: 0.0240 | PLoss: -0.0051 | Ent: 0.1821
+[ 80.3%] Iter 7840/9765 | SPS:  2619 | Reward:  0.069 | Value:  0.786 | VLoss: 0.0178 | PLoss: -0.0033 | Ent: 0.2055
+[ 80.3%] Iter 7841/9765 | SPS:  2619 | Reward:  0.067 | Value:  0.749 | VLoss: 0.0221 | PLoss: -0.0014 | Ent: 0.2014
+[ 80.3%] Iter 7842/9765 | SPS:  2619 | Reward:  0.054 | Value:  0.747 | VLoss: 0.0166 | PLoss: -0.0031 | Ent: 0.2770
+[ 80.3%] Iter 7843/9765 | SPS:  2619 | Reward:  0.057 | Value:  0.662 | VLoss: 0.0223 | PLoss: -0.0015 | Ent: 0.2135
+[ 80.3%] Iter 7844/9765 | SPS:  2619 | Reward:  0.053 | Value:  0.636 | VLoss: 0.0207 | PLoss: -0.0072 | Ent: 0.2450
+[ 80.3%] Iter 7845/9765 | SPS:  2619 | Reward:  0.045 | Value:  0.657 | VLoss: 0.0185 | PLoss: -0.0041 | Ent: 0.1916
+[ 80.3%] Iter 7846/9765 | SPS:  2619 | Reward:  0.058 | Value:  0.676 | VLoss: 0.0245 | PLoss: -0.0046 | Ent: 0.2368
+[ 80.4%] Iter 7847/9765 | SPS:  2620 | Reward:  0.045 | Value:  0.634 | VLoss: 0.0194 | PLoss: -0.0036 | Ent: 0.2202
+[ 80.4%] Iter 7848/9765 | SPS:  2620 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0183 | PLoss: -0.0062 | Ent: 0.2219
+[ 80.4%] Iter 7849/9765 | SPS:  2620 | Reward:  0.070 | Value:  0.715 | VLoss: 0.0250 | PLoss: -0.0005 | Ent: 0.2383
+[ 80.4%] Iter 7850/9765 | SPS:  2620 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0266 | PLoss: -0.0064 | Ent: 0.2359
+[ 80.4%] Iter 7851/9765 | SPS:  2620 | Reward:  0.055 | Value:  0.677 | VLoss: 0.0175 | PLoss: -0.0055 | Ent: 0.2786
+[ 80.4%] Iter 7852/9765 | SPS:  2620 | Reward:  0.061 | Value:  0.713 | VLoss: 0.0203 | PLoss: -0.0010 | Ent: 0.2255
+[ 80.4%] Iter 7853/9765 | SPS:  2620 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0137 | PLoss: -0.0073 | Ent: 0.2162
+[ 80.4%] Iter 7854/9765 | SPS:  2620 | Reward:  0.062 | Value:  0.754 | VLoss: 0.0217 | PLoss: -0.0111 | Ent: 0.2177
+[ 80.4%] Iter 7855/9765 | SPS:  2620 | Reward:  0.050 | Value:  0.688 | VLoss: 0.0163 | PLoss: -0.0072 | Ent: 0.2455
+[ 80.5%] Iter 7856/9765 | SPS:  2620 | Reward:  0.064 | Value:  0.778 | VLoss: 0.0198 | PLoss: -0.0048 | Ent: 0.1866
+[ 80.5%] Iter 7857/9765 | SPS:  2621 | Reward:  0.065 | Value:  0.724 | VLoss: 0.0236 | PLoss: -0.0027 | Ent: 0.1896
+[ 80.5%] Iter 7858/9765 | SPS:  2621 | Reward:  0.069 | Value:  0.784 | VLoss: 0.0242 | PLoss: -0.0044 | Ent: 0.2279
+[ 80.5%] Iter 7859/9765 | SPS:  2621 | Reward:  0.052 | Value:  0.750 | VLoss: 0.0249 | PLoss: -0.0084 | Ent: 0.1983
+[ 80.5%] Iter 7860/9765 | SPS:  2621 | Reward:  0.063 | Value:  0.707 | VLoss: 0.0164 | PLoss: -0.0007 | Ent: 0.2272
+[ 80.5%] Iter 7861/9765 | SPS:  2621 | Reward:  0.048 | Value:  0.708 | VLoss: 0.0174 | PLoss: -0.0115 | Ent: 0.2432
+[ 80.5%] Iter 7862/9765 | SPS:  2621 | Reward:  0.051 | Value:  0.752 | VLoss: 0.0208 | PLoss: -0.0014 | Ent: 0.2161
+[ 80.5%] Iter 7863/9765 | SPS:  2621 | Reward:  0.071 | Value:  0.731 | VLoss: 0.0209 | PLoss: -0.0070 | Ent: 0.2093
+[ 80.5%] Iter 7864/9765 | SPS:  2621 | Reward:  0.058 | Value:  0.758 | VLoss: 0.0171 | PLoss: -0.0111 | Ent: 0.2377
+[ 80.5%] Iter 7865/9765 | SPS:  2621 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0163 | PLoss: -0.0099 | Ent: 0.2483
+[ 80.6%] Iter 7866/9765 | SPS:  2621 | Reward:  0.060 | Value:  0.762 | VLoss: 0.0218 | PLoss: -0.0085 | Ent: 0.2729
+[ 80.6%] Iter 7867/9765 | SPS:  2622 | Reward:  0.055 | Value:  0.672 | VLoss: 0.0201 | PLoss: -0.0031 | Ent: 0.1969
+[ 80.6%] Iter 7868/9765 | SPS:  2622 | Reward:  0.081 | Value:  0.691 | VLoss: 0.0217 | PLoss: -0.0062 | Ent: 0.2231
+[ 80.6%] Iter 7869/9765 | SPS:  2622 | Reward:  0.062 | Value:  0.736 | VLoss: 0.0224 | PLoss: -0.0076 | Ent: 0.2318
+[ 80.6%] Iter 7870/9765 | SPS:  2622 | Reward:  0.060 | Value:  0.756 | VLoss: 0.0180 | PLoss: -0.0050 | Ent: 0.1987
+[ 80.6%] Iter 7871/9765 | SPS:  2622 | Reward:  0.058 | Value:  0.774 | VLoss: 0.0213 | PLoss: -0.0043 | Ent: 0.2166
+[ 80.6%] Iter 7872/9765 | SPS:  2622 | Reward:  0.051 | Value:  0.713 | VLoss: 0.0204 | PLoss: -0.0047 | Ent: 0.1929
+[ 80.6%] Iter 7873/9765 | SPS:  2622 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0242 | PLoss: -0.0030 | Ent: 0.1924
+[ 80.6%] Iter 7874/9765 | SPS:  2622 | Reward:  0.064 | Value:  0.765 | VLoss: 0.0199 | PLoss: -0.0005 | Ent: 0.2163
+[ 80.6%] Iter 7875/9765 | SPS:  2622 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0200 | PLoss: -0.0088 | Ent: 0.1955
+[ 80.7%] Iter 7876/9765 | SPS:  2622 | Reward:  0.060 | Value:  0.691 | VLoss: 0.0299 | PLoss: -0.0034 | Ent: 0.2583
+[ 80.7%] Iter 7877/9765 | SPS:  2623 | Reward:  0.065 | Value:  0.761 | VLoss: 0.0160 | PLoss: -0.0032 | Ent: 0.2106
+[ 80.7%] Iter 7878/9765 | SPS:  2623 | Reward:  0.072 | Value:  0.741 | VLoss: 0.0241 | PLoss: -0.0059 | Ent: 0.2041
+[ 80.7%] Iter 7879/9765 | SPS:  2623 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0194 | PLoss: -0.0018 | Ent: 0.2370
+[ 80.7%] Iter 7880/9765 | SPS:  2623 | Reward:  0.062 | Value:  0.690 | VLoss: 0.0289 | PLoss: -0.0050 | Ent: 0.2169
+[ 80.7%] Iter 7881/9765 | SPS:  2623 | Reward:  0.059 | Value:  0.693 | VLoss: 0.0179 | PLoss: -0.0027 | Ent: 0.1901
+[ 80.7%] Iter 7882/9765 | SPS:  2623 | Reward:  0.059 | Value:  0.771 | VLoss: 0.0218 | PLoss: -0.0098 | Ent: 0.2529
+[ 80.7%] Iter 7883/9765 | SPS:  2623 | Reward:  0.069 | Value:  0.716 | VLoss: 0.0192 | PLoss: -0.0113 | Ent: 0.2127
+[ 80.7%] Iter 7884/9765 | SPS:  2623 | Reward:  0.058 | Value:  0.710 | VLoss: 0.0335 | PLoss: -0.0084 | Ent: 0.1753
+[ 80.7%] Iter 7885/9765 | SPS:  2623 | Reward:  0.064 | Value:  0.709 | VLoss: 0.0216 | PLoss: -0.0015 | Ent: 0.2091
+[ 80.8%] Iter 7886/9765 | SPS:  2623 | Reward:  0.063 | Value:  0.677 | VLoss: 0.0256 | PLoss: -0.0006 | Ent: 0.2183
+[ 80.8%] Iter 7887/9765 | SPS:  2624 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0175 | PLoss: -0.0031 | Ent: 0.1952
+[ 80.8%] Iter 7888/9765 | SPS:  2624 | Reward:  0.059 | Value:  0.687 | VLoss: 0.0315 | PLoss: -0.0101 | Ent: 0.3028
+[ 80.8%] Iter 7889/9765 | SPS:  2624 | Reward:  0.044 | Value:  0.707 | VLoss: 0.0164 | PLoss: -0.0069 | Ent: 0.1983
+[ 80.8%] Iter 7890/9765 | SPS:  2624 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0214 | PLoss: -0.0078 | Ent: 0.2319
+[ 80.8%] Iter 7891/9765 | SPS:  2624 | Reward:  0.061 | Value:  0.723 | VLoss: 0.0222 | PLoss: -0.0070 | Ent: 0.2248
+[ 80.8%] Iter 7892/9765 | SPS:  2624 | Reward:  0.063 | Value:  0.755 | VLoss: 0.0249 | PLoss: -0.0082 | Ent: 0.2309
+[ 80.8%] Iter 7893/9765 | SPS:  2624 | Reward:  0.049 | Value:  0.721 | VLoss: 0.0187 | PLoss: -0.0096 | Ent: 0.2319
+[ 80.8%] Iter 7894/9765 | SPS:  2624 | Reward:  0.060 | Value:  0.644 | VLoss: 0.0247 | PLoss: -0.0044 | Ent: 0.2283
+[ 80.8%] Iter 7895/9765 | SPS:  2624 | Reward:  0.073 | Value:  0.716 | VLoss: 0.0219 | PLoss: -0.0040 | Ent: 0.2034
+[ 80.9%] Iter 7896/9765 | SPS:  2624 | Reward:  0.063 | Value:  0.804 | VLoss: 0.0188 | PLoss: -0.0007 | Ent: 0.2361
+[ 80.9%] Iter 7897/9765 | SPS:  2625 | Reward:  0.071 | Value:  0.742 | VLoss: 0.0300 | PLoss: 0.0008 | Ent: 0.2376
+[ 80.9%] Iter 7898/9765 | SPS:  2625 | Reward:  0.069 | Value:  0.742 | VLoss: 0.0238 | PLoss: -0.0108 | Ent: 0.2308
+[ 80.9%] Iter 7899/9765 | SPS:  2625 | Reward:  0.058 | Value:  0.730 | VLoss: 0.0258 | PLoss: -0.0020 | Ent: 0.2367
+[ 80.9%] Iter 7900/9765 | SPS:  2625 | Reward:  0.069 | Value:  0.697 | VLoss: 0.0179 | PLoss: -0.0044 | Ent: 0.2193
+[ 80.9%] Iter 7901/9765 | SPS:  2625 | Reward:  0.067 | Value:  0.768 | VLoss: 0.0172 | PLoss: -0.0067 | Ent: 0.2076
+[ 80.9%] Iter 7902/9765 | SPS:  2625 | Reward:  0.067 | Value:  0.739 | VLoss: 0.0202 | PLoss: -0.0030 | Ent: 0.2086
+[ 80.9%] Iter 7903/9765 | SPS:  2625 | Reward:  0.059 | Value:  0.779 | VLoss: 0.0168 | PLoss: -0.0053 | Ent: 0.2643
+[ 80.9%] Iter 7904/9765 | SPS:  2625 | Reward:  0.059 | Value:  0.738 | VLoss: 0.0234 | PLoss: -0.0049 | Ent: 0.2703
+[ 81.0%] Iter 7905/9765 | SPS:  2625 | Reward:  0.049 | Value:  0.669 | VLoss: 0.0162 | PLoss: -0.0054 | Ent: 0.2383
+[ 81.0%] Iter 7906/9765 | SPS:  2625 | Reward:  0.061 | Value:  0.723 | VLoss: 0.0224 | PLoss: 0.0019 | Ent: 0.2554
+[ 81.0%] Iter 7907/9765 | SPS:  2626 | Reward:  0.058 | Value:  0.715 | VLoss: 0.0289 | PLoss: -0.0062 | Ent: 0.2310
+[ 81.0%] Iter 7908/9765 | SPS:  2626 | Reward:  0.061 | Value:  0.737 | VLoss: 0.0157 | PLoss: 0.0006 | Ent: 0.2092
+[ 81.0%] Iter 7909/9765 | SPS:  2626 | Reward:  0.054 | Value:  0.721 | VLoss: 0.0186 | PLoss: -0.0010 | Ent: 0.1765
+[ 81.0%] Iter 7910/9765 | SPS:  2626 | Reward:  0.057 | Value:  0.750 | VLoss: 0.0207 | PLoss: -0.0077 | Ent: 0.2217
+[ 81.0%] Iter 7911/9765 | SPS:  2626 | Reward:  0.079 | Value:  0.703 | VLoss: 0.0190 | PLoss: -0.0085 | Ent: 0.2123
+[ 81.0%] Iter 7912/9765 | SPS:  2626 | Reward:  0.064 | Value:  0.748 | VLoss: 0.0272 | PLoss: -0.0029 | Ent: 0.2170
+[ 81.0%] Iter 7913/9765 | SPS:  2626 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0256 | PLoss: -0.0050 | Ent: 0.1807
+[ 81.0%] Iter 7914/9765 | SPS:  2626 | Reward:  0.051 | Value:  0.638 | VLoss: 0.0329 | PLoss: -0.0067 | Ent: 0.2536
+[ 81.1%] Iter 7915/9765 | SPS:  2626 | Reward:  0.056 | Value:  0.734 | VLoss: 0.0157 | PLoss: -0.0055 | Ent: 0.2184
+[ 81.1%] Iter 7916/9765 | SPS:  2626 | Reward:  0.063 | Value:  0.738 | VLoss: 0.0225 | PLoss: -0.0081 | Ent: 0.2339
+[ 81.1%] Iter 7917/9765 | SPS:  2627 | Reward:  0.071 | Value:  0.716 | VLoss: 0.0204 | PLoss: -0.0075 | Ent: 0.1973
+[ 81.1%] Iter 7918/9765 | SPS:  2627 | Reward:  0.055 | Value:  0.744 | VLoss: 0.0180 | PLoss: -0.0037 | Ent: 0.1560
+[ 81.1%] Iter 7919/9765 | SPS:  2627 | Reward:  0.057 | Value:  0.696 | VLoss: 0.0174 | PLoss: -0.0076 | Ent: 0.2347
+[ 81.1%] Iter 7920/9765 | SPS:  2627 | Reward:  0.075 | Value:  0.774 | VLoss: 0.0264 | PLoss: -0.0072 | Ent: 0.2272
+[ 81.1%] Iter 7921/9765 | SPS:  2627 | Reward:  0.058 | Value:  0.784 | VLoss: 0.0163 | PLoss: -0.0021 | Ent: 0.2224
+[ 81.1%] Iter 7922/9765 | SPS:  2627 | Reward:  0.053 | Value:  0.619 | VLoss: 0.0237 | PLoss: -0.0039 | Ent: 0.2611
+[ 81.1%] Iter 7923/9765 | SPS:  2627 | Reward:  0.050 | Value:  0.607 | VLoss: 0.0221 | PLoss: -0.0041 | Ent: 0.1971
+[ 81.1%] Iter 7924/9765 | SPS:  2627 | Reward:  0.048 | Value:  0.650 | VLoss: 0.0162 | PLoss: -0.0002 | Ent: 0.2502
+[ 81.2%] Iter 7925/9765 | SPS:  2627 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0169 | PLoss: -0.0047 | Ent: 0.1980
+[ 81.2%] Iter 7926/9765 | SPS:  2627 | Reward:  0.057 | Value:  0.684 | VLoss: 0.0266 | PLoss: -0.0030 | Ent: 0.1733
+[ 81.2%] Iter 7927/9765 | SPS:  2627 | Reward:  0.054 | Value:  0.726 | VLoss: 0.0181 | PLoss: -0.0024 | Ent: 0.2146
+[ 81.2%] Iter 7928/9765 | SPS:  2628 | Reward:  0.053 | Value:  0.748 | VLoss: 0.0190 | PLoss: -0.0087 | Ent: 0.2020
+[ 81.2%] Iter 7929/9765 | SPS:  2628 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0273 | PLoss: -0.0019 | Ent: 0.2510
+[ 81.2%] Iter 7930/9765 | SPS:  2628 | Reward:  0.048 | Value:  0.667 | VLoss: 0.0181 | PLoss: -0.0046 | Ent: 0.1992
+[ 81.2%] Iter 7931/9765 | SPS:  2628 | Reward:  0.055 | Value:  0.611 | VLoss: 0.0296 | PLoss: -0.0044 | Ent: 0.2267
+[ 81.2%] Iter 7932/9765 | SPS:  2628 | Reward:  0.061 | Value:  0.785 | VLoss: 0.0152 | PLoss: -0.0056 | Ent: 0.1955
+[ 81.2%] Iter 7933/9765 | SPS:  2628 | Reward:  0.057 | Value:  0.698 | VLoss: 0.0243 | PLoss: -0.0013 | Ent: 0.2478
+[ 81.2%] Iter 7934/9765 | SPS:  2628 | Reward:  0.072 | Value:  0.708 | VLoss: 0.0255 | PLoss: -0.0057 | Ent: 0.2114
+[ 81.3%] Iter 7935/9765 | SPS:  2628 | Reward:  0.044 | Value:  0.767 | VLoss: 0.0131 | PLoss: -0.0024 | Ent: 0.2507
+[ 81.3%] Iter 7936/9765 | SPS:  2628 | Reward:  0.059 | Value:  0.724 | VLoss: 0.0244 | PLoss: -0.0012 | Ent: 0.2200
+[ 81.3%] Iter 7937/9765 | SPS:  2629 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0169 | PLoss: -0.0029 | Ent: 0.2053
+[ 81.3%] Iter 7938/9765 | SPS:  2629 | Reward:  0.052 | Value:  0.674 | VLoss: 0.0192 | PLoss: -0.0050 | Ent: 0.1982
+[ 81.3%] Iter 7939/9765 | SPS:  2629 | Reward:  0.060 | Value:  0.717 | VLoss: 0.0253 | PLoss: -0.0046 | Ent: 0.1876
+[ 81.3%] Iter 7940/9765 | SPS:  2629 | Reward:  0.054 | Value:  0.739 | VLoss: 0.0233 | PLoss: -0.0057 | Ent: 0.2434
+[ 81.3%] Iter 7941/9765 | SPS:  2629 | Reward:  0.048 | Value:  0.634 | VLoss: 0.0248 | PLoss: -0.0012 | Ent: 0.2475
+[ 81.3%] Iter 7942/9765 | SPS:  2629 | Reward:  0.052 | Value:  0.681 | VLoss: 0.0214 | PLoss: -0.0025 | Ent: 0.2181
+[ 81.3%] Iter 7943/9765 | SPS:  2629 | Reward:  0.059 | Value:  0.635 | VLoss: 0.0242 | PLoss: -0.0055 | Ent: 0.2312
+[ 81.4%] Iter 7944/9765 | SPS:  2629 | Reward:  0.054 | Value:  0.667 | VLoss: 0.0266 | PLoss: -0.0073 | Ent: 0.1884
+[ 81.4%] Iter 7945/9765 | SPS:  2629 | Reward:  0.061 | Value:  0.671 | VLoss: 0.0189 | PLoss: -0.0032 | Ent: 0.2218
+[ 81.4%] Iter 7946/9765 | SPS:  2629 | Reward:  0.062 | Value:  0.736 | VLoss: 0.0239 | PLoss: -0.0025 | Ent: 0.2219
+[ 81.4%] Iter 7947/9765 | SPS:  2630 | Reward:  0.058 | Value:  0.741 | VLoss: 0.0207 | PLoss: -0.0008 | Ent: 0.1814
+[ 81.4%] Iter 7948/9765 | SPS:  2630 | Reward:  0.072 | Value:  0.760 | VLoss: 0.0251 | PLoss: -0.0068 | Ent: 0.1888
+[ 81.4%] Iter 7949/9765 | SPS:  2630 | Reward:  0.057 | Value:  0.755 | VLoss: 0.0209 | PLoss: -0.0040 | Ent: 0.1780
+[ 81.4%] Iter 7950/9765 | SPS:  2630 | Reward:  0.065 | Value:  0.743 | VLoss: 0.0206 | PLoss: -0.0074 | Ent: 0.2264
+[ 81.4%] Iter 7951/9765 | SPS:  2630 | Reward:  0.058 | Value:  0.730 | VLoss: 0.0169 | PLoss: -0.0089 | Ent: 0.2187
+[ 81.4%] Iter 7952/9765 | SPS:  2630 | Reward:  0.045 | Value:  0.752 | VLoss: 0.0213 | PLoss: -0.0029 | Ent: 0.2213
+[ 81.4%] Iter 7953/9765 | SPS:  2630 | Reward:  0.055 | Value:  0.639 | VLoss: 0.0250 | PLoss: -0.0070 | Ent: 0.2150
+[ 81.5%] Iter 7954/9765 | SPS:  2630 | Reward:  0.069 | Value:  0.685 | VLoss: 0.0179 | PLoss: -0.0071 | Ent: 0.1990
+[ 81.5%] Iter 7955/9765 | SPS:  2630 | Reward:  0.064 | Value:  0.750 | VLoss: 0.0243 | PLoss: -0.0040 | Ent: 0.1934
+[ 81.5%] Iter 7956/9765 | SPS:  2630 | Reward:  0.063 | Value:  0.740 | VLoss: 0.0236 | PLoss: -0.0038 | Ent: 0.1964
+[ 81.5%] Iter 7957/9765 | SPS:  2631 | Reward:  0.056 | Value:  0.714 | VLoss: 0.0172 | PLoss: -0.0039 | Ent: 0.1972
+[ 81.5%] Iter 7958/9765 | SPS:  2631 | Reward:  0.064 | Value:  0.697 | VLoss: 0.0185 | PLoss: -0.0050 | Ent: 0.2398
+[ 81.5%] Iter 7959/9765 | SPS:  2631 | Reward:  0.056 | Value:  0.749 | VLoss: 0.0171 | PLoss: -0.0042 | Ent: 0.2207
+[ 81.5%] Iter 7960/9765 | SPS:  2631 | Reward:  0.070 | Value:  0.740 | VLoss: 0.0134 | PLoss: -0.0047 | Ent: 0.2150
+[ 81.5%] Iter 7961/9765 | SPS:  2631 | Reward:  0.048 | Value:  0.744 | VLoss: 0.0267 | PLoss: -0.0039 | Ent: 0.2652
+[ 81.5%] Iter 7962/9765 | SPS:  2631 | Reward:  0.051 | Value:  0.650 | VLoss: 0.0143 | PLoss: -0.0029 | Ent: 0.2300
+[ 81.5%] Iter 7963/9765 | SPS:  2631 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0238 | PLoss: -0.0051 | Ent: 0.2149
+[ 81.6%] Iter 7964/9765 | SPS:  2631 | Reward:  0.048 | Value:  0.746 | VLoss: 0.0162 | PLoss: -0.0037 | Ent: 0.2363
+[ 81.6%] Iter 7965/9765 | SPS:  2631 | Reward:  0.059 | Value:  0.719 | VLoss: 0.0207 | PLoss: -0.0026 | Ent: 0.2125
+[ 81.6%] Iter 7966/9765 | SPS:  2631 | Reward:  0.060 | Value:  0.734 | VLoss: 0.0276 | PLoss: -0.0015 | Ent: 0.2032
+[ 81.6%] Iter 7967/9765 | SPS:  2632 | Reward:  0.080 | Value:  0.741 | VLoss: 0.0181 | PLoss: -0.0086 | Ent: 0.2047
+[ 81.6%] Iter 7968/9765 | SPS:  2632 | Reward:  0.065 | Value:  0.741 | VLoss: 0.0199 | PLoss: -0.0018 | Ent: 0.2180
+[ 81.6%] Iter 7969/9765 | SPS:  2632 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0212 | PLoss: -0.0033 | Ent: 0.2031
+[ 81.6%] Iter 7970/9765 | SPS:  2632 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0260 | PLoss: -0.0032 | Ent: 0.2494
+[ 81.6%] Iter 7971/9765 | SPS:  2632 | Reward:  0.046 | Value:  0.674 | VLoss: 0.0170 | PLoss: -0.0026 | Ent: 0.2293
+[ 81.6%] Iter 7972/9765 | SPS:  2632 | Reward:  0.054 | Value:  0.669 | VLoss: 0.0143 | PLoss: -0.0053 | Ent: 0.2547
+[ 81.6%] Iter 7973/9765 | SPS:  2632 | Reward:  0.056 | Value:  0.683 | VLoss: 0.0248 | PLoss: -0.0090 | Ent: 0.2122
+[ 81.7%] Iter 7974/9765 | SPS:  2632 | Reward:  0.045 | Value:  0.748 | VLoss: 0.0182 | PLoss: -0.0022 | Ent: 0.1823
+[ 81.7%] Iter 7975/9765 | SPS:  2632 | Reward:  0.061 | Value:  0.721 | VLoss: 0.0188 | PLoss: -0.0018 | Ent: 0.2056
+[ 81.7%] Iter 7976/9765 | SPS:  2633 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0177 | PLoss: 0.0006 | Ent: 0.2060
+[ 81.7%] Iter 7977/9765 | SPS:  2633 | Reward:  0.072 | Value:  0.756 | VLoss: 0.0191 | PLoss: -0.0037 | Ent: 0.2374
+[ 81.7%] Iter 7978/9765 | SPS:  2633 | Reward:  0.066 | Value:  0.794 | VLoss: 0.0196 | PLoss: -0.0041 | Ent: 0.2160
+[ 81.7%] Iter 7979/9765 | SPS:  2633 | Reward:  0.073 | Value:  0.751 | VLoss: 0.0206 | PLoss: -0.0032 | Ent: 0.2248
+[ 81.7%] Iter 7980/9765 | SPS:  2633 | Reward:  0.050 | Value:  0.782 | VLoss: 0.0159 | PLoss: -0.0048 | Ent: 0.2527
+[ 81.7%] Iter 7981/9765 | SPS:  2633 | Reward:  0.069 | Value:  0.717 | VLoss: 0.0311 | PLoss: -0.0052 | Ent: 0.2134
+[ 81.7%] Iter 7982/9765 | SPS:  2633 | Reward:  0.060 | Value:  0.689 | VLoss: 0.0252 | PLoss: -0.0016 | Ent: 0.2061
+[ 81.8%] Iter 7983/9765 | SPS:  2633 | Reward:  0.065 | Value:  0.676 | VLoss: 0.0236 | PLoss: -0.0057 | Ent: 0.2364
+[ 81.8%] Iter 7984/9765 | SPS:  2633 | Reward:  0.046 | Value:  0.738 | VLoss: 0.0150 | PLoss: -0.0100 | Ent: 0.2883
+[ 81.8%] Iter 7985/9765 | SPS:  2633 | Reward:  0.057 | Value:  0.694 | VLoss: 0.0251 | PLoss: -0.0090 | Ent: 0.2027
+[ 81.8%] Iter 7986/9765 | SPS:  2633 | Reward:  0.069 | Value:  0.736 | VLoss: 0.0198 | PLoss: -0.0064 | Ent: 0.1720
+[ 81.8%] Iter 7987/9765 | SPS:  2634 | Reward:  0.066 | Value:  0.682 | VLoss: 0.0230 | PLoss: -0.0074 | Ent: 0.1852
+[ 81.8%] Iter 7988/9765 | SPS:  2634 | Reward:  0.056 | Value:  0.713 | VLoss: 0.0277 | PLoss: -0.0029 | Ent: 0.2063
+[ 81.8%] Iter 7989/9765 | SPS:  2634 | Reward:  0.063 | Value:  0.704 | VLoss: 0.0189 | PLoss: -0.0058 | Ent: 0.1931
+[ 81.8%] Iter 7990/9765 | SPS:  2634 | Reward:  0.075 | Value:  0.730 | VLoss: 0.0264 | PLoss: -0.0072 | Ent: 0.1916
+[ 81.8%] Iter 7991/9765 | SPS:  2634 | Reward:  0.062 | Value:  0.797 | VLoss: 0.0179 | PLoss: -0.0035 | Ent: 0.2056
+[ 81.8%] Iter 7992/9765 | SPS:  2634 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0174 | PLoss: -0.0058 | Ent: 0.2094
+[ 81.9%] Iter 7993/9765 | SPS:  2634 | Reward:  0.061 | Value:  0.700 | VLoss: 0.0241 | PLoss: -0.0026 | Ent: 0.2059
+[ 81.9%] Iter 7994/9765 | SPS:  2634 | Reward:  0.052 | Value:  0.722 | VLoss: 0.0163 | PLoss: -0.0014 | Ent: 0.1815
+[ 81.9%] Iter 7995/9765 | SPS:  2634 | Reward:  0.070 | Value:  0.705 | VLoss: 0.0202 | PLoss: -0.0071 | Ent: 0.2413
+[ 81.9%] Iter 7996/9765 | SPS:  2634 | Reward:  0.062 | Value:  0.758 | VLoss: 0.0251 | PLoss: -0.0030 | Ent: 0.2178
+[ 81.9%] Iter 7997/9765 | SPS:  2635 | Reward:  0.068 | Value:  0.754 | VLoss: 0.0129 | PLoss: 0.0000 | Ent: 0.2177
+[ 81.9%] Iter 7998/9765 | SPS:  2635 | Reward:  0.069 | Value:  0.701 | VLoss: 0.0231 | PLoss: -0.0047 | Ent: 0.1751
+[ 81.9%] Iter 7999/9765 | SPS:  2635 | Reward:  0.049 | Value:  0.656 | VLoss: 0.0192 | PLoss: -0.0043 | Ent: 0.2431
+[ 81.9%] Iter 8000/9765 | SPS:  2635 | Reward:  0.056 | Value:  0.760 | VLoss: 0.0269 | PLoss: -0.0078 | Ent: 0.1999
+[ 81.9%] Iter 8001/9765 | SPS:  2635 | Reward:  0.066 | Value:  0.681 | VLoss: 0.0212 | PLoss: -0.0058 | Ent: 0.2008
+[ 81.9%] Iter 8002/9765 | SPS:  2635 | Reward:  0.071 | Value:  0.778 | VLoss: 0.0174 | PLoss: -0.0057 | Ent: 0.2273
+[ 82.0%] Iter 8003/9765 | SPS:  2635 | Reward:  0.046 | Value:  0.684 | VLoss: 0.0196 | PLoss: -0.0092 | Ent: 0.2474
+[ 82.0%] Iter 8004/9765 | SPS:  2635 | Reward:  0.060 | Value:  0.722 | VLoss: 0.0264 | PLoss: -0.0072 | Ent: 0.2645
+[ 82.0%] Iter 8005/9765 | SPS:  2635 | Reward:  0.049 | Value:  0.675 | VLoss: 0.0207 | PLoss: -0.0074 | Ent: 0.2284
+[ 82.0%] Iter 8006/9765 | SPS:  2635 | Reward:  0.054 | Value:  0.675 | VLoss: 0.0245 | PLoss: -0.0098 | Ent: 0.1730
+[ 82.0%] Iter 8007/9765 | SPS:  2636 | Reward:  0.063 | Value:  0.726 | VLoss: 0.0154 | PLoss: -0.0036 | Ent: 0.2090
+[ 82.0%] Iter 8008/9765 | SPS:  2636 | Reward:  0.064 | Value:  0.738 | VLoss: 0.0268 | PLoss: -0.0073 | Ent: 0.2077
+[ 82.0%] Iter 8009/9765 | SPS:  2636 | Reward:  0.062 | Value:  0.777 | VLoss: 0.0318 | PLoss: -0.0084 | Ent: 0.2649
+[ 82.0%] Iter 8010/9765 | SPS:  2636 | Reward:  0.063 | Value:  0.781 | VLoss: 0.0301 | PLoss: -0.0115 | Ent: 0.1809
+[ 82.0%] Iter 8011/9765 | SPS:  2636 | Reward:  0.060 | Value:  0.682 | VLoss: 0.0247 | PLoss: -0.0134 | Ent: 0.2312
+[ 82.0%] Iter 8012/9765 | SPS:  2636 | Reward:  0.071 | Value:  0.703 | VLoss: 0.0195 | PLoss: -0.0062 | Ent: 0.2171
+[ 82.1%] Iter 8013/9765 | SPS:  2636 | Reward:  0.051 | Value:  0.752 | VLoss: 0.0229 | PLoss: -0.0054 | Ent: 0.2424
+[ 82.1%] Iter 8014/9765 | SPS:  2636 | Reward:  0.075 | Value:  0.713 | VLoss: 0.0341 | PLoss: -0.0083 | Ent: 0.1965
+[ 82.1%] Iter 8015/9765 | SPS:  2636 | Reward:  0.066 | Value:  0.742 | VLoss: 0.0310 | PLoss: -0.0113 | Ent: 0.2383
+[ 82.1%] Iter 8016/9765 | SPS:  2636 | Reward:  0.072 | Value:  0.676 | VLoss: 0.0276 | PLoss: -0.0016 | Ent: 0.2617
+[ 82.1%] Iter 8017/9765 | SPS:  2636 | Reward:  0.055 | Value:  0.674 | VLoss: 0.0233 | PLoss: -0.0011 | Ent: 0.2143
+[ 82.1%] Iter 8018/9765 | SPS:  2637 | Reward:  0.066 | Value:  0.647 | VLoss: 0.0263 | PLoss: -0.0010 | Ent: 0.2367
+[ 82.1%] Iter 8019/9765 | SPS:  2637 | Reward:  0.057 | Value:  0.776 | VLoss: 0.0151 | PLoss: -0.0075 | Ent: 0.1814
+[ 82.1%] Iter 8020/9765 | SPS:  2637 | Reward:  0.053 | Value:  0.726 | VLoss: 0.0178 | PLoss: -0.0067 | Ent: 0.1829
+[ 82.1%] Iter 8021/9765 | SPS:  2637 | Reward:  0.053 | Value:  0.670 | VLoss: 0.0184 | PLoss: -0.0053 | Ent: 0.2330
+[ 82.2%] Iter 8022/9765 | SPS:  2637 | Reward:  0.049 | Value:  0.681 | VLoss: 0.0228 | PLoss: -0.0074 | Ent: 0.2251
+[ 82.2%] Iter 8023/9765 | SPS:  2637 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0180 | PLoss: -0.0049 | Ent: 0.2078
+[ 82.2%] Iter 8024/9765 | SPS:  2637 | Reward:  0.067 | Value:  0.753 | VLoss: 0.0173 | PLoss: -0.0018 | Ent: 0.1980
+[ 82.2%] Iter 8025/9765 | SPS:  2637 | Reward:  0.056 | Value:  0.738 | VLoss: 0.0311 | PLoss: -0.0010 | Ent: 0.1777
+[ 82.2%] Iter 8026/9765 | SPS:  2637 | Reward:  0.063 | Value:  0.725 | VLoss: 0.0227 | PLoss: -0.0030 | Ent: 0.2520
+[ 82.2%] Iter 8027/9765 | SPS:  2637 | Reward:  0.064 | Value:  0.732 | VLoss: 0.0207 | PLoss: -0.0052 | Ent: 0.2028
+[ 82.2%] Iter 8028/9765 | SPS:  2638 | Reward:  0.049 | Value:  0.651 | VLoss: 0.0148 | PLoss: -0.0039 | Ent: 0.2268
+[ 82.2%] Iter 8029/9765 | SPS:  2638 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0237 | PLoss: -0.0015 | Ent: 0.2336
+[ 82.2%] Iter 8030/9765 | SPS:  2638 | Reward:  0.077 | Value:  0.738 | VLoss: 0.0274 | PLoss: -0.0045 | Ent: 0.2058
+[ 82.2%] Iter 8031/9765 | SPS:  2638 | Reward:  0.067 | Value:  0.723 | VLoss: 0.0226 | PLoss: -0.0038 | Ent: 0.2368
+[ 82.3%] Iter 8032/9765 | SPS:  2638 | Reward:  0.055 | Value:  0.739 | VLoss: 0.0235 | PLoss: 0.0005 | Ent: 0.2154
+[ 82.3%] Iter 8033/9765 | SPS:  2638 | Reward:  0.074 | Value:  0.738 | VLoss: 0.0225 | PLoss: -0.0034 | Ent: 0.1948
+[ 82.3%] Iter 8034/9765 | SPS:  2638 | Reward:  0.058 | Value:  0.759 | VLoss: 0.0170 | PLoss: -0.0076 | Ent: 0.2329
+[ 82.3%] Iter 8035/9765 | SPS:  2638 | Reward:  0.055 | Value:  0.762 | VLoss: 0.0173 | PLoss: -0.0040 | Ent: 0.2710
+[ 82.3%] Iter 8036/9765 | SPS:  2638 | Reward:  0.050 | Value:  0.704 | VLoss: 0.0268 | PLoss: 0.0005 | Ent: 0.2183
+[ 82.3%] Iter 8037/9765 | SPS:  2638 | Reward:  0.064 | Value:  0.664 | VLoss: 0.0227 | PLoss: -0.0023 | Ent: 0.2011
+[ 82.3%] Iter 8038/9765 | SPS:  2638 | Reward:  0.060 | Value:  0.754 | VLoss: 0.0229 | PLoss: -0.0051 | Ent: 0.2670
+[ 82.3%] Iter 8039/9765 | SPS:  2639 | Reward:  0.056 | Value:  0.752 | VLoss: 0.0190 | PLoss: -0.0037 | Ent: 0.1855
+[ 82.3%] Iter 8040/9765 | SPS:  2639 | Reward:  0.072 | Value:  0.702 | VLoss: 0.0227 | PLoss: -0.0080 | Ent: 0.2386
+[ 82.3%] Iter 8041/9765 | SPS:  2639 | Reward:  0.050 | Value:  0.661 | VLoss: 0.0169 | PLoss: -0.0083 | Ent: 0.2488
+[ 82.4%] Iter 8042/9765 | SPS:  2639 | Reward:  0.052 | Value:  0.780 | VLoss: 0.0128 | PLoss: -0.0026 | Ent: 0.2041
+[ 82.4%] Iter 8043/9765 | SPS:  2639 | Reward:  0.063 | Value:  0.810 | VLoss: 0.0195 | PLoss: -0.0025 | Ent: 0.2327
+[ 82.4%] Iter 8044/9765 | SPS:  2639 | Reward:  0.064 | Value:  0.757 | VLoss: 0.0196 | PLoss: -0.0043 | Ent: 0.2044
+[ 82.4%] Iter 8045/9765 | SPS:  2639 | Reward:  0.059 | Value:  0.715 | VLoss: 0.0220 | PLoss: -0.0008 | Ent: 0.2579
+[ 82.4%] Iter 8046/9765 | SPS:  2639 | Reward:  0.044 | Value:  0.689 | VLoss: 0.0163 | PLoss: -0.0036 | Ent: 0.2214
+[ 82.4%] Iter 8047/9765 | SPS:  2639 | Reward:  0.056 | Value:  0.697 | VLoss: 0.0229 | PLoss: -0.0046 | Ent: 0.2158
+[ 82.4%] Iter 8048/9765 | SPS:  2639 | Reward:  0.063 | Value:  0.760 | VLoss: 0.0203 | PLoss: -0.0048 | Ent: 0.2037
+[ 82.4%] Iter 8049/9765 | SPS:  2640 | Reward:  0.054 | Value:  0.692 | VLoss: 0.0203 | PLoss: -0.0029 | Ent: 0.1933
+[ 82.4%] Iter 8050/9765 | SPS:  2640 | Reward:  0.054 | Value:  0.640 | VLoss: 0.0272 | PLoss: -0.0042 | Ent: 0.2071
+[ 82.4%] Iter 8051/9765 | SPS:  2640 | Reward:  0.052 | Value:  0.668 | VLoss: 0.0175 | PLoss: -0.0051 | Ent: 0.2250
+[ 82.5%] Iter 8052/9765 | SPS:  2640 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0293 | PLoss: -0.0088 | Ent: 0.2112
+[ 82.5%] Iter 8053/9765 | SPS:  2640 | Reward:  0.053 | Value:  0.675 | VLoss: 0.0253 | PLoss: -0.0049 | Ent: 0.1996
+[ 82.5%] Iter 8054/9765 | SPS:  2640 | Reward:  0.064 | Value:  0.685 | VLoss: 0.0295 | PLoss: -0.0060 | Ent: 0.2182
+[ 82.5%] Iter 8055/9765 | SPS:  2640 | Reward:  0.058 | Value:  0.767 | VLoss: 0.0124 | PLoss: -0.0089 | Ent: 0.2461
+[ 82.5%] Iter 8056/9765 | SPS:  2640 | Reward:  0.048 | Value:  0.706 | VLoss: 0.0209 | PLoss: -0.0087 | Ent: 0.1964
+[ 82.5%] Iter 8057/9765 | SPS:  2640 | Reward:  0.055 | Value:  0.678 | VLoss: 0.0239 | PLoss: -0.0043 | Ent: 0.2057
+[ 82.5%] Iter 8058/9765 | SPS:  2640 | Reward:  0.058 | Value:  0.720 | VLoss: 0.0267 | PLoss: -0.0087 | Ent: 0.2224
+[ 82.5%] Iter 8059/9765 | SPS:  2641 | Reward:  0.063 | Value:  0.713 | VLoss: 0.0176 | PLoss: -0.0033 | Ent: 0.2197
+[ 82.5%] Iter 8060/9765 | SPS:  2641 | Reward:  0.051 | Value:  0.750 | VLoss: 0.0154 | PLoss: -0.0075 | Ent: 0.2198
+[ 82.5%] Iter 8061/9765 | SPS:  2641 | Reward:  0.055 | Value:  0.697 | VLoss: 0.0220 | PLoss: -0.0056 | Ent: 0.2485
+[ 82.6%] Iter 8062/9765 | SPS:  2641 | Reward:  0.050 | Value:  0.642 | VLoss: 0.0181 | PLoss: -0.0039 | Ent: 0.1944
+[ 82.6%] Iter 8063/9765 | SPS:  2641 | Reward:  0.064 | Value:  0.767 | VLoss: 0.0212 | PLoss: -0.0038 | Ent: 0.2191
+[ 82.6%] Iter 8064/9765 | SPS:  2641 | Reward:  0.062 | Value:  0.797 | VLoss: 0.0148 | PLoss: -0.0035 | Ent: 0.1669
+[ 82.6%] Iter 8065/9765 | SPS:  2641 | Reward:  0.051 | Value:  0.698 | VLoss: 0.0206 | PLoss: -0.0048 | Ent: 0.2067
+[ 82.6%] Iter 8066/9765 | SPS:  2641 | Reward:  0.049 | Value:  0.731 | VLoss: 0.0170 | PLoss: -0.0092 | Ent: 0.2282
+[ 82.6%] Iter 8067/9765 | SPS:  2641 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0184 | PLoss: -0.0056 | Ent: 0.2697
+[ 82.6%] Iter 8068/9765 | SPS:  2641 | Reward:  0.048 | Value:  0.710 | VLoss: 0.0204 | PLoss: -0.0051 | Ent: 0.2601
+[ 82.6%] Iter 8069/9765 | SPS:  2642 | Reward:  0.061 | Value:  0.685 | VLoss: 0.0181 | PLoss: -0.0037 | Ent: 0.2311
+[ 82.6%] Iter 8070/9765 | SPS:  2642 | Reward:  0.057 | Value:  0.668 | VLoss: 0.0212 | PLoss: -0.0040 | Ent: 0.2160
+[ 82.7%] Iter 8071/9765 | SPS:  2642 | Reward:  0.054 | Value:  0.678 | VLoss: 0.0172 | PLoss: -0.0016 | Ent: 0.2022
+[ 82.7%] Iter 8072/9765 | SPS:  2642 | Reward:  0.060 | Value:  0.626 | VLoss: 0.0186 | PLoss: -0.0066 | Ent: 0.2666
+[ 82.7%] Iter 8073/9765 | SPS:  2642 | Reward:  0.065 | Value:  0.719 | VLoss: 0.0224 | PLoss: -0.0045 | Ent: 0.2376
+[ 82.7%] Iter 8074/9765 | SPS:  2642 | Reward:  0.043 | Value:  0.706 | VLoss: 0.0132 | PLoss: -0.0038 | Ent: 0.2727
+[ 82.7%] Iter 8075/9765 | SPS:  2642 | Reward:  0.050 | Value:  0.739 | VLoss: 0.0192 | PLoss: -0.0048 | Ent: 0.1697
+[ 82.7%] Iter 8076/9765 | SPS:  2642 | Reward:  0.040 | Value:  0.597 | VLoss: 0.0213 | PLoss: -0.0157 | Ent: 0.3160
+[ 82.7%] Iter 8077/9765 | SPS:  2642 | Reward:  0.064 | Value:  0.661 | VLoss: 0.0230 | PLoss: -0.0048 | Ent: 0.2194
+[ 82.7%] Iter 8078/9765 | SPS:  2642 | Reward:  0.056 | Value:  0.720 | VLoss: 0.0247 | PLoss: -0.0028 | Ent: 0.1682
+[ 82.7%] Iter 8079/9765 | SPS:  2643 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0243 | PLoss: -0.0033 | Ent: 0.2263
+[ 82.7%] Iter 8080/9765 | SPS:  2643 | Reward:  0.054 | Value:  0.776 | VLoss: 0.0226 | PLoss: -0.0035 | Ent: 0.1892
+[ 82.8%] Iter 8081/9765 | SPS:  2643 | Reward:  0.059 | Value:  0.761 | VLoss: 0.0242 | PLoss: -0.0061 | Ent: 0.2018
+[ 82.8%] Iter 8082/9765 | SPS:  2643 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0188 | PLoss: -0.0033 | Ent: 0.1915
+[ 82.8%] Iter 8083/9765 | SPS:  2643 | Reward:  0.070 | Value:  0.745 | VLoss: 0.0268 | PLoss: -0.0063 | Ent: 0.2138
+[ 82.8%] Iter 8084/9765 | SPS:  2643 | Reward:  0.052 | Value:  0.753 | VLoss: 0.0251 | PLoss: -0.0015 | Ent: 0.2503
+[ 82.8%] Iter 8085/9765 | SPS:  2643 | Reward:  0.059 | Value:  0.733 | VLoss: 0.0229 | PLoss: -0.0058 | Ent: 0.2146
+[ 82.8%] Iter 8086/9765 | SPS:  2643 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0172 | PLoss: -0.0048 | Ent: 0.2480
+[ 82.8%] Iter 8087/9765 | SPS:  2643 | Reward:  0.069 | Value:  0.742 | VLoss: 0.0230 | PLoss: -0.0058 | Ent: 0.2265
+[ 82.8%] Iter 8088/9765 | SPS:  2643 | Reward:  0.059 | Value:  0.671 | VLoss: 0.0206 | PLoss: -0.0041 | Ent: 0.2342
+[ 82.8%] Iter 8089/9765 | SPS:  2644 | Reward:  0.066 | Value:  0.758 | VLoss: 0.0195 | PLoss: -0.0031 | Ent: 0.2092
+[ 82.8%] Iter 8090/9765 | SPS:  2644 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0178 | PLoss: -0.0022 | Ent: 0.2205
+[ 82.9%] Iter 8091/9765 | SPS:  2644 | Reward:  0.039 | Value:  0.729 | VLoss: 0.0167 | PLoss: -0.0142 | Ent: 0.2027
+[ 82.9%] Iter 8092/9765 | SPS:  2644 | Reward:  0.049 | Value:  0.619 | VLoss: 0.0147 | PLoss: -0.0040 | Ent: 0.2628
+[ 82.9%] Iter 8093/9765 | SPS:  2644 | Reward:  0.060 | Value:  0.723 | VLoss: 0.0213 | PLoss: -0.0030 | Ent: 0.1900
+[ 82.9%] Iter 8094/9765 | SPS:  2644 | Reward:  0.049 | Value:  0.686 | VLoss: 0.0195 | PLoss: -0.0070 | Ent: 0.2070
+[ 82.9%] Iter 8095/9765 | SPS:  2644 | Reward:  0.054 | Value:  0.703 | VLoss: 0.0212 | PLoss: -0.0047 | Ent: 0.2745
+[ 82.9%] Iter 8096/9765 | SPS:  2644 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0173 | PLoss: -0.0019 | Ent: 0.1986
+[ 82.9%] Iter 8097/9765 | SPS:  2644 | Reward:  0.040 | Value:  0.655 | VLoss: 0.0142 | PLoss: -0.0001 | Ent: 0.1820
+[ 82.9%] Iter 8098/9765 | SPS:  2644 | Reward:  0.065 | Value:  0.665 | VLoss: 0.0214 | PLoss: -0.0079 | Ent: 0.2115
+[ 82.9%] Iter 8099/9765 | SPS:  2644 | Reward:  0.061 | Value:  0.734 | VLoss: 0.0237 | PLoss: -0.0042 | Ent: 0.2390
+[ 82.9%] Iter 8100/9765 | SPS:  2645 | Reward:  0.050 | Value:  0.738 | VLoss: 0.0166 | PLoss: -0.0064 | Ent: 0.2138
+[ 83.0%] Iter 8101/9765 | SPS:  2645 | Reward:  0.055 | Value:  0.684 | VLoss: 0.0190 | PLoss: -0.0072 | Ent: 0.2141
+[ 83.0%] Iter 8102/9765 | SPS:  2645 | Reward:  0.059 | Value:  0.699 | VLoss: 0.0311 | PLoss: -0.0052 | Ent: 0.2553
+[ 83.0%] Iter 8103/9765 | SPS:  2645 | Reward:  0.057 | Value:  0.715 | VLoss: 0.0274 | PLoss: -0.0074 | Ent: 0.2051
+[ 83.0%] Iter 8104/9765 | SPS:  2645 | Reward:  0.076 | Value:  0.775 | VLoss: 0.0180 | PLoss: -0.0077 | Ent: 0.2464
+[ 83.0%] Iter 8105/9765 | SPS:  2645 | Reward:  0.055 | Value:  0.742 | VLoss: 0.0159 | PLoss: -0.0017 | Ent: 0.2013
+[ 83.0%] Iter 8106/9765 | SPS:  2645 | Reward:  0.056 | Value:  0.687 | VLoss: 0.0182 | PLoss: -0.0064 | Ent: 0.2473
+[ 83.0%] Iter 8107/9765 | SPS:  2645 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0237 | PLoss: -0.0030 | Ent: 0.2428
+[ 83.0%] Iter 8108/9765 | SPS:  2645 | Reward:  0.062 | Value:  0.756 | VLoss: 0.0171 | PLoss: -0.0043 | Ent: 0.1938
+[ 83.0%] Iter 8109/9765 | SPS:  2645 | Reward:  0.073 | Value:  0.831 | VLoss: 0.0204 | PLoss: -0.0040 | Ent: 0.2383
+[ 83.1%] Iter 8110/9765 | SPS:  2646 | Reward:  0.056 | Value:  0.792 | VLoss: 0.0237 | PLoss: -0.0020 | Ent: 0.2199
+[ 83.1%] Iter 8111/9765 | SPS:  2646 | Reward:  0.060 | Value:  0.742 | VLoss: 0.0262 | PLoss: -0.0096 | Ent: 0.2240
+[ 83.1%] Iter 8112/9765 | SPS:  2646 | Reward:  0.049 | Value:  0.704 | VLoss: 0.0154 | PLoss: -0.0058 | Ent: 0.2625
+[ 83.1%] Iter 8113/9765 | SPS:  2646 | Reward:  0.058 | Value:  0.660 | VLoss: 0.0218 | PLoss: -0.0084 | Ent: 0.2263
+[ 83.1%] Iter 8114/9765 | SPS:  2646 | Reward:  0.061 | Value:  0.734 | VLoss: 0.0203 | PLoss: -0.0035 | Ent: 0.1837
+[ 83.1%] Iter 8115/9765 | SPS:  2646 | Reward:  0.061 | Value:  0.735 | VLoss: 0.0209 | PLoss: -0.0040 | Ent: 0.2138
+[ 83.1%] Iter 8116/9765 | SPS:  2646 | Reward:  0.056 | Value:  0.705 | VLoss: 0.0269 | PLoss: -0.0071 | Ent: 0.2569
+[ 83.1%] Iter 8117/9765 | SPS:  2646 | Reward:  0.060 | Value:  0.693 | VLoss: 0.0207 | PLoss: -0.0041 | Ent: 0.2232
+[ 83.1%] Iter 8118/9765 | SPS:  2646 | Reward:  0.070 | Value:  0.720 | VLoss: 0.0224 | PLoss: -0.0057 | Ent: 0.2083
+[ 83.1%] Iter 8119/9765 | SPS:  2646 | Reward:  0.053 | Value:  0.770 | VLoss: 0.0182 | PLoss: -0.0013 | Ent: 0.2322
+[ 83.2%] Iter 8120/9765 | SPS:  2646 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0276 | PLoss: -0.0010 | Ent: 0.2126
+[ 83.2%] Iter 8121/9765 | SPS:  2647 | Reward:  0.052 | Value:  0.687 | VLoss: 0.0246 | PLoss: -0.0067 | Ent: 0.1741
+[ 83.2%] Iter 8122/9765 | SPS:  2647 | Reward:  0.074 | Value:  0.673 | VLoss: 0.0365 | PLoss: -0.0085 | Ent: 0.2142
+[ 83.2%] Iter 8123/9765 | SPS:  2647 | Reward:  0.067 | Value:  0.759 | VLoss: 0.0210 | PLoss: -0.0043 | Ent: 0.1996
+[ 83.2%] Iter 8124/9765 | SPS:  2647 | Reward:  0.056 | Value:  0.740 | VLoss: 0.0177 | PLoss: -0.0009 | Ent: 0.2169
+[ 83.2%] Iter 8125/9765 | SPS:  2647 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0211 | PLoss: -0.0044 | Ent: 0.2059
+[ 83.2%] Iter 8126/9765 | SPS:  2647 | Reward:  0.064 | Value:  0.764 | VLoss: 0.0283 | PLoss: 0.0000 | Ent: 0.2294
+[ 83.2%] Iter 8127/9765 | SPS:  2647 | Reward:  0.050 | Value:  0.728 | VLoss: 0.0137 | PLoss: -0.0086 | Ent: 0.2280
+[ 83.2%] Iter 8128/9765 | SPS:  2647 | Reward:  0.066 | Value:  0.739 | VLoss: 0.0218 | PLoss: -0.0024 | Ent: 0.2318
+[ 83.2%] Iter 8129/9765 | SPS:  2647 | Reward:  0.061 | Value:  0.778 | VLoss: 0.0257 | PLoss: -0.0023 | Ent: 0.2217
+[ 83.3%] Iter 8130/9765 | SPS:  2647 | Reward:  0.050 | Value:  0.684 | VLoss: 0.0158 | PLoss: -0.0045 | Ent: 0.1936
+[ 83.3%] Iter 8131/9765 | SPS:  2648 | Reward:  0.050 | Value:  0.669 | VLoss: 0.0166 | PLoss: -0.0023 | Ent: 0.1742
+[ 83.3%] Iter 8132/9765 | SPS:  2648 | Reward:  0.056 | Value:  0.695 | VLoss: 0.0170 | PLoss: -0.0080 | Ent: 0.1895
+[ 83.3%] Iter 8133/9765 | SPS:  2648 | Reward:  0.054 | Value:  0.622 | VLoss: 0.0181 | PLoss: -0.0041 | Ent: 0.2751
+[ 83.3%] Iter 8134/9765 | SPS:  2648 | Reward:  0.063 | Value:  0.654 | VLoss: 0.0318 | PLoss: -0.0130 | Ent: 0.2341
+[ 83.3%] Iter 8135/9765 | SPS:  2648 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0252 | PLoss: -0.0051 | Ent: 0.2165
+[ 83.3%] Iter 8136/9765 | SPS:  2648 | Reward:  0.078 | Value:  0.720 | VLoss: 0.0284 | PLoss: -0.0023 | Ent: 0.2164
+[ 83.3%] Iter 8137/9765 | SPS:  2648 | Reward:  0.047 | Value:  0.724 | VLoss: 0.0200 | PLoss: -0.0044 | Ent: 0.2347
+[ 83.3%] Iter 8138/9765 | SPS:  2648 | Reward:  0.059 | Value:  0.714 | VLoss: 0.0218 | PLoss: -0.0029 | Ent: 0.2565
+[ 83.3%] Iter 8139/9765 | SPS:  2648 | Reward:  0.055 | Value:  0.691 | VLoss: 0.0165 | PLoss: -0.0050 | Ent: 0.2246
+[ 83.4%] Iter 8140/9765 | SPS:  2648 | Reward:  0.048 | Value:  0.660 | VLoss: 0.0188 | PLoss: -0.0017 | Ent: 0.1991
+[ 83.4%] Iter 8141/9765 | SPS:  2649 | Reward:  0.058 | Value:  0.602 | VLoss: 0.0222 | PLoss: -0.0034 | Ent: 0.2298
+[ 83.4%] Iter 8142/9765 | SPS:  2649 | Reward:  0.067 | Value:  0.686 | VLoss: 0.0336 | PLoss: -0.0062 | Ent: 0.2284
+[ 83.4%] Iter 8143/9765 | SPS:  2649 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0315 | PLoss: -0.0044 | Ent: 0.2300
+[ 83.4%] Iter 8144/9765 | SPS:  2649 | Reward:  0.054 | Value:  0.727 | VLoss: 0.0174 | PLoss: -0.0088 | Ent: 0.2445
+[ 83.4%] Iter 8145/9765 | SPS:  2649 | Reward:  0.057 | Value:  0.695 | VLoss: 0.0264 | PLoss: -0.0007 | Ent: 0.2498
+[ 83.4%] Iter 8146/9765 | SPS:  2649 | Reward:  0.070 | Value:  0.650 | VLoss: 0.0302 | PLoss: -0.0046 | Ent: 0.2852
+[ 83.4%] Iter 8147/9765 | SPS:  2649 | Reward:  0.054 | Value:  0.700 | VLoss: 0.0207 | PLoss: -0.0066 | Ent: 0.2093
+[ 83.4%] Iter 8148/9765 | SPS:  2649 | Reward:  0.062 | Value:  0.706 | VLoss: 0.0259 | PLoss: -0.0079 | Ent: 0.2441
+[ 83.5%] Iter 8149/9765 | SPS:  2649 | Reward:  0.059 | Value:  0.746 | VLoss: 0.0168 | PLoss: -0.0040 | Ent: 0.1954
+[ 83.5%] Iter 8150/9765 | SPS:  2649 | Reward:  0.065 | Value:  0.777 | VLoss: 0.0146 | PLoss: -0.0068 | Ent: 0.2292
+[ 83.5%] Iter 8151/9765 | SPS:  2649 | Reward:  0.055 | Value:  0.731 | VLoss: 0.0264 | PLoss: -0.0083 | Ent: 0.2084
+[ 83.5%] Iter 8152/9765 | SPS:  2650 | Reward:  0.050 | Value:  0.747 | VLoss: 0.0182 | PLoss: -0.0049 | Ent: 0.2108
+[ 83.5%] Iter 8153/9765 | SPS:  2650 | Reward:  0.075 | Value:  0.696 | VLoss: 0.0232 | PLoss: -0.0024 | Ent: 0.1933
+[ 83.5%] Iter 8154/9765 | SPS:  2650 | Reward:  0.056 | Value:  0.746 | VLoss: 0.0351 | PLoss: -0.0079 | Ent: 0.2004
+[ 83.5%] Iter 8155/9765 | SPS:  2650 | Reward:  0.072 | Value:  0.696 | VLoss: 0.0210 | PLoss: -0.0039 | Ent: 0.1754
+[ 83.5%] Iter 8156/9765 | SPS:  2650 | Reward:  0.049 | Value:  0.700 | VLoss: 0.0174 | PLoss: -0.0100 | Ent: 0.2056
+[ 83.5%] Iter 8157/9765 | SPS:  2650 | Reward:  0.053 | Value:  0.640 | VLoss: 0.0242 | PLoss: -0.0049 | Ent: 0.2109
+[ 83.5%] Iter 8158/9765 | SPS:  2650 | Reward:  0.031 | Value:  0.613 | VLoss: 0.0136 | PLoss: -0.0053 | Ent: 0.2496
+[ 83.6%] Iter 8159/9765 | SPS:  2650 | Reward:  0.059 | Value:  0.615 | VLoss: 0.0290 | PLoss: -0.0077 | Ent: 0.1998
+[ 83.6%] Iter 8160/9765 | SPS:  2650 | Reward:  0.072 | Value:  0.707 | VLoss: 0.0251 | PLoss: -0.0046 | Ent: 0.2150
+[ 83.6%] Iter 8161/9765 | SPS:  2650 | Reward:  0.046 | Value:  0.738 | VLoss: 0.0208 | PLoss: -0.0041 | Ent: 0.2355
+[ 83.6%] Iter 8162/9765 | SPS:  2651 | Reward:  0.055 | Value:  0.636 | VLoss: 0.0156 | PLoss: -0.0030 | Ent: 0.1946
+[ 83.6%] Iter 8163/9765 | SPS:  2651 | Reward:  0.045 | Value:  0.664 | VLoss: 0.0235 | PLoss: -0.0011 | Ent: 0.2202
+[ 83.6%] Iter 8164/9765 | SPS:  2651 | Reward:  0.048 | Value:  0.700 | VLoss: 0.0138 | PLoss: -0.0023 | Ent: 0.2273
+[ 83.6%] Iter 8165/9765 | SPS:  2651 | Reward:  0.063 | Value:  0.671 | VLoss: 0.0216 | PLoss: -0.0041 | Ent: 0.2033
+[ 83.6%] Iter 8166/9765 | SPS:  2651 | Reward:  0.067 | Value:  0.711 | VLoss: 0.0203 | PLoss: -0.0062 | Ent: 0.2412
+[ 83.6%] Iter 8167/9765 | SPS:  2651 | Reward:  0.050 | Value:  0.732 | VLoss: 0.0200 | PLoss: -0.0029 | Ent: 0.2336
+[ 83.6%] Iter 8168/9765 | SPS:  2651 | Reward:  0.047 | Value:  0.686 | VLoss: 0.0202 | PLoss: -0.0016 | Ent: 0.2048
+[ 83.7%] Iter 8169/9765 | SPS:  2651 | Reward:  0.054 | Value:  0.587 | VLoss: 0.0180 | PLoss: -0.0042 | Ent: 0.2422
+[ 83.7%] Iter 8170/9765 | SPS:  2651 | Reward:  0.062 | Value:  0.692 | VLoss: 0.0202 | PLoss: -0.0053 | Ent: 0.2345
+[ 83.7%] Iter 8171/9765 | SPS:  2651 | Reward:  0.055 | Value:  0.695 | VLoss: 0.0250 | PLoss: -0.0077 | Ent: 0.2135
+[ 83.7%] Iter 8172/9765 | SPS:  2652 | Reward:  0.053 | Value:  0.746 | VLoss: 0.0200 | PLoss: -0.0078 | Ent: 0.2175
+[ 83.7%] Iter 8173/9765 | SPS:  2652 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0147 | PLoss: -0.0003 | Ent: 0.1962
+[ 83.7%] Iter 8174/9765 | SPS:  2652 | Reward:  0.049 | Value:  0.618 | VLoss: 0.0303 | PLoss: -0.0071 | Ent: 0.2138
+[ 83.7%] Iter 8175/9765 | SPS:  2652 | Reward:  0.050 | Value:  0.696 | VLoss: 0.0314 | PLoss: 0.0001 | Ent: 0.2138
+[ 83.7%] Iter 8176/9765 | SPS:  2652 | Reward:  0.063 | Value:  0.655 | VLoss: 0.0251 | PLoss: -0.0050 | Ent: 0.1743
+[ 83.7%] Iter 8177/9765 | SPS:  2652 | Reward:  0.059 | Value:  0.742 | VLoss: 0.0159 | PLoss: -0.0062 | Ent: 0.2422
+[ 83.7%] Iter 8178/9765 | SPS:  2652 | Reward:  0.068 | Value:  0.707 | VLoss: 0.0208 | PLoss: -0.0054 | Ent: 0.2341
+[ 83.8%] Iter 8179/9765 | SPS:  2652 | Reward:  0.074 | Value:  0.708 | VLoss: 0.0247 | PLoss: -0.0064 | Ent: 0.2068
+[ 83.8%] Iter 8180/9765 | SPS:  2652 | Reward:  0.055 | Value:  0.696 | VLoss: 0.0251 | PLoss: -0.0058 | Ent: 0.2156
+[ 83.8%] Iter 8181/9765 | SPS:  2652 | Reward:  0.070 | Value:  0.725 | VLoss: 0.0194 | PLoss: -0.0034 | Ent: 0.2267
+[ 83.8%] Iter 8182/9765 | SPS:  2652 | Reward:  0.053 | Value:  0.744 | VLoss: 0.0174 | PLoss: -0.0018 | Ent: 0.2211
+[ 83.8%] Iter 8183/9765 | SPS:  2653 | Reward:  0.070 | Value:  0.654 | VLoss: 0.0187 | PLoss: -0.0079 | Ent: 0.2092
+[ 83.8%] Iter 8184/9765 | SPS:  2653 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0202 | PLoss: -0.0076 | Ent: 0.2189
+[ 83.8%] Iter 8185/9765 | SPS:  2653 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0250 | PLoss: -0.0026 | Ent: 0.2286
+[ 83.8%] Iter 8186/9765 | SPS:  2653 | Reward:  0.072 | Value:  0.711 | VLoss: 0.0133 | PLoss: -0.0016 | Ent: 0.1751
+[ 83.8%] Iter 8187/9765 | SPS:  2653 | Reward:  0.065 | Value:  0.715 | VLoss: 0.0251 | PLoss: -0.0034 | Ent: 0.2130
+[ 83.9%] Iter 8188/9765 | SPS:  2653 | Reward:  0.069 | Value:  0.755 | VLoss: 0.0233 | PLoss: -0.0089 | Ent: 0.2337
+[ 83.9%] Iter 8189/9765 | SPS:  2653 | Reward:  0.054 | Value:  0.724 | VLoss: 0.0225 | PLoss: -0.0075 | Ent: 0.2281
+[ 83.9%] Iter 8190/9765 | SPS:  2653 | Reward:  0.052 | Value:  0.620 | VLoss: 0.0292 | PLoss: -0.0097 | Ent: 0.2248
+[ 83.9%] Iter 8191/9765 | SPS:  2653 | Reward:  0.073 | Value:  0.686 | VLoss: 0.0243 | PLoss: -0.0021 | Ent: 0.1836
+[ 83.9%] Iter 8192/9765 | SPS:  2653 | Reward:  0.058 | Value:  0.721 | VLoss: 0.0247 | PLoss: -0.0059 | Ent: 0.1877
+[ 83.9%] Iter 8193/9765 | SPS:  2653 | Reward:  0.055 | Value:  0.669 | VLoss: 0.0195 | PLoss: 0.0026 | Ent: 0.1834
+[ 83.9%] Iter 8194/9765 | SPS:  2654 | Reward:  0.066 | Value:  0.764 | VLoss: 0.0161 | PLoss: -0.0056 | Ent: 0.1905
+[ 83.9%] Iter 8195/9765 | SPS:  2654 | Reward:  0.071 | Value:  0.791 | VLoss: 0.0201 | PLoss: -0.0027 | Ent: 0.1831
+[ 83.9%] Iter 8196/9765 | SPS:  2654 | Reward:  0.062 | Value:  0.766 | VLoss: 0.0203 | PLoss: -0.0011 | Ent: 0.2605
+[ 83.9%] Iter 8197/9765 | SPS:  2654 | Reward:  0.074 | Value:  0.687 | VLoss: 0.0338 | PLoss: -0.0046 | Ent: 0.2487
+[ 84.0%] Iter 8198/9765 | SPS:  2654 | Reward:  0.060 | Value:  0.666 | VLoss: 0.0260 | PLoss: -0.0051 | Ent: 0.2599
+[ 84.0%] Iter 8199/9765 | SPS:  2654 | Reward:  0.071 | Value:  0.691 | VLoss: 0.0291 | PLoss: -0.0023 | Ent: 0.1696
+[ 84.0%] Iter 8200/9765 | SPS:  2654 | Reward:  0.062 | Value:  0.771 | VLoss: 0.0188 | PLoss: -0.0045 | Ent: 0.2503
+[ 84.0%] Iter 8201/9765 | SPS:  2654 | Reward:  0.062 | Value:  0.743 | VLoss: 0.0252 | PLoss: -0.0057 | Ent: 0.2080
+[ 84.0%] Iter 8202/9765 | SPS:  2654 | Reward:  0.063 | Value:  0.778 | VLoss: 0.0210 | PLoss: 0.0014 | Ent: 0.2350
+[ 84.0%] Iter 8203/9765 | SPS:  2654 | Reward:  0.065 | Value:  0.686 | VLoss: 0.0283 | PLoss: -0.0063 | Ent: 0.2442
+[ 84.0%] Iter 8204/9765 | SPS:  2654 | Reward:  0.043 | Value:  0.674 | VLoss: 0.0229 | PLoss: -0.0027 | Ent: 0.2197
+[ 84.0%] Iter 8205/9765 | SPS:  2655 | Reward:  0.052 | Value:  0.693 | VLoss: 0.0226 | PLoss: -0.0096 | Ent: 0.2433
+[ 84.0%] Iter 8206/9765 | SPS:  2655 | Reward:  0.053 | Value:  0.707 | VLoss: 0.0239 | PLoss: -0.0035 | Ent: 0.2395
+[ 84.0%] Iter 8207/9765 | SPS:  2655 | Reward:  0.057 | Value:  0.749 | VLoss: 0.0189 | PLoss: -0.0017 | Ent: 0.1918
+[ 84.1%] Iter 8208/9765 | SPS:  2655 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0203 | PLoss: -0.0017 | Ent: 0.2175
+[ 84.1%] Iter 8209/9765 | SPS:  2655 | Reward:  0.059 | Value:  0.730 | VLoss: 0.0209 | PLoss: -0.0026 | Ent: 0.2650
+[ 84.1%] Iter 8210/9765 | SPS:  2655 | Reward:  0.066 | Value:  0.746 | VLoss: 0.0199 | PLoss: -0.0034 | Ent: 0.2135
+[ 84.1%] Iter 8211/9765 | SPS:  2655 | Reward:  0.059 | Value:  0.711 | VLoss: 0.0276 | PLoss: -0.0076 | Ent: 0.2279
+[ 84.1%] Iter 8212/9765 | SPS:  2655 | Reward:  0.048 | Value:  0.711 | VLoss: 0.0132 | PLoss: -0.0036 | Ent: 0.2103
+[ 84.1%] Iter 8213/9765 | SPS:  2655 | Reward:  0.066 | Value:  0.687 | VLoss: 0.0275 | PLoss: -0.0046 | Ent: 0.2059
+[ 84.1%] Iter 8214/9765 | SPS:  2655 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0215 | PLoss: 0.0009 | Ent: 0.2200
+[ 84.1%] Iter 8215/9765 | SPS:  2656 | Reward:  0.064 | Value:  0.731 | VLoss: 0.0227 | PLoss: -0.0067 | Ent: 0.2737
+[ 84.1%] Iter 8216/9765 | SPS:  2656 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0222 | PLoss: -0.0066 | Ent: 0.2244
+[ 84.1%] Iter 8217/9765 | SPS:  2656 | Reward:  0.062 | Value:  0.749 | VLoss: 0.0196 | PLoss: -0.0086 | Ent: 0.1857
+[ 84.2%] Iter 8218/9765 | SPS:  2656 | Reward:  0.073 | Value:  0.674 | VLoss: 0.0322 | PLoss: -0.0040 | Ent: 0.1841
+[ 84.2%] Iter 8219/9765 | SPS:  2656 | Reward:  0.061 | Value:  0.784 | VLoss: 0.0161 | PLoss: -0.0030 | Ent: 0.2122
+[ 84.2%] Iter 8220/9765 | SPS:  2656 | Reward:  0.057 | Value:  0.750 | VLoss: 0.0210 | PLoss: -0.0031 | Ent: 0.2602
+[ 84.2%] Iter 8221/9765 | SPS:  2656 | Reward:  0.054 | Value:  0.741 | VLoss: 0.0243 | PLoss: -0.0057 | Ent: 0.2287
+[ 84.2%] Iter 8222/9765 | SPS:  2656 | Reward:  0.066 | Value:  0.713 | VLoss: 0.0179 | PLoss: -0.0021 | Ent: 0.1938
+[ 84.2%] Iter 8223/9765 | SPS:  2656 | Reward:  0.056 | Value:  0.773 | VLoss: 0.0258 | PLoss: -0.0055 | Ent: 0.2453
+[ 84.2%] Iter 8224/9765 | SPS:  2656 | Reward:  0.056 | Value:  0.686 | VLoss: 0.0236 | PLoss: -0.0051 | Ent: 0.2482
+[ 84.2%] Iter 8225/9765 | SPS:  2656 | Reward:  0.072 | Value:  0.688 | VLoss: 0.0255 | PLoss: -0.0042 | Ent: 0.1931
+[ 84.2%] Iter 8226/9765 | SPS:  2657 | Reward:  0.059 | Value:  0.657 | VLoss: 0.0268 | PLoss: -0.0100 | Ent: 0.2179
+[ 84.2%] Iter 8227/9765 | SPS:  2657 | Reward:  0.054 | Value:  0.666 | VLoss: 0.0248 | PLoss: -0.0012 | Ent: 0.1845
+[ 84.3%] Iter 8228/9765 | SPS:  2657 | Reward:  0.065 | Value:  0.714 | VLoss: 0.0281 | PLoss: -0.0088 | Ent: 0.2201
+[ 84.3%] Iter 8229/9765 | SPS:  2657 | Reward:  0.056 | Value:  0.753 | VLoss: 0.0215 | PLoss: -0.0041 | Ent: 0.2339
+[ 84.3%] Iter 8230/9765 | SPS:  2657 | Reward:  0.060 | Value:  0.696 | VLoss: 0.0172 | PLoss: -0.0087 | Ent: 0.2870
+[ 84.3%] Iter 8231/9765 | SPS:  2657 | Reward:  0.053 | Value:  0.686 | VLoss: 0.0188 | PLoss: -0.0050 | Ent: 0.2337
+[ 84.3%] Iter 8232/9765 | SPS:  2657 | Reward:  0.055 | Value:  0.724 | VLoss: 0.0230 | PLoss: -0.0060 | Ent: 0.2332
+[ 84.3%] Iter 8233/9765 | SPS:  2657 | Reward:  0.065 | Value:  0.681 | VLoss: 0.0249 | PLoss: -0.0050 | Ent: 0.2035
+[ 84.3%] Iter 8234/9765 | SPS:  2657 | Reward:  0.043 | Value:  0.659 | VLoss: 0.0177 | PLoss: -0.0034 | Ent: 0.1842
+[ 84.3%] Iter 8235/9765 | SPS:  2657 | Reward:  0.053 | Value:  0.738 | VLoss: 0.0206 | PLoss: -0.0032 | Ent: 0.2103
+[ 84.3%] Iter 8236/9765 | SPS:  2658 | Reward:  0.063 | Value:  0.671 | VLoss: 0.0228 | PLoss: -0.0038 | Ent: 0.2120
+[ 84.4%] Iter 8237/9765 | SPS:  2658 | Reward:  0.056 | Value:  0.721 | VLoss: 0.0200 | PLoss: -0.0040 | Ent: 0.2344
+[ 84.4%] Iter 8238/9765 | SPS:  2658 | Reward:  0.059 | Value:  0.689 | VLoss: 0.0205 | PLoss: -0.0039 | Ent: 0.2084
+[ 84.4%] Iter 8239/9765 | SPS:  2658 | Reward:  0.061 | Value:  0.716 | VLoss: 0.0151 | PLoss: -0.0027 | Ent: 0.2178
+[ 84.4%] Iter 8240/9765 | SPS:  2658 | Reward:  0.068 | Value:  0.775 | VLoss: 0.0329 | PLoss: -0.0091 | Ent: 0.2266
+[ 84.4%] Iter 8241/9765 | SPS:  2658 | Reward:  0.059 | Value:  0.837 | VLoss: 0.0114 | PLoss: -0.0048 | Ent: 0.2407
+[ 84.4%] Iter 8242/9765 | SPS:  2658 | Reward:  0.061 | Value:  0.756 | VLoss: 0.0247 | PLoss: -0.0081 | Ent: 0.2095
+[ 84.4%] Iter 8243/9765 | SPS:  2658 | Reward:  0.044 | Value:  0.784 | VLoss: 0.0196 | PLoss: -0.0041 | Ent: 0.2330
+[ 84.4%] Iter 8244/9765 | SPS:  2658 | Reward:  0.058 | Value:  0.693 | VLoss: 0.0177 | PLoss: -0.0084 | Ent: 0.2336
+[ 84.4%] Iter 8245/9765 | SPS:  2658 | Reward:  0.061 | Value:  0.709 | VLoss: 0.0218 | PLoss: -0.0005 | Ent: 0.2371
+[ 84.4%] Iter 8246/9765 | SPS:  2658 | Reward:  0.059 | Value:  0.688 | VLoss: 0.0188 | PLoss: -0.0025 | Ent: 0.2190
+[ 84.5%] Iter 8247/9765 | SPS:  2659 | Reward:  0.056 | Value:  0.737 | VLoss: 0.0167 | PLoss: -0.0008 | Ent: 0.2794
+[ 84.5%] Iter 8248/9765 | SPS:  2659 | Reward:  0.060 | Value:  0.707 | VLoss: 0.0154 | PLoss: -0.0038 | Ent: 0.2013
+[ 84.5%] Iter 8249/9765 | SPS:  2659 | Reward:  0.066 | Value:  0.687 | VLoss: 0.0324 | PLoss: -0.0093 | Ent: 0.2205
+[ 84.5%] Iter 8250/9765 | SPS:  2659 | Reward:  0.062 | Value:  0.726 | VLoss: 0.0168 | PLoss: -0.0096 | Ent: 0.2672
+[ 84.5%] Iter 8251/9765 | SPS:  2659 | Reward:  0.064 | Value:  0.722 | VLoss: 0.0231 | PLoss: -0.0027 | Ent: 0.2236
+[ 84.5%] Iter 8252/9765 | SPS:  2659 | Reward:  0.065 | Value:  0.806 | VLoss: 0.0207 | PLoss: -0.0025 | Ent: 0.2133
+[ 84.5%] Iter 8253/9765 | SPS:  2659 | Reward:  0.057 | Value:  0.728 | VLoss: 0.0228 | PLoss: -0.0045 | Ent: 0.2449
+[ 84.5%] Iter 8254/9765 | SPS:  2659 | Reward:  0.051 | Value:  0.761 | VLoss: 0.0238 | PLoss: -0.0031 | Ent: 0.2858
+[ 84.5%] Iter 8255/9765 | SPS:  2659 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0213 | PLoss: -0.0041 | Ent: 0.2340
+[ 84.5%] Iter 8256/9765 | SPS:  2659 | Reward:  0.061 | Value:  0.682 | VLoss: 0.0221 | PLoss: -0.0054 | Ent: 0.2369
+[ 84.6%] Iter 8257/9765 | SPS:  2659 | Reward:  0.053 | Value:  0.679 | VLoss: 0.0226 | PLoss: -0.0021 | Ent: 0.1842
+[ 84.6%] Iter 8258/9765 | SPS:  2660 | Reward:  0.068 | Value:  0.705 | VLoss: 0.0223 | PLoss: -0.0045 | Ent: 0.2150
+[ 84.6%] Iter 8259/9765 | SPS:  2660 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0170 | PLoss: -0.0039 | Ent: 0.2065
+[ 84.6%] Iter 8260/9765 | SPS:  2660 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0238 | PLoss: -0.0068 | Ent: 0.2127
+[ 84.6%] Iter 8261/9765 | SPS:  2660 | Reward:  0.059 | Value:  0.634 | VLoss: 0.0238 | PLoss: -0.0040 | Ent: 0.2422
+[ 84.6%] Iter 8262/9765 | SPS:  2660 | Reward:  0.072 | Value:  0.689 | VLoss: 0.0215 | PLoss: -0.0029 | Ent: 0.2224
+[ 84.6%] Iter 8263/9765 | SPS:  2660 | Reward:  0.057 | Value:  0.755 | VLoss: 0.0231 | PLoss: -0.0066 | Ent: 0.2691
+[ 84.6%] Iter 8264/9765 | SPS:  2660 | Reward:  0.059 | Value:  0.686 | VLoss: 0.0188 | PLoss: -0.0107 | Ent: 0.2432
+[ 84.6%] Iter 8265/9765 | SPS:  2660 | Reward:  0.064 | Value:  0.730 | VLoss: 0.0283 | PLoss: -0.0054 | Ent: 0.2014
+[ 84.6%] Iter 8266/9765 | SPS:  2660 | Reward:  0.050 | Value:  0.717 | VLoss: 0.0188 | PLoss: -0.0020 | Ent: 0.1948
+[ 84.7%] Iter 8267/9765 | SPS:  2660 | Reward:  0.050 | Value:  0.745 | VLoss: 0.0203 | PLoss: -0.0045 | Ent: 0.2323
+[ 84.7%] Iter 8268/9765 | SPS:  2661 | Reward:  0.064 | Value:  0.678 | VLoss: 0.0269 | PLoss: -0.0061 | Ent: 0.2318
+[ 84.7%] Iter 8269/9765 | SPS:  2661 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0143 | PLoss: -0.0069 | Ent: 0.2493
+[ 84.7%] Iter 8270/9765 | SPS:  2661 | Reward:  0.065 | Value:  0.807 | VLoss: 0.0167 | PLoss: -0.0045 | Ent: 0.1986
+[ 84.7%] Iter 8271/9765 | SPS:  2661 | Reward:  0.064 | Value:  0.758 | VLoss: 0.0161 | PLoss: -0.0054 | Ent: 0.2672
+[ 84.7%] Iter 8272/9765 | SPS:  2661 | Reward:  0.060 | Value:  0.722 | VLoss: 0.0195 | PLoss: -0.0064 | Ent: 0.2243
+[ 84.7%] Iter 8273/9765 | SPS:  2661 | Reward:  0.055 | Value:  0.735 | VLoss: 0.0198 | PLoss: -0.0030 | Ent: 0.2074
+[ 84.7%] Iter 8274/9765 | SPS:  2661 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0182 | PLoss: -0.0006 | Ent: 0.1862
+[ 84.7%] Iter 8275/9765 | SPS:  2661 | Reward:  0.067 | Value:  0.791 | VLoss: 0.0309 | PLoss: -0.0057 | Ent: 0.2197
+[ 84.8%] Iter 8276/9765 | SPS:  2661 | Reward:  0.052 | Value:  0.731 | VLoss: 0.0305 | PLoss: -0.0004 | Ent: 0.2406
+[ 84.8%] Iter 8277/9765 | SPS:  2661 | Reward:  0.059 | Value:  0.651 | VLoss: 0.0203 | PLoss: -0.0032 | Ent: 0.1928
+[ 84.8%] Iter 8278/9765 | SPS:  2661 | Reward:  0.061 | Value:  0.724 | VLoss: 0.0196 | PLoss: -0.0084 | Ent: 0.2247
+[ 84.8%] Iter 8279/9765 | SPS:  2662 | Reward:  0.052 | Value:  0.754 | VLoss: 0.0208 | PLoss: -0.0004 | Ent: 0.2285
+[ 84.8%] Iter 8280/9765 | SPS:  2662 | Reward:  0.066 | Value:  0.742 | VLoss: 0.0186 | PLoss: -0.0041 | Ent: 0.1997
+[ 84.8%] Iter 8281/9765 | SPS:  2662 | Reward:  0.062 | Value:  0.726 | VLoss: 0.0237 | PLoss: -0.0080 | Ent: 0.2426
+[ 84.8%] Iter 8282/9765 | SPS:  2662 | Reward:  0.055 | Value:  0.692 | VLoss: 0.0257 | PLoss: -0.0061 | Ent: 0.2078
+[ 84.8%] Iter 8283/9765 | SPS:  2662 | Reward:  0.062 | Value:  0.607 | VLoss: 0.0193 | PLoss: -0.0097 | Ent: 0.2396
+[ 84.8%] Iter 8284/9765 | SPS:  2662 | Reward:  0.049 | Value:  0.730 | VLoss: 0.0193 | PLoss: -0.0048 | Ent: 0.2025
+[ 84.8%] Iter 8285/9765 | SPS:  2662 | Reward:  0.056 | Value:  0.686 | VLoss: 0.0221 | PLoss: -0.0030 | Ent: 0.2003
+[ 84.9%] Iter 8286/9765 | SPS:  2662 | Reward:  0.069 | Value:  0.718 | VLoss: 0.0233 | PLoss: -0.0082 | Ent: 0.2328
+[ 84.9%] Iter 8287/9765 | SPS:  2662 | Reward:  0.066 | Value:  0.733 | VLoss: 0.0235 | PLoss: -0.0053 | Ent: 0.2283
+[ 84.9%] Iter 8288/9765 | SPS:  2662 | Reward:  0.071 | Value:  0.745 | VLoss: 0.0217 | PLoss: -0.0063 | Ent: 0.2660
+[ 84.9%] Iter 8289/9765 | SPS:  2662 | Reward:  0.070 | Value:  0.766 | VLoss: 0.0247 | PLoss: -0.0016 | Ent: 0.1569
+[ 84.9%] Iter 8290/9765 | SPS:  2663 | Reward:  0.054 | Value:  0.826 | VLoss: 0.0227 | PLoss: -0.0044 | Ent: 0.2159
+[ 84.9%] Iter 8291/9765 | SPS:  2663 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0148 | PLoss: -0.0039 | Ent: 0.2552
+[ 84.9%] Iter 8292/9765 | SPS:  2663 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0198 | PLoss: -0.0090 | Ent: 0.1626
+[ 84.9%] Iter 8293/9765 | SPS:  2663 | Reward:  0.048 | Value:  0.689 | VLoss: 0.0207 | PLoss: -0.0050 | Ent: 0.1993
+[ 84.9%] Iter 8294/9765 | SPS:  2663 | Reward:  0.060 | Value:  0.657 | VLoss: 0.0250 | PLoss: -0.0015 | Ent: 0.2233
+[ 84.9%] Iter 8295/9765 | SPS:  2663 | Reward:  0.062 | Value:  0.655 | VLoss: 0.0279 | PLoss: -0.0057 | Ent: 0.2129
+[ 85.0%] Iter 8296/9765 | SPS:  2663 | Reward:  0.057 | Value:  0.653 | VLoss: 0.0216 | PLoss: -0.0048 | Ent: 0.2248
+[ 85.0%] Iter 8297/9765 | SPS:  2663 | Reward:  0.038 | Value:  0.633 | VLoss: 0.0192 | PLoss: -0.0061 | Ent: 0.2409
+[ 85.0%] Iter 8298/9765 | SPS:  2663 | Reward:  0.073 | Value:  0.722 | VLoss: 0.0211 | PLoss: -0.0010 | Ent: 0.2191
+[ 85.0%] Iter 8299/9765 | SPS:  2663 | Reward:  0.071 | Value:  0.698 | VLoss: 0.0317 | PLoss: -0.0018 | Ent: 0.2317
+[ 85.0%] Iter 8300/9765 | SPS:  2664 | Reward:  0.065 | Value:  0.775 | VLoss: 0.0310 | PLoss: -0.0047 | Ent: 0.2475
+[ 85.0%] Iter 8301/9765 | SPS:  2664 | Reward:  0.049 | Value:  0.721 | VLoss: 0.0226 | PLoss: -0.0014 | Ent: 0.2188
+[ 85.0%] Iter 8302/9765 | SPS:  2664 | Reward:  0.061 | Value:  0.695 | VLoss: 0.0170 | PLoss: 0.0006 | Ent: 0.1961
+[ 85.0%] Iter 8303/9765 | SPS:  2664 | Reward:  0.059 | Value:  0.703 | VLoss: 0.0195 | PLoss: -0.0012 | Ent: 0.2146
+[ 85.0%] Iter 8304/9765 | SPS:  2664 | Reward:  0.059 | Value:  0.731 | VLoss: 0.0123 | PLoss: -0.0023 | Ent: 0.2327
+[ 85.0%] Iter 8305/9765 | SPS:  2664 | Reward:  0.077 | Value:  0.737 | VLoss: 0.0155 | PLoss: -0.0049 | Ent: 0.2025
+[ 85.1%] Iter 8306/9765 | SPS:  2664 | Reward:  0.071 | Value:  0.774 | VLoss: 0.0256 | PLoss: -0.0000 | Ent: 0.2052
+[ 85.1%] Iter 8307/9765 | SPS:  2664 | Reward:  0.056 | Value:  0.779 | VLoss: 0.0168 | PLoss: -0.0017 | Ent: 0.2005
+[ 85.1%] Iter 8308/9765 | SPS:  2664 | Reward:  0.062 | Value:  0.634 | VLoss: 0.0246 | PLoss: -0.0060 | Ent: 0.2239
+[ 85.1%] Iter 8309/9765 | SPS:  2664 | Reward:  0.061 | Value:  0.663 | VLoss: 0.0204 | PLoss: -0.0037 | Ent: 0.2435
+[ 85.1%] Iter 8310/9765 | SPS:  2664 | Reward:  0.060 | Value:  0.767 | VLoss: 0.0236 | PLoss: -0.0030 | Ent: 0.2539
+[ 85.1%] Iter 8311/9765 | SPS:  2665 | Reward:  0.056 | Value:  0.657 | VLoss: 0.0227 | PLoss: -0.0043 | Ent: 0.2064
+[ 85.1%] Iter 8312/9765 | SPS:  2665 | Reward:  0.069 | Value:  0.739 | VLoss: 0.0223 | PLoss: -0.0052 | Ent: 0.2246
+[ 85.1%] Iter 8313/9765 | SPS:  2665 | Reward:  0.075 | Value:  0.782 | VLoss: 0.0172 | PLoss: -0.0012 | Ent: 0.1716
+[ 85.1%] Iter 8314/9765 | SPS:  2665 | Reward:  0.055 | Value:  0.767 | VLoss: 0.0177 | PLoss: -0.0065 | Ent: 0.1873
+[ 85.2%] Iter 8315/9765 | SPS:  2665 | Reward:  0.060 | Value:  0.760 | VLoss: 0.0240 | PLoss: -0.0041 | Ent: 0.1665
+[ 85.2%] Iter 8316/9765 | SPS:  2665 | Reward:  0.064 | Value:  0.698 | VLoss: 0.0186 | PLoss: -0.0032 | Ent: 0.1934
+[ 85.2%] Iter 8317/9765 | SPS:  2665 | Reward:  0.068 | Value:  0.715 | VLoss: 0.0281 | PLoss: -0.0100 | Ent: 0.2760
+[ 85.2%] Iter 8318/9765 | SPS:  2665 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0178 | PLoss: -0.0005 | Ent: 0.2162
+[ 85.2%] Iter 8319/9765 | SPS:  2665 | Reward:  0.052 | Value:  0.731 | VLoss: 0.0186 | PLoss: -0.0010 | Ent: 0.2564
+[ 85.2%] Iter 8320/9765 | SPS:  2665 | Reward:  0.058 | Value:  0.722 | VLoss: 0.0178 | PLoss: -0.0059 | Ent: 0.2358
+[ 85.2%] Iter 8321/9765 | SPS:  2665 | Reward:  0.058 | Value:  0.700 | VLoss: 0.0182 | PLoss: -0.0057 | Ent: 0.2372
+[ 85.2%] Iter 8322/9765 | SPS:  2666 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0195 | PLoss: -0.0086 | Ent: 0.2011
+[ 85.2%] Iter 8323/9765 | SPS:  2666 | Reward:  0.056 | Value:  0.701 | VLoss: 0.0196 | PLoss: -0.0036 | Ent: 0.2306
+[ 85.2%] Iter 8324/9765 | SPS:  2666 | Reward:  0.048 | Value:  0.672 | VLoss: 0.0199 | PLoss: -0.0030 | Ent: 0.2166
+[ 85.3%] Iter 8325/9765 | SPS:  2666 | Reward:  0.067 | Value:  0.713 | VLoss: 0.0224 | PLoss: -0.0041 | Ent: 0.2189
+[ 85.3%] Iter 8326/9765 | SPS:  2666 | Reward:  0.059 | Value:  0.739 | VLoss: 0.0233 | PLoss: -0.0039 | Ent: 0.2054
+[ 85.3%] Iter 8327/9765 | SPS:  2666 | Reward:  0.054 | Value:  0.655 | VLoss: 0.0183 | PLoss: -0.0057 | Ent: 0.2077
+[ 85.3%] Iter 8328/9765 | SPS:  2666 | Reward:  0.046 | Value:  0.646 | VLoss: 0.0208 | PLoss: -0.0039 | Ent: 0.1800
+[ 85.3%] Iter 8329/9765 | SPS:  2666 | Reward:  0.053 | Value:  0.673 | VLoss: 0.0195 | PLoss: -0.0070 | Ent: 0.1981
+[ 85.3%] Iter 8330/9765 | SPS:  2666 | Reward:  0.063 | Value:  0.706 | VLoss: 0.0186 | PLoss: -0.0115 | Ent: 0.2147
+[ 85.3%] Iter 8331/9765 | SPS:  2666 | Reward:  0.066 | Value:  0.697 | VLoss: 0.0223 | PLoss: -0.0008 | Ent: 0.1924
+[ 85.3%] Iter 8332/9765 | SPS:  2666 | Reward:  0.067 | Value:  0.721 | VLoss: 0.0237 | PLoss: -0.0042 | Ent: 0.1769
+[ 85.3%] Iter 8333/9765 | SPS:  2667 | Reward:  0.055 | Value:  0.747 | VLoss: 0.0128 | PLoss: -0.0095 | Ent: 0.2018
+[ 85.3%] Iter 8334/9765 | SPS:  2667 | Reward:  0.065 | Value:  0.699 | VLoss: 0.0214 | PLoss: -0.0021 | Ent: 0.1682
+[ 85.4%] Iter 8335/9765 | SPS:  2667 | Reward:  0.064 | Value:  0.742 | VLoss: 0.0196 | PLoss: -0.0015 | Ent: 0.1913
+[ 85.4%] Iter 8336/9765 | SPS:  2667 | Reward:  0.064 | Value:  0.777 | VLoss: 0.0195 | PLoss: -0.0003 | Ent: 0.2115
+[ 85.4%] Iter 8337/9765 | SPS:  2667 | Reward:  0.048 | Value:  0.688 | VLoss: 0.0198 | PLoss: 0.0003 | Ent: 0.2497
+[ 85.4%] Iter 8338/9765 | SPS:  2667 | Reward:  0.058 | Value:  0.680 | VLoss: 0.0202 | PLoss: -0.0057 | Ent: 0.2351
+[ 85.4%] Iter 8339/9765 | SPS:  2667 | Reward:  0.066 | Value:  0.719 | VLoss: 0.0150 | PLoss: -0.0026 | Ent: 0.2109
+[ 85.4%] Iter 8340/9765 | SPS:  2667 | Reward:  0.055 | Value:  0.690 | VLoss: 0.0186 | PLoss: 0.0005 | Ent: 0.1654
+[ 85.4%] Iter 8341/9765 | SPS:  2667 | Reward:  0.064 | Value:  0.685 | VLoss: 0.0184 | PLoss: -0.0011 | Ent: 0.1392
+[ 85.4%] Iter 8342/9765 | SPS:  2667 | Reward:  0.062 | Value:  0.758 | VLoss: 0.0195 | PLoss: -0.0024 | Ent: 0.2110
+[ 85.4%] Iter 8343/9765 | SPS:  2668 | Reward:  0.067 | Value:  0.676 | VLoss: 0.0240 | PLoss: -0.0035 | Ent: 0.2341
+[ 85.4%] Iter 8344/9765 | SPS:  2668 | Reward:  0.076 | Value:  0.708 | VLoss: 0.0270 | PLoss: -0.0050 | Ent: 0.1918
+[ 85.5%] Iter 8345/9765 | SPS:  2668 | Reward:  0.047 | Value:  0.778 | VLoss: 0.0150 | PLoss: -0.0010 | Ent: 0.2434
+[ 85.5%] Iter 8346/9765 | SPS:  2668 | Reward:  0.058 | Value:  0.772 | VLoss: 0.0175 | PLoss: -0.0040 | Ent: 0.2059
+[ 85.5%] Iter 8347/9765 | SPS:  2668 | Reward:  0.063 | Value:  0.762 | VLoss: 0.0189 | PLoss: -0.0115 | Ent: 0.1835
+[ 85.5%] Iter 8348/9765 | SPS:  2668 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0155 | PLoss: -0.0053 | Ent: 0.1788
+[ 85.5%] Iter 8349/9765 | SPS:  2668 | Reward:  0.069 | Value:  0.756 | VLoss: 0.0203 | PLoss: -0.0046 | Ent: 0.2106
+[ 85.5%] Iter 8350/9765 | SPS:  2668 | Reward:  0.087 | Value:  0.720 | VLoss: 0.0291 | PLoss: -0.0056 | Ent: 0.2296
+[ 85.5%] Iter 8351/9765 | SPS:  2668 | Reward:  0.066 | Value:  0.726 | VLoss: 0.0260 | PLoss: -0.0095 | Ent: 0.2333
+[ 85.5%] Iter 8352/9765 | SPS:  2668 | Reward:  0.055 | Value:  0.741 | VLoss: 0.0231 | PLoss: -0.0047 | Ent: 0.2135
+[ 85.5%] Iter 8353/9765 | SPS:  2668 | Reward:  0.052 | Value:  0.715 | VLoss: 0.0266 | PLoss: -0.0014 | Ent: 0.1940
+[ 85.6%] Iter 8354/9765 | SPS:  2668 | Reward:  0.061 | Value:  0.634 | VLoss: 0.0336 | PLoss: -0.0046 | Ent: 0.2229
+[ 85.6%] Iter 8355/9765 | SPS:  2669 | Reward:  0.067 | Value:  0.728 | VLoss: 0.0198 | PLoss: -0.0033 | Ent: 0.2409
+[ 85.6%] Iter 8356/9765 | SPS:  2669 | Reward:  0.056 | Value:  0.664 | VLoss: 0.0207 | PLoss: -0.0064 | Ent: 0.2536
+[ 85.6%] Iter 8357/9765 | SPS:  2669 | Reward:  0.062 | Value:  0.783 | VLoss: 0.0263 | PLoss: -0.0015 | Ent: 0.1896
+[ 85.6%] Iter 8358/9765 | SPS:  2669 | Reward:  0.059 | Value:  0.761 | VLoss: 0.0180 | PLoss: -0.0033 | Ent: 0.2509
+[ 85.6%] Iter 8359/9765 | SPS:  2669 | Reward:  0.075 | Value:  0.746 | VLoss: 0.0174 | PLoss: -0.0004 | Ent: 0.1837
+[ 85.6%] Iter 8360/9765 | SPS:  2669 | Reward:  0.052 | Value:  0.747 | VLoss: 0.0212 | PLoss: -0.0064 | Ent: 0.2494
+[ 85.6%] Iter 8361/9765 | SPS:  2669 | Reward:  0.066 | Value:  0.688 | VLoss: 0.0216 | PLoss: -0.0048 | Ent: 0.2078
+[ 85.6%] Iter 8362/9765 | SPS:  2669 | Reward:  0.051 | Value:  0.760 | VLoss: 0.0198 | PLoss: -0.0064 | Ent: 0.1914
+[ 85.6%] Iter 8363/9765 | SPS:  2669 | Reward:  0.063 | Value:  0.742 | VLoss: 0.0218 | PLoss: -0.0007 | Ent: 0.1780
+[ 85.7%] Iter 8364/9765 | SPS:  2669 | Reward:  0.059 | Value:  0.704 | VLoss: 0.0124 | PLoss: -0.0076 | Ent: 0.2091
+[ 85.7%] Iter 8365/9765 | SPS:  2670 | Reward:  0.060 | Value:  0.741 | VLoss: 0.0257 | PLoss: -0.0005 | Ent: 0.2118
+[ 85.7%] Iter 8366/9765 | SPS:  2670 | Reward:  0.070 | Value:  0.766 | VLoss: 0.0222 | PLoss: -0.0033 | Ent: 0.2530
+[ 85.7%] Iter 8367/9765 | SPS:  2670 | Reward:  0.059 | Value:  0.665 | VLoss: 0.0192 | PLoss: -0.0041 | Ent: 0.2156
+[ 85.7%] Iter 8368/9765 | SPS:  2670 | Reward:  0.072 | Value:  0.678 | VLoss: 0.0277 | PLoss: -0.0004 | Ent: 0.2469
+[ 85.7%] Iter 8369/9765 | SPS:  2670 | Reward:  0.057 | Value:  0.725 | VLoss: 0.0186 | PLoss: -0.0031 | Ent: 0.2002
+[ 85.7%] Iter 8370/9765 | SPS:  2670 | Reward:  0.065 | Value:  0.769 | VLoss: 0.0235 | PLoss: -0.0021 | Ent: 0.1660
+[ 85.7%] Iter 8371/9765 | SPS:  2670 | Reward:  0.053 | Value:  0.667 | VLoss: 0.0188 | PLoss: -0.0050 | Ent: 0.2109
+[ 85.7%] Iter 8372/9765 | SPS:  2670 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0164 | PLoss: -0.0031 | Ent: 0.2236
+[ 85.7%] Iter 8373/9765 | SPS:  2670 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0162 | PLoss: -0.0104 | Ent: 0.2126
+[ 85.8%] Iter 8374/9765 | SPS:  2670 | Reward:  0.061 | Value:  0.786 | VLoss: 0.0203 | PLoss: -0.0052 | Ent: 0.2250
+[ 85.8%] Iter 8375/9765 | SPS:  2670 | Reward:  0.059 | Value:  0.761 | VLoss: 0.0229 | PLoss: -0.0068 | Ent: 0.2065
+[ 85.8%] Iter 8376/9765 | SPS:  2671 | Reward:  0.051 | Value:  0.709 | VLoss: 0.0210 | PLoss: -0.0041 | Ent: 0.2731
+[ 85.8%] Iter 8377/9765 | SPS:  2671 | Reward:  0.047 | Value:  0.699 | VLoss: 0.0175 | PLoss: -0.0029 | Ent: 0.2329
+[ 85.8%] Iter 8378/9765 | SPS:  2671 | Reward:  0.060 | Value:  0.673 | VLoss: 0.0219 | PLoss: -0.0035 | Ent: 0.2118
+[ 85.8%] Iter 8379/9765 | SPS:  2671 | Reward:  0.061 | Value:  0.715 | VLoss: 0.0195 | PLoss: -0.0032 | Ent: 0.1873
+[ 85.8%] Iter 8380/9765 | SPS:  2671 | Reward:  0.060 | Value:  0.685 | VLoss: 0.0236 | PLoss: -0.0050 | Ent: 0.1968
+[ 85.8%] Iter 8381/9765 | SPS:  2671 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0234 | PLoss: -0.0021 | Ent: 0.1817
+[ 85.8%] Iter 8382/9765 | SPS:  2671 | Reward:  0.059 | Value:  0.726 | VLoss: 0.0207 | PLoss: -0.0018 | Ent: 0.1686
+[ 85.8%] Iter 8383/9765 | SPS:  2671 | Reward:  0.065 | Value:  0.697 | VLoss: 0.0217 | PLoss: -0.0084 | Ent: 0.2261
+[ 85.9%] Iter 8384/9765 | SPS:  2671 | Reward:  0.075 | Value:  0.807 | VLoss: 0.0186 | PLoss: -0.0037 | Ent: 0.2484
+[ 85.9%] Iter 8385/9765 | SPS:  2671 | Reward:  0.040 | Value:  0.827 | VLoss: 0.0119 | PLoss: -0.0015 | Ent: 0.2212
+[ 85.9%] Iter 8386/9765 | SPS:  2671 | Reward:  0.054 | Value:  0.690 | VLoss: 0.0190 | PLoss: -0.0037 | Ent: 0.2247
+[ 85.9%] Iter 8387/9765 | SPS:  2672 | Reward:  0.048 | Value:  0.611 | VLoss: 0.0262 | PLoss: -0.0043 | Ent: 0.2102
+[ 85.9%] Iter 8388/9765 | SPS:  2672 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0147 | PLoss: -0.0038 | Ent: 0.2334
+[ 85.9%] Iter 8389/9765 | SPS:  2672 | Reward:  0.071 | Value:  0.752 | VLoss: 0.0207 | PLoss: -0.0062 | Ent: 0.2578
+[ 85.9%] Iter 8390/9765 | SPS:  2672 | Reward:  0.065 | Value:  0.773 | VLoss: 0.0201 | PLoss: -0.0016 | Ent: 0.2404
+[ 85.9%] Iter 8391/9765 | SPS:  2672 | Reward:  0.065 | Value:  0.772 | VLoss: 0.0254 | PLoss: -0.0094 | Ent: 0.2271
+[ 85.9%] Iter 8392/9765 | SPS:  2672 | Reward:  0.048 | Value:  0.710 | VLoss: 0.0223 | PLoss: -0.0059 | Ent: 0.2455
+[ 85.9%] Iter 8393/9765 | SPS:  2672 | Reward:  0.065 | Value:  0.728 | VLoss: 0.0178 | PLoss: -0.0015 | Ent: 0.2016
+[ 86.0%] Iter 8394/9765 | SPS:  2672 | Reward:  0.058 | Value:  0.667 | VLoss: 0.0208 | PLoss: -0.0042 | Ent: 0.2144
+[ 86.0%] Iter 8395/9765 | SPS:  2672 | Reward:  0.070 | Value:  0.677 | VLoss: 0.0200 | PLoss: 0.0016 | Ent: 0.2368
+[ 86.0%] Iter 8396/9765 | SPS:  2672 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0221 | PLoss: -0.0028 | Ent: 0.1950
+[ 86.0%] Iter 8397/9765 | SPS:  2672 | Reward:  0.050 | Value:  0.662 | VLoss: 0.0198 | PLoss: -0.0049 | Ent: 0.2352
+[ 86.0%] Iter 8398/9765 | SPS:  2673 | Reward:  0.051 | Value:  0.790 | VLoss: 0.0167 | PLoss: -0.0022 | Ent: 0.1577
+[ 86.0%] Iter 8399/9765 | SPS:  2673 | Reward:  0.059 | Value:  0.748 | VLoss: 0.0188 | PLoss: -0.0087 | Ent: 0.2066
+[ 86.0%] Iter 8400/9765 | SPS:  2673 | Reward:  0.059 | Value:  0.759 | VLoss: 0.0269 | PLoss: -0.0074 | Ent: 0.2507
+[ 86.0%] Iter 8401/9765 | SPS:  2673 | Reward:  0.059 | Value:  0.713 | VLoss: 0.0217 | PLoss: -0.0017 | Ent: 0.2506
+[ 86.0%] Iter 8402/9765 | SPS:  2673 | Reward:  0.069 | Value:  0.724 | VLoss: 0.0202 | PLoss: -0.0014 | Ent: 0.2073
+[ 86.1%] Iter 8403/9765 | SPS:  2673 | Reward:  0.056 | Value:  0.781 | VLoss: 0.0244 | PLoss: -0.0053 | Ent: 0.1563
+[ 86.1%] Iter 8404/9765 | SPS:  2673 | Reward:  0.064 | Value:  0.745 | VLoss: 0.0227 | PLoss: -0.0034 | Ent: 0.1945
+[ 86.1%] Iter 8405/9765 | SPS:  2673 | Reward:  0.053 | Value:  0.710 | VLoss: 0.0189 | PLoss: -0.0048 | Ent: 0.2648
+[ 86.1%] Iter 8406/9765 | SPS:  2673 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0145 | PLoss: -0.0038 | Ent: 0.2186
+[ 86.1%] Iter 8407/9765 | SPS:  2673 | Reward:  0.070 | Value:  0.679 | VLoss: 0.0211 | PLoss: -0.0039 | Ent: 0.1904
+[ 86.1%] Iter 8408/9765 | SPS:  2673 | Reward:  0.054 | Value:  0.751 | VLoss: 0.0223 | PLoss: -0.0021 | Ent: 0.1751
+[ 86.1%] Iter 8409/9765 | SPS:  2674 | Reward:  0.063 | Value:  0.592 | VLoss: 0.0265 | PLoss: -0.0040 | Ent: 0.2734
+[ 86.1%] Iter 8410/9765 | SPS:  2674 | Reward:  0.058 | Value:  0.773 | VLoss: 0.0183 | PLoss: -0.0091 | Ent: 0.1824
+[ 86.1%] Iter 8411/9765 | SPS:  2674 | Reward:  0.054 | Value:  0.793 | VLoss: 0.0133 | PLoss: -0.0031 | Ent: 0.2090
+[ 86.1%] Iter 8412/9765 | SPS:  2674 | Reward:  0.054 | Value:  0.695 | VLoss: 0.0146 | PLoss: -0.0013 | Ent: 0.2278
+[ 86.2%] Iter 8413/9765 | SPS:  2674 | Reward:  0.047 | Value:  0.631 | VLoss: 0.0121 | PLoss: -0.0044 | Ent: 0.2088
+[ 86.2%] Iter 8414/9765 | SPS:  2674 | Reward:  0.059 | Value:  0.659 | VLoss: 0.0196 | PLoss: -0.0054 | Ent: 0.1968
+[ 86.2%] Iter 8415/9765 | SPS:  2674 | Reward:  0.060 | Value:  0.668 | VLoss: 0.0257 | PLoss: -0.0068 | Ent: 0.1983
+[ 86.2%] Iter 8416/9765 | SPS:  2674 | Reward:  0.054 | Value:  0.668 | VLoss: 0.0171 | PLoss: -0.0052 | Ent: 0.1863
+[ 86.2%] Iter 8417/9765 | SPS:  2674 | Reward:  0.067 | Value:  0.766 | VLoss: 0.0130 | PLoss: -0.0012 | Ent: 0.2053
+[ 86.2%] Iter 8418/9765 | SPS:  2674 | Reward:  0.054 | Value:  0.648 | VLoss: 0.0197 | PLoss: -0.0008 | Ent: 0.1921
+[ 86.2%] Iter 8419/9765 | SPS:  2675 | Reward:  0.070 | Value:  0.719 | VLoss: 0.0205 | PLoss: -0.0072 | Ent: 0.2059
+[ 86.2%] Iter 8420/9765 | SPS:  2675 | Reward:  0.059 | Value:  0.679 | VLoss: 0.0264 | PLoss: -0.0027 | Ent: 0.2299
+[ 86.2%] Iter 8421/9765 | SPS:  2675 | Reward:  0.055 | Value:  0.687 | VLoss: 0.0164 | PLoss: -0.0017 | Ent: 0.1989
+[ 86.2%] Iter 8422/9765 | SPS:  2675 | Reward:  0.067 | Value:  0.668 | VLoss: 0.0242 | PLoss: -0.0022 | Ent: 0.2778
+[ 86.3%] Iter 8423/9765 | SPS:  2675 | Reward:  0.068 | Value:  0.732 | VLoss: 0.0238 | PLoss: -0.0018 | Ent: 0.2208
+[ 86.3%] Iter 8424/9765 | SPS:  2675 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0261 | PLoss: -0.0039 | Ent: 0.2356
+[ 86.3%] Iter 8425/9765 | SPS:  2675 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0178 | PLoss: -0.0034 | Ent: 0.2392
+[ 86.3%] Iter 8426/9765 | SPS:  2675 | Reward:  0.059 | Value:  0.691 | VLoss: 0.0237 | PLoss: -0.0115 | Ent: 0.2767
+[ 86.3%] Iter 8427/9765 | SPS:  2675 | Reward:  0.068 | Value:  0.749 | VLoss: 0.0158 | PLoss: -0.0007 | Ent: 0.2646
+[ 86.3%] Iter 8428/9765 | SPS:  2675 | Reward:  0.064 | Value:  0.740 | VLoss: 0.0308 | PLoss: -0.0007 | Ent: 0.2294
+[ 86.3%] Iter 8429/9765 | SPS:  2675 | Reward:  0.066 | Value:  0.719 | VLoss: 0.0329 | PLoss: -0.0029 | Ent: 0.2256
+[ 86.3%] Iter 8430/9765 | SPS:  2676 | Reward:  0.058 | Value:  0.704 | VLoss: 0.0234 | PLoss: -0.0045 | Ent: 0.2289
+[ 86.3%] Iter 8431/9765 | SPS:  2676 | Reward:  0.064 | Value:  0.701 | VLoss: 0.0200 | PLoss: -0.0036 | Ent: 0.2150
+[ 86.3%] Iter 8432/9765 | SPS:  2676 | Reward:  0.046 | Value:  0.654 | VLoss: 0.0198 | PLoss: -0.0033 | Ent: 0.1514
+[ 86.4%] Iter 8433/9765 | SPS:  2676 | Reward:  0.061 | Value:  0.697 | VLoss: 0.0189 | PLoss: -0.0023 | Ent: 0.2097
+[ 86.4%] Iter 8434/9765 | SPS:  2676 | Reward:  0.058 | Value:  0.701 | VLoss: 0.0230 | PLoss: -0.0031 | Ent: 0.2391
+[ 86.4%] Iter 8435/9765 | SPS:  2676 | Reward:  0.061 | Value:  0.778 | VLoss: 0.0195 | PLoss: -0.0079 | Ent: 0.1930
+[ 86.4%] Iter 8436/9765 | SPS:  2676 | Reward:  0.061 | Value:  0.649 | VLoss: 0.0307 | PLoss: -0.0044 | Ent: 0.2019
+[ 86.4%] Iter 8437/9765 | SPS:  2676 | Reward:  0.060 | Value:  0.646 | VLoss: 0.0321 | PLoss: -0.0057 | Ent: 0.2160
+[ 86.4%] Iter 8438/9765 | SPS:  2676 | Reward:  0.070 | Value:  0.675 | VLoss: 0.0218 | PLoss: -0.0007 | Ent: 0.2611
+[ 86.4%] Iter 8439/9765 | SPS:  2676 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0293 | PLoss: -0.0003 | Ent: 0.2073
+[ 86.4%] Iter 8440/9765 | SPS:  2676 | Reward:  0.067 | Value:  0.774 | VLoss: 0.0217 | PLoss: -0.0014 | Ent: 0.2005
+[ 86.4%] Iter 8441/9765 | SPS:  2677 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0172 | PLoss: -0.0037 | Ent: 0.2872
+[ 86.5%] Iter 8442/9765 | SPS:  2677 | Reward:  0.054 | Value:  0.834 | VLoss: 0.0149 | PLoss: -0.0038 | Ent: 0.2214
+[ 86.5%] Iter 8443/9765 | SPS:  2677 | Reward:  0.048 | Value:  0.694 | VLoss: 0.0237 | PLoss: -0.0036 | Ent: 0.2019
+[ 86.5%] Iter 8444/9765 | SPS:  2677 | Reward:  0.061 | Value:  0.675 | VLoss: 0.0196 | PLoss: -0.0053 | Ent: 0.2151
+[ 86.5%] Iter 8445/9765 | SPS:  2677 | Reward:  0.056 | Value:  0.700 | VLoss: 0.0219 | PLoss: -0.0071 | Ent: 0.3029
+[ 86.5%] Iter 8446/9765 | SPS:  2677 | Reward:  0.049 | Value:  0.701 | VLoss: 0.0210 | PLoss: -0.0033 | Ent: 0.2389
+[ 86.5%] Iter 8447/9765 | SPS:  2677 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0261 | PLoss: -0.0071 | Ent: 0.2433
+[ 86.5%] Iter 8448/9765 | SPS:  2677 | Reward:  0.064 | Value:  0.779 | VLoss: 0.0164 | PLoss: -0.0007 | Ent: 0.2285
+[ 86.5%] Iter 8449/9765 | SPS:  2677 | Reward:  0.069 | Value:  0.766 | VLoss: 0.0199 | PLoss: -0.0041 | Ent: 0.2124
+[ 86.5%] Iter 8450/9765 | SPS:  2677 | Reward:  0.062 | Value:  0.714 | VLoss: 0.0258 | PLoss: -0.0092 | Ent: 0.1956
+[ 86.5%] Iter 8451/9765 | SPS:  2677 | Reward:  0.070 | Value:  0.730 | VLoss: 0.0206 | PLoss: -0.0024 | Ent: 0.2419
+[ 86.6%] Iter 8452/9765 | SPS:  2678 | Reward:  0.069 | Value:  0.663 | VLoss: 0.0248 | PLoss: -0.0040 | Ent: 0.2302
+[ 86.6%] Iter 8453/9765 | SPS:  2678 | Reward:  0.058 | Value:  0.796 | VLoss: 0.0215 | PLoss: -0.0035 | Ent: 0.2505
+[ 86.6%] Iter 8454/9765 | SPS:  2678 | Reward:  0.043 | Value:  0.683 | VLoss: 0.0189 | PLoss: -0.0046 | Ent: 0.2377
+[ 86.6%] Iter 8455/9765 | SPS:  2678 | Reward:  0.060 | Value:  0.619 | VLoss: 0.0277 | PLoss: -0.0026 | Ent: 0.2351
+[ 86.6%] Iter 8456/9765 | SPS:  2678 | Reward:  0.062 | Value:  0.719 | VLoss: 0.0157 | PLoss: -0.0033 | Ent: 0.1909
+[ 86.6%] Iter 8457/9765 | SPS:  2678 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0206 | PLoss: -0.0041 | Ent: 0.2155
+[ 86.6%] Iter 8458/9765 | SPS:  2678 | Reward:  0.058 | Value:  0.679 | VLoss: 0.0218 | PLoss: -0.0043 | Ent: 0.1863
+[ 86.6%] Iter 8459/9765 | SPS:  2678 | Reward:  0.067 | Value:  0.726 | VLoss: 0.0166 | PLoss: -0.0049 | Ent: 0.2230
+[ 86.6%] Iter 8460/9765 | SPS:  2678 | Reward:  0.065 | Value:  0.781 | VLoss: 0.0196 | PLoss: -0.0039 | Ent: 0.2251
+[ 86.6%] Iter 8461/9765 | SPS:  2678 | Reward:  0.054 | Value:  0.699 | VLoss: 0.0274 | PLoss: -0.0022 | Ent: 0.1960
+[ 86.7%] Iter 8462/9765 | SPS:  2678 | Reward:  0.077 | Value:  0.763 | VLoss: 0.0172 | PLoss: -0.0040 | Ent: 0.1981
+[ 86.7%] Iter 8463/9765 | SPS:  2679 | Reward:  0.053 | Value:  0.762 | VLoss: 0.0145 | PLoss: -0.0043 | Ent: 0.2520
+[ 86.7%] Iter 8464/9765 | SPS:  2679 | Reward:  0.055 | Value:  0.747 | VLoss: 0.0204 | PLoss: -0.0085 | Ent: 0.2591
+[ 86.7%] Iter 8465/9765 | SPS:  2679 | Reward:  0.070 | Value:  0.736 | VLoss: 0.0222 | PLoss: -0.0013 | Ent: 0.1925
+[ 86.7%] Iter 8466/9765 | SPS:  2679 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0181 | PLoss: -0.0026 | Ent: 0.1987
+[ 86.7%] Iter 8467/9765 | SPS:  2679 | Reward:  0.052 | Value:  0.734 | VLoss: 0.0263 | PLoss: -0.0029 | Ent: 0.2500
+[ 86.7%] Iter 8468/9765 | SPS:  2679 | Reward:  0.053 | Value:  0.788 | VLoss: 0.0220 | PLoss: -0.0085 | Ent: 0.2359
+[ 86.7%] Iter 8469/9765 | SPS:  2679 | Reward:  0.059 | Value:  0.713 | VLoss: 0.0253 | PLoss: -0.0031 | Ent: 0.2471
+[ 86.7%] Iter 8470/9765 | SPS:  2679 | Reward:  0.062 | Value:  0.733 | VLoss: 0.0177 | PLoss: -0.0060 | Ent: 0.2268
+[ 86.7%] Iter 8471/9765 | SPS:  2679 | Reward:  0.062 | Value:  0.657 | VLoss: 0.0246 | PLoss: -0.0032 | Ent: 0.2346
+[ 86.8%] Iter 8472/9765 | SPS:  2679 | Reward:  0.068 | Value:  0.691 | VLoss: 0.0183 | PLoss: -0.0021 | Ent: 0.2509
+[ 86.8%] Iter 8473/9765 | SPS:  2679 | Reward:  0.059 | Value:  0.744 | VLoss: 0.0191 | PLoss: -0.0047 | Ent: 0.2349
+[ 86.8%] Iter 8474/9765 | SPS:  2680 | Reward:  0.052 | Value:  0.648 | VLoss: 0.0237 | PLoss: -0.0099 | Ent: 0.2313
+[ 86.8%] Iter 8475/9765 | SPS:  2680 | Reward:  0.057 | Value:  0.715 | VLoss: 0.0250 | PLoss: -0.0056 | Ent: 0.2391
+[ 86.8%] Iter 8476/9765 | SPS:  2680 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0187 | PLoss: -0.0053 | Ent: 0.2084
+[ 86.8%] Iter 8477/9765 | SPS:  2680 | Reward:  0.061 | Value:  0.730 | VLoss: 0.0107 | PLoss: -0.0055 | Ent: 0.1777
+[ 86.8%] Iter 8478/9765 | SPS:  2680 | Reward:  0.046 | Value:  0.651 | VLoss: 0.0176 | PLoss: -0.0123 | Ent: 0.2443
+[ 86.8%] Iter 8479/9765 | SPS:  2680 | Reward:  0.053 | Value:  0.649 | VLoss: 0.0164 | PLoss: -0.0016 | Ent: 0.2086
+[ 86.8%] Iter 8480/9765 | SPS:  2680 | Reward:  0.053 | Value:  0.593 | VLoss: 0.0205 | PLoss: -0.0027 | Ent: 0.1775
+[ 86.9%] Iter 8481/9765 | SPS:  2680 | Reward:  0.047 | Value:  0.650 | VLoss: 0.0245 | PLoss: -0.0017 | Ent: 0.2270
+[ 86.9%] Iter 8482/9765 | SPS:  2680 | Reward:  0.059 | Value:  0.699 | VLoss: 0.0175 | PLoss: -0.0027 | Ent: 0.2344
+[ 86.9%] Iter 8483/9765 | SPS:  2680 | Reward:  0.060 | Value:  0.685 | VLoss: 0.0279 | PLoss: -0.0053 | Ent: 0.2434
+[ 86.9%] Iter 8484/9765 | SPS:  2680 | Reward:  0.064 | Value:  0.760 | VLoss: 0.0238 | PLoss: -0.0046 | Ent: 0.2155
+[ 86.9%] Iter 8485/9765 | SPS:  2681 | Reward:  0.064 | Value:  0.714 | VLoss: 0.0302 | PLoss: -0.0054 | Ent: 0.2093
+[ 86.9%] Iter 8486/9765 | SPS:  2681 | Reward:  0.065 | Value:  0.705 | VLoss: 0.0302 | PLoss: -0.0060 | Ent: 0.2205
+[ 86.9%] Iter 8487/9765 | SPS:  2681 | Reward:  0.060 | Value:  0.697 | VLoss: 0.0283 | PLoss: -0.0052 | Ent: 0.2437
+[ 86.9%] Iter 8488/9765 | SPS:  2681 | Reward:  0.062 | Value:  0.653 | VLoss: 0.0315 | PLoss: -0.0043 | Ent: 0.2109
+[ 86.9%] Iter 8489/9765 | SPS:  2681 | Reward:  0.063 | Value:  0.684 | VLoss: 0.0213 | PLoss: -0.0010 | Ent: 0.1880
+[ 86.9%] Iter 8490/9765 | SPS:  2681 | Reward:  0.057 | Value:  0.633 | VLoss: 0.0199 | PLoss: -0.0042 | Ent: 0.2066
+[ 87.0%] Iter 8491/9765 | SPS:  2681 | Reward:  0.061 | Value:  0.681 | VLoss: 0.0205 | PLoss: -0.0011 | Ent: 0.2753
+[ 87.0%] Iter 8492/9765 | SPS:  2681 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0228 | PLoss: -0.0039 | Ent: 0.2262
+[ 87.0%] Iter 8493/9765 | SPS:  2681 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0237 | PLoss: -0.0049 | Ent: 0.2052
+[ 87.0%] Iter 8494/9765 | SPS:  2681 | Reward:  0.066 | Value:  0.724 | VLoss: 0.0218 | PLoss: -0.0034 | Ent: 0.2227
+[ 87.0%] Iter 8495/9765 | SPS:  2681 | Reward:  0.065 | Value:  0.732 | VLoss: 0.0200 | PLoss: -0.0040 | Ent: 0.1839
+[ 87.0%] Iter 8496/9765 | SPS:  2682 | Reward:  0.074 | Value:  0.722 | VLoss: 0.0291 | PLoss: -0.0053 | Ent: 0.2332
+[ 87.0%] Iter 8497/9765 | SPS:  2682 | Reward:  0.056 | Value:  0.661 | VLoss: 0.0194 | PLoss: -0.0023 | Ent: 0.2149
+[ 87.0%] Iter 8498/9765 | SPS:  2682 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0216 | PLoss: -0.0047 | Ent: 0.2596
+[ 87.0%] Iter 8499/9765 | SPS:  2682 | Reward:  0.081 | Value:  0.748 | VLoss: 0.0251 | PLoss: -0.0056 | Ent: 0.2338
+[ 87.0%] Iter 8500/9765 | SPS:  2682 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0163 | PLoss: -0.0028 | Ent: 0.2310
+[ 87.1%] Iter 8501/9765 | SPS:  2682 | Reward:  0.057 | Value:  0.704 | VLoss: 0.0206 | PLoss: 0.0017 | Ent: 0.2459
+[ 87.1%] Iter 8502/9765 | SPS:  2682 | Reward:  0.056 | Value:  0.704 | VLoss: 0.0189 | PLoss: -0.0091 | Ent: 0.2805
+[ 87.1%] Iter 8503/9765 | SPS:  2682 | Reward:  0.080 | Value:  0.725 | VLoss: 0.0258 | PLoss: -0.0034 | Ent: 0.2311
+[ 87.1%] Iter 8504/9765 | SPS:  2682 | Reward:  0.055 | Value:  0.718 | VLoss: 0.0154 | PLoss: -0.0066 | Ent: 0.1865
+[ 87.1%] Iter 8505/9765 | SPS:  2682 | Reward:  0.048 | Value:  0.707 | VLoss: 0.0162 | PLoss: -0.0063 | Ent: 0.1975
+[ 87.1%] Iter 8506/9765 | SPS:  2682 | Reward:  0.062 | Value:  0.708 | VLoss: 0.0344 | PLoss: -0.0067 | Ent: 0.2607
+[ 87.1%] Iter 8507/9765 | SPS:  2682 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0230 | PLoss: -0.0059 | Ent: 0.2045
+[ 87.1%] Iter 8508/9765 | SPS:  2683 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0239 | PLoss: -0.0034 | Ent: 0.2302
+[ 87.1%] Iter 8509/9765 | SPS:  2683 | Reward:  0.054 | Value:  0.782 | VLoss: 0.0175 | PLoss: -0.0048 | Ent: 0.2258
+[ 87.1%] Iter 8510/9765 | SPS:  2683 | Reward:  0.076 | Value:  0.762 | VLoss: 0.0215 | PLoss: -0.0017 | Ent: 0.2430
+[ 87.2%] Iter 8511/9765 | SPS:  2683 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0201 | PLoss: -0.0046 | Ent: 0.2294
+[ 87.2%] Iter 8512/9765 | SPS:  2683 | Reward:  0.062 | Value:  0.652 | VLoss: 0.0301 | PLoss: -0.0028 | Ent: 0.2614
+[ 87.2%] Iter 8513/9765 | SPS:  2683 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0222 | PLoss: -0.0062 | Ent: 0.2210
+[ 87.2%] Iter 8514/9765 | SPS:  2683 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0228 | PLoss: -0.0033 | Ent: 0.2583
+[ 87.2%] Iter 8515/9765 | SPS:  2683 | Reward:  0.066 | Value:  0.710 | VLoss: 0.0266 | PLoss: -0.0055 | Ent: 0.2515
+[ 87.2%] Iter 8516/9765 | SPS:  2683 | Reward:  0.051 | Value:  0.673 | VLoss: 0.0316 | PLoss: -0.0052 | Ent: 0.2566
+[ 87.2%] Iter 8517/9765 | SPS:  2683 | Reward:  0.052 | Value:  0.729 | VLoss: 0.0184 | PLoss: -0.0040 | Ent: 0.2297
+[ 87.2%] Iter 8518/9765 | SPS:  2683 | Reward:  0.058 | Value:  0.674 | VLoss: 0.0252 | PLoss: -0.0034 | Ent: 0.2441
+[ 87.2%] Iter 8519/9765 | SPS:  2684 | Reward:  0.062 | Value:  0.759 | VLoss: 0.0217 | PLoss: -0.0014 | Ent: 0.1746
+[ 87.3%] Iter 8520/9765 | SPS:  2684 | Reward:  0.063 | Value:  0.731 | VLoss: 0.0232 | PLoss: -0.0012 | Ent: 0.2317
+[ 87.3%] Iter 8521/9765 | SPS:  2684 | Reward:  0.052 | Value:  0.723 | VLoss: 0.0154 | PLoss: 0.0004 | Ent: 0.2061
+[ 87.3%] Iter 8522/9765 | SPS:  2684 | Reward:  0.059 | Value:  0.683 | VLoss: 0.0232 | PLoss: -0.0016 | Ent: 0.2376
+[ 87.3%] Iter 8523/9765 | SPS:  2684 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0259 | PLoss: -0.0039 | Ent: 0.2213
+[ 87.3%] Iter 8524/9765 | SPS:  2684 | Reward:  0.049 | Value:  0.763 | VLoss: 0.0192 | PLoss: -0.0027 | Ent: 0.1919
+[ 87.3%] Iter 8525/9765 | SPS:  2684 | Reward:  0.057 | Value:  0.722 | VLoss: 0.0215 | PLoss: -0.0042 | Ent: 0.2111
+[ 87.3%] Iter 8526/9765 | SPS:  2684 | Reward:  0.057 | Value:  0.757 | VLoss: 0.0215 | PLoss: -0.0047 | Ent: 0.2469
+[ 87.3%] Iter 8527/9765 | SPS:  2684 | Reward:  0.080 | Value:  0.801 | VLoss: 0.0207 | PLoss: -0.0048 | Ent: 0.2220
+[ 87.3%] Iter 8528/9765 | SPS:  2684 | Reward:  0.057 | Value:  0.747 | VLoss: 0.0170 | PLoss: -0.0021 | Ent: 0.2192
+[ 87.3%] Iter 8529/9765 | SPS:  2684 | Reward:  0.043 | Value:  0.768 | VLoss: 0.0155 | PLoss: -0.0033 | Ent: 0.2133
+[ 87.4%] Iter 8530/9765 | SPS:  2685 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0234 | PLoss: -0.0022 | Ent: 0.2641
+[ 87.4%] Iter 8531/9765 | SPS:  2685 | Reward:  0.062 | Value:  0.723 | VLoss: 0.0264 | PLoss: -0.0039 | Ent: 0.1802
+[ 87.4%] Iter 8532/9765 | SPS:  2685 | Reward:  0.062 | Value:  0.735 | VLoss: 0.0305 | PLoss: -0.0056 | Ent: 0.2125
+[ 87.4%] Iter 8533/9765 | SPS:  2685 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0187 | PLoss: -0.0039 | Ent: 0.2471
+[ 87.4%] Iter 8534/9765 | SPS:  2685 | Reward:  0.053 | Value:  0.639 | VLoss: 0.0260 | PLoss: -0.0065 | Ent: 0.2400
+[ 87.4%] Iter 8535/9765 | SPS:  2685 | Reward:  0.048 | Value:  0.651 | VLoss: 0.0237 | PLoss: -0.0056 | Ent: 0.2124
+[ 87.4%] Iter 8536/9765 | SPS:  2685 | Reward:  0.057 | Value:  0.689 | VLoss: 0.0232 | PLoss: -0.0006 | Ent: 0.1671
+[ 87.4%] Iter 8537/9765 | SPS:  2685 | Reward:  0.048 | Value:  0.642 | VLoss: 0.0169 | PLoss: -0.0046 | Ent: 0.2153
+[ 87.4%] Iter 8538/9765 | SPS:  2685 | Reward:  0.054 | Value:  0.767 | VLoss: 0.0166 | PLoss: -0.0055 | Ent: 0.1966
+[ 87.4%] Iter 8539/9765 | SPS:  2685 | Reward:  0.047 | Value:  0.734 | VLoss: 0.0191 | PLoss: -0.0075 | Ent: 0.2146
+[ 87.5%] Iter 8540/9765 | SPS:  2686 | Reward:  0.055 | Value:  0.642 | VLoss: 0.0239 | PLoss: -0.0070 | Ent: 0.2323
+[ 87.5%] Iter 8541/9765 | SPS:  2686 | Reward:  0.057 | Value:  0.636 | VLoss: 0.0202 | PLoss: -0.0027 | Ent: 0.2713
+[ 87.5%] Iter 8542/9765 | SPS:  2686 | Reward:  0.051 | Value:  0.608 | VLoss: 0.0267 | PLoss: -0.0024 | Ent: 0.2671
+[ 87.5%] Iter 8543/9765 | SPS:  2686 | Reward:  0.059 | Value:  0.661 | VLoss: 0.0257 | PLoss: -0.0016 | Ent: 0.2217
+[ 87.5%] Iter 8544/9765 | SPS:  2686 | Reward:  0.068 | Value:  0.700 | VLoss: 0.0193 | PLoss: -0.0022 | Ent: 0.2453
+[ 87.5%] Iter 8545/9765 | SPS:  2686 | Reward:  0.071 | Value:  0.643 | VLoss: 0.0223 | PLoss: -0.0021 | Ent: 0.2103
+[ 87.5%] Iter 8546/9765 | SPS:  2686 | Reward:  0.067 | Value:  0.766 | VLoss: 0.0209 | PLoss: -0.0072 | Ent: 0.2273
+[ 87.5%] Iter 8547/9765 | SPS:  2686 | Reward:  0.056 | Value:  0.720 | VLoss: 0.0274 | PLoss: -0.0021 | Ent: 0.2176
+[ 87.5%] Iter 8548/9765 | SPS:  2686 | Reward:  0.065 | Value:  0.631 | VLoss: 0.0276 | PLoss: -0.0030 | Ent: 0.2513
+[ 87.5%] Iter 8549/9765 | SPS:  2686 | Reward:  0.068 | Value:  0.723 | VLoss: 0.0208 | PLoss: -0.0033 | Ent: 0.2107
+[ 87.6%] Iter 8550/9765 | SPS:  2686 | Reward:  0.058 | Value:  0.766 | VLoss: 0.0153 | PLoss: -0.0059 | Ent: 0.1990
+[ 87.6%] Iter 8551/9765 | SPS:  2687 | Reward:  0.059 | Value:  0.696 | VLoss: 0.0266 | PLoss: -0.0035 | Ent: 0.2321
+[ 87.6%] Iter 8552/9765 | SPS:  2687 | Reward:  0.057 | Value:  0.681 | VLoss: 0.0251 | PLoss: -0.0050 | Ent: 0.2136
+[ 87.6%] Iter 8553/9765 | SPS:  2687 | Reward:  0.065 | Value:  0.700 | VLoss: 0.0255 | PLoss: -0.0037 | Ent: 0.2203
+[ 87.6%] Iter 8554/9765 | SPS:  2687 | Reward:  0.054 | Value:  0.661 | VLoss: 0.0187 | PLoss: -0.0020 | Ent: 0.2763
+[ 87.6%] Iter 8555/9765 | SPS:  2687 | Reward:  0.069 | Value:  0.722 | VLoss: 0.0203 | PLoss: -0.0079 | Ent: 0.2451
+[ 87.6%] Iter 8556/9765 | SPS:  2687 | Reward:  0.071 | Value:  0.733 | VLoss: 0.0156 | PLoss: -0.0054 | Ent: 0.2418
+[ 87.6%] Iter 8557/9765 | SPS:  2687 | Reward:  0.056 | Value:  0.785 | VLoss: 0.0231 | PLoss: -0.0022 | Ent: 0.2358
+[ 87.6%] Iter 8558/9765 | SPS:  2687 | Reward:  0.063 | Value:  0.754 | VLoss: 0.0354 | PLoss: 0.0008 | Ent: 0.1873
+[ 87.6%] Iter 8559/9765 | SPS:  2687 | Reward:  0.053 | Value:  0.681 | VLoss: 0.0211 | PLoss: -0.0049 | Ent: 0.2369
+[ 87.7%] Iter 8560/9765 | SPS:  2687 | Reward:  0.054 | Value:  0.605 | VLoss: 0.0160 | PLoss: -0.0056 | Ent: 0.1942
+[ 87.7%] Iter 8561/9765 | SPS:  2687 | Reward:  0.069 | Value:  0.673 | VLoss: 0.0215 | PLoss: -0.0026 | Ent: 0.1930
+[ 87.7%] Iter 8562/9765 | SPS:  2687 | Reward:  0.066 | Value:  0.684 | VLoss: 0.0223 | PLoss: -0.0002 | Ent: 0.2675
+[ 87.7%] Iter 8563/9765 | SPS:  2688 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0198 | PLoss: -0.0020 | Ent: 0.2121
+[ 87.7%] Iter 8564/9765 | SPS:  2688 | Reward:  0.052 | Value:  0.681 | VLoss: 0.0179 | PLoss: 0.0008 | Ent: 0.2344
+[ 87.7%] Iter 8565/9765 | SPS:  2688 | Reward:  0.047 | Value:  0.725 | VLoss: 0.0199 | PLoss: -0.0043 | Ent: 0.2387
+[ 87.7%] Iter 8566/9765 | SPS:  2688 | Reward:  0.062 | Value:  0.683 | VLoss: 0.0217 | PLoss: -0.0036 | Ent: 0.2700
+[ 87.7%] Iter 8567/9765 | SPS:  2688 | Reward:  0.057 | Value:  0.722 | VLoss: 0.0160 | PLoss: -0.0023 | Ent: 0.2025
+[ 87.7%] Iter 8568/9765 | SPS:  2688 | Reward:  0.066 | Value:  0.747 | VLoss: 0.0186 | PLoss: -0.0032 | Ent: 0.2046
+[ 87.8%] Iter 8569/9765 | SPS:  2688 | Reward:  0.066 | Value:  0.744 | VLoss: 0.0235 | PLoss: -0.0074 | Ent: 0.2118
+[ 87.8%] Iter 8570/9765 | SPS:  2688 | Reward:  0.055 | Value:  0.728 | VLoss: 0.0219 | PLoss: -0.0015 | Ent: 0.2195
+[ 87.8%] Iter 8571/9765 | SPS:  2688 | Reward:  0.077 | Value:  0.740 | VLoss: 0.0248 | PLoss: -0.0060 | Ent: 0.2420
+[ 87.8%] Iter 8572/9765 | SPS:  2688 | Reward:  0.049 | Value:  0.702 | VLoss: 0.0205 | PLoss: -0.0084 | Ent: 0.1841
+[ 87.8%] Iter 8573/9765 | SPS:  2688 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0214 | PLoss: -0.0077 | Ent: 0.2270
+[ 87.8%] Iter 8574/9765 | SPS:  2689 | Reward:  0.056 | Value:  0.769 | VLoss: 0.0173 | PLoss: -0.0056 | Ent: 0.2022
+[ 87.8%] Iter 8575/9765 | SPS:  2689 | Reward:  0.052 | Value:  0.702 | VLoss: 0.0179 | PLoss: -0.0043 | Ent: 0.1766
+[ 87.8%] Iter 8576/9765 | SPS:  2689 | Reward:  0.066 | Value:  0.689 | VLoss: 0.0264 | PLoss: -0.0067 | Ent: 0.2012
+[ 87.8%] Iter 8577/9765 | SPS:  2689 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0221 | PLoss: -0.0050 | Ent: 0.2344
+[ 87.8%] Iter 8578/9765 | SPS:  2689 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0289 | PLoss: -0.0075 | Ent: 0.2026
+[ 87.9%] Iter 8579/9765 | SPS:  2689 | Reward:  0.050 | Value:  0.724 | VLoss: 0.0212 | PLoss: -0.0007 | Ent: 0.2490
+[ 87.9%] Iter 8580/9765 | SPS:  2689 | Reward:  0.058 | Value:  0.611 | VLoss: 0.0248 | PLoss: -0.0008 | Ent: 0.2080
+[ 87.9%] Iter 8581/9765 | SPS:  2689 | Reward:  0.066 | Value:  0.732 | VLoss: 0.0207 | PLoss: -0.0007 | Ent: 0.2013
+[ 87.9%] Iter 8582/9765 | SPS:  2689 | Reward:  0.065 | Value:  0.681 | VLoss: 0.0288 | PLoss: -0.0025 | Ent: 0.2173
+[ 87.9%] Iter 8583/9765 | SPS:  2689 | Reward:  0.048 | Value:  0.753 | VLoss: 0.0161 | PLoss: -0.0031 | Ent: 0.2213
+[ 87.9%] Iter 8584/9765 | SPS:  2689 | Reward:  0.054 | Value:  0.676 | VLoss: 0.0156 | PLoss: -0.0035 | Ent: 0.2738
+[ 87.9%] Iter 8585/9765 | SPS:  2690 | Reward:  0.051 | Value:  0.685 | VLoss: 0.0162 | PLoss: -0.0008 | Ent: 0.2597
+[ 87.9%] Iter 8586/9765 | SPS:  2690 | Reward:  0.059 | Value:  0.737 | VLoss: 0.0190 | PLoss: -0.0060 | Ent: 0.2007
+[ 87.9%] Iter 8587/9765 | SPS:  2690 | Reward:  0.067 | Value:  0.690 | VLoss: 0.0206 | PLoss: -0.0025 | Ent: 0.2121
+[ 87.9%] Iter 8588/9765 | SPS:  2690 | Reward:  0.069 | Value:  0.753 | VLoss: 0.0223 | PLoss: -0.0001 | Ent: 0.2171
+[ 88.0%] Iter 8589/9765 | SPS:  2690 | Reward:  0.066 | Value:  0.746 | VLoss: 0.0252 | PLoss: -0.0018 | Ent: 0.2306
+[ 88.0%] Iter 8590/9765 | SPS:  2690 | Reward:  0.062 | Value:  0.775 | VLoss: 0.0212 | PLoss: -0.0029 | Ent: 0.2093
+[ 88.0%] Iter 8591/9765 | SPS:  2690 | Reward:  0.054 | Value:  0.694 | VLoss: 0.0276 | PLoss: -0.0056 | Ent: 0.2573
+[ 88.0%] Iter 8592/9765 | SPS:  2690 | Reward:  0.062 | Value:  0.772 | VLoss: 0.0152 | PLoss: -0.0026 | Ent: 0.1823
+[ 88.0%] Iter 8593/9765 | SPS:  2690 | Reward:  0.068 | Value:  0.707 | VLoss: 0.0195 | PLoss: 0.0002 | Ent: 0.2409
+[ 88.0%] Iter 8594/9765 | SPS:  2690 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0165 | PLoss: -0.0018 | Ent: 0.2358
+[ 88.0%] Iter 8595/9765 | SPS:  2690 | Reward:  0.047 | Value:  0.703 | VLoss: 0.0241 | PLoss: -0.0031 | Ent: 0.2348
+[ 88.0%] Iter 8596/9765 | SPS:  2691 | Reward:  0.058 | Value:  0.686 | VLoss: 0.0185 | PLoss: -0.0034 | Ent: 0.2131
+[ 88.0%] Iter 8597/9765 | SPS:  2691 | Reward:  0.046 | Value:  0.750 | VLoss: 0.0178 | PLoss: -0.0066 | Ent: 0.2343
+[ 88.0%] Iter 8598/9765 | SPS:  2691 | Reward:  0.063 | Value:  0.702 | VLoss: 0.0177 | PLoss: -0.0066 | Ent: 0.1969
+[ 88.1%] Iter 8599/9765 | SPS:  2691 | Reward:  0.072 | Value:  0.728 | VLoss: 0.0234 | PLoss: -0.0018 | Ent: 0.1929
+[ 88.1%] Iter 8600/9765 | SPS:  2691 | Reward:  0.075 | Value:  0.831 | VLoss: 0.0231 | PLoss: -0.0017 | Ent: 0.2196
+[ 88.1%] Iter 8601/9765 | SPS:  2691 | Reward:  0.073 | Value:  0.781 | VLoss: 0.0260 | PLoss: -0.0050 | Ent: 0.2119
+[ 88.1%] Iter 8602/9765 | SPS:  2691 | Reward:  0.067 | Value:  0.739 | VLoss: 0.0222 | PLoss: -0.0066 | Ent: 0.1889
+[ 88.1%] Iter 8603/9765 | SPS:  2691 | Reward:  0.051 | Value:  0.667 | VLoss: 0.0266 | PLoss: -0.0068 | Ent: 0.2278
+[ 88.1%] Iter 8604/9765 | SPS:  2691 | Reward:  0.054 | Value:  0.707 | VLoss: 0.0105 | PLoss: -0.0055 | Ent: 0.2833
+[ 88.1%] Iter 8605/9765 | SPS:  2691 | Reward:  0.050 | Value:  0.669 | VLoss: 0.0222 | PLoss: -0.0057 | Ent: 0.2407
+[ 88.1%] Iter 8606/9765 | SPS:  2691 | Reward:  0.058 | Value:  0.643 | VLoss: 0.0197 | PLoss: -0.0007 | Ent: 0.2739
+[ 88.1%] Iter 8607/9765 | SPS:  2691 | Reward:  0.058 | Value:  0.757 | VLoss: 0.0146 | PLoss: -0.0022 | Ent: 0.2781
+[ 88.2%] Iter 8608/9765 | SPS:  2692 | Reward:  0.064 | Value:  0.763 | VLoss: 0.0267 | PLoss: -0.0070 | Ent: 0.2093
+[ 88.2%] Iter 8609/9765 | SPS:  2692 | Reward:  0.062 | Value:  0.776 | VLoss: 0.0152 | PLoss: -0.0045 | Ent: 0.2480
+[ 88.2%] Iter 8610/9765 | SPS:  2692 | Reward:  0.067 | Value:  0.752 | VLoss: 0.0228 | PLoss: -0.0026 | Ent: 0.2217
+[ 88.2%] Iter 8611/9765 | SPS:  2692 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0192 | PLoss: -0.0039 | Ent: 0.2165
+[ 88.2%] Iter 8612/9765 | SPS:  2692 | Reward:  0.078 | Value:  0.748 | VLoss: 0.0155 | PLoss: -0.0038 | Ent: 0.1808
+[ 88.2%] Iter 8613/9765 | SPS:  2692 | Reward:  0.069 | Value:  0.796 | VLoss: 0.0186 | PLoss: -0.0028 | Ent: 0.2232
+[ 88.2%] Iter 8614/9765 | SPS:  2692 | Reward:  0.069 | Value:  0.764 | VLoss: 0.0171 | PLoss: -0.0020 | Ent: 0.1841
+[ 88.2%] Iter 8615/9765 | SPS:  2692 | Reward:  0.064 | Value:  0.728 | VLoss: 0.0248 | PLoss: -0.0026 | Ent: 0.2180
+[ 88.2%] Iter 8616/9765 | SPS:  2692 | Reward:  0.060 | Value:  0.719 | VLoss: 0.0277 | PLoss: -0.0031 | Ent: 0.2163
+[ 88.2%] Iter 8617/9765 | SPS:  2692 | Reward:  0.080 | Value:  0.747 | VLoss: 0.0316 | PLoss: -0.0051 | Ent: 0.2334
+[ 88.3%] Iter 8618/9765 | SPS:  2692 | Reward:  0.074 | Value:  0.773 | VLoss: 0.0222 | PLoss: -0.0045 | Ent: 0.2200
+[ 88.3%] Iter 8619/9765 | SPS:  2692 | Reward:  0.061 | Value:  0.766 | VLoss: 0.0268 | PLoss: -0.0029 | Ent: 0.2387
+[ 88.3%] Iter 8620/9765 | SPS:  2693 | Reward:  0.061 | Value:  0.727 | VLoss: 0.0223 | PLoss: -0.0005 | Ent: 0.1794
+[ 88.3%] Iter 8621/9765 | SPS:  2693 | Reward:  0.057 | Value:  0.603 | VLoss: 0.0306 | PLoss: -0.0004 | Ent: 0.2543
+[ 88.3%] Iter 8622/9765 | SPS:  2693 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0306 | PLoss: -0.0094 | Ent: 0.2410
+[ 88.3%] Iter 8623/9765 | SPS:  2693 | Reward:  0.057 | Value:  0.723 | VLoss: 0.0192 | PLoss: -0.0057 | Ent: 0.2613
+[ 88.3%] Iter 8624/9765 | SPS:  2693 | Reward:  0.056 | Value:  0.680 | VLoss: 0.0219 | PLoss: -0.0056 | Ent: 0.2829
+[ 88.3%] Iter 8625/9765 | SPS:  2693 | Reward:  0.069 | Value:  0.725 | VLoss: 0.0249 | PLoss: -0.0077 | Ent: 0.2142
+[ 88.3%] Iter 8626/9765 | SPS:  2693 | Reward:  0.060 | Value:  0.718 | VLoss: 0.0256 | PLoss: -0.0063 | Ent: 0.2197
+[ 88.3%] Iter 8627/9765 | SPS:  2693 | Reward:  0.054 | Value:  0.645 | VLoss: 0.0290 | PLoss: -0.0029 | Ent: 0.2757
+[ 88.4%] Iter 8628/9765 | SPS:  2693 | Reward:  0.074 | Value:  0.742 | VLoss: 0.0257 | PLoss: -0.0039 | Ent: 0.2273
+[ 88.4%] Iter 8629/9765 | SPS:  2693 | Reward:  0.048 | Value:  0.766 | VLoss: 0.0223 | PLoss: -0.0027 | Ent: 0.2170
+[ 88.4%] Iter 8630/9765 | SPS:  2693 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0264 | PLoss: -0.0038 | Ent: 0.2589
+[ 88.4%] Iter 8631/9765 | SPS:  2693 | Reward:  0.047 | Value:  0.706 | VLoss: 0.0242 | PLoss: -0.0024 | Ent: 0.2230
+[ 88.4%] Iter 8632/9765 | SPS:  2694 | Reward:  0.079 | Value:  0.731 | VLoss: 0.0234 | PLoss: -0.0034 | Ent: 0.2073
+[ 88.4%] Iter 8633/9765 | SPS:  2694 | Reward:  0.052 | Value:  0.686 | VLoss: 0.0190 | PLoss: -0.0025 | Ent: 0.1828
+[ 88.4%] Iter 8634/9765 | SPS:  2694 | Reward:  0.065 | Value:  0.711 | VLoss: 0.0186 | PLoss: -0.0039 | Ent: 0.2175
+[ 88.4%] Iter 8635/9765 | SPS:  2694 | Reward:  0.055 | Value:  0.678 | VLoss: 0.0258 | PLoss: -0.0047 | Ent: 0.2482
+[ 88.4%] Iter 8636/9765 | SPS:  2694 | Reward:  0.051 | Value:  0.691 | VLoss: 0.0164 | PLoss: -0.0040 | Ent: 0.2696
+[ 88.4%] Iter 8637/9765 | SPS:  2694 | Reward:  0.056 | Value:  0.616 | VLoss: 0.0259 | PLoss: -0.0027 | Ent: 0.2027
+[ 88.5%] Iter 8638/9765 | SPS:  2694 | Reward:  0.070 | Value:  0.745 | VLoss: 0.0169 | PLoss: -0.0038 | Ent: 0.1983
+[ 88.5%] Iter 8639/9765 | SPS:  2694 | Reward:  0.051 | Value:  0.744 | VLoss: 0.0175 | PLoss: -0.0034 | Ent: 0.2073
+[ 88.5%] Iter 8640/9765 | SPS:  2694 | Reward:  0.066 | Value:  0.738 | VLoss: 0.0218 | PLoss: -0.0043 | Ent: 0.2131
+[ 88.5%] Iter 8641/9765 | SPS:  2694 | Reward:  0.063 | Value:  0.781 | VLoss: 0.0163 | PLoss: -0.0007 | Ent: 0.2172
+[ 88.5%] Iter 8642/9765 | SPS:  2694 | Reward:  0.072 | Value:  0.733 | VLoss: 0.0284 | PLoss: -0.0070 | Ent: 0.2215
+[ 88.5%] Iter 8643/9765 | SPS:  2694 | Reward:  0.060 | Value:  0.721 | VLoss: 0.0248 | PLoss: -0.0027 | Ent: 0.1791
+[ 88.5%] Iter 8644/9765 | SPS:  2695 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0173 | PLoss: -0.0028 | Ent: 0.1806
+[ 88.5%] Iter 8645/9765 | SPS:  2695 | Reward:  0.055 | Value:  0.743 | VLoss: 0.0184 | PLoss: -0.0021 | Ent: 0.2656
+[ 88.5%] Iter 8646/9765 | SPS:  2695 | Reward:  0.066 | Value:  0.715 | VLoss: 0.0225 | PLoss: -0.0046 | Ent: 0.2085
+[ 88.6%] Iter 8647/9765 | SPS:  2695 | Reward:  0.062 | Value:  0.679 | VLoss: 0.0214 | PLoss: -0.0052 | Ent: 0.2335
+[ 88.6%] Iter 8648/9765 | SPS:  2695 | Reward:  0.059 | Value:  0.692 | VLoss: 0.0206 | PLoss: -0.0034 | Ent: 0.2176
+[ 88.6%] Iter 8649/9765 | SPS:  2695 | Reward:  0.055 | Value:  0.686 | VLoss: 0.0172 | PLoss: -0.0030 | Ent: 0.2132
+[ 88.6%] Iter 8650/9765 | SPS:  2695 | Reward:  0.068 | Value:  0.689 | VLoss: 0.0252 | PLoss: -0.0033 | Ent: 0.2443
+[ 88.6%] Iter 8651/9765 | SPS:  2695 | Reward:  0.058 | Value:  0.689 | VLoss: 0.0202 | PLoss: -0.0003 | Ent: 0.2444
+[ 88.6%] Iter 8652/9765 | SPS:  2695 | Reward:  0.061 | Value:  0.768 | VLoss: 0.0181 | PLoss: -0.0048 | Ent: 0.2469
+[ 88.6%] Iter 8653/9765 | SPS:  2695 | Reward:  0.059 | Value:  0.780 | VLoss: 0.0207 | PLoss: -0.0067 | Ent: 0.1683
+[ 88.6%] Iter 8654/9765 | SPS:  2695 | Reward:  0.062 | Value:  0.654 | VLoss: 0.0112 | PLoss: -0.0050 | Ent: 0.2162
+[ 88.6%] Iter 8655/9765 | SPS:  2695 | Reward:  0.066 | Value:  0.728 | VLoss: 0.0260 | PLoss: -0.0027 | Ent: 0.1419
+[ 88.6%] Iter 8656/9765 | SPS:  2696 | Reward:  0.076 | Value:  0.824 | VLoss: 0.0246 | PLoss: -0.0024 | Ent: 0.1654
+[ 88.7%] Iter 8657/9765 | SPS:  2696 | Reward:  0.046 | Value:  0.801 | VLoss: 0.0202 | PLoss: -0.0027 | Ent: 0.2404
+[ 88.7%] Iter 8658/9765 | SPS:  2696 | Reward:  0.067 | Value:  0.768 | VLoss: 0.0192 | PLoss: -0.0079 | Ent: 0.2769
+[ 88.7%] Iter 8659/9765 | SPS:  2696 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0235 | PLoss: -0.0031 | Ent: 0.2037
+[ 88.7%] Iter 8660/9765 | SPS:  2696 | Reward:  0.054 | Value:  0.754 | VLoss: 0.0141 | PLoss: -0.0029 | Ent: 0.2295
+[ 88.7%] Iter 8661/9765 | SPS:  2696 | Reward:  0.065 | Value:  0.788 | VLoss: 0.0230 | PLoss: -0.0058 | Ent: 0.2189
+[ 88.7%] Iter 8662/9765 | SPS:  2696 | Reward:  0.043 | Value:  0.637 | VLoss: 0.0215 | PLoss: -0.0040 | Ent: 0.2438
+[ 88.7%] Iter 8663/9765 | SPS:  2696 | Reward:  0.052 | Value:  0.700 | VLoss: 0.0234 | PLoss: -0.0054 | Ent: 0.2346
+[ 88.7%] Iter 8664/9765 | SPS:  2696 | Reward:  0.071 | Value:  0.675 | VLoss: 0.0172 | PLoss: -0.0023 | Ent: 0.2351
+[ 88.7%] Iter 8665/9765 | SPS:  2696 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0200 | PLoss: -0.0029 | Ent: 0.2578
+[ 88.7%] Iter 8666/9765 | SPS:  2697 | Reward:  0.057 | Value:  0.738 | VLoss: 0.0245 | PLoss: -0.0022 | Ent: 0.2142
+[ 88.8%] Iter 8667/9765 | SPS:  2697 | Reward:  0.068 | Value:  0.745 | VLoss: 0.0200 | PLoss: -0.0087 | Ent: 0.2270
+[ 88.8%] Iter 8668/9765 | SPS:  2697 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0226 | PLoss: -0.0027 | Ent: 0.2049
+[ 88.8%] Iter 8669/9765 | SPS:  2697 | Reward:  0.063 | Value:  0.667 | VLoss: 0.0284 | PLoss: -0.0026 | Ent: 0.2696
+[ 88.8%] Iter 8670/9765 | SPS:  2697 | Reward:  0.050 | Value:  0.699 | VLoss: 0.0241 | PLoss: -0.0069 | Ent: 0.2208
+[ 88.8%] Iter 8671/9765 | SPS:  2697 | Reward:  0.053 | Value:  0.668 | VLoss: 0.0255 | PLoss: -0.0007 | Ent: 0.2015
+[ 88.8%] Iter 8672/9765 | SPS:  2697 | Reward:  0.054 | Value:  0.711 | VLoss: 0.0185 | PLoss: -0.0029 | Ent: 0.2157
+[ 88.8%] Iter 8673/9765 | SPS:  2697 | Reward:  0.060 | Value:  0.675 | VLoss: 0.0242 | PLoss: -0.0046 | Ent: 0.2503
+[ 88.8%] Iter 8674/9765 | SPS:  2697 | Reward:  0.061 | Value:  0.701 | VLoss: 0.0216 | PLoss: -0.0033 | Ent: 0.2753
+[ 88.8%] Iter 8675/9765 | SPS:  2697 | Reward:  0.056 | Value:  0.728 | VLoss: 0.0270 | PLoss: -0.0051 | Ent: 0.2605
+[ 88.8%] Iter 8676/9765 | SPS:  2697 | Reward:  0.061 | Value:  0.682 | VLoss: 0.0229 | PLoss: -0.0035 | Ent: 0.2471
+[ 88.9%] Iter 8677/9765 | SPS:  2698 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0238 | PLoss: -0.0045 | Ent: 0.2022
+[ 88.9%] Iter 8678/9765 | SPS:  2698 | Reward:  0.056 | Value:  0.711 | VLoss: 0.0204 | PLoss: -0.0011 | Ent: 0.2463
+[ 88.9%] Iter 8679/9765 | SPS:  2698 | Reward:  0.050 | Value:  0.736 | VLoss: 0.0214 | PLoss: -0.0054 | Ent: 0.2243
+[ 88.9%] Iter 8680/9765 | SPS:  2698 | Reward:  0.068 | Value:  0.697 | VLoss: 0.0288 | PLoss: -0.0049 | Ent: 0.1923
+[ 88.9%] Iter 8681/9765 | SPS:  2698 | Reward:  0.056 | Value:  0.662 | VLoss: 0.0204 | PLoss: -0.0054 | Ent: 0.2729
+[ 88.9%] Iter 8682/9765 | SPS:  2698 | Reward:  0.046 | Value:  0.666 | VLoss: 0.0188 | PLoss: -0.0013 | Ent: 0.2282
+[ 88.9%] Iter 8683/9765 | SPS:  2698 | Reward:  0.045 | Value:  0.663 | VLoss: 0.0145 | PLoss: -0.0011 | Ent: 0.2970
+[ 88.9%] Iter 8684/9765 | SPS:  2698 | Reward:  0.044 | Value:  0.669 | VLoss: 0.0163 | PLoss: -0.0023 | Ent: 0.2576
+[ 88.9%] Iter 8685/9765 | SPS:  2698 | Reward:  0.068 | Value:  0.674 | VLoss: 0.0161 | PLoss: -0.0048 | Ent: 0.2500
+[ 89.0%] Iter 8686/9765 | SPS:  2698 | Reward:  0.071 | Value:  0.657 | VLoss: 0.0242 | PLoss: -0.0031 | Ent: 0.2546
+[ 89.0%] Iter 8687/9765 | SPS:  2698 | Reward:  0.061 | Value:  0.741 | VLoss: 0.0239 | PLoss: -0.0058 | Ent: 0.2485
+[ 89.0%] Iter 8688/9765 | SPS:  2698 | Reward:  0.072 | Value:  0.756 | VLoss: 0.0264 | PLoss: -0.0026 | Ent: 0.2660
+[ 89.0%] Iter 8689/9765 | SPS:  2699 | Reward:  0.069 | Value:  0.778 | VLoss: 0.0183 | PLoss: -0.0007 | Ent: 0.2120
+[ 89.0%] Iter 8690/9765 | SPS:  2699 | Reward:  0.076 | Value:  0.813 | VLoss: 0.0173 | PLoss: -0.0016 | Ent: 0.2486
+[ 89.0%] Iter 8691/9765 | SPS:  2699 | Reward:  0.065 | Value:  0.756 | VLoss: 0.0227 | PLoss: -0.0019 | Ent: 0.2331
+[ 89.0%] Iter 8692/9765 | SPS:  2699 | Reward:  0.065 | Value:  0.718 | VLoss: 0.0225 | PLoss: -0.0054 | Ent: 0.2586
+[ 89.0%] Iter 8693/9765 | SPS:  2699 | Reward:  0.048 | Value:  0.739 | VLoss: 0.0180 | PLoss: -0.0040 | Ent: 0.2345
+[ 89.0%] Iter 8694/9765 | SPS:  2699 | Reward:  0.053 | Value:  0.628 | VLoss: 0.0275 | PLoss: -0.0029 | Ent: 0.2279
+[ 89.0%] Iter 8695/9765 | SPS:  2699 | Reward:  0.066 | Value:  0.742 | VLoss: 0.0174 | PLoss: -0.0014 | Ent: 0.2124
+[ 89.1%] Iter 8696/9765 | SPS:  2699 | Reward:  0.081 | Value:  0.793 | VLoss: 0.0208 | PLoss: -0.0029 | Ent: 0.1764
+[ 89.1%] Iter 8697/9765 | SPS:  2699 | Reward:  0.063 | Value:  0.811 | VLoss: 0.0211 | PLoss: -0.0047 | Ent: 0.2183
+[ 89.1%] Iter 8698/9765 | SPS:  2699 | Reward:  0.079 | Value:  0.810 | VLoss: 0.0209 | PLoss: -0.0027 | Ent: 0.1887
+[ 89.1%] Iter 8699/9765 | SPS:  2699 | Reward:  0.072 | Value:  0.799 | VLoss: 0.0159 | PLoss: -0.0022 | Ent: 0.2523
+[ 89.1%] Iter 8700/9765 | SPS:  2699 | Reward:  0.068 | Value:  0.773 | VLoss: 0.0277 | PLoss: -0.0050 | Ent: 0.2288
+[ 89.1%] Iter 8701/9765 | SPS:  2700 | Reward:  0.070 | Value:  0.698 | VLoss: 0.0223 | PLoss: -0.0071 | Ent: 0.2309
+[ 89.1%] Iter 8702/9765 | SPS:  2700 | Reward:  0.057 | Value:  0.649 | VLoss: 0.0238 | PLoss: -0.0061 | Ent: 0.2413
+[ 89.1%] Iter 8703/9765 | SPS:  2700 | Reward:  0.062 | Value:  0.775 | VLoss: 0.0189 | PLoss: -0.0016 | Ent: 0.2074
+[ 89.1%] Iter 8704/9765 | SPS:  2700 | Reward:  0.060 | Value:  0.742 | VLoss: 0.0202 | PLoss: -0.0007 | Ent: 0.1802
+[ 89.1%] Iter 8705/9765 | SPS:  2700 | Reward:  0.067 | Value:  0.712 | VLoss: 0.0217 | PLoss: -0.0032 | Ent: 0.2355
+[ 89.2%] Iter 8706/9765 | SPS:  2700 | Reward:  0.054 | Value:  0.719 | VLoss: 0.0282 | PLoss: -0.0050 | Ent: 0.2334
+[ 89.2%] Iter 8707/9765 | SPS:  2700 | Reward:  0.067 | Value:  0.660 | VLoss: 0.0257 | PLoss: -0.0025 | Ent: 0.2756
+[ 89.2%] Iter 8708/9765 | SPS:  2700 | Reward:  0.071 | Value:  0.698 | VLoss: 0.0197 | PLoss: -0.0024 | Ent: 0.1833
+[ 89.2%] Iter 8709/9765 | SPS:  2700 | Reward:  0.074 | Value:  0.762 | VLoss: 0.0171 | PLoss: -0.0037 | Ent: 0.2405
+[ 89.2%] Iter 8710/9765 | SPS:  2700 | Reward:  0.074 | Value:  0.795 | VLoss: 0.0188 | PLoss: -0.0001 | Ent: 0.2647
+[ 89.2%] Iter 8711/9765 | SPS:  2700 | Reward:  0.057 | Value:  0.827 | VLoss: 0.0113 | PLoss: -0.0013 | Ent: 0.2418
+[ 89.2%] Iter 8712/9765 | SPS:  2701 | Reward:  0.050 | Value:  0.755 | VLoss: 0.0201 | PLoss: -0.0025 | Ent: 0.1982
+[ 89.2%] Iter 8713/9765 | SPS:  2701 | Reward:  0.060 | Value:  0.750 | VLoss: 0.0194 | PLoss: -0.0011 | Ent: 0.2313
+[ 89.2%] Iter 8714/9765 | SPS:  2701 | Reward:  0.047 | Value:  0.683 | VLoss: 0.0124 | PLoss: -0.0049 | Ent: 0.2165
+[ 89.2%] Iter 8715/9765 | SPS:  2701 | Reward:  0.057 | Value:  0.710 | VLoss: 0.0185 | PLoss: -0.0063 | Ent: 0.2109
+[ 89.3%] Iter 8716/9765 | SPS:  2701 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0185 | PLoss: -0.0048 | Ent: 0.1838
+[ 89.3%] Iter 8717/9765 | SPS:  2701 | Reward:  0.063 | Value:  0.765 | VLoss: 0.0268 | PLoss: -0.0039 | Ent: 0.2195
+[ 89.3%] Iter 8718/9765 | SPS:  2701 | Reward:  0.066 | Value:  0.744 | VLoss: 0.0237 | PLoss: -0.0024 | Ent: 0.3095
+[ 89.3%] Iter 8719/9765 | SPS:  2701 | Reward:  0.072 | Value:  0.786 | VLoss: 0.0205 | PLoss: -0.0023 | Ent: 0.2051
+[ 89.3%] Iter 8720/9765 | SPS:  2701 | Reward:  0.062 | Value:  0.788 | VLoss: 0.0234 | PLoss: -0.0024 | Ent: 0.2008
+[ 89.3%] Iter 8721/9765 | SPS:  2701 | Reward:  0.055 | Value:  0.767 | VLoss: 0.0245 | PLoss: -0.0021 | Ent: 0.2363
+[ 89.3%] Iter 8722/9765 | SPS:  2701 | Reward:  0.079 | Value:  0.779 | VLoss: 0.0216 | PLoss: -0.0046 | Ent: 0.2116
+[ 89.3%] Iter 8723/9765 | SPS:  2702 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0201 | PLoss: -0.0027 | Ent: 0.2074
+[ 89.3%] Iter 8724/9765 | SPS:  2702 | Reward:  0.071 | Value:  0.739 | VLoss: 0.0220 | PLoss: -0.0010 | Ent: 0.2425
+[ 89.3%] Iter 8725/9765 | SPS:  2702 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0233 | PLoss: -0.0039 | Ent: 0.2542
+[ 89.4%] Iter 8726/9765 | SPS:  2702 | Reward:  0.053 | Value:  0.718 | VLoss: 0.0250 | PLoss: -0.0027 | Ent: 0.2352
+[ 89.4%] Iter 8727/9765 | SPS:  2702 | Reward:  0.074 | Value:  0.755 | VLoss: 0.0156 | PLoss: -0.0055 | Ent: 0.2168
+[ 89.4%] Iter 8728/9765 | SPS:  2702 | Reward:  0.057 | Value:  0.687 | VLoss: 0.0271 | PLoss: -0.0034 | Ent: 0.2129
+[ 89.4%] Iter 8729/9765 | SPS:  2702 | Reward:  0.066 | Value:  0.688 | VLoss: 0.0238 | PLoss: -0.0035 | Ent: 0.2046
+[ 89.4%] Iter 8730/9765 | SPS:  2702 | Reward:  0.069 | Value:  0.711 | VLoss: 0.0186 | PLoss: -0.0025 | Ent: 0.2432
+[ 89.4%] Iter 8731/9765 | SPS:  2702 | Reward:  0.061 | Value:  0.769 | VLoss: 0.0226 | PLoss: -0.0047 | Ent: 0.2557
+[ 89.4%] Iter 8732/9765 | SPS:  2702 | Reward:  0.066 | Value:  0.781 | VLoss: 0.0198 | PLoss: -0.0004 | Ent: 0.2283
+[ 89.4%] Iter 8733/9765 | SPS:  2702 | Reward:  0.064 | Value:  0.787 | VLoss: 0.0103 | PLoss: -0.0026 | Ent: 0.2081
+[ 89.4%] Iter 8734/9765 | SPS:  2702 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0183 | PLoss: -0.0017 | Ent: 0.1904
+[ 89.5%] Iter 8735/9765 | SPS:  2703 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0293 | PLoss: -0.0016 | Ent: 0.2004
+[ 89.5%] Iter 8736/9765 | SPS:  2703 | Reward:  0.062 | Value:  0.687 | VLoss: 0.0224 | PLoss: -0.0038 | Ent: 0.2222
+[ 89.5%] Iter 8737/9765 | SPS:  2703 | Reward:  0.068 | Value:  0.787 | VLoss: 0.0285 | PLoss: -0.0036 | Ent: 0.2483
+[ 89.5%] Iter 8738/9765 | SPS:  2703 | Reward:  0.055 | Value:  0.728 | VLoss: 0.0170 | PLoss: -0.0037 | Ent: 0.2556
+[ 89.5%] Iter 8739/9765 | SPS:  2703 | Reward:  0.043 | Value:  0.719 | VLoss: 0.0175 | PLoss: -0.0011 | Ent: 0.2223
+[ 89.5%] Iter 8740/9765 | SPS:  2703 | Reward:  0.045 | Value:  0.693 | VLoss: 0.0196 | PLoss: -0.0036 | Ent: 0.1893
+[ 89.5%] Iter 8741/9765 | SPS:  2703 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0154 | PLoss: -0.0052 | Ent: 0.2038
+[ 89.5%] Iter 8742/9765 | SPS:  2703 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0244 | PLoss: -0.0066 | Ent: 0.2154
+[ 89.5%] Iter 8743/9765 | SPS:  2703 | Reward:  0.057 | Value:  0.770 | VLoss: 0.0168 | PLoss: -0.0019 | Ent: 0.2389
+[ 89.5%] Iter 8744/9765 | SPS:  2703 | Reward:  0.064 | Value:  0.743 | VLoss: 0.0217 | PLoss: -0.0040 | Ent: 0.2514
+[ 89.6%] Iter 8745/9765 | SPS:  2703 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0215 | PLoss: -0.0036 | Ent: 0.2525
+[ 89.6%] Iter 8746/9765 | SPS:  2704 | Reward:  0.064 | Value:  0.713 | VLoss: 0.0229 | PLoss: -0.0091 | Ent: 0.2639
+[ 89.6%] Iter 8747/9765 | SPS:  2704 | Reward:  0.057 | Value:  0.739 | VLoss: 0.0183 | PLoss: -0.0014 | Ent: 0.2253
+[ 89.6%] Iter 8748/9765 | SPS:  2704 | Reward:  0.067 | Value:  0.740 | VLoss: 0.0223 | PLoss: -0.0022 | Ent: 0.2347
+[ 89.6%] Iter 8749/9765 | SPS:  2704 | Reward:  0.069 | Value:  0.735 | VLoss: 0.0284 | PLoss: -0.0052 | Ent: 0.2479
+[ 89.6%] Iter 8750/9765 | SPS:  2704 | Reward:  0.054 | Value:  0.718 | VLoss: 0.0185 | PLoss: -0.0008 | Ent: 0.2322
+[ 89.6%] Iter 8751/9765 | SPS:  2704 | Reward:  0.063 | Value:  0.726 | VLoss: 0.0226 | PLoss: -0.0054 | Ent: 0.2288
+[ 89.6%] Iter 8752/9765 | SPS:  2704 | Reward:  0.043 | Value:  0.625 | VLoss: 0.0223 | PLoss: -0.0024 | Ent: 0.2920
+[ 89.6%] Iter 8753/9765 | SPS:  2704 | Reward:  0.058 | Value:  0.754 | VLoss: 0.0167 | PLoss: -0.0041 | Ent: 0.2474
+[ 89.6%] Iter 8754/9765 | SPS:  2704 | Reward:  0.069 | Value:  0.781 | VLoss: 0.0212 | PLoss: -0.0023 | Ent: 0.2819
+[ 89.7%] Iter 8755/9765 | SPS:  2704 | Reward:  0.053 | Value:  0.739 | VLoss: 0.0244 | PLoss: 0.0014 | Ent: 0.3075
+[ 89.7%] Iter 8756/9765 | SPS:  2704 | Reward:  0.052 | Value:  0.651 | VLoss: 0.0226 | PLoss: -0.0015 | Ent: 0.2689
+[ 89.7%] Iter 8757/9765 | SPS:  2705 | Reward:  0.059 | Value:  0.663 | VLoss: 0.0229 | PLoss: -0.0021 | Ent: 0.2717
+[ 89.7%] Iter 8758/9765 | SPS:  2705 | Reward:  0.057 | Value:  0.706 | VLoss: 0.0182 | PLoss: -0.0002 | Ent: 0.2183
+[ 89.7%] Iter 8759/9765 | SPS:  2705 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0201 | PLoss: -0.0016 | Ent: 0.2590
+[ 89.7%] Iter 8760/9765 | SPS:  2705 | Reward:  0.061 | Value:  0.778 | VLoss: 0.0186 | PLoss: -0.0037 | Ent: 0.2468
+[ 89.7%] Iter 8761/9765 | SPS:  2705 | Reward:  0.049 | Value:  0.733 | VLoss: 0.0177 | PLoss: -0.0014 | Ent: 0.2082
+[ 89.7%] Iter 8762/9765 | SPS:  2705 | Reward:  0.059 | Value:  0.723 | VLoss: 0.0178 | PLoss: -0.0034 | Ent: 0.2337
+[ 89.7%] Iter 8763/9765 | SPS:  2705 | Reward:  0.067 | Value:  0.758 | VLoss: 0.0265 | PLoss: -0.0054 | Ent: 0.1866
+[ 89.7%] Iter 8764/9765 | SPS:  2705 | Reward:  0.072 | Value:  0.717 | VLoss: 0.0279 | PLoss: -0.0017 | Ent: 0.2387
+[ 89.8%] Iter 8765/9765 | SPS:  2705 | Reward:  0.055 | Value:  0.639 | VLoss: 0.0191 | PLoss: -0.0005 | Ent: 0.2325
+[ 89.8%] Iter 8766/9765 | SPS:  2705 | Reward:  0.059 | Value:  0.719 | VLoss: 0.0167 | PLoss: -0.0022 | Ent: 0.2071
+[ 89.8%] Iter 8767/9765 | SPS:  2705 | Reward:  0.074 | Value:  0.732 | VLoss: 0.0181 | PLoss: -0.0050 | Ent: 0.2253
+[ 89.8%] Iter 8768/9765 | SPS:  2705 | Reward:  0.054 | Value:  0.738 | VLoss: 0.0180 | PLoss: -0.0016 | Ent: 0.2568
+[ 89.8%] Iter 8769/9765 | SPS:  2706 | Reward:  0.057 | Value:  0.639 | VLoss: 0.0299 | PLoss: -0.0093 | Ent: 0.2363
+[ 89.8%] Iter 8770/9765 | SPS:  2706 | Reward:  0.063 | Value:  0.742 | VLoss: 0.0211 | PLoss: -0.0064 | Ent: 0.2132
+[ 89.8%] Iter 8771/9765 | SPS:  2706 | Reward:  0.048 | Value:  0.680 | VLoss: 0.0225 | PLoss: -0.0053 | Ent: 0.2217
+[ 89.8%] Iter 8772/9765 | SPS:  2706 | Reward:  0.070 | Value:  0.611 | VLoss: 0.0242 | PLoss: -0.0009 | Ent: 0.2626
+[ 89.8%] Iter 8773/9765 | SPS:  2706 | Reward:  0.063 | Value:  0.721 | VLoss: 0.0227 | PLoss: -0.0013 | Ent: 0.2146
+[ 89.9%] Iter 8774/9765 | SPS:  2706 | Reward:  0.075 | Value:  0.797 | VLoss: 0.0215 | PLoss: -0.0009 | Ent: 0.2151
+[ 89.9%] Iter 8775/9765 | SPS:  2706 | Reward:  0.055 | Value:  0.744 | VLoss: 0.0219 | PLoss: -0.0032 | Ent: 0.1993
+[ 89.9%] Iter 8776/9765 | SPS:  2706 | Reward:  0.066 | Value:  0.715 | VLoss: 0.0184 | PLoss: -0.0021 | Ent: 0.1966
+[ 89.9%] Iter 8777/9765 | SPS:  2706 | Reward:  0.060 | Value:  0.765 | VLoss: 0.0141 | PLoss: -0.0040 | Ent: 0.2489
+[ 89.9%] Iter 8778/9765 | SPS:  2706 | Reward:  0.067 | Value:  0.764 | VLoss: 0.0225 | PLoss: -0.0031 | Ent: 0.2112
+[ 89.9%] Iter 8779/9765 | SPS:  2706 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0209 | PLoss: -0.0027 | Ent: 0.2371
+[ 89.9%] Iter 8780/9765 | SPS:  2706 | Reward:  0.059 | Value:  0.726 | VLoss: 0.0166 | PLoss: -0.0032 | Ent: 0.2185
+[ 89.9%] Iter 8781/9765 | SPS:  2707 | Reward:  0.055 | Value:  0.680 | VLoss: 0.0236 | PLoss: -0.0011 | Ent: 0.2599
+[ 89.9%] Iter 8782/9765 | SPS:  2707 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0231 | PLoss: -0.0008 | Ent: 0.1977
+[ 89.9%] Iter 8783/9765 | SPS:  2707 | Reward:  0.062 | Value:  0.711 | VLoss: 0.0199 | PLoss: -0.0044 | Ent: 0.1847
+[ 90.0%] Iter 8784/9765 | SPS:  2707 | Reward:  0.063 | Value:  0.744 | VLoss: 0.0240 | PLoss: -0.0092 | Ent: 0.1883
+[ 90.0%] Iter 8785/9765 | SPS:  2707 | Reward:  0.073 | Value:  0.750 | VLoss: 0.0279 | PLoss: -0.0032 | Ent: 0.1821
+[ 90.0%] Iter 8786/9765 | SPS:  2707 | Reward:  0.058 | Value:  0.759 | VLoss: 0.0251 | PLoss: -0.0027 | Ent: 0.2057
+[ 90.0%] Iter 8787/9765 | SPS:  2707 | Reward:  0.070 | Value:  0.722 | VLoss: 0.0223 | PLoss: -0.0034 | Ent: 0.2248
+[ 90.0%] Iter 8788/9765 | SPS:  2707 | Reward:  0.051 | Value:  0.679 | VLoss: 0.0256 | PLoss: -0.0055 | Ent: 0.2202
+[ 90.0%] Iter 8789/9765 | SPS:  2707 | Reward:  0.065 | Value:  0.713 | VLoss: 0.0268 | PLoss: -0.0059 | Ent: 0.2454
+[ 90.0%] Iter 8790/9765 | SPS:  2707 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0191 | PLoss: -0.0019 | Ent: 0.2146
+[ 90.0%] Iter 8791/9765 | SPS:  2707 | Reward:  0.047 | Value:  0.694 | VLoss: 0.0283 | PLoss: -0.0025 | Ent: 0.2750
+[ 90.0%] Iter 8792/9765 | SPS:  2708 | Reward:  0.060 | Value:  0.599 | VLoss: 0.0162 | PLoss: -0.0060 | Ent: 0.2280
+[ 90.0%] Iter 8793/9765 | SPS:  2708 | Reward:  0.059 | Value:  0.690 | VLoss: 0.0203 | PLoss: -0.0044 | Ent: 0.2647
+[ 90.1%] Iter 8794/9765 | SPS:  2708 | Reward:  0.073 | Value:  0.732 | VLoss: 0.0181 | PLoss: -0.0025 | Ent: 0.2484
+[ 90.1%] Iter 8795/9765 | SPS:  2708 | Reward:  0.054 | Value:  0.699 | VLoss: 0.0177 | PLoss: -0.0032 | Ent: 0.2286
+[ 90.1%] Iter 8796/9765 | SPS:  2708 | Reward:  0.049 | Value:  0.703 | VLoss: 0.0165 | PLoss: -0.0044 | Ent: 0.2919
+[ 90.1%] Iter 8797/9765 | SPS:  2708 | Reward:  0.056 | Value:  0.668 | VLoss: 0.0217 | PLoss: -0.0030 | Ent: 0.2486
+[ 90.1%] Iter 8798/9765 | SPS:  2708 | Reward:  0.054 | Value:  0.732 | VLoss: 0.0124 | PLoss: -0.0015 | Ent: 0.2194
+[ 90.1%] Iter 8799/9765 | SPS:  2708 | Reward:  0.069 | Value:  0.712 | VLoss: 0.0238 | PLoss: -0.0026 | Ent: 0.2468
+[ 90.1%] Iter 8800/9765 | SPS:  2708 | Reward:  0.059 | Value:  0.750 | VLoss: 0.0321 | PLoss: 0.0001 | Ent: 0.2156
+[ 90.1%] Iter 8801/9765 | SPS:  2708 | Reward:  0.071 | Value:  0.852 | VLoss: 0.0157 | PLoss: -0.0039 | Ent: 0.2595
+[ 90.1%] Iter 8802/9765 | SPS:  2708 | Reward:  0.063 | Value:  0.736 | VLoss: 0.0152 | PLoss: -0.0014 | Ent: 0.2462
+[ 90.1%] Iter 8803/9765 | SPS:  2709 | Reward:  0.053 | Value:  0.728 | VLoss: 0.0237 | PLoss: -0.0048 | Ent: 0.2279
+[ 90.2%] Iter 8804/9765 | SPS:  2709 | Reward:  0.065 | Value:  0.762 | VLoss: 0.0188 | PLoss: -0.0013 | Ent: 0.1934
+[ 90.2%] Iter 8805/9765 | SPS:  2709 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0229 | PLoss: -0.0077 | Ent: 0.2308
+[ 90.2%] Iter 8806/9765 | SPS:  2709 | Reward:  0.055 | Value:  0.722 | VLoss: 0.0181 | PLoss: -0.0023 | Ent: 0.2207
+[ 90.2%] Iter 8807/9765 | SPS:  2709 | Reward:  0.051 | Value:  0.636 | VLoss: 0.0260 | PLoss: -0.0041 | Ent: 0.2157
+[ 90.2%] Iter 8808/9765 | SPS:  2709 | Reward:  0.058 | Value:  0.613 | VLoss: 0.0292 | PLoss: -0.0042 | Ent: 0.2121
+[ 90.2%] Iter 8809/9765 | SPS:  2709 | Reward:  0.051 | Value:  0.655 | VLoss: 0.0206 | PLoss: -0.0055 | Ent: 0.2447
+[ 90.2%] Iter 8810/9765 | SPS:  2709 | Reward:  0.054 | Value:  0.679 | VLoss: 0.0211 | PLoss: -0.0006 | Ent: 0.2277
+[ 90.2%] Iter 8811/9765 | SPS:  2709 | Reward:  0.060 | Value:  0.719 | VLoss: 0.0179 | PLoss: 0.0008 | Ent: 0.2881
+[ 90.2%] Iter 8812/9765 | SPS:  2709 | Reward:  0.052 | Value:  0.725 | VLoss: 0.0207 | PLoss: -0.0022 | Ent: 0.2401
+[ 90.3%] Iter 8813/9765 | SPS:  2709 | Reward:  0.052 | Value:  0.703 | VLoss: 0.0216 | PLoss: -0.0041 | Ent: 0.2294
+[ 90.3%] Iter 8814/9765 | SPS:  2709 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0179 | PLoss: -0.0019 | Ent: 0.2010
+[ 90.3%] Iter 8815/9765 | SPS:  2710 | Reward:  0.056 | Value:  0.727 | VLoss: 0.0257 | PLoss: -0.0038 | Ent: 0.2166
+[ 90.3%] Iter 8816/9765 | SPS:  2710 | Reward:  0.058 | Value:  0.707 | VLoss: 0.0191 | PLoss: -0.0002 | Ent: 0.2048
+[ 90.3%] Iter 8817/9765 | SPS:  2710 | Reward:  0.062 | Value:  0.684 | VLoss: 0.0249 | PLoss: -0.0027 | Ent: 0.2522
+[ 90.3%] Iter 8818/9765 | SPS:  2710 | Reward:  0.063 | Value:  0.739 | VLoss: 0.0220 | PLoss: -0.0046 | Ent: 0.1798
+[ 90.3%] Iter 8819/9765 | SPS:  2710 | Reward:  0.063 | Value:  0.712 | VLoss: 0.0236 | PLoss: -0.0039 | Ent: 0.2164
+[ 90.3%] Iter 8820/9765 | SPS:  2710 | Reward:  0.054 | Value:  0.687 | VLoss: 0.0174 | PLoss: -0.0048 | Ent: 0.2515
+[ 90.3%] Iter 8821/9765 | SPS:  2710 | Reward:  0.065 | Value:  0.701 | VLoss: 0.0196 | PLoss: -0.0024 | Ent: 0.1984
+[ 90.3%] Iter 8822/9765 | SPS:  2710 | Reward:  0.056 | Value:  0.760 | VLoss: 0.0129 | PLoss: -0.0037 | Ent: 0.1970
+[ 90.4%] Iter 8823/9765 | SPS:  2710 | Reward:  0.072 | Value:  0.750 | VLoss: 0.0252 | PLoss: -0.0009 | Ent: 0.2623
+[ 90.4%] Iter 8824/9765 | SPS:  2710 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0272 | PLoss: -0.0055 | Ent: 0.2445
+[ 90.4%] Iter 8825/9765 | SPS:  2710 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0262 | PLoss: -0.0027 | Ent: 0.2198
+[ 90.4%] Iter 8826/9765 | SPS:  2711 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0231 | PLoss: -0.0019 | Ent: 0.2443
+[ 90.4%] Iter 8827/9765 | SPS:  2711 | Reward:  0.063 | Value:  0.699 | VLoss: 0.0153 | PLoss: 0.0001 | Ent: 0.2633
+[ 90.4%] Iter 8828/9765 | SPS:  2711 | Reward:  0.068 | Value:  0.747 | VLoss: 0.0189 | PLoss: -0.0013 | Ent: 0.2528
+[ 90.4%] Iter 8829/9765 | SPS:  2711 | Reward:  0.059 | Value:  0.768 | VLoss: 0.0251 | PLoss: -0.0037 | Ent: 0.2078
+[ 90.4%] Iter 8830/9765 | SPS:  2711 | Reward:  0.062 | Value:  0.754 | VLoss: 0.0181 | PLoss: -0.0050 | Ent: 0.2384
+[ 90.4%] Iter 8831/9765 | SPS:  2711 | Reward:  0.061 | Value:  0.672 | VLoss: 0.0330 | PLoss: -0.0023 | Ent: 0.2840
+[ 90.4%] Iter 8832/9765 | SPS:  2711 | Reward:  0.063 | Value:  0.703 | VLoss: 0.0261 | PLoss: -0.0063 | Ent: 0.2152
+[ 90.5%] Iter 8833/9765 | SPS:  2711 | Reward:  0.057 | Value:  0.733 | VLoss: 0.0172 | PLoss: -0.0018 | Ent: 0.2315
+[ 90.5%] Iter 8834/9765 | SPS:  2711 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0151 | PLoss: -0.0018 | Ent: 0.1853
+[ 90.5%] Iter 8835/9765 | SPS:  2711 | Reward:  0.067 | Value:  0.766 | VLoss: 0.0184 | PLoss: -0.0018 | Ent: 0.2277
+[ 90.5%] Iter 8836/9765 | SPS:  2711 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0223 | PLoss: -0.0045 | Ent: 0.2252
+[ 90.5%] Iter 8837/9765 | SPS:  2711 | Reward:  0.062 | Value:  0.757 | VLoss: 0.0187 | PLoss: -0.0058 | Ent: 0.2265
+[ 90.5%] Iter 8838/9765 | SPS:  2712 | Reward:  0.066 | Value:  0.741 | VLoss: 0.0216 | PLoss: -0.0044 | Ent: 0.2000
+[ 90.5%] Iter 8839/9765 | SPS:  2712 | Reward:  0.062 | Value:  0.722 | VLoss: 0.0157 | PLoss: -0.0033 | Ent: 0.2085
+[ 90.5%] Iter 8840/9765 | SPS:  2712 | Reward:  0.066 | Value:  0.784 | VLoss: 0.0201 | PLoss: -0.0003 | Ent: 0.1779
+[ 90.5%] Iter 8841/9765 | SPS:  2712 | Reward:  0.061 | Value:  0.761 | VLoss: 0.0250 | PLoss: -0.0038 | Ent: 0.2369
+[ 90.5%] Iter 8842/9765 | SPS:  2712 | Reward:  0.065 | Value:  0.676 | VLoss: 0.0206 | PLoss: -0.0031 | Ent: 0.2487
+[ 90.6%] Iter 8843/9765 | SPS:  2712 | Reward:  0.065 | Value:  0.730 | VLoss: 0.0146 | PLoss: -0.0009 | Ent: 0.2516
+[ 90.6%] Iter 8844/9765 | SPS:  2712 | Reward:  0.064 | Value:  0.677 | VLoss: 0.0233 | PLoss: -0.0025 | Ent: 0.2315
+[ 90.6%] Iter 8845/9765 | SPS:  2712 | Reward:  0.056 | Value:  0.718 | VLoss: 0.0178 | PLoss: -0.0014 | Ent: 0.2446
+[ 90.6%] Iter 8846/9765 | SPS:  2712 | Reward:  0.065 | Value:  0.695 | VLoss: 0.0317 | PLoss: -0.0029 | Ent: 0.2877
+[ 90.6%] Iter 8847/9765 | SPS:  2712 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0160 | PLoss: -0.0059 | Ent: 0.2450
+[ 90.6%] Iter 8848/9765 | SPS:  2712 | Reward:  0.060 | Value:  0.747 | VLoss: 0.0121 | PLoss: -0.0031 | Ent: 0.2744
+[ 90.6%] Iter 8849/9765 | SPS:  2712 | Reward:  0.054 | Value:  0.754 | VLoss: 0.0196 | PLoss: -0.0030 | Ent: 0.2147
+[ 90.6%] Iter 8850/9765 | SPS:  2713 | Reward:  0.062 | Value:  0.695 | VLoss: 0.0196 | PLoss: -0.0009 | Ent: 0.2824
+[ 90.6%] Iter 8851/9765 | SPS:  2713 | Reward:  0.059 | Value:  0.732 | VLoss: 0.0218 | PLoss: -0.0012 | Ent: 0.2430
+[ 90.7%] Iter 8852/9765 | SPS:  2713 | Reward:  0.060 | Value:  0.675 | VLoss: 0.0220 | PLoss: -0.0027 | Ent: 0.2239
+[ 90.7%] Iter 8853/9765 | SPS:  2713 | Reward:  0.066 | Value:  0.742 | VLoss: 0.0173 | PLoss: -0.0017 | Ent: 0.2699
+[ 90.7%] Iter 8854/9765 | SPS:  2713 | Reward:  0.070 | Value:  0.763 | VLoss: 0.0280 | PLoss: -0.0018 | Ent: 0.2605
+[ 90.7%] Iter 8855/9765 | SPS:  2713 | Reward:  0.065 | Value:  0.748 | VLoss: 0.0260 | PLoss: -0.0039 | Ent: 0.2561
+[ 90.7%] Iter 8856/9765 | SPS:  2713 | Reward:  0.053 | Value:  0.738 | VLoss: 0.0148 | PLoss: -0.0030 | Ent: 0.2703
+[ 90.7%] Iter 8857/9765 | SPS:  2713 | Reward:  0.062 | Value:  0.739 | VLoss: 0.0208 | PLoss: -0.0013 | Ent: 0.2120
+[ 90.7%] Iter 8858/9765 | SPS:  2713 | Reward:  0.061 | Value:  0.718 | VLoss: 0.0152 | PLoss: -0.0028 | Ent: 0.2571
+[ 90.7%] Iter 8859/9765 | SPS:  2713 | Reward:  0.059 | Value:  0.611 | VLoss: 0.0220 | PLoss: -0.0031 | Ent: 0.2045
+[ 90.7%] Iter 8860/9765 | SPS:  2713 | Reward:  0.048 | Value:  0.638 | VLoss: 0.0185 | PLoss: -0.0026 | Ent: 0.2916
+[ 90.7%] Iter 8861/9765 | SPS:  2714 | Reward:  0.066 | Value:  0.743 | VLoss: 0.0219 | PLoss: -0.0027 | Ent: 0.1924
+[ 90.8%] Iter 8862/9765 | SPS:  2714 | Reward:  0.059 | Value:  0.779 | VLoss: 0.0170 | PLoss: -0.0065 | Ent: 0.2281
+[ 90.8%] Iter 8863/9765 | SPS:  2714 | Reward:  0.064 | Value:  0.749 | VLoss: 0.0204 | PLoss: -0.0073 | Ent: 0.2870
+[ 90.8%] Iter 8864/9765 | SPS:  2714 | Reward:  0.057 | Value:  0.759 | VLoss: 0.0151 | PLoss: -0.0037 | Ent: 0.2155
+[ 90.8%] Iter 8865/9765 | SPS:  2714 | Reward:  0.069 | Value:  0.752 | VLoss: 0.0289 | PLoss: -0.0013 | Ent: 0.2034
+[ 90.8%] Iter 8866/9765 | SPS:  2714 | Reward:  0.067 | Value:  0.760 | VLoss: 0.0196 | PLoss: -0.0041 | Ent: 0.2158
+[ 90.8%] Iter 8867/9765 | SPS:  2714 | Reward:  0.054 | Value:  0.695 | VLoss: 0.0191 | PLoss: -0.0080 | Ent: 0.2576
+[ 90.8%] Iter 8868/9765 | SPS:  2714 | Reward:  0.061 | Value:  0.753 | VLoss: 0.0176 | PLoss: -0.0030 | Ent: 0.2325
+[ 90.8%] Iter 8869/9765 | SPS:  2714 | Reward:  0.049 | Value:  0.685 | VLoss: 0.0257 | PLoss: -0.0017 | Ent: 0.2465
+[ 90.8%] Iter 8870/9765 | SPS:  2714 | Reward:  0.061 | Value:  0.747 | VLoss: 0.0236 | PLoss: 0.0015 | Ent: 0.2260
+[ 90.8%] Iter 8871/9765 | SPS:  2714 | Reward:  0.059 | Value:  0.755 | VLoss: 0.0132 | PLoss: -0.0021 | Ent: 0.2307
+[ 90.9%] Iter 8872/9765 | SPS:  2715 | Reward:  0.047 | Value:  0.700 | VLoss: 0.0165 | PLoss: -0.0059 | Ent: 0.2254
+[ 90.9%] Iter 8873/9765 | SPS:  2715 | Reward:  0.070 | Value:  0.679 | VLoss: 0.0252 | PLoss: -0.0037 | Ent: 0.2302
+[ 90.9%] Iter 8874/9765 | SPS:  2715 | Reward:  0.058 | Value:  0.711 | VLoss: 0.0188 | PLoss: -0.0022 | Ent: 0.2101
+[ 90.9%] Iter 8875/9765 | SPS:  2715 | Reward:  0.046 | Value:  0.709 | VLoss: 0.0210 | PLoss: -0.0011 | Ent: 0.2209
+[ 90.9%] Iter 8876/9765 | SPS:  2715 | Reward:  0.069 | Value:  0.699 | VLoss: 0.0277 | PLoss: -0.0029 | Ent: 0.2683
+[ 90.9%] Iter 8877/9765 | SPS:  2715 | Reward:  0.068 | Value:  0.721 | VLoss: 0.0243 | PLoss: -0.0044 | Ent: 0.2089
+[ 90.9%] Iter 8878/9765 | SPS:  2715 | Reward:  0.060 | Value:  0.785 | VLoss: 0.0219 | PLoss: -0.0038 | Ent: 0.1765
+[ 90.9%] Iter 8879/9765 | SPS:  2715 | Reward:  0.062 | Value:  0.697 | VLoss: 0.0204 | PLoss: -0.0019 | Ent: 0.2149
+[ 90.9%] Iter 8880/9765 | SPS:  2715 | Reward:  0.059 | Value:  0.725 | VLoss: 0.0169 | PLoss: -0.0041 | Ent: 0.2349
+[ 90.9%] Iter 8881/9765 | SPS:  2715 | Reward:  0.060 | Value:  0.726 | VLoss: 0.0229 | PLoss: -0.0032 | Ent: 0.2248
+[ 91.0%] Iter 8882/9765 | SPS:  2715 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0222 | PLoss: -0.0017 | Ent: 0.2315
+[ 91.0%] Iter 8883/9765 | SPS:  2715 | Reward:  0.066 | Value:  0.699 | VLoss: 0.0212 | PLoss: -0.0017 | Ent: 0.2223
+[ 91.0%] Iter 8884/9765 | SPS:  2715 | Reward:  0.051 | Value:  0.655 | VLoss: 0.0181 | PLoss: -0.0042 | Ent: 0.2876
+[ 91.0%] Iter 8885/9765 | SPS:  2716 | Reward:  0.056 | Value:  0.733 | VLoss: 0.0135 | PLoss: -0.0027 | Ent: 0.2097
+[ 91.0%] Iter 8886/9765 | SPS:  2716 | Reward:  0.052 | Value:  0.714 | VLoss: 0.0214 | PLoss: -0.0036 | Ent: 0.2183
+[ 91.0%] Iter 8887/9765 | SPS:  2716 | Reward:  0.055 | Value:  0.637 | VLoss: 0.0174 | PLoss: -0.0023 | Ent: 0.2305
+[ 91.0%] Iter 8888/9765 | SPS:  2716 | Reward:  0.059 | Value:  0.737 | VLoss: 0.0244 | PLoss: -0.0049 | Ent: 0.2521
+[ 91.0%] Iter 8889/9765 | SPS:  2716 | Reward:  0.063 | Value:  0.724 | VLoss: 0.0238 | PLoss: -0.0029 | Ent: 0.1945
+[ 91.0%] Iter 8890/9765 | SPS:  2716 | Reward:  0.064 | Value:  0.716 | VLoss: 0.0200 | PLoss: -0.0038 | Ent: 0.2682
+[ 91.0%] Iter 8891/9765 | SPS:  2716 | Reward:  0.055 | Value:  0.751 | VLoss: 0.0160 | PLoss: -0.0100 | Ent: 0.2173
+[ 91.1%] Iter 8892/9765 | SPS:  2716 | Reward:  0.049 | Value:  0.730 | VLoss: 0.0246 | PLoss: -0.0035 | Ent: 0.2125
+[ 91.1%] Iter 8893/9765 | SPS:  2716 | Reward:  0.062 | Value:  0.742 | VLoss: 0.0251 | PLoss: -0.0048 | Ent: 0.2236
+[ 91.1%] Iter 8894/9765 | SPS:  2716 | Reward:  0.061 | Value:  0.701 | VLoss: 0.0141 | PLoss: -0.0042 | Ent: 0.2149
+[ 91.1%] Iter 8895/9765 | SPS:  2716 | Reward:  0.060 | Value:  0.649 | VLoss: 0.0188 | PLoss: -0.0030 | Ent: 0.2475
+[ 91.1%] Iter 8896/9765 | SPS:  2717 | Reward:  0.040 | Value:  0.543 | VLoss: 0.0147 | PLoss: -0.0028 | Ent: 0.2690
+[ 91.1%] Iter 8897/9765 | SPS:  2717 | Reward:  0.049 | Value:  0.619 | VLoss: 0.0175 | PLoss: -0.0030 | Ent: 0.2286
+[ 91.1%] Iter 8898/9765 | SPS:  2717 | Reward:  0.057 | Value:  0.651 | VLoss: 0.0225 | PLoss: -0.0030 | Ent: 0.2502
+[ 91.1%] Iter 8899/9765 | SPS:  2717 | Reward:  0.061 | Value:  0.679 | VLoss: 0.0164 | PLoss: -0.0018 | Ent: 0.2926
+[ 91.1%] Iter 8900/9765 | SPS:  2717 | Reward:  0.061 | Value:  0.636 | VLoss: 0.0329 | PLoss: -0.0029 | Ent: 0.2251
+[ 91.2%] Iter 8901/9765 | SPS:  2717 | Reward:  0.068 | Value:  0.685 | VLoss: 0.0205 | PLoss: -0.0022 | Ent: 0.1979
+[ 91.2%] Iter 8902/9765 | SPS:  2717 | Reward:  0.056 | Value:  0.731 | VLoss: 0.0223 | PLoss: 0.0003 | Ent: 0.1962
+[ 91.2%] Iter 8903/9765 | SPS:  2717 | Reward:  0.063 | Value:  0.754 | VLoss: 0.0230 | PLoss: -0.0045 | Ent: 0.2728
+[ 91.2%] Iter 8904/9765 | SPS:  2717 | Reward:  0.058 | Value:  0.704 | VLoss: 0.0187 | PLoss: -0.0047 | Ent: 0.2111
+[ 91.2%] Iter 8905/9765 | SPS:  2717 | Reward:  0.056 | Value:  0.710 | VLoss: 0.0144 | PLoss: -0.0024 | Ent: 0.2845
+[ 91.2%] Iter 8906/9765 | SPS:  2717 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0201 | PLoss: -0.0015 | Ent: 0.2129
+[ 91.2%] Iter 8907/9765 | SPS:  2717 | Reward:  0.061 | Value:  0.746 | VLoss: 0.0191 | PLoss: -0.0021 | Ent: 0.2785
+[ 91.2%] Iter 8908/9765 | SPS:  2718 | Reward:  0.049 | Value:  0.707 | VLoss: 0.0225 | PLoss: -0.0047 | Ent: 0.2264
+[ 91.2%] Iter 8909/9765 | SPS:  2718 | Reward:  0.068 | Value:  0.658 | VLoss: 0.0254 | PLoss: -0.0030 | Ent: 0.2491
+[ 91.2%] Iter 8910/9765 | SPS:  2718 | Reward:  0.056 | Value:  0.757 | VLoss: 0.0199 | PLoss: -0.0036 | Ent: 0.2106
+[ 91.3%] Iter 8911/9765 | SPS:  2718 | Reward:  0.069 | Value:  0.721 | VLoss: 0.0315 | PLoss: -0.0049 | Ent: 0.2273
+[ 91.3%] Iter 8912/9765 | SPS:  2718 | Reward:  0.046 | Value:  0.686 | VLoss: 0.0136 | PLoss: -0.0036 | Ent: 0.2264
+[ 91.3%] Iter 8913/9765 | SPS:  2718 | Reward:  0.061 | Value:  0.716 | VLoss: 0.0263 | PLoss: -0.0022 | Ent: 0.2487
+[ 91.3%] Iter 8914/9765 | SPS:  2718 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0273 | PLoss: -0.0079 | Ent: 0.2439
+[ 91.3%] Iter 8915/9765 | SPS:  2718 | Reward:  0.057 | Value:  0.700 | VLoss: 0.0174 | PLoss: -0.0029 | Ent: 0.2409
+[ 91.3%] Iter 8916/9765 | SPS:  2718 | Reward:  0.059 | Value:  0.703 | VLoss: 0.0218 | PLoss: -0.0059 | Ent: 0.2457
+[ 91.3%] Iter 8917/9765 | SPS:  2718 | Reward:  0.048 | Value:  0.703 | VLoss: 0.0157 | PLoss: -0.0030 | Ent: 0.2515
+[ 91.3%] Iter 8918/9765 | SPS:  2718 | Reward:  0.063 | Value:  0.711 | VLoss: 0.0153 | PLoss: -0.0023 | Ent: 0.2395
+[ 91.3%] Iter 8919/9765 | SPS:  2718 | Reward:  0.057 | Value:  0.690 | VLoss: 0.0195 | PLoss: -0.0003 | Ent: 0.2467
+[ 91.3%] Iter 8920/9765 | SPS:  2719 | Reward:  0.057 | Value:  0.678 | VLoss: 0.0187 | PLoss: -0.0052 | Ent: 0.2210
+[ 91.4%] Iter 8921/9765 | SPS:  2719 | Reward:  0.065 | Value:  0.734 | VLoss: 0.0222 | PLoss: -0.0037 | Ent: 0.2277
+[ 91.4%] Iter 8922/9765 | SPS:  2719 | Reward:  0.063 | Value:  0.683 | VLoss: 0.0226 | PLoss: -0.0044 | Ent: 0.2207
+[ 91.4%] Iter 8923/9765 | SPS:  2719 | Reward:  0.058 | Value:  0.659 | VLoss: 0.0272 | PLoss: -0.0045 | Ent: 0.2171
+[ 91.4%] Iter 8924/9765 | SPS:  2719 | Reward:  0.060 | Value:  0.646 | VLoss: 0.0225 | PLoss: -0.0005 | Ent: 0.2572
+[ 91.4%] Iter 8925/9765 | SPS:  2719 | Reward:  0.057 | Value:  0.723 | VLoss: 0.0175 | PLoss: -0.0024 | Ent: 0.2106
+[ 91.4%] Iter 8926/9765 | SPS:  2719 | Reward:  0.067 | Value:  0.754 | VLoss: 0.0178 | PLoss: -0.0025 | Ent: 0.2049
+[ 91.4%] Iter 8927/9765 | SPS:  2719 | Reward:  0.056 | Value:  0.683 | VLoss: 0.0256 | PLoss: -0.0044 | Ent: 0.2401
+[ 91.4%] Iter 8928/9765 | SPS:  2719 | Reward:  0.068 | Value:  0.727 | VLoss: 0.0264 | PLoss: -0.0015 | Ent: 0.2302
+[ 91.4%] Iter 8929/9765 | SPS:  2719 | Reward:  0.069 | Value:  0.712 | VLoss: 0.0222 | PLoss: -0.0017 | Ent: 0.2222
+[ 91.4%] Iter 8930/9765 | SPS:  2719 | Reward:  0.054 | Value:  0.668 | VLoss: 0.0228 | PLoss: -0.0011 | Ent: 0.2382
+[ 91.5%] Iter 8931/9765 | SPS:  2719 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0194 | PLoss: -0.0013 | Ent: 0.2423
+[ 91.5%] Iter 8932/9765 | SPS:  2720 | Reward:  0.067 | Value:  0.698 | VLoss: 0.0284 | PLoss: 0.0000 | Ent: 0.2588
+[ 91.5%] Iter 8933/9765 | SPS:  2720 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0251 | PLoss: -0.0018 | Ent: 0.2355
+[ 91.5%] Iter 8934/9765 | SPS:  2720 | Reward:  0.046 | Value:  0.631 | VLoss: 0.0249 | PLoss: -0.0051 | Ent: 0.2740
+[ 91.5%] Iter 8935/9765 | SPS:  2720 | Reward:  0.074 | Value:  0.780 | VLoss: 0.0185 | PLoss: -0.0010 | Ent: 0.2200
+[ 91.5%] Iter 8936/9765 | SPS:  2720 | Reward:  0.070 | Value:  0.736 | VLoss: 0.0299 | PLoss: -0.0057 | Ent: 0.2272
+[ 91.5%] Iter 8937/9765 | SPS:  2720 | Reward:  0.073 | Value:  0.766 | VLoss: 0.0250 | PLoss: 0.0002 | Ent: 0.2248
+[ 91.5%] Iter 8938/9765 | SPS:  2720 | Reward:  0.063 | Value:  0.748 | VLoss: 0.0205 | PLoss: -0.0030 | Ent: 0.2258
+[ 91.5%] Iter 8939/9765 | SPS:  2720 | Reward:  0.068 | Value:  0.809 | VLoss: 0.0203 | PLoss: -0.0026 | Ent: 0.2100
+[ 91.6%] Iter 8940/9765 | SPS:  2720 | Reward:  0.075 | Value:  0.738 | VLoss: 0.0302 | PLoss: -0.0019 | Ent: 0.2202
+[ 91.6%] Iter 8941/9765 | SPS:  2720 | Reward:  0.064 | Value:  0.753 | VLoss: 0.0236 | PLoss: -0.0034 | Ent: 0.2261
+[ 91.6%] Iter 8942/9765 | SPS:  2720 | Reward:  0.052 | Value:  0.732 | VLoss: 0.0273 | PLoss: -0.0016 | Ent: 0.2600
+[ 91.6%] Iter 8943/9765 | SPS:  2720 | Reward:  0.056 | Value:  0.705 | VLoss: 0.0166 | PLoss: -0.0020 | Ent: 0.1969
+[ 91.6%] Iter 8944/9765 | SPS:  2721 | Reward:  0.039 | Value:  0.648 | VLoss: 0.0140 | PLoss: -0.0078 | Ent: 0.3165
+[ 91.6%] Iter 8945/9765 | SPS:  2721 | Reward:  0.046 | Value:  0.685 | VLoss: 0.0346 | PLoss: -0.0047 | Ent: 0.2155
+[ 91.6%] Iter 8946/9765 | SPS:  2721 | Reward:  0.069 | Value:  0.733 | VLoss: 0.0231 | PLoss: -0.0082 | Ent: 0.2436
+[ 91.6%] Iter 8947/9765 | SPS:  2721 | Reward:  0.046 | Value:  0.702 | VLoss: 0.0220 | PLoss: -0.0049 | Ent: 0.2228
+[ 91.6%] Iter 8948/9765 | SPS:  2721 | Reward:  0.048 | Value:  0.692 | VLoss: 0.0219 | PLoss: -0.0021 | Ent: 0.2373
+[ 91.6%] Iter 8949/9765 | SPS:  2721 | Reward:  0.067 | Value:  0.701 | VLoss: 0.0199 | PLoss: -0.0006 | Ent: 0.2056
+[ 91.7%] Iter 8950/9765 | SPS:  2721 | Reward:  0.066 | Value:  0.727 | VLoss: 0.0232 | PLoss: -0.0030 | Ent: 0.2192
+[ 91.7%] Iter 8951/9765 | SPS:  2721 | Reward:  0.055 | Value:  0.724 | VLoss: 0.0204 | PLoss: -0.0004 | Ent: 0.2226
+[ 91.7%] Iter 8952/9765 | SPS:  2721 | Reward:  0.050 | Value:  0.711 | VLoss: 0.0148 | PLoss: -0.0034 | Ent: 0.2623
+[ 91.7%] Iter 8953/9765 | SPS:  2721 | Reward:  0.057 | Value:  0.721 | VLoss: 0.0236 | PLoss: -0.0006 | Ent: 0.2525
+[ 91.7%] Iter 8954/9765 | SPS:  2721 | Reward:  0.067 | Value:  0.679 | VLoss: 0.0269 | PLoss: -0.0015 | Ent: 0.2502
+[ 91.7%] Iter 8955/9765 | SPS:  2721 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0255 | PLoss: -0.0017 | Ent: 0.2178
+[ 91.7%] Iter 8956/9765 | SPS:  2722 | Reward:  0.058 | Value:  0.749 | VLoss: 0.0207 | PLoss: -0.0004 | Ent: 0.1782
+[ 91.7%] Iter 8957/9765 | SPS:  2722 | Reward:  0.066 | Value:  0.723 | VLoss: 0.0284 | PLoss: -0.0034 | Ent: 0.2437
+[ 91.7%] Iter 8958/9765 | SPS:  2722 | Reward:  0.063 | Value:  0.732 | VLoss: 0.0239 | PLoss: -0.0015 | Ent: 0.2237
+[ 91.7%] Iter 8959/9765 | SPS:  2722 | Reward:  0.056 | Value:  0.723 | VLoss: 0.0175 | PLoss: -0.0038 | Ent: 0.2325
+[ 91.8%] Iter 8960/9765 | SPS:  2722 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0229 | PLoss: -0.0040 | Ent: 0.2272
+[ 91.8%] Iter 8961/9765 | SPS:  2722 | Reward:  0.057 | Value:  0.773 | VLoss: 0.0245 | PLoss: -0.0060 | Ent: 0.2420
+[ 91.8%] Iter 8962/9765 | SPS:  2722 | Reward:  0.067 | Value:  0.736 | VLoss: 0.0243 | PLoss: -0.0084 | Ent: 0.2296
+[ 91.8%] Iter 8963/9765 | SPS:  2722 | Reward:  0.066 | Value:  0.733 | VLoss: 0.0279 | PLoss: -0.0004 | Ent: 0.2233
+[ 91.8%] Iter 8964/9765 | SPS:  2722 | Reward:  0.053 | Value:  0.748 | VLoss: 0.0213 | PLoss: -0.0016 | Ent: 0.2477
+[ 91.8%] Iter 8965/9765 | SPS:  2722 | Reward:  0.070 | Value:  0.704 | VLoss: 0.0273 | PLoss: -0.0032 | Ent: 0.2172
+[ 91.8%] Iter 8966/9765 | SPS:  2722 | Reward:  0.055 | Value:  0.659 | VLoss: 0.0209 | PLoss: -0.0023 | Ent: 0.2633
+[ 91.8%] Iter 8967/9765 | SPS:  2723 | Reward:  0.066 | Value:  0.733 | VLoss: 0.0184 | PLoss: -0.0046 | Ent: 0.2317
+[ 91.8%] Iter 8968/9765 | SPS:  2723 | Reward:  0.061 | Value:  0.720 | VLoss: 0.0204 | PLoss: -0.0025 | Ent: 0.2362
+[ 91.8%] Iter 8969/9765 | SPS:  2723 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0224 | PLoss: -0.0063 | Ent: 0.2609
+[ 91.9%] Iter 8970/9765 | SPS:  2723 | Reward:  0.065 | Value:  0.765 | VLoss: 0.0161 | PLoss: -0.0012 | Ent: 0.2769
+[ 91.9%] Iter 8971/9765 | SPS:  2723 | Reward:  0.060 | Value:  0.716 | VLoss: 0.0281 | PLoss: -0.0025 | Ent: 0.2591
+[ 91.9%] Iter 8972/9765 | SPS:  2723 | Reward:  0.055 | Value:  0.710 | VLoss: 0.0168 | PLoss: -0.0016 | Ent: 0.2252
+[ 91.9%] Iter 8973/9765 | SPS:  2723 | Reward:  0.066 | Value:  0.809 | VLoss: 0.0235 | PLoss: -0.0020 | Ent: 0.2303
+[ 91.9%] Iter 8974/9765 | SPS:  2723 | Reward:  0.049 | Value:  0.697 | VLoss: 0.0259 | PLoss: -0.0022 | Ent: 0.1899
+[ 91.9%] Iter 8975/9765 | SPS:  2723 | Reward:  0.047 | Value:  0.691 | VLoss: 0.0227 | PLoss: -0.0011 | Ent: 0.2556
+[ 91.9%] Iter 8976/9765 | SPS:  2723 | Reward:  0.060 | Value:  0.691 | VLoss: 0.0253 | PLoss: -0.0050 | Ent: 0.2281
+[ 91.9%] Iter 8977/9765 | SPS:  2723 | Reward:  0.055 | Value:  0.622 | VLoss: 0.0274 | PLoss: -0.0027 | Ent: 0.1830
+[ 91.9%] Iter 8978/9765 | SPS:  2723 | Reward:  0.074 | Value:  0.669 | VLoss: 0.0248 | PLoss: -0.0046 | Ent: 0.2471
+[ 92.0%] Iter 8979/9765 | SPS:  2724 | Reward:  0.047 | Value:  0.683 | VLoss: 0.0214 | PLoss: -0.0014 | Ent: 0.2188
+[ 92.0%] Iter 8980/9765 | SPS:  2724 | Reward:  0.061 | Value:  0.724 | VLoss: 0.0189 | PLoss: -0.0012 | Ent: 0.1916
+[ 92.0%] Iter 8981/9765 | SPS:  2724 | Reward:  0.061 | Value:  0.680 | VLoss: 0.0288 | PLoss: -0.0023 | Ent: 0.2195
+[ 92.0%] Iter 8982/9765 | SPS:  2724 | Reward:  0.067 | Value:  0.764 | VLoss: 0.0202 | PLoss: -0.0020 | Ent: 0.1834
+[ 92.0%] Iter 8983/9765 | SPS:  2724 | Reward:  0.058 | Value:  0.681 | VLoss: 0.0229 | PLoss: -0.0022 | Ent: 0.1996
+[ 92.0%] Iter 8984/9765 | SPS:  2724 | Reward:  0.054 | Value:  0.749 | VLoss: 0.0175 | PLoss: 0.0007 | Ent: 0.2193
+[ 92.0%] Iter 8985/9765 | SPS:  2724 | Reward:  0.070 | Value:  0.752 | VLoss: 0.0205 | PLoss: -0.0027 | Ent: 0.2335
+[ 92.0%] Iter 8986/9765 | SPS:  2724 | Reward:  0.062 | Value:  0.732 | VLoss: 0.0181 | PLoss: -0.0038 | Ent: 0.1860
+[ 92.0%] Iter 8987/9765 | SPS:  2724 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0221 | PLoss: -0.0044 | Ent: 0.2436
+[ 92.0%] Iter 8988/9765 | SPS:  2724 | Reward:  0.060 | Value:  0.689 | VLoss: 0.0174 | PLoss: -0.0041 | Ent: 0.2087
+[ 92.1%] Iter 8989/9765 | SPS:  2724 | Reward:  0.064 | Value:  0.725 | VLoss: 0.0177 | PLoss: -0.0027 | Ent: 0.2442
+[ 92.1%] Iter 8990/9765 | SPS:  2724 | Reward:  0.059 | Value:  0.767 | VLoss: 0.0238 | PLoss: -0.0070 | Ent: 0.2553
+[ 92.1%] Iter 8991/9765 | SPS:  2725 | Reward:  0.062 | Value:  0.727 | VLoss: 0.0166 | PLoss: -0.0023 | Ent: 0.2323
+[ 92.1%] Iter 8992/9765 | SPS:  2725 | Reward:  0.056 | Value:  0.673 | VLoss: 0.0216 | PLoss: -0.0029 | Ent: 0.2462
+[ 92.1%] Iter 8993/9765 | SPS:  2725 | Reward:  0.057 | Value:  0.659 | VLoss: 0.0251 | PLoss: -0.0052 | Ent: 0.2664
+[ 92.1%] Iter 8994/9765 | SPS:  2725 | Reward:  0.047 | Value:  0.746 | VLoss: 0.0109 | PLoss: -0.0026 | Ent: 0.2002
+[ 92.1%] Iter 8995/9765 | SPS:  2725 | Reward:  0.060 | Value:  0.720 | VLoss: 0.0169 | PLoss: -0.0042 | Ent: 0.2624
+[ 92.1%] Iter 8996/9765 | SPS:  2725 | Reward:  0.056 | Value:  0.747 | VLoss: 0.0259 | PLoss: -0.0041 | Ent: 0.1787
+[ 92.1%] Iter 8997/9765 | SPS:  2725 | Reward:  0.059 | Value:  0.796 | VLoss: 0.0141 | PLoss: -0.0025 | Ent: 0.1383
+[ 92.1%] Iter 8998/9765 | SPS:  2725 | Reward:  0.054 | Value:  0.690 | VLoss: 0.0215 | PLoss: -0.0011 | Ent: 0.2458
+[ 92.2%] Iter 8999/9765 | SPS:  2725 | Reward:  0.055 | Value:  0.695 | VLoss: 0.0124 | PLoss: -0.0013 | Ent: 0.1903
+[ 92.2%] Iter 9000/9765 | SPS:  2725 | Reward:  0.060 | Value:  0.696 | VLoss: 0.0206 | PLoss: -0.0057 | Ent: 0.2245
+[ 92.2%] Iter 9001/9765 | SPS:  2725 | Reward:  0.058 | Value:  0.780 | VLoss: 0.0227 | PLoss: -0.0022 | Ent: 0.2226
+[ 92.2%] Iter 9002/9765 | SPS:  2725 | Reward:  0.069 | Value:  0.749 | VLoss: 0.0236 | PLoss: -0.0037 | Ent: 0.1958
+[ 92.2%] Iter 9003/9765 | SPS:  2726 | Reward:  0.068 | Value:  0.786 | VLoss: 0.0267 | PLoss: -0.0035 | Ent: 0.2110
+[ 92.2%] Iter 9004/9765 | SPS:  2726 | Reward:  0.069 | Value:  0.772 | VLoss: 0.0227 | PLoss: -0.0052 | Ent: 0.2209
+[ 92.2%] Iter 9005/9765 | SPS:  2726 | Reward:  0.060 | Value:  0.756 | VLoss: 0.0225 | PLoss: -0.0049 | Ent: 0.2684
+[ 92.2%] Iter 9006/9765 | SPS:  2726 | Reward:  0.060 | Value:  0.745 | VLoss: 0.0168 | PLoss: -0.0023 | Ent: 0.2637
+[ 92.2%] Iter 9007/9765 | SPS:  2726 | Reward:  0.039 | Value:  0.683 | VLoss: 0.0189 | PLoss: -0.0040 | Ent: 0.2414
+[ 92.2%] Iter 9008/9765 | SPS:  2726 | Reward:  0.064 | Value:  0.695 | VLoss: 0.0207 | PLoss: -0.0026 | Ent: 0.2521
+[ 92.3%] Iter 9009/9765 | SPS:  2726 | Reward:  0.054 | Value:  0.710 | VLoss: 0.0241 | PLoss: -0.0026 | Ent: 0.2120
+[ 92.3%] Iter 9010/9765 | SPS:  2726 | Reward:  0.057 | Value:  0.718 | VLoss: 0.0210 | PLoss: -0.0024 | Ent: 0.2698
+[ 92.3%] Iter 9011/9765 | SPS:  2726 | Reward:  0.050 | Value:  0.739 | VLoss: 0.0146 | PLoss: -0.0007 | Ent: 0.2743
+[ 92.3%] Iter 9012/9765 | SPS:  2726 | Reward:  0.082 | Value:  0.774 | VLoss: 0.0195 | PLoss: -0.0046 | Ent: 0.2247
+[ 92.3%] Iter 9013/9765 | SPS:  2726 | Reward:  0.069 | Value:  0.810 | VLoss: 0.0204 | PLoss: -0.0032 | Ent: 0.2286
+[ 92.3%] Iter 9014/9765 | SPS:  2726 | Reward:  0.065 | Value:  0.766 | VLoss: 0.0161 | PLoss: -0.0032 | Ent: 0.2115
+[ 92.3%] Iter 9015/9765 | SPS:  2726 | Reward:  0.074 | Value:  0.751 | VLoss: 0.0231 | PLoss: -0.0031 | Ent: 0.1962
+[ 92.3%] Iter 9016/9765 | SPS:  2727 | Reward:  0.051 | Value:  0.781 | VLoss: 0.0305 | PLoss: -0.0031 | Ent: 0.2543
+[ 92.3%] Iter 9017/9765 | SPS:  2727 | Reward:  0.068 | Value:  0.771 | VLoss: 0.0236 | PLoss: -0.0040 | Ent: 0.1951
+[ 92.4%] Iter 9018/9765 | SPS:  2727 | Reward:  0.063 | Value:  0.678 | VLoss: 0.0246 | PLoss: -0.0019 | Ent: 0.1962
+[ 92.4%] Iter 9019/9765 | SPS:  2727 | Reward:  0.073 | Value:  0.746 | VLoss: 0.0280 | PLoss: -0.0028 | Ent: 0.2611
+[ 92.4%] Iter 9020/9765 | SPS:  2727 | Reward:  0.055 | Value:  0.714 | VLoss: 0.0189 | PLoss: -0.0020 | Ent: 0.2616
+[ 92.4%] Iter 9021/9765 | SPS:  2727 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0161 | PLoss: -0.0016 | Ent: 0.1629
+[ 92.4%] Iter 9022/9765 | SPS:  2727 | Reward:  0.053 | Value:  0.682 | VLoss: 0.0208 | PLoss: -0.0033 | Ent: 0.2366
+[ 92.4%] Iter 9023/9765 | SPS:  2727 | Reward:  0.075 | Value:  0.740 | VLoss: 0.0225 | PLoss: -0.0029 | Ent: 0.2697
+[ 92.4%] Iter 9024/9765 | SPS:  2727 | Reward:  0.043 | Value:  0.701 | VLoss: 0.0198 | PLoss: -0.0031 | Ent: 0.2560
+[ 92.4%] Iter 9025/9765 | SPS:  2727 | Reward:  0.063 | Value:  0.741 | VLoss: 0.0178 | PLoss: -0.0033 | Ent: 0.2154
+[ 92.4%] Iter 9026/9765 | SPS:  2727 | Reward:  0.071 | Value:  0.748 | VLoss: 0.0239 | PLoss: -0.0031 | Ent: 0.2324
+[ 92.4%] Iter 9027/9765 | SPS:  2727 | Reward:  0.067 | Value:  0.703 | VLoss: 0.0248 | PLoss: -0.0035 | Ent: 0.2703
+[ 92.5%] Iter 9028/9765 | SPS:  2728 | Reward:  0.063 | Value:  0.719 | VLoss: 0.0150 | PLoss: -0.0065 | Ent: 0.1926
+[ 92.5%] Iter 9029/9765 | SPS:  2728 | Reward:  0.058 | Value:  0.732 | VLoss: 0.0168 | PLoss: -0.0014 | Ent: 0.2266
+[ 92.5%] Iter 9030/9765 | SPS:  2728 | Reward:  0.066 | Value:  0.730 | VLoss: 0.0223 | PLoss: -0.0020 | Ent: 0.2226
+[ 92.5%] Iter 9031/9765 | SPS:  2728 | Reward:  0.056 | Value:  0.697 | VLoss: 0.0201 | PLoss: -0.0026 | Ent: 0.1946
+[ 92.5%] Iter 9032/9765 | SPS:  2728 | Reward:  0.067 | Value:  0.746 | VLoss: 0.0192 | PLoss: -0.0033 | Ent: 0.2437
+[ 92.5%] Iter 9033/9765 | SPS:  2728 | Reward:  0.067 | Value:  0.754 | VLoss: 0.0174 | PLoss: -0.0034 | Ent: 0.1825
+[ 92.5%] Iter 9034/9765 | SPS:  2728 | Reward:  0.056 | Value:  0.694 | VLoss: 0.0220 | PLoss: -0.0054 | Ent: 0.2524
+[ 92.5%] Iter 9035/9765 | SPS:  2728 | Reward:  0.066 | Value:  0.659 | VLoss: 0.0298 | PLoss: -0.0012 | Ent: 0.2478
+[ 92.5%] Iter 9036/9765 | SPS:  2728 | Reward:  0.064 | Value:  0.797 | VLoss: 0.0201 | PLoss: -0.0011 | Ent: 0.2233
+[ 92.5%] Iter 9037/9765 | SPS:  2728 | Reward:  0.079 | Value:  0.721 | VLoss: 0.0218 | PLoss: -0.0024 | Ent: 0.2151
+[ 92.6%] Iter 9038/9765 | SPS:  2728 | Reward:  0.063 | Value:  0.700 | VLoss: 0.0236 | PLoss: -0.0014 | Ent: 0.2394
+[ 92.6%] Iter 9039/9765 | SPS:  2728 | Reward:  0.046 | Value:  0.732 | VLoss: 0.0186 | PLoss: -0.0037 | Ent: 0.2549
+[ 92.6%] Iter 9040/9765 | SPS:  2728 | Reward:  0.074 | Value:  0.780 | VLoss: 0.0147 | PLoss: -0.0031 | Ent: 0.2069
+[ 92.6%] Iter 9041/9765 | SPS:  2729 | Reward:  0.074 | Value:  0.768 | VLoss: 0.0225 | PLoss: -0.0030 | Ent: 0.1991
+[ 92.6%] Iter 9042/9765 | SPS:  2729 | Reward:  0.069 | Value:  0.795 | VLoss: 0.0148 | PLoss: -0.0023 | Ent: 0.2417
+[ 92.6%] Iter 9043/9765 | SPS:  2729 | Reward:  0.064 | Value:  0.770 | VLoss: 0.0167 | PLoss: -0.0015 | Ent: 0.2056
+[ 92.6%] Iter 9044/9765 | SPS:  2729 | Reward:  0.062 | Value:  0.682 | VLoss: 0.0289 | PLoss: -0.0008 | Ent: 0.2504
+[ 92.6%] Iter 9045/9765 | SPS:  2729 | Reward:  0.049 | Value:  0.675 | VLoss: 0.0271 | PLoss: -0.0026 | Ent: 0.2436
+[ 92.6%] Iter 9046/9765 | SPS:  2729 | Reward:  0.056 | Value:  0.716 | VLoss: 0.0202 | PLoss: -0.0021 | Ent: 0.2710
+[ 92.6%] Iter 9047/9765 | SPS:  2729 | Reward:  0.057 | Value:  0.689 | VLoss: 0.0213 | PLoss: -0.0034 | Ent: 0.2036
+[ 92.7%] Iter 9048/9765 | SPS:  2729 | Reward:  0.053 | Value:  0.688 | VLoss: 0.0124 | PLoss: -0.0044 | Ent: 0.2628
+[ 92.7%] Iter 9049/9765 | SPS:  2729 | Reward:  0.058 | Value:  0.734 | VLoss: 0.0165 | PLoss: -0.0010 | Ent: 0.2541
+[ 92.7%] Iter 9050/9765 | SPS:  2729 | Reward:  0.065 | Value:  0.772 | VLoss: 0.0172 | PLoss: -0.0023 | Ent: 0.2246
+[ 92.7%] Iter 9051/9765 | SPS:  2729 | Reward:  0.057 | Value:  0.732 | VLoss: 0.0266 | PLoss: -0.0065 | Ent: 0.2520
+[ 92.7%] Iter 9052/9765 | SPS:  2729 | Reward:  0.059 | Value:  0.706 | VLoss: 0.0199 | PLoss: -0.0015 | Ent: 0.1978
+[ 92.7%] Iter 9053/9765 | SPS:  2730 | Reward:  0.065 | Value:  0.765 | VLoss: 0.0278 | PLoss: -0.0019 | Ent: 0.2172
+[ 92.7%] Iter 9054/9765 | SPS:  2730 | Reward:  0.076 | Value:  0.720 | VLoss: 0.0287 | PLoss: -0.0012 | Ent: 0.2719
+[ 92.7%] Iter 9055/9765 | SPS:  2730 | Reward:  0.061 | Value:  0.711 | VLoss: 0.0259 | PLoss: -0.0046 | Ent: 0.2327
+[ 92.7%] Iter 9056/9765 | SPS:  2730 | Reward:  0.062 | Value:  0.714 | VLoss: 0.0246 | PLoss: -0.0011 | Ent: 0.2207
+[ 92.7%] Iter 9057/9765 | SPS:  2730 | Reward:  0.076 | Value:  0.732 | VLoss: 0.0255 | PLoss: -0.0002 | Ent: 0.2397
+[ 92.8%] Iter 9058/9765 | SPS:  2730 | Reward:  0.062 | Value:  0.669 | VLoss: 0.0174 | PLoss: -0.0019 | Ent: 0.2747
+[ 92.8%] Iter 9059/9765 | SPS:  2730 | Reward:  0.059 | Value:  0.754 | VLoss: 0.0162 | PLoss: 0.0001 | Ent: 0.2138
+[ 92.8%] Iter 9060/9765 | SPS:  2730 | Reward:  0.062 | Value:  0.775 | VLoss: 0.0193 | PLoss: -0.0014 | Ent: 0.2430
+[ 92.8%] Iter 9061/9765 | SPS:  2730 | Reward:  0.069 | Value:  0.710 | VLoss: 0.0232 | PLoss: -0.0008 | Ent: 0.1762
+[ 92.8%] Iter 9062/9765 | SPS:  2730 | Reward:  0.045 | Value:  0.701 | VLoss: 0.0222 | PLoss: -0.0006 | Ent: 0.2610
+[ 92.8%] Iter 9063/9765 | SPS:  2730 | Reward:  0.060 | Value:  0.714 | VLoss: 0.0126 | PLoss: -0.0036 | Ent: 0.2260
+[ 92.8%] Iter 9064/9765 | SPS:  2730 | Reward:  0.062 | Value:  0.717 | VLoss: 0.0194 | PLoss: -0.0056 | Ent: 0.2559
+[ 92.8%] Iter 9065/9765 | SPS:  2731 | Reward:  0.062 | Value:  0.760 | VLoss: 0.0196 | PLoss: -0.0019 | Ent: 0.2564
+[ 92.8%] Iter 9066/9765 | SPS:  2731 | Reward:  0.055 | Value:  0.704 | VLoss: 0.0186 | PLoss: -0.0024 | Ent: 0.2278
+[ 92.9%] Iter 9067/9765 | SPS:  2731 | Reward:  0.064 | Value:  0.717 | VLoss: 0.0234 | PLoss: -0.0041 | Ent: 0.2142
+[ 92.9%] Iter 9068/9765 | SPS:  2731 | Reward:  0.060 | Value:  0.742 | VLoss: 0.0288 | PLoss: -0.0015 | Ent: 0.1957
+[ 92.9%] Iter 9069/9765 | SPS:  2731 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0264 | PLoss: -0.0009 | Ent: 0.2419
+[ 92.9%] Iter 9070/9765 | SPS:  2731 | Reward:  0.046 | Value:  0.732 | VLoss: 0.0198 | PLoss: -0.0042 | Ent: 0.2899
+[ 92.9%] Iter 9071/9765 | SPS:  2731 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0270 | PLoss: 0.0003 | Ent: 0.2220
+[ 92.9%] Iter 9072/9765 | SPS:  2731 | Reward:  0.073 | Value:  0.744 | VLoss: 0.0164 | PLoss: -0.0031 | Ent: 0.1993
+[ 92.9%] Iter 9073/9765 | SPS:  2731 | Reward:  0.052 | Value:  0.675 | VLoss: 0.0195 | PLoss: -0.0025 | Ent: 0.1425
+[ 92.9%] Iter 9074/9765 | SPS:  2731 | Reward:  0.059 | Value:  0.708 | VLoss: 0.0187 | PLoss: -0.0010 | Ent: 0.2261
+[ 92.9%] Iter 9075/9765 | SPS:  2731 | Reward:  0.062 | Value:  0.702 | VLoss: 0.0198 | PLoss: -0.0018 | Ent: 0.2378
+[ 92.9%] Iter 9076/9765 | SPS:  2731 | Reward:  0.070 | Value:  0.635 | VLoss: 0.0206 | PLoss: -0.0007 | Ent: 0.2656
+[ 93.0%] Iter 9077/9765 | SPS:  2732 | Reward:  0.055 | Value:  0.655 | VLoss: 0.0223 | PLoss: -0.0046 | Ent: 0.2552
+[ 93.0%] Iter 9078/9765 | SPS:  2732 | Reward:  0.057 | Value:  0.674 | VLoss: 0.0272 | PLoss: -0.0004 | Ent: 0.1936
+[ 93.0%] Iter 9079/9765 | SPS:  2732 | Reward:  0.075 | Value:  0.729 | VLoss: 0.0252 | PLoss: -0.0041 | Ent: 0.1833
+[ 93.0%] Iter 9080/9765 | SPS:  2732 | Reward:  0.063 | Value:  0.749 | VLoss: 0.0162 | PLoss: -0.0062 | Ent: 0.2397
+[ 93.0%] Iter 9081/9765 | SPS:  2732 | Reward:  0.071 | Value:  0.761 | VLoss: 0.0261 | PLoss: -0.0019 | Ent: 0.2229
+[ 93.0%] Iter 9082/9765 | SPS:  2732 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0225 | PLoss: -0.0022 | Ent: 0.1901
+[ 93.0%] Iter 9083/9765 | SPS:  2732 | Reward:  0.052 | Value:  0.732 | VLoss: 0.0133 | PLoss: -0.0024 | Ent: 0.2004
+[ 93.0%] Iter 9084/9765 | SPS:  2732 | Reward:  0.065 | Value:  0.764 | VLoss: 0.0215 | PLoss: -0.0016 | Ent: 0.2064
+[ 93.0%] Iter 9085/9765 | SPS:  2732 | Reward:  0.059 | Value:  0.671 | VLoss: 0.0220 | PLoss: -0.0035 | Ent: 0.2170
+[ 93.0%] Iter 9086/9765 | SPS:  2732 | Reward:  0.062 | Value:  0.694 | VLoss: 0.0172 | PLoss: -0.0025 | Ent: 0.1943
+[ 93.1%] Iter 9087/9765 | SPS:  2732 | Reward:  0.067 | Value:  0.713 | VLoss: 0.0235 | PLoss: -0.0033 | Ent: 0.2265
+[ 93.1%] Iter 9088/9765 | SPS:  2732 | Reward:  0.073 | Value:  0.770 | VLoss: 0.0202 | PLoss: -0.0011 | Ent: 0.2052
+[ 93.1%] Iter 9089/9765 | SPS:  2732 | Reward:  0.066 | Value:  0.773 | VLoss: 0.0170 | PLoss: -0.0015 | Ent: 0.2234
+[ 93.1%] Iter 9090/9765 | SPS:  2733 | Reward:  0.058 | Value:  0.706 | VLoss: 0.0305 | PLoss: -0.0027 | Ent: 0.2479
+[ 93.1%] Iter 9091/9765 | SPS:  2733 | Reward:  0.059 | Value:  0.744 | VLoss: 0.0253 | PLoss: -0.0007 | Ent: 0.2310
+[ 93.1%] Iter 9092/9765 | SPS:  2733 | Reward:  0.053 | Value:  0.693 | VLoss: 0.0202 | PLoss: -0.0016 | Ent: 0.2008
+[ 93.1%] Iter 9093/9765 | SPS:  2733 | Reward:  0.070 | Value:  0.716 | VLoss: 0.0215 | PLoss: 0.0001 | Ent: 0.2376
+[ 93.1%] Iter 9094/9765 | SPS:  2733 | Reward:  0.061 | Value:  0.766 | VLoss: 0.0199 | PLoss: -0.0016 | Ent: 0.2011
+[ 93.1%] Iter 9095/9765 | SPS:  2733 | Reward:  0.071 | Value:  0.723 | VLoss: 0.0221 | PLoss: -0.0025 | Ent: 0.1943
+[ 93.1%] Iter 9096/9765 | SPS:  2733 | Reward:  0.053 | Value:  0.639 | VLoss: 0.0208 | PLoss: -0.0018 | Ent: 0.2104
+[ 93.2%] Iter 9097/9765 | SPS:  2733 | Reward:  0.050 | Value:  0.753 | VLoss: 0.0184 | PLoss: -0.0017 | Ent: 0.1757
+[ 93.2%] Iter 9098/9765 | SPS:  2733 | Reward:  0.065 | Value:  0.739 | VLoss: 0.0207 | PLoss: -0.0014 | Ent: 0.1932
+[ 93.2%] Iter 9099/9765 | SPS:  2733 | Reward:  0.058 | Value:  0.751 | VLoss: 0.0215 | PLoss: -0.0035 | Ent: 0.2555
+[ 93.2%] Iter 9100/9765 | SPS:  2733 | Reward:  0.060 | Value:  0.764 | VLoss: 0.0148 | PLoss: -0.0005 | Ent: 0.2680
+[ 93.2%] Iter 9101/9765 | SPS:  2733 | Reward:  0.069 | Value:  0.754 | VLoss: 0.0184 | PLoss: -0.0067 | Ent: 0.2725
+[ 93.2%] Iter 9102/9765 | SPS:  2734 | Reward:  0.067 | Value:  0.742 | VLoss: 0.0183 | PLoss: -0.0030 | Ent: 0.1790
+[ 93.2%] Iter 9103/9765 | SPS:  2734 | Reward:  0.066 | Value:  0.715 | VLoss: 0.0257 | PLoss: -0.0011 | Ent: 0.1819
+[ 93.2%] Iter 9104/9765 | SPS:  2734 | Reward:  0.060 | Value:  0.736 | VLoss: 0.0188 | PLoss: -0.0007 | Ent: 0.2917
+[ 93.2%] Iter 9105/9765 | SPS:  2734 | Reward:  0.057 | Value:  0.688 | VLoss: 0.0232 | PLoss: -0.0020 | Ent: 0.2645
+[ 93.3%] Iter 9106/9765 | SPS:  2734 | Reward:  0.048 | Value:  0.655 | VLoss: 0.0161 | PLoss: -0.0014 | Ent: 0.2834
+[ 93.3%] Iter 9107/9765 | SPS:  2734 | Reward:  0.049 | Value:  0.669 | VLoss: 0.0254 | PLoss: -0.0014 | Ent: 0.2542
+[ 93.3%] Iter 9108/9765 | SPS:  2734 | Reward:  0.070 | Value:  0.719 | VLoss: 0.0249 | PLoss: -0.0027 | Ent: 0.2591
+[ 93.3%] Iter 9109/9765 | SPS:  2734 | Reward:  0.055 | Value:  0.664 | VLoss: 0.0256 | PLoss: -0.0031 | Ent: 0.2315
+[ 93.3%] Iter 9110/9765 | SPS:  2734 | Reward:  0.063 | Value:  0.714 | VLoss: 0.0214 | PLoss: -0.0011 | Ent: 0.2460
+[ 93.3%] Iter 9111/9765 | SPS:  2734 | Reward:  0.069 | Value:  0.751 | VLoss: 0.0320 | PLoss: -0.0015 | Ent: 0.2358
+[ 93.3%] Iter 9112/9765 | SPS:  2734 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0281 | PLoss: -0.0032 | Ent: 0.2527
+[ 93.3%] Iter 9113/9765 | SPS:  2734 | Reward:  0.058 | Value:  0.836 | VLoss: 0.0128 | PLoss: -0.0016 | Ent: 0.2220
+[ 93.3%] Iter 9114/9765 | SPS:  2735 | Reward:  0.060 | Value:  0.734 | VLoss: 0.0203 | PLoss: -0.0021 | Ent: 0.2366
+[ 93.3%] Iter 9115/9765 | SPS:  2735 | Reward:  0.067 | Value:  0.722 | VLoss: 0.0258 | PLoss: -0.0030 | Ent: 0.2342
+[ 93.4%] Iter 9116/9765 | SPS:  2735 | Reward:  0.059 | Value:  0.759 | VLoss: 0.0169 | PLoss: -0.0028 | Ent: 0.1934
+[ 93.4%] Iter 9117/9765 | SPS:  2735 | Reward:  0.060 | Value:  0.740 | VLoss: 0.0255 | PLoss: -0.0033 | Ent: 0.2262
+[ 93.4%] Iter 9118/9765 | SPS:  2735 | Reward:  0.070 | Value:  0.725 | VLoss: 0.0236 | PLoss: -0.0021 | Ent: 0.1971
+[ 93.4%] Iter 9119/9765 | SPS:  2735 | Reward:  0.066 | Value:  0.737 | VLoss: 0.0276 | PLoss: -0.0024 | Ent: 0.1750
+[ 93.4%] Iter 9120/9765 | SPS:  2735 | Reward:  0.061 | Value:  0.733 | VLoss: 0.0249 | PLoss: -0.0021 | Ent: 0.1959
+[ 93.4%] Iter 9121/9765 | SPS:  2735 | Reward:  0.056 | Value:  0.686 | VLoss: 0.0214 | PLoss: -0.0009 | Ent: 0.2042
+[ 93.4%] Iter 9122/9765 | SPS:  2735 | Reward:  0.060 | Value:  0.751 | VLoss: 0.0222 | PLoss: -0.0014 | Ent: 0.2139
+[ 93.4%] Iter 9123/9765 | SPS:  2735 | Reward:  0.052 | Value:  0.740 | VLoss: 0.0231 | PLoss: -0.0020 | Ent: 0.2614
+[ 93.4%] Iter 9124/9765 | SPS:  2735 | Reward:  0.058 | Value:  0.747 | VLoss: 0.0228 | PLoss: -0.0027 | Ent: 0.2133
+[ 93.4%] Iter 9125/9765 | SPS:  2735 | Reward:  0.065 | Value:  0.712 | VLoss: 0.0282 | PLoss: -0.0035 | Ent: 0.3138
+[ 93.5%] Iter 9126/9765 | SPS:  2736 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0267 | PLoss: -0.0055 | Ent: 0.2844
+[ 93.5%] Iter 9127/9765 | SPS:  2736 | Reward:  0.054 | Value:  0.719 | VLoss: 0.0236 | PLoss: -0.0017 | Ent: 0.1980
+[ 93.5%] Iter 9128/9765 | SPS:  2736 | Reward:  0.058 | Value:  0.752 | VLoss: 0.0236 | PLoss: -0.0018 | Ent: 0.1907
+[ 93.5%] Iter 9129/9765 | SPS:  2736 | Reward:  0.063 | Value:  0.643 | VLoss: 0.0288 | PLoss: -0.0055 | Ent: 0.2537
+[ 93.5%] Iter 9130/9765 | SPS:  2736 | Reward:  0.054 | Value:  0.658 | VLoss: 0.0189 | PLoss: -0.0027 | Ent: 0.1969
+[ 93.5%] Iter 9131/9765 | SPS:  2736 | Reward:  0.065 | Value:  0.774 | VLoss: 0.0199 | PLoss: -0.0018 | Ent: 0.2206
+[ 93.5%] Iter 9132/9765 | SPS:  2736 | Reward:  0.056 | Value:  0.695 | VLoss: 0.0192 | PLoss: 0.0007 | Ent: 0.2253
+[ 93.5%] Iter 9133/9765 | SPS:  2736 | Reward:  0.060 | Value:  0.739 | VLoss: 0.0217 | PLoss: -0.0023 | Ent: 0.1803
+[ 93.5%] Iter 9134/9765 | SPS:  2736 | Reward:  0.069 | Value:  0.726 | VLoss: 0.0216 | PLoss: -0.0026 | Ent: 0.2307
+[ 93.5%] Iter 9135/9765 | SPS:  2736 | Reward:  0.047 | Value:  0.702 | VLoss: 0.0162 | PLoss: -0.0008 | Ent: 0.2351
+[ 93.6%] Iter 9136/9765 | SPS:  2736 | Reward:  0.055 | Value:  0.659 | VLoss: 0.0227 | PLoss: -0.0025 | Ent: 0.2212
+[ 93.6%] Iter 9137/9765 | SPS:  2737 | Reward:  0.048 | Value:  0.667 | VLoss: 0.0156 | PLoss: -0.0023 | Ent: 0.2132
+[ 93.6%] Iter 9138/9765 | SPS:  2737 | Reward:  0.044 | Value:  0.590 | VLoss: 0.0270 | PLoss: -0.0010 | Ent: 0.2341
+[ 93.6%] Iter 9139/9765 | SPS:  2737 | Reward:  0.058 | Value:  0.613 | VLoss: 0.0198 | PLoss: -0.0005 | Ent: 0.2750
+[ 93.6%] Iter 9140/9765 | SPS:  2737 | Reward:  0.050 | Value:  0.643 | VLoss: 0.0207 | PLoss: -0.0047 | Ent: 0.2445
+[ 93.6%] Iter 9141/9765 | SPS:  2737 | Reward:  0.065 | Value:  0.682 | VLoss: 0.0277 | PLoss: -0.0002 | Ent: 0.2767
+[ 93.6%] Iter 9142/9765 | SPS:  2737 | Reward:  0.064 | Value:  0.692 | VLoss: 0.0260 | PLoss: -0.0020 | Ent: 0.2252
+[ 93.6%] Iter 9143/9765 | SPS:  2737 | Reward:  0.061 | Value:  0.710 | VLoss: 0.0188 | PLoss: -0.0022 | Ent: 0.2210
+[ 93.6%] Iter 9144/9765 | SPS:  2737 | Reward:  0.058 | Value:  0.758 | VLoss: 0.0123 | PLoss: -0.0020 | Ent: 0.2090
+[ 93.7%] Iter 9145/9765 | SPS:  2737 | Reward:  0.057 | Value:  0.639 | VLoss: 0.0132 | PLoss: -0.0008 | Ent: 0.2382
+[ 93.7%] Iter 9146/9765 | SPS:  2737 | Reward:  0.060 | Value:  0.689 | VLoss: 0.0238 | PLoss: -0.0041 | Ent: 0.2703
+[ 93.7%] Iter 9147/9765 | SPS:  2737 | Reward:  0.054 | Value:  0.700 | VLoss: 0.0374 | PLoss: -0.0004 | Ent: 0.2377
+[ 93.7%] Iter 9148/9765 | SPS:  2737 | Reward:  0.066 | Value:  0.714 | VLoss: 0.0294 | PLoss: -0.0019 | Ent: 0.2325
+[ 93.7%] Iter 9149/9765 | SPS:  2738 | Reward:  0.063 | Value:  0.664 | VLoss: 0.0326 | PLoss: -0.0023 | Ent: 0.2081
+[ 93.7%] Iter 9150/9765 | SPS:  2738 | Reward:  0.061 | Value:  0.746 | VLoss: 0.0194 | PLoss: -0.0021 | Ent: 0.2508
+[ 93.7%] Iter 9151/9765 | SPS:  2738 | Reward:  0.070 | Value:  0.720 | VLoss: 0.0264 | PLoss: -0.0018 | Ent: 0.2462
+[ 93.7%] Iter 9152/9765 | SPS:  2738 | Reward:  0.062 | Value:  0.678 | VLoss: 0.0308 | PLoss: -0.0030 | Ent: 0.2520
+[ 93.7%] Iter 9153/9765 | SPS:  2738 | Reward:  0.064 | Value:  0.759 | VLoss: 0.0150 | PLoss: -0.0027 | Ent: 0.1927
+[ 93.7%] Iter 9154/9765 | SPS:  2738 | Reward:  0.056 | Value:  0.769 | VLoss: 0.0185 | PLoss: -0.0022 | Ent: 0.2000
+[ 93.8%] Iter 9155/9765 | SPS:  2738 | Reward:  0.077 | Value:  0.733 | VLoss: 0.0292 | PLoss: -0.0035 | Ent: 0.2362
+[ 93.8%] Iter 9156/9765 | SPS:  2738 | Reward:  0.048 | Value:  0.706 | VLoss: 0.0234 | PLoss: -0.0019 | Ent: 0.2864
+[ 93.8%] Iter 9157/9765 | SPS:  2738 | Reward:  0.058 | Value:  0.737 | VLoss: 0.0251 | PLoss: -0.0004 | Ent: 0.2895
+[ 93.8%] Iter 9158/9765 | SPS:  2738 | Reward:  0.054 | Value:  0.736 | VLoss: 0.0122 | PLoss: -0.0010 | Ent: 0.2200
+[ 93.8%] Iter 9159/9765 | SPS:  2738 | Reward:  0.066 | Value:  0.716 | VLoss: 0.0227 | PLoss: -0.0020 | Ent: 0.1785
+[ 93.8%] Iter 9160/9765 | SPS:  2738 | Reward:  0.058 | Value:  0.744 | VLoss: 0.0134 | PLoss: -0.0005 | Ent: 0.2476
+[ 93.8%] Iter 9161/9765 | SPS:  2739 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0246 | PLoss: -0.0003 | Ent: 0.2496
+[ 93.8%] Iter 9162/9765 | SPS:  2739 | Reward:  0.056 | Value:  0.717 | VLoss: 0.0209 | PLoss: -0.0015 | Ent: 0.2466
+[ 93.8%] Iter 9163/9765 | SPS:  2739 | Reward:  0.065 | Value:  0.717 | VLoss: 0.0241 | PLoss: -0.0004 | Ent: 0.2625
+[ 93.8%] Iter 9164/9765 | SPS:  2739 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0223 | PLoss: -0.0013 | Ent: 0.2261
+[ 93.9%] Iter 9165/9765 | SPS:  2739 | Reward:  0.067 | Value:  0.763 | VLoss: 0.0197 | PLoss: -0.0012 | Ent: 0.2014
+[ 93.9%] Iter 9166/9765 | SPS:  2739 | Reward:  0.068 | Value:  0.749 | VLoss: 0.0329 | PLoss: -0.0040 | Ent: 0.2103
+[ 93.9%] Iter 9167/9765 | SPS:  2739 | Reward:  0.059 | Value:  0.743 | VLoss: 0.0206 | PLoss: -0.0026 | Ent: 0.2237
+[ 93.9%] Iter 9168/9765 | SPS:  2739 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0273 | PLoss: -0.0015 | Ent: 0.1984
+[ 93.9%] Iter 9169/9765 | SPS:  2739 | Reward:  0.074 | Value:  0.731 | VLoss: 0.0189 | PLoss: -0.0007 | Ent: 0.2339
+[ 93.9%] Iter 9170/9765 | SPS:  2739 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0293 | PLoss: -0.0018 | Ent: 0.1879
+[ 93.9%] Iter 9171/9765 | SPS:  2739 | Reward:  0.067 | Value:  0.752 | VLoss: 0.0199 | PLoss: -0.0009 | Ent: 0.2004
+[ 93.9%] Iter 9172/9765 | SPS:  2739 | Reward:  0.052 | Value:  0.772 | VLoss: 0.0227 | PLoss: -0.0014 | Ent: 0.2261
+[ 93.9%] Iter 9173/9765 | SPS:  2740 | Reward:  0.063 | Value:  0.704 | VLoss: 0.0216 | PLoss: -0.0024 | Ent: 0.2614
+[ 93.9%] Iter 9174/9765 | SPS:  2740 | Reward:  0.060 | Value:  0.746 | VLoss: 0.0206 | PLoss: -0.0015 | Ent: 0.2308
+[ 94.0%] Iter 9175/9765 | SPS:  2740 | Reward:  0.057 | Value:  0.644 | VLoss: 0.0296 | PLoss: -0.0006 | Ent: 0.2244
+[ 94.0%] Iter 9176/9765 | SPS:  2740 | Reward:  0.075 | Value:  0.752 | VLoss: 0.0238 | PLoss: -0.0018 | Ent: 0.2105
+[ 94.0%] Iter 9177/9765 | SPS:  2740 | Reward:  0.065 | Value:  0.782 | VLoss: 0.0206 | PLoss: -0.0041 | Ent: 0.2573
+[ 94.0%] Iter 9178/9765 | SPS:  2740 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0188 | PLoss: -0.0010 | Ent: 0.2340
+[ 94.0%] Iter 9179/9765 | SPS:  2740 | Reward:  0.055 | Value:  0.673 | VLoss: 0.0226 | PLoss: -0.0014 | Ent: 0.2575
+[ 94.0%] Iter 9180/9765 | SPS:  2740 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0151 | PLoss: -0.0024 | Ent: 0.1986
+[ 94.0%] Iter 9181/9765 | SPS:  2740 | Reward:  0.058 | Value:  0.678 | VLoss: 0.0274 | PLoss: -0.0032 | Ent: 0.2629
+[ 94.0%] Iter 9182/9765 | SPS:  2740 | Reward:  0.059 | Value:  0.689 | VLoss: 0.0251 | PLoss: -0.0011 | Ent: 0.2291
+[ 94.0%] Iter 9183/9765 | SPS:  2740 | Reward:  0.068 | Value:  0.804 | VLoss: 0.0204 | PLoss: -0.0021 | Ent: 0.2281
+[ 94.1%] Iter 9184/9765 | SPS:  2741 | Reward:  0.066 | Value:  0.717 | VLoss: 0.0235 | PLoss: -0.0009 | Ent: 0.2221
+[ 94.1%] Iter 9185/9765 | SPS:  2741 | Reward:  0.065 | Value:  0.666 | VLoss: 0.0263 | PLoss: -0.0011 | Ent: 0.2260
+[ 94.1%] Iter 9186/9765 | SPS:  2741 | Reward:  0.062 | Value:  0.745 | VLoss: 0.0103 | PLoss: -0.0008 | Ent: 0.2038
+[ 94.1%] Iter 9187/9765 | SPS:  2741 | Reward:  0.062 | Value:  0.770 | VLoss: 0.0161 | PLoss: -0.0016 | Ent: 0.1982
+[ 94.1%] Iter 9188/9765 | SPS:  2741 | Reward:  0.060 | Value:  0.701 | VLoss: 0.0250 | PLoss: -0.0016 | Ent: 0.2156
+[ 94.1%] Iter 9189/9765 | SPS:  2741 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0213 | PLoss: -0.0015 | Ent: 0.2113
+[ 94.1%] Iter 9190/9765 | SPS:  2741 | Reward:  0.062 | Value:  0.714 | VLoss: 0.0241 | PLoss: -0.0007 | Ent: 0.2379
+[ 94.1%] Iter 9191/9765 | SPS:  2741 | Reward:  0.056 | Value:  0.681 | VLoss: 0.0160 | PLoss: -0.0031 | Ent: 0.2529
+[ 94.1%] Iter 9192/9765 | SPS:  2741 | Reward:  0.061 | Value:  0.717 | VLoss: 0.0196 | PLoss: -0.0022 | Ent: 0.2092
+[ 94.1%] Iter 9193/9765 | SPS:  2741 | Reward:  0.050 | Value:  0.628 | VLoss: 0.0265 | PLoss: -0.0048 | Ent: 0.2424
+[ 94.2%] Iter 9194/9765 | SPS:  2741 | Reward:  0.062 | Value:  0.707 | VLoss: 0.0177 | PLoss: -0.0021 | Ent: 0.2564
+[ 94.2%] Iter 9195/9765 | SPS:  2741 | Reward:  0.067 | Value:  0.728 | VLoss: 0.0262 | PLoss: -0.0033 | Ent: 0.2806
+[ 94.2%] Iter 9196/9765 | SPS:  2742 | Reward:  0.055 | Value:  0.717 | VLoss: 0.0187 | PLoss: -0.0014 | Ent: 0.2449
+[ 94.2%] Iter 9197/9765 | SPS:  2742 | Reward:  0.070 | Value:  0.777 | VLoss: 0.0145 | PLoss: -0.0004 | Ent: 0.2408
+[ 94.2%] Iter 9198/9765 | SPS:  2742 | Reward:  0.052 | Value:  0.739 | VLoss: 0.0206 | PLoss: -0.0016 | Ent: 0.2519
+[ 94.2%] Iter 9199/9765 | SPS:  2742 | Reward:  0.067 | Value:  0.700 | VLoss: 0.0225 | PLoss: -0.0023 | Ent: 0.2159
+[ 94.2%] Iter 9200/9765 | SPS:  2742 | Reward:  0.064 | Value:  0.687 | VLoss: 0.0213 | PLoss: -0.0030 | Ent: 0.2360
+[ 94.2%] Iter 9201/9765 | SPS:  2742 | Reward:  0.042 | Value:  0.730 | VLoss: 0.0184 | PLoss: -0.0005 | Ent: 0.2144
+[ 94.2%] Iter 9202/9765 | SPS:  2742 | Reward:  0.062 | Value:  0.703 | VLoss: 0.0236 | PLoss: -0.0031 | Ent: 0.2323
+[ 94.2%] Iter 9203/9765 | SPS:  2742 | Reward:  0.057 | Value:  0.717 | VLoss: 0.0139 | PLoss: -0.0006 | Ent: 0.2161
+[ 94.3%] Iter 9204/9765 | SPS:  2742 | Reward:  0.061 | Value:  0.701 | VLoss: 0.0317 | PLoss: -0.0019 | Ent: 0.2493
+[ 94.3%] Iter 9205/9765 | SPS:  2742 | Reward:  0.057 | Value:  0.709 | VLoss: 0.0176 | PLoss: -0.0008 | Ent: 0.2573
+[ 94.3%] Iter 9206/9765 | SPS:  2742 | Reward:  0.057 | Value:  0.710 | VLoss: 0.0208 | PLoss: -0.0013 | Ent: 0.2660
+[ 94.3%] Iter 9207/9765 | SPS:  2742 | Reward:  0.076 | Value:  0.703 | VLoss: 0.0289 | PLoss: -0.0008 | Ent: 0.2016
+[ 94.3%] Iter 9208/9765 | SPS:  2743 | Reward:  0.061 | Value:  0.717 | VLoss: 0.0219 | PLoss: -0.0006 | Ent: 0.2457
+[ 94.3%] Iter 9209/9765 | SPS:  2743 | Reward:  0.055 | Value:  0.678 | VLoss: 0.0192 | PLoss: -0.0009 | Ent: 0.2373
+[ 94.3%] Iter 9210/9765 | SPS:  2743 | Reward:  0.063 | Value:  0.713 | VLoss: 0.0230 | PLoss: -0.0036 | Ent: 0.2824
+[ 94.3%] Iter 9211/9765 | SPS:  2743 | Reward:  0.062 | Value:  0.704 | VLoss: 0.0227 | PLoss: -0.0016 | Ent: 0.2672
+[ 94.3%] Iter 9212/9765 | SPS:  2743 | Reward:  0.047 | Value:  0.700 | VLoss: 0.0200 | PLoss: -0.0029 | Ent: 0.2065
+[ 94.3%] Iter 9213/9765 | SPS:  2743 | Reward:  0.062 | Value:  0.767 | VLoss: 0.0216 | PLoss: -0.0028 | Ent: 0.2209
+[ 94.4%] Iter 9214/9765 | SPS:  2743 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0268 | PLoss: -0.0015 | Ent: 0.2401
+[ 94.4%] Iter 9215/9765 | SPS:  2743 | Reward:  0.057 | Value:  0.732 | VLoss: 0.0177 | PLoss: -0.0011 | Ent: 0.2217
+[ 94.4%] Iter 9216/9765 | SPS:  2743 | Reward:  0.070 | Value:  0.712 | VLoss: 0.0214 | PLoss: -0.0013 | Ent: 0.2491
+[ 94.4%] Iter 9217/9765 | SPS:  2743 | Reward:  0.063 | Value:  0.674 | VLoss: 0.0259 | PLoss: -0.0013 | Ent: 0.2429
+[ 94.4%] Iter 9218/9765 | SPS:  2743 | Reward:  0.052 | Value:  0.709 | VLoss: 0.0275 | PLoss: -0.0010 | Ent: 0.2353
+[ 94.4%] Iter 9219/9765 | SPS:  2743 | Reward:  0.057 | Value:  0.724 | VLoss: 0.0167 | PLoss: -0.0027 | Ent: 0.2523
+[ 94.4%] Iter 9220/9765 | SPS:  2744 | Reward:  0.062 | Value:  0.744 | VLoss: 0.0210 | PLoss: -0.0048 | Ent: 0.1984
+[ 94.4%] Iter 9221/9765 | SPS:  2744 | Reward:  0.057 | Value:  0.683 | VLoss: 0.0249 | PLoss: -0.0012 | Ent: 0.2299
+[ 94.4%] Iter 9222/9765 | SPS:  2744 | Reward:  0.060 | Value:  0.716 | VLoss: 0.0220 | PLoss: -0.0025 | Ent: 0.2147
+[ 94.4%] Iter 9223/9765 | SPS:  2744 | Reward:  0.062 | Value:  0.738 | VLoss: 0.0218 | PLoss: -0.0019 | Ent: 0.2027
+[ 94.5%] Iter 9224/9765 | SPS:  2744 | Reward:  0.071 | Value:  0.712 | VLoss: 0.0156 | PLoss: -0.0003 | Ent: 0.2171
+[ 94.5%] Iter 9225/9765 | SPS:  2744 | Reward:  0.067 | Value:  0.777 | VLoss: 0.0232 | PLoss: -0.0014 | Ent: 0.2556
+[ 94.5%] Iter 9226/9765 | SPS:  2744 | Reward:  0.064 | Value:  0.692 | VLoss: 0.0250 | PLoss: -0.0004 | Ent: 0.2136
+[ 94.5%] Iter 9227/9765 | SPS:  2744 | Reward:  0.053 | Value:  0.734 | VLoss: 0.0222 | PLoss: -0.0012 | Ent: 0.2041
+[ 94.5%] Iter 9228/9765 | SPS:  2744 | Reward:  0.046 | Value:  0.664 | VLoss: 0.0155 | PLoss: -0.0028 | Ent: 0.1848
+[ 94.5%] Iter 9229/9765 | SPS:  2744 | Reward:  0.048 | Value:  0.698 | VLoss: 0.0162 | PLoss: -0.0027 | Ent: 0.2039
+[ 94.5%] Iter 9230/9765 | SPS:  2744 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0260 | PLoss: -0.0021 | Ent: 0.2379
+[ 94.5%] Iter 9231/9765 | SPS:  2744 | Reward:  0.058 | Value:  0.716 | VLoss: 0.0254 | PLoss: -0.0017 | Ent: 0.2236
+[ 94.5%] Iter 9232/9765 | SPS:  2745 | Reward:  0.048 | Value:  0.731 | VLoss: 0.0307 | PLoss: -0.0011 | Ent: 0.2260
+[ 94.6%] Iter 9233/9765 | SPS:  2745 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0259 | PLoss: -0.0002 | Ent: 0.2210
+[ 94.6%] Iter 9234/9765 | SPS:  2745 | Reward:  0.069 | Value:  0.647 | VLoss: 0.0280 | PLoss: -0.0030 | Ent: 0.2701
+[ 94.6%] Iter 9235/9765 | SPS:  2745 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0220 | PLoss: -0.0024 | Ent: 0.2204
+[ 94.6%] Iter 9236/9765 | SPS:  2745 | Reward:  0.058 | Value:  0.669 | VLoss: 0.0246 | PLoss: -0.0009 | Ent: 0.2238
+[ 94.6%] Iter 9237/9765 | SPS:  2745 | Reward:  0.060 | Value:  0.749 | VLoss: 0.0165 | PLoss: -0.0013 | Ent: 0.1942
+[ 94.6%] Iter 9238/9765 | SPS:  2745 | Reward:  0.067 | Value:  0.703 | VLoss: 0.0231 | PLoss: -0.0011 | Ent: 0.2266
+[ 94.6%] Iter 9239/9765 | SPS:  2745 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0212 | PLoss: -0.0010 | Ent: 0.2153
+[ 94.6%] Iter 9240/9765 | SPS:  2745 | Reward:  0.064 | Value:  0.718 | VLoss: 0.0225 | PLoss: -0.0001 | Ent: 0.2492
+[ 94.6%] Iter 9241/9765 | SPS:  2745 | Reward:  0.066 | Value:  0.708 | VLoss: 0.0283 | PLoss: -0.0023 | Ent: 0.2092
+[ 94.6%] Iter 9242/9765 | SPS:  2745 | Reward:  0.062 | Value:  0.758 | VLoss: 0.0112 | PLoss: -0.0013 | Ent: 0.1607
+[ 94.7%] Iter 9243/9765 | SPS:  2745 | Reward:  0.048 | Value:  0.730 | VLoss: 0.0225 | PLoss: -0.0010 | Ent: 0.2649
+[ 94.7%] Iter 9244/9765 | SPS:  2745 | Reward:  0.054 | Value:  0.733 | VLoss: 0.0160 | PLoss: -0.0003 | Ent: 0.2177
+[ 94.7%] Iter 9245/9765 | SPS:  2746 | Reward:  0.056 | Value:  0.769 | VLoss: 0.0208 | PLoss: -0.0036 | Ent: 0.1780
+[ 94.7%] Iter 9246/9765 | SPS:  2746 | Reward:  0.058 | Value:  0.711 | VLoss: 0.0191 | PLoss: -0.0014 | Ent: 0.2260
+[ 94.7%] Iter 9247/9765 | SPS:  2746 | Reward:  0.043 | Value:  0.604 | VLoss: 0.0209 | PLoss: -0.0010 | Ent: 0.2496
+[ 94.7%] Iter 9248/9765 | SPS:  2746 | Reward:  0.052 | Value:  0.647 | VLoss: 0.0212 | PLoss: -0.0009 | Ent: 0.2002
+[ 94.7%] Iter 9249/9765 | SPS:  2746 | Reward:  0.056 | Value:  0.664 | VLoss: 0.0218 | PLoss: -0.0015 | Ent: 0.2743
+[ 94.7%] Iter 9250/9765 | SPS:  2746 | Reward:  0.047 | Value:  0.731 | VLoss: 0.0164 | PLoss: -0.0010 | Ent: 0.2190
+[ 94.7%] Iter 9251/9765 | SPS:  2746 | Reward:  0.063 | Value:  0.692 | VLoss: 0.0303 | PLoss: -0.0019 | Ent: 0.2709
+[ 94.7%] Iter 9252/9765 | SPS:  2746 | Reward:  0.057 | Value:  0.704 | VLoss: 0.0313 | PLoss: -0.0020 | Ent: 0.2284
+[ 94.8%] Iter 9253/9765 | SPS:  2746 | Reward:  0.052 | Value:  0.656 | VLoss: 0.0179 | PLoss: -0.0031 | Ent: 0.2124
+[ 94.8%] Iter 9254/9765 | SPS:  2746 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0161 | PLoss: -0.0019 | Ent: 0.2786
+[ 94.8%] Iter 9255/9765 | SPS:  2746 | Reward:  0.068 | Value:  0.721 | VLoss: 0.0234 | PLoss: -0.0013 | Ent: 0.2134
+[ 94.8%] Iter 9256/9765 | SPS:  2746 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0297 | PLoss: -0.0059 | Ent: 0.2017
+[ 94.8%] Iter 9257/9765 | SPS:  2747 | Reward:  0.062 | Value:  0.680 | VLoss: 0.0246 | PLoss: -0.0010 | Ent: 0.2262
+[ 94.8%] Iter 9258/9765 | SPS:  2747 | Reward:  0.061 | Value:  0.745 | VLoss: 0.0204 | PLoss: -0.0014 | Ent: 0.2518
+[ 94.8%] Iter 9259/9765 | SPS:  2747 | Reward:  0.054 | Value:  0.683 | VLoss: 0.0254 | PLoss: -0.0019 | Ent: 0.2197
+[ 94.8%] Iter 9260/9765 | SPS:  2747 | Reward:  0.060 | Value:  0.781 | VLoss: 0.0152 | PLoss: -0.0039 | Ent: 0.2256
+[ 94.8%] Iter 9261/9765 | SPS:  2747 | Reward:  0.064 | Value:  0.710 | VLoss: 0.0295 | PLoss: -0.0015 | Ent: 0.2793
+[ 94.8%] Iter 9262/9765 | SPS:  2747 | Reward:  0.064 | Value:  0.806 | VLoss: 0.0168 | PLoss: -0.0022 | Ent: 0.1882
+[ 94.9%] Iter 9263/9765 | SPS:  2747 | Reward:  0.059 | Value:  0.745 | VLoss: 0.0219 | PLoss: -0.0006 | Ent: 0.2061
+[ 94.9%] Iter 9264/9765 | SPS:  2747 | Reward:  0.048 | Value:  0.742 | VLoss: 0.0173 | PLoss: -0.0012 | Ent: 0.2527
+[ 94.9%] Iter 9265/9765 | SPS:  2747 | Reward:  0.067 | Value:  0.716 | VLoss: 0.0179 | PLoss: -0.0006 | Ent: 0.2164
+[ 94.9%] Iter 9266/9765 | SPS:  2747 | Reward:  0.079 | Value:  0.780 | VLoss: 0.0164 | PLoss: -0.0007 | Ent: 0.2048
+[ 94.9%] Iter 9267/9765 | SPS:  2747 | Reward:  0.049 | Value:  0.684 | VLoss: 0.0270 | PLoss: -0.0019 | Ent: 0.2643
+[ 94.9%] Iter 9268/9765 | SPS:  2747 | Reward:  0.063 | Value:  0.715 | VLoss: 0.0200 | PLoss: -0.0021 | Ent: 0.2437
+[ 94.9%] Iter 9269/9765 | SPS:  2748 | Reward:  0.051 | Value:  0.716 | VLoss: 0.0192 | PLoss: -0.0005 | Ent: 0.2151
+[ 94.9%] Iter 9270/9765 | SPS:  2748 | Reward:  0.061 | Value:  0.726 | VLoss: 0.0247 | PLoss: -0.0014 | Ent: 0.2282
+[ 94.9%] Iter 9271/9765 | SPS:  2748 | Reward:  0.063 | Value:  0.693 | VLoss: 0.0257 | PLoss: -0.0016 | Ent: 0.2033
+[ 95.0%] Iter 9272/9765 | SPS:  2748 | Reward:  0.060 | Value:  0.651 | VLoss: 0.0233 | PLoss: -0.0025 | Ent: 0.2229
+[ 95.0%] Iter 9273/9765 | SPS:  2748 | Reward:  0.064 | Value:  0.664 | VLoss: 0.0247 | PLoss: -0.0018 | Ent: 0.2452
+[ 95.0%] Iter 9274/9765 | SPS:  2748 | Reward:  0.069 | Value:  0.758 | VLoss: 0.0205 | PLoss: -0.0014 | Ent: 0.2289
+[ 95.0%] Iter 9275/9765 | SPS:  2748 | Reward:  0.042 | Value:  0.655 | VLoss: 0.0337 | PLoss: -0.0027 | Ent: 0.2139
+[ 95.0%] Iter 9276/9765 | SPS:  2748 | Reward:  0.068 | Value:  0.775 | VLoss: 0.0192 | PLoss: -0.0007 | Ent: 0.1574
+[ 95.0%] Iter 9277/9765 | SPS:  2748 | Reward:  0.063 | Value:  0.717 | VLoss: 0.0226 | PLoss: -0.0009 | Ent: 0.2175
+[ 95.0%] Iter 9278/9765 | SPS:  2748 | Reward:  0.061 | Value:  0.712 | VLoss: 0.0199 | PLoss: -0.0006 | Ent: 0.2183
+[ 95.0%] Iter 9279/9765 | SPS:  2748 | Reward:  0.061 | Value:  0.749 | VLoss: 0.0203 | PLoss: -0.0010 | Ent: 0.1940
+[ 95.0%] Iter 9280/9765 | SPS:  2748 | Reward:  0.063 | Value:  0.729 | VLoss: 0.0214 | PLoss: -0.0018 | Ent: 0.2514
+[ 95.0%] Iter 9281/9765 | SPS:  2749 | Reward:  0.059 | Value:  0.773 | VLoss: 0.0219 | PLoss: -0.0021 | Ent: 0.2414
+[ 95.1%] Iter 9282/9765 | SPS:  2749 | Reward:  0.054 | Value:  0.693 | VLoss: 0.0189 | PLoss: -0.0011 | Ent: 0.2852
+[ 95.1%] Iter 9283/9765 | SPS:  2749 | Reward:  0.061 | Value:  0.786 | VLoss: 0.0178 | PLoss: -0.0006 | Ent: 0.2359
+[ 95.1%] Iter 9284/9765 | SPS:  2749 | Reward:  0.053 | Value:  0.676 | VLoss: 0.0139 | PLoss: -0.0008 | Ent: 0.2659
+[ 95.1%] Iter 9285/9765 | SPS:  2749 | Reward:  0.058 | Value:  0.649 | VLoss: 0.0209 | PLoss: -0.0009 | Ent: 0.2247
+[ 95.1%] Iter 9286/9765 | SPS:  2749 | Reward:  0.043 | Value:  0.632 | VLoss: 0.0221 | PLoss: -0.0044 | Ent: 0.2189
+[ 95.1%] Iter 9287/9765 | SPS:  2749 | Reward:  0.059 | Value:  0.756 | VLoss: 0.0138 | PLoss: -0.0015 | Ent: 0.1853
+[ 95.1%] Iter 9288/9765 | SPS:  2749 | Reward:  0.067 | Value:  0.695 | VLoss: 0.0212 | PLoss: -0.0011 | Ent: 0.2616
+[ 95.1%] Iter 9289/9765 | SPS:  2749 | Reward:  0.059 | Value:  0.774 | VLoss: 0.0186 | PLoss: -0.0036 | Ent: 0.2246
+[ 95.1%] Iter 9290/9765 | SPS:  2749 | Reward:  0.051 | Value:  0.670 | VLoss: 0.0158 | PLoss: -0.0027 | Ent: 0.2638
+[ 95.1%] Iter 9291/9765 | SPS:  2749 | Reward:  0.062 | Value:  0.721 | VLoss: 0.0210 | PLoss: -0.0016 | Ent: 0.2238
+[ 95.2%] Iter 9292/9765 | SPS:  2749 | Reward:  0.053 | Value:  0.731 | VLoss: 0.0214 | PLoss: -0.0022 | Ent: 0.2717
+[ 95.2%] Iter 9293/9765 | SPS:  2750 | Reward:  0.064 | Value:  0.743 | VLoss: 0.0162 | PLoss: -0.0027 | Ent: 0.1798
+[ 95.2%] Iter 9294/9765 | SPS:  2750 | Reward:  0.065 | Value:  0.628 | VLoss: 0.0288 | PLoss: -0.0010 | Ent: 0.2023
+[ 95.2%] Iter 9295/9765 | SPS:  2750 | Reward:  0.051 | Value:  0.688 | VLoss: 0.0219 | PLoss: 0.0002 | Ent: 0.2602
+[ 95.2%] Iter 9296/9765 | SPS:  2750 | Reward:  0.062 | Value:  0.690 | VLoss: 0.0251 | PLoss: -0.0039 | Ent: 0.2114
+[ 95.2%] Iter 9297/9765 | SPS:  2750 | Reward:  0.049 | Value:  0.699 | VLoss: 0.0216 | PLoss: -0.0025 | Ent: 0.2824
+[ 95.2%] Iter 9298/9765 | SPS:  2750 | Reward:  0.058 | Value:  0.735 | VLoss: 0.0195 | PLoss: -0.0005 | Ent: 0.1953
+[ 95.2%] Iter 9299/9765 | SPS:  2750 | Reward:  0.068 | Value:  0.684 | VLoss: 0.0288 | PLoss: -0.0013 | Ent: 0.2149
+[ 95.2%] Iter 9300/9765 | SPS:  2750 | Reward:  0.058 | Value:  0.739 | VLoss: 0.0250 | PLoss: -0.0049 | Ent: 0.2283
+[ 95.2%] Iter 9301/9765 | SPS:  2750 | Reward:  0.063 | Value:  0.713 | VLoss: 0.0160 | PLoss: -0.0040 | Ent: 0.2442
+[ 95.3%] Iter 9302/9765 | SPS:  2750 | Reward:  0.052 | Value:  0.770 | VLoss: 0.0181 | PLoss: -0.0005 | Ent: 0.1898
+[ 95.3%] Iter 9303/9765 | SPS:  2750 | Reward:  0.063 | Value:  0.717 | VLoss: 0.0238 | PLoss: -0.0003 | Ent: 0.2232
+[ 95.3%] Iter 9304/9765 | SPS:  2750 | Reward:  0.062 | Value:  0.737 | VLoss: 0.0277 | PLoss: -0.0006 | Ent: 0.2222
+[ 95.3%] Iter 9305/9765 | SPS:  2751 | Reward:  0.071 | Value:  0.744 | VLoss: 0.0228 | PLoss: -0.0015 | Ent: 0.2021
+[ 95.3%] Iter 9306/9765 | SPS:  2751 | Reward:  0.058 | Value:  0.734 | VLoss: 0.0261 | PLoss: -0.0009 | Ent: 0.2385
+[ 95.3%] Iter 9307/9765 | SPS:  2751 | Reward:  0.068 | Value:  0.719 | VLoss: 0.0240 | PLoss: -0.0016 | Ent: 0.2068
+[ 95.3%] Iter 9308/9765 | SPS:  2751 | Reward:  0.049 | Value:  0.721 | VLoss: 0.0294 | PLoss: -0.0017 | Ent: 0.2093
+[ 95.3%] Iter 9309/9765 | SPS:  2751 | Reward:  0.069 | Value:  0.743 | VLoss: 0.0266 | PLoss: -0.0027 | Ent: 0.2536
+[ 95.3%] Iter 9310/9765 | SPS:  2751 | Reward:  0.053 | Value:  0.769 | VLoss: 0.0220 | PLoss: -0.0015 | Ent: 0.1989
+[ 95.4%] Iter 9311/9765 | SPS:  2751 | Reward:  0.056 | Value:  0.701 | VLoss: 0.0206 | PLoss: -0.0024 | Ent: 0.2072
+[ 95.4%] Iter 9312/9765 | SPS:  2751 | Reward:  0.062 | Value:  0.696 | VLoss: 0.0171 | PLoss: -0.0024 | Ent: 0.2457
+[ 95.4%] Iter 9313/9765 | SPS:  2751 | Reward:  0.064 | Value:  0.743 | VLoss: 0.0195 | PLoss: -0.0018 | Ent: 0.1788
+[ 95.4%] Iter 9314/9765 | SPS:  2751 | Reward:  0.062 | Value:  0.712 | VLoss: 0.0223 | PLoss: -0.0014 | Ent: 0.2636
+[ 95.4%] Iter 9315/9765 | SPS:  2751 | Reward:  0.069 | Value:  0.720 | VLoss: 0.0247 | PLoss: -0.0016 | Ent: 0.2433
+[ 95.4%] Iter 9316/9765 | SPS:  2751 | Reward:  0.063 | Value:  0.686 | VLoss: 0.0190 | PLoss: -0.0011 | Ent: 0.2641
+[ 95.4%] Iter 9317/9765 | SPS:  2752 | Reward:  0.057 | Value:  0.754 | VLoss: 0.0197 | PLoss: -0.0015 | Ent: 0.1629
+[ 95.4%] Iter 9318/9765 | SPS:  2752 | Reward:  0.062 | Value:  0.699 | VLoss: 0.0226 | PLoss: -0.0020 | Ent: 0.2381
+[ 95.4%] Iter 9319/9765 | SPS:  2752 | Reward:  0.066 | Value:  0.712 | VLoss: 0.0302 | PLoss: -0.0065 | Ent: 0.2433
+[ 95.4%] Iter 9320/9765 | SPS:  2752 | Reward:  0.058 | Value:  0.688 | VLoss: 0.0229 | PLoss: -0.0023 | Ent: 0.2559
+[ 95.5%] Iter 9321/9765 | SPS:  2752 | Reward:  0.057 | Value:  0.603 | VLoss: 0.0320 | PLoss: -0.0037 | Ent: 0.2442
+[ 95.5%] Iter 9322/9765 | SPS:  2752 | Reward:  0.073 | Value:  0.734 | VLoss: 0.0245 | PLoss: -0.0014 | Ent: 0.2016
+[ 95.5%] Iter 9323/9765 | SPS:  2752 | Reward:  0.054 | Value:  0.671 | VLoss: 0.0213 | PLoss: -0.0004 | Ent: 0.2683
+[ 95.5%] Iter 9324/9765 | SPS:  2752 | Reward:  0.062 | Value:  0.716 | VLoss: 0.0218 | PLoss: -0.0011 | Ent: 0.2608
+[ 95.5%] Iter 9325/9765 | SPS:  2752 | Reward:  0.058 | Value:  0.754 | VLoss: 0.0176 | PLoss: -0.0016 | Ent: 0.2563
+[ 95.5%] Iter 9326/9765 | SPS:  2752 | Reward:  0.058 | Value:  0.790 | VLoss: 0.0164 | PLoss: -0.0014 | Ent: 0.2302
+[ 95.5%] Iter 9327/9765 | SPS:  2752 | Reward:  0.062 | Value:  0.758 | VLoss: 0.0178 | PLoss: -0.0006 | Ent: 0.2201
+[ 95.5%] Iter 9328/9765 | SPS:  2752 | Reward:  0.056 | Value:  0.721 | VLoss: 0.0207 | PLoss: -0.0011 | Ent: 0.2471
+[ 95.5%] Iter 9329/9765 | SPS:  2752 | Reward:  0.066 | Value:  0.720 | VLoss: 0.0273 | PLoss: -0.0003 | Ent: 0.2436
+[ 95.5%] Iter 9330/9765 | SPS:  2753 | Reward:  0.059 | Value:  0.680 | VLoss: 0.0260 | PLoss: -0.0018 | Ent: 0.2677
+[ 95.6%] Iter 9331/9765 | SPS:  2753 | Reward:  0.062 | Value:  0.746 | VLoss: 0.0234 | PLoss: -0.0012 | Ent: 0.2697
+[ 95.6%] Iter 9332/9765 | SPS:  2753 | Reward:  0.066 | Value:  0.729 | VLoss: 0.0210 | PLoss: -0.0008 | Ent: 0.2537
+[ 95.6%] Iter 9333/9765 | SPS:  2753 | Reward:  0.050 | Value:  0.694 | VLoss: 0.0194 | PLoss: -0.0007 | Ent: 0.2879
+[ 95.6%] Iter 9334/9765 | SPS:  2753 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0192 | PLoss: -0.0041 | Ent: 0.2380
+[ 95.6%] Iter 9335/9765 | SPS:  2753 | Reward:  0.056 | Value:  0.682 | VLoss: 0.0293 | PLoss: -0.0010 | Ent: 0.2258
+[ 95.6%] Iter 9336/9765 | SPS:  2753 | Reward:  0.066 | Value:  0.762 | VLoss: 0.0228 | PLoss: -0.0020 | Ent: 0.2284
+[ 95.6%] Iter 9337/9765 | SPS:  2753 | Reward:  0.064 | Value:  0.734 | VLoss: 0.0166 | PLoss: -0.0006 | Ent: 0.2108
+[ 95.6%] Iter 9338/9765 | SPS:  2753 | Reward:  0.059 | Value:  0.743 | VLoss: 0.0223 | PLoss: -0.0009 | Ent: 0.2072
+[ 95.6%] Iter 9339/9765 | SPS:  2753 | Reward:  0.054 | Value:  0.711 | VLoss: 0.0150 | PLoss: -0.0002 | Ent: 0.2459
+[ 95.6%] Iter 9340/9765 | SPS:  2753 | Reward:  0.066 | Value:  0.737 | VLoss: 0.0188 | PLoss: -0.0019 | Ent: 0.2065
+[ 95.7%] Iter 9341/9765 | SPS:  2753 | Reward:  0.062 | Value:  0.655 | VLoss: 0.0326 | PLoss: -0.0021 | Ent: 0.2444
+[ 95.7%] Iter 9342/9765 | SPS:  2754 | Reward:  0.056 | Value:  0.696 | VLoss: 0.0214 | PLoss: -0.0010 | Ent: 0.2167
+[ 95.7%] Iter 9343/9765 | SPS:  2754 | Reward:  0.062 | Value:  0.670 | VLoss: 0.0226 | PLoss: -0.0007 | Ent: 0.2603
+[ 95.7%] Iter 9344/9765 | SPS:  2754 | Reward:  0.048 | Value:  0.707 | VLoss: 0.0191 | PLoss: -0.0028 | Ent: 0.3071
+[ 95.7%] Iter 9345/9765 | SPS:  2754 | Reward:  0.070 | Value:  0.775 | VLoss: 0.0225 | PLoss: -0.0019 | Ent: 0.2802
+[ 95.7%] Iter 9346/9765 | SPS:  2754 | Reward:  0.069 | Value:  0.715 | VLoss: 0.0213 | PLoss: -0.0031 | Ent: 0.2244
+[ 95.7%] Iter 9347/9765 | SPS:  2754 | Reward:  0.058 | Value:  0.681 | VLoss: 0.0190 | PLoss: -0.0023 | Ent: 0.2410
+[ 95.7%] Iter 9348/9765 | SPS:  2754 | Reward:  0.054 | Value:  0.721 | VLoss: 0.0188 | PLoss: -0.0014 | Ent: 0.1690
+[ 95.7%] Iter 9349/9765 | SPS:  2754 | Reward:  0.067 | Value:  0.761 | VLoss: 0.0236 | PLoss: -0.0018 | Ent: 0.1948
+[ 95.8%] Iter 9350/9765 | SPS:  2754 | Reward:  0.062 | Value:  0.754 | VLoss: 0.0212 | PLoss: -0.0024 | Ent: 0.1974
+[ 95.8%] Iter 9351/9765 | SPS:  2754 | Reward:  0.064 | Value:  0.739 | VLoss: 0.0245 | PLoss: -0.0004 | Ent: 0.2363
+[ 95.8%] Iter 9352/9765 | SPS:  2754 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0294 | PLoss: -0.0012 | Ent: 0.2237
+[ 95.8%] Iter 9353/9765 | SPS:  2754 | Reward:  0.062 | Value:  0.788 | VLoss: 0.0218 | PLoss: -0.0007 | Ent: 0.1922
+[ 95.8%] Iter 9354/9765 | SPS:  2755 | Reward:  0.061 | Value:  0.707 | VLoss: 0.0253 | PLoss: -0.0009 | Ent: 0.2102
+[ 95.8%] Iter 9355/9765 | SPS:  2755 | Reward:  0.055 | Value:  0.716 | VLoss: 0.0175 | PLoss: -0.0013 | Ent: 0.2443
+[ 95.8%] Iter 9356/9765 | SPS:  2755 | Reward:  0.059 | Value:  0.726 | VLoss: 0.0242 | PLoss: -0.0017 | Ent: 0.2582
+[ 95.8%] Iter 9357/9765 | SPS:  2755 | Reward:  0.068 | Value:  0.688 | VLoss: 0.0218 | PLoss: -0.0014 | Ent: 0.2296
+[ 95.8%] Iter 9358/9765 | SPS:  2755 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0174 | PLoss: -0.0010 | Ent: 0.2350
+[ 95.8%] Iter 9359/9765 | SPS:  2755 | Reward:  0.061 | Value:  0.729 | VLoss: 0.0224 | PLoss: -0.0030 | Ent: 0.2318
+[ 95.9%] Iter 9360/9765 | SPS:  2755 | Reward:  0.058 | Value:  0.703 | VLoss: 0.0207 | PLoss: -0.0019 | Ent: 0.2319
+[ 95.9%] Iter 9361/9765 | SPS:  2755 | Reward:  0.066 | Value:  0.645 | VLoss: 0.0265 | PLoss: -0.0015 | Ent: 0.2190
+[ 95.9%] Iter 9362/9765 | SPS:  2755 | Reward:  0.068 | Value:  0.717 | VLoss: 0.0209 | PLoss: -0.0012 | Ent: 0.2631
+[ 95.9%] Iter 9363/9765 | SPS:  2755 | Reward:  0.052 | Value:  0.755 | VLoss: 0.0197 | PLoss: -0.0020 | Ent: 0.2388
+[ 95.9%] Iter 9364/9765 | SPS:  2755 | Reward:  0.069 | Value:  0.799 | VLoss: 0.0310 | PLoss: -0.0013 | Ent: 0.2205
+[ 95.9%] Iter 9365/9765 | SPS:  2755 | Reward:  0.057 | Value:  0.679 | VLoss: 0.0252 | PLoss: -0.0022 | Ent: 0.1941
+[ 95.9%] Iter 9366/9765 | SPS:  2755 | Reward:  0.052 | Value:  0.703 | VLoss: 0.0193 | PLoss: -0.0016 | Ent: 0.2232
+[ 95.9%] Iter 9367/9765 | SPS:  2756 | Reward:  0.049 | Value:  0.687 | VLoss: 0.0126 | PLoss: -0.0016 | Ent: 0.2034
+[ 95.9%] Iter 9368/9765 | SPS:  2756 | Reward:  0.072 | Value:  0.700 | VLoss: 0.0225 | PLoss: -0.0004 | Ent: 0.2242
+[ 95.9%] Iter 9369/9765 | SPS:  2756 | Reward:  0.054 | Value:  0.741 | VLoss: 0.0259 | PLoss: -0.0006 | Ent: 0.2273
+[ 96.0%] Iter 9370/9765 | SPS:  2756 | Reward:  0.079 | Value:  0.738 | VLoss: 0.0201 | PLoss: -0.0035 | Ent: 0.2514
+[ 96.0%] Iter 9371/9765 | SPS:  2756 | Reward:  0.071 | Value:  0.739 | VLoss: 0.0168 | PLoss: -0.0021 | Ent: 0.2441
+[ 96.0%] Iter 9372/9765 | SPS:  2756 | Reward:  0.065 | Value:  0.755 | VLoss: 0.0221 | PLoss: -0.0020 | Ent: 0.2359
+[ 96.0%] Iter 9373/9765 | SPS:  2756 | Reward:  0.044 | Value:  0.716 | VLoss: 0.0310 | PLoss: -0.0014 | Ent: 0.2725
+[ 96.0%] Iter 9374/9765 | SPS:  2756 | Reward:  0.066 | Value:  0.685 | VLoss: 0.0212 | PLoss: -0.0002 | Ent: 0.2107
+[ 96.0%] Iter 9375/9765 | SPS:  2756 | Reward:  0.050 | Value:  0.730 | VLoss: 0.0230 | PLoss: -0.0025 | Ent: 0.1968
+[ 96.0%] Iter 9376/9765 | SPS:  2756 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0187 | PLoss: -0.0014 | Ent: 0.1893
+[ 96.0%] Iter 9377/9765 | SPS:  2756 | Reward:  0.060 | Value:  0.706 | VLoss: 0.0231 | PLoss: -0.0021 | Ent: 0.2056
+[ 96.0%] Iter 9378/9765 | SPS:  2756 | Reward:  0.058 | Value:  0.691 | VLoss: 0.0173 | PLoss: -0.0011 | Ent: 0.2676
+[ 96.0%] Iter 9379/9765 | SPS:  2757 | Reward:  0.054 | Value:  0.687 | VLoss: 0.0183 | PLoss: -0.0021 | Ent: 0.2250
+[ 96.1%] Iter 9380/9765 | SPS:  2757 | Reward:  0.071 | Value:  0.683 | VLoss: 0.0332 | PLoss: -0.0009 | Ent: 0.2521
+[ 96.1%] Iter 9381/9765 | SPS:  2757 | Reward:  0.058 | Value:  0.722 | VLoss: 0.0233 | PLoss: -0.0007 | Ent: 0.2203
+[ 96.1%] Iter 9382/9765 | SPS:  2757 | Reward:  0.067 | Value:  0.681 | VLoss: 0.0360 | PLoss: -0.0021 | Ent: 0.2495
+[ 96.1%] Iter 9383/9765 | SPS:  2757 | Reward:  0.054 | Value:  0.699 | VLoss: 0.0197 | PLoss: -0.0011 | Ent: 0.2527
+[ 96.1%] Iter 9384/9765 | SPS:  2757 | Reward:  0.059 | Value:  0.740 | VLoss: 0.0212 | PLoss: -0.0003 | Ent: 0.2018
+[ 96.1%] Iter 9385/9765 | SPS:  2757 | Reward:  0.063 | Value:  0.712 | VLoss: 0.0225 | PLoss: -0.0009 | Ent: 0.2540
+[ 96.1%] Iter 9386/9765 | SPS:  2757 | Reward:  0.051 | Value:  0.651 | VLoss: 0.0182 | PLoss: -0.0009 | Ent: 0.2648
+[ 96.1%] Iter 9387/9765 | SPS:  2757 | Reward:  0.056 | Value:  0.732 | VLoss: 0.0262 | PLoss: -0.0014 | Ent: 0.2124
+[ 96.1%] Iter 9388/9765 | SPS:  2757 | Reward:  0.049 | Value:  0.774 | VLoss: 0.0147 | PLoss: -0.0020 | Ent: 0.1988
+[ 96.1%] Iter 9389/9765 | SPS:  2757 | Reward:  0.072 | Value:  0.679 | VLoss: 0.0245 | PLoss: -0.0014 | Ent: 0.2736
+[ 96.2%] Iter 9390/9765 | SPS:  2757 | Reward:  0.053 | Value:  0.664 | VLoss: 0.0217 | PLoss: -0.0015 | Ent: 0.1886
+[ 96.2%] Iter 9391/9765 | SPS:  2757 | Reward:  0.065 | Value:  0.650 | VLoss: 0.0203 | PLoss: -0.0024 | Ent: 0.2498
+[ 96.2%] Iter 9392/9765 | SPS:  2758 | Reward:  0.058 | Value:  0.706 | VLoss: 0.0246 | PLoss: -0.0011 | Ent: 0.2183
+[ 96.2%] Iter 9393/9765 | SPS:  2758 | Reward:  0.049 | Value:  0.722 | VLoss: 0.0223 | PLoss: -0.0011 | Ent: 0.2729
+[ 96.2%] Iter 9394/9765 | SPS:  2758 | Reward:  0.072 | Value:  0.733 | VLoss: 0.0218 | PLoss: -0.0010 | Ent: 0.2069
+[ 96.2%] Iter 9395/9765 | SPS:  2758 | Reward:  0.058 | Value:  0.723 | VLoss: 0.0242 | PLoss: -0.0009 | Ent: 0.2302
+[ 96.2%] Iter 9396/9765 | SPS:  2758 | Reward:  0.068 | Value:  0.717 | VLoss: 0.0203 | PLoss: -0.0008 | Ent: 0.2429
+[ 96.2%] Iter 9397/9765 | SPS:  2758 | Reward:  0.053 | Value:  0.700 | VLoss: 0.0193 | PLoss: -0.0025 | Ent: 0.2422
+[ 96.2%] Iter 9398/9765 | SPS:  2758 | Reward:  0.064 | Value:  0.723 | VLoss: 0.0251 | PLoss: -0.0013 | Ent: 0.2869
+[ 96.3%] Iter 9399/9765 | SPS:  2758 | Reward:  0.056 | Value:  0.763 | VLoss: 0.0179 | PLoss: -0.0019 | Ent: 0.2338
+[ 96.3%] Iter 9400/9765 | SPS:  2758 | Reward:  0.051 | Value:  0.665 | VLoss: 0.0121 | PLoss: -0.0015 | Ent: 0.2121
+[ 96.3%] Iter 9401/9765 | SPS:  2758 | Reward:  0.047 | Value:  0.708 | VLoss: 0.0161 | PLoss: -0.0020 | Ent: 0.2116
+[ 96.3%] Iter 9402/9765 | SPS:  2758 | Reward:  0.068 | Value:  0.677 | VLoss: 0.0268 | PLoss: -0.0018 | Ent: 0.2871
+[ 96.3%] Iter 9403/9765 | SPS:  2758 | Reward:  0.058 | Value:  0.669 | VLoss: 0.0249 | PLoss: -0.0006 | Ent: 0.2075
+[ 96.3%] Iter 9404/9765 | SPS:  2759 | Reward:  0.058 | Value:  0.670 | VLoss: 0.0247 | PLoss: -0.0015 | Ent: 0.2527
+[ 96.3%] Iter 9405/9765 | SPS:  2759 | Reward:  0.069 | Value:  0.726 | VLoss: 0.0298 | PLoss: -0.0029 | Ent: 0.2784
+[ 96.3%] Iter 9406/9765 | SPS:  2759 | Reward:  0.043 | Value:  0.647 | VLoss: 0.0177 | PLoss: 0.0001 | Ent: 0.2033
+[ 96.3%] Iter 9407/9765 | SPS:  2759 | Reward:  0.035 | Value:  0.639 | VLoss: 0.0168 | PLoss: -0.0010 | Ent: 0.2526
+[ 96.3%] Iter 9408/9765 | SPS:  2759 | Reward:  0.042 | Value:  0.712 | VLoss: 0.0154 | PLoss: -0.0013 | Ent: 0.2710
+[ 96.4%] Iter 9409/9765 | SPS:  2759 | Reward:  0.059 | Value:  0.653 | VLoss: 0.0260 | PLoss: -0.0015 | Ent: 0.2919
+[ 96.4%] Iter 9410/9765 | SPS:  2759 | Reward:  0.065 | Value:  0.748 | VLoss: 0.0167 | PLoss: -0.0009 | Ent: 0.1888
+[ 96.4%] Iter 9411/9765 | SPS:  2759 | Reward:  0.070 | Value:  0.709 | VLoss: 0.0244 | PLoss: -0.0009 | Ent: 0.2324
+[ 96.4%] Iter 9412/9765 | SPS:  2759 | Reward:  0.056 | Value:  0.744 | VLoss: 0.0183 | PLoss: -0.0007 | Ent: 0.2114
+[ 96.4%] Iter 9413/9765 | SPS:  2759 | Reward:  0.065 | Value:  0.726 | VLoss: 0.0241 | PLoss: -0.0001 | Ent: 0.2766
+[ 96.4%] Iter 9414/9765 | SPS:  2759 | Reward:  0.058 | Value:  0.704 | VLoss: 0.0195 | PLoss: -0.0017 | Ent: 0.2234
+[ 96.4%] Iter 9415/9765 | SPS:  2759 | Reward:  0.050 | Value:  0.774 | VLoss: 0.0151 | PLoss: -0.0014 | Ent: 0.2326
+[ 96.4%] Iter 9416/9765 | SPS:  2760 | Reward:  0.067 | Value:  0.782 | VLoss: 0.0224 | PLoss: -0.0012 | Ent: 0.2542
+[ 96.4%] Iter 9417/9765 | SPS:  2760 | Reward:  0.055 | Value:  0.738 | VLoss: 0.0255 | PLoss: -0.0006 | Ent: 0.2449
+[ 96.4%] Iter 9418/9765 | SPS:  2760 | Reward:  0.052 | Value:  0.732 | VLoss: 0.0204 | PLoss: -0.0010 | Ent: 0.2101
+[ 96.5%] Iter 9419/9765 | SPS:  2760 | Reward:  0.064 | Value:  0.721 | VLoss: 0.0188 | PLoss: -0.0009 | Ent: 0.2254
+[ 96.5%] Iter 9420/9765 | SPS:  2760 | Reward:  0.041 | Value:  0.725 | VLoss: 0.0232 | PLoss: -0.0011 | Ent: 0.2730
+[ 96.5%] Iter 9421/9765 | SPS:  2760 | Reward:  0.074 | Value:  0.714 | VLoss: 0.0218 | PLoss: -0.0026 | Ent: 0.2429
+[ 96.5%] Iter 9422/9765 | SPS:  2760 | Reward:  0.055 | Value:  0.668 | VLoss: 0.0251 | PLoss: -0.0003 | Ent: 0.2335
+[ 96.5%] Iter 9423/9765 | SPS:  2760 | Reward:  0.051 | Value:  0.669 | VLoss: 0.0175 | PLoss: -0.0024 | Ent: 0.2588
+[ 96.5%] Iter 9424/9765 | SPS:  2760 | Reward:  0.057 | Value:  0.647 | VLoss: 0.0193 | PLoss: -0.0002 | Ent: 0.2507
+[ 96.5%] Iter 9425/9765 | SPS:  2760 | Reward:  0.054 | Value:  0.665 | VLoss: 0.0208 | PLoss: -0.0012 | Ent: 0.2240
+[ 96.5%] Iter 9426/9765 | SPS:  2760 | Reward:  0.072 | Value:  0.686 | VLoss: 0.0274 | PLoss: -0.0013 | Ent: 0.2151
+[ 96.5%] Iter 9427/9765 | SPS:  2760 | Reward:  0.061 | Value:  0.719 | VLoss: 0.0250 | PLoss: -0.0010 | Ent: 0.2388
+[ 96.5%] Iter 9428/9765 | SPS:  2761 | Reward:  0.050 | Value:  0.729 | VLoss: 0.0230 | PLoss: -0.0019 | Ent: 0.2493
+[ 96.6%] Iter 9429/9765 | SPS:  2761 | Reward:  0.059 | Value:  0.774 | VLoss: 0.0236 | PLoss: -0.0026 | Ent: 0.2420
+[ 96.6%] Iter 9430/9765 | SPS:  2761 | Reward:  0.052 | Value:  0.747 | VLoss: 0.0163 | PLoss: -0.0040 | Ent: 0.2679
+[ 96.6%] Iter 9431/9765 | SPS:  2761 | Reward:  0.071 | Value:  0.707 | VLoss: 0.0321 | PLoss: -0.0012 | Ent: 0.2449
+[ 96.6%] Iter 9432/9765 | SPS:  2761 | Reward:  0.055 | Value:  0.754 | VLoss: 0.0154 | PLoss: -0.0022 | Ent: 0.2209
+[ 96.6%] Iter 9433/9765 | SPS:  2761 | Reward:  0.062 | Value:  0.790 | VLoss: 0.0157 | PLoss: -0.0008 | Ent: 0.2248
+[ 96.6%] Iter 9434/9765 | SPS:  2761 | Reward:  0.068 | Value:  0.753 | VLoss: 0.0217 | PLoss: -0.0003 | Ent: 0.1997
+[ 96.6%] Iter 9435/9765 | SPS:  2761 | Reward:  0.047 | Value:  0.752 | VLoss: 0.0193 | PLoss: -0.0016 | Ent: 0.2414
+[ 96.6%] Iter 9436/9765 | SPS:  2761 | Reward:  0.071 | Value:  0.688 | VLoss: 0.0174 | PLoss: 0.0001 | Ent: 0.2236
+[ 96.6%] Iter 9437/9765 | SPS:  2761 | Reward:  0.073 | Value:  0.704 | VLoss: 0.0257 | PLoss: -0.0013 | Ent: 0.2264
+[ 96.7%] Iter 9438/9765 | SPS:  2761 | Reward:  0.064 | Value:  0.727 | VLoss: 0.0184 | PLoss: -0.0001 | Ent: 0.1770
+[ 96.7%] Iter 9439/9765 | SPS:  2761 | Reward:  0.059 | Value:  0.733 | VLoss: 0.0223 | PLoss: -0.0019 | Ent: 0.2495
+[ 96.7%] Iter 9440/9765 | SPS:  2761 | Reward:  0.057 | Value:  0.778 | VLoss: 0.0207 | PLoss: -0.0005 | Ent: 0.2088
+[ 96.7%] Iter 9441/9765 | SPS:  2762 | Reward:  0.062 | Value:  0.715 | VLoss: 0.0204 | PLoss: -0.0003 | Ent: 0.2372
+[ 96.7%] Iter 9442/9765 | SPS:  2762 | Reward:  0.063 | Value:  0.701 | VLoss: 0.0261 | PLoss: -0.0009 | Ent: 0.1903
+[ 96.7%] Iter 9443/9765 | SPS:  2762 | Reward:  0.070 | Value:  0.753 | VLoss: 0.0148 | PLoss: -0.0006 | Ent: 0.2034
+[ 96.7%] Iter 9444/9765 | SPS:  2762 | Reward:  0.061 | Value:  0.765 | VLoss: 0.0165 | PLoss: -0.0009 | Ent: 0.2087
+[ 96.7%] Iter 9445/9765 | SPS:  2762 | Reward:  0.053 | Value:  0.716 | VLoss: 0.0193 | PLoss: -0.0010 | Ent: 0.2536
+[ 96.7%] Iter 9446/9765 | SPS:  2762 | Reward:  0.069 | Value:  0.724 | VLoss: 0.0175 | PLoss: -0.0019 | Ent: 0.2343
+[ 96.7%] Iter 9447/9765 | SPS:  2762 | Reward:  0.057 | Value:  0.719 | VLoss: 0.0218 | PLoss: -0.0010 | Ent: 0.2501
+[ 96.8%] Iter 9448/9765 | SPS:  2762 | Reward:  0.060 | Value:  0.702 | VLoss: 0.0210 | PLoss: -0.0017 | Ent: 0.3101
+[ 96.8%] Iter 9449/9765 | SPS:  2762 | Reward:  0.055 | Value:  0.668 | VLoss: 0.0180 | PLoss: -0.0014 | Ent: 0.2668
+[ 96.8%] Iter 9450/9765 | SPS:  2762 | Reward:  0.062 | Value:  0.701 | VLoss: 0.0168 | PLoss: -0.0014 | Ent: 0.2249
+[ 96.8%] Iter 9451/9765 | SPS:  2762 | Reward:  0.056 | Value:  0.712 | VLoss: 0.0204 | PLoss: -0.0011 | Ent: 0.2144
+[ 96.8%] Iter 9452/9765 | SPS:  2762 | Reward:  0.050 | Value:  0.701 | VLoss: 0.0170 | PLoss: -0.0015 | Ent: 0.2317
+[ 96.8%] Iter 9453/9765 | SPS:  2763 | Reward:  0.043 | Value:  0.695 | VLoss: 0.0171 | PLoss: 0.0001 | Ent: 0.2460
+[ 96.8%] Iter 9454/9765 | SPS:  2763 | Reward:  0.056 | Value:  0.684 | VLoss: 0.0203 | PLoss: -0.0021 | Ent: 0.2862
+[ 96.8%] Iter 9455/9765 | SPS:  2763 | Reward:  0.059 | Value:  0.754 | VLoss: 0.0210 | PLoss: -0.0002 | Ent: 0.2446
+[ 96.8%] Iter 9456/9765 | SPS:  2763 | Reward:  0.066 | Value:  0.737 | VLoss: 0.0212 | PLoss: -0.0005 | Ent: 0.2325
+[ 96.8%] Iter 9457/9765 | SPS:  2763 | Reward:  0.053 | Value:  0.694 | VLoss: 0.0180 | PLoss: -0.0010 | Ent: 0.1974
+[ 96.9%] Iter 9458/9765 | SPS:  2763 | Reward:  0.058 | Value:  0.687 | VLoss: 0.0284 | PLoss: -0.0013 | Ent: 0.2178
+[ 96.9%] Iter 9459/9765 | SPS:  2763 | Reward:  0.059 | Value:  0.720 | VLoss: 0.0236 | PLoss: -0.0014 | Ent: 0.1818
+[ 96.9%] Iter 9460/9765 | SPS:  2763 | Reward:  0.055 | Value:  0.723 | VLoss: 0.0217 | PLoss: -0.0009 | Ent: 0.2676
+[ 96.9%] Iter 9461/9765 | SPS:  2763 | Reward:  0.059 | Value:  0.670 | VLoss: 0.0202 | PLoss: -0.0017 | Ent: 0.2047
+[ 96.9%] Iter 9462/9765 | SPS:  2763 | Reward:  0.060 | Value:  0.640 | VLoss: 0.0202 | PLoss: -0.0007 | Ent: 0.2338
+[ 96.9%] Iter 9463/9765 | SPS:  2763 | Reward:  0.051 | Value:  0.750 | VLoss: 0.0142 | PLoss: -0.0014 | Ent: 0.1801
+[ 96.9%] Iter 9464/9765 | SPS:  2763 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0251 | PLoss: -0.0016 | Ent: 0.2592
+[ 96.9%] Iter 9465/9765 | SPS:  2763 | Reward:  0.057 | Value:  0.700 | VLoss: 0.0280 | PLoss: -0.0003 | Ent: 0.2070
+[ 96.9%] Iter 9466/9765 | SPS:  2764 | Reward:  0.063 | Value:  0.722 | VLoss: 0.0210 | PLoss: -0.0009 | Ent: 0.2444
+[ 96.9%] Iter 9467/9765 | SPS:  2764 | Reward:  0.064 | Value:  0.715 | VLoss: 0.0287 | PLoss: -0.0003 | Ent: 0.2378
+[ 97.0%] Iter 9468/9765 | SPS:  2764 | Reward:  0.064 | Value:  0.670 | VLoss: 0.0263 | PLoss: -0.0015 | Ent: 0.2146
+[ 97.0%] Iter 9469/9765 | SPS:  2764 | Reward:  0.057 | Value:  0.687 | VLoss: 0.0242 | PLoss: -0.0006 | Ent: 0.2055
+[ 97.0%] Iter 9470/9765 | SPS:  2764 | Reward:  0.058 | Value:  0.724 | VLoss: 0.0216 | PLoss: -0.0008 | Ent: 0.2409
+[ 97.0%] Iter 9471/9765 | SPS:  2764 | Reward:  0.063 | Value:  0.701 | VLoss: 0.0274 | PLoss: -0.0011 | Ent: 0.2150
+[ 97.0%] Iter 9472/9765 | SPS:  2764 | Reward:  0.061 | Value:  0.673 | VLoss: 0.0260 | PLoss: -0.0020 | Ent: 0.2038
+[ 97.0%] Iter 9473/9765 | SPS:  2764 | Reward:  0.052 | Value:  0.728 | VLoss: 0.0158 | PLoss: -0.0009 | Ent: 0.2040
+[ 97.0%] Iter 9474/9765 | SPS:  2764 | Reward:  0.067 | Value:  0.720 | VLoss: 0.0209 | PLoss: -0.0009 | Ent: 0.2437
+[ 97.0%] Iter 9475/9765 | SPS:  2764 | Reward:  0.064 | Value:  0.729 | VLoss: 0.0140 | PLoss: -0.0004 | Ent: 0.2195
+[ 97.0%] Iter 9476/9765 | SPS:  2764 | Reward:  0.063 | Value:  0.720 | VLoss: 0.0222 | PLoss: -0.0010 | Ent: 0.2779
+[ 97.1%] Iter 9477/9765 | SPS:  2764 | Reward:  0.062 | Value:  0.723 | VLoss: 0.0160 | PLoss: -0.0029 | Ent: 0.2430
+[ 97.1%] Iter 9478/9765 | SPS:  2765 | Reward:  0.060 | Value:  0.731 | VLoss: 0.0237 | PLoss: -0.0015 | Ent: 0.2693
+[ 97.1%] Iter 9479/9765 | SPS:  2765 | Reward:  0.058 | Value:  0.776 | VLoss: 0.0192 | PLoss: -0.0009 | Ent: 0.1932
+[ 97.1%] Iter 9480/9765 | SPS:  2765 | Reward:  0.058 | Value:  0.764 | VLoss: 0.0165 | PLoss: -0.0009 | Ent: 0.1950
+[ 97.1%] Iter 9481/9765 | SPS:  2765 | Reward:  0.061 | Value:  0.675 | VLoss: 0.0236 | PLoss: -0.0009 | Ent: 0.2241
+[ 97.1%] Iter 9482/9765 | SPS:  2765 | Reward:  0.059 | Value:  0.693 | VLoss: 0.0198 | PLoss: -0.0026 | Ent: 0.2484
+[ 97.1%] Iter 9483/9765 | SPS:  2765 | Reward:  0.049 | Value:  0.643 | VLoss: 0.0209 | PLoss: -0.0010 | Ent: 0.3001
+[ 97.1%] Iter 9484/9765 | SPS:  2765 | Reward:  0.064 | Value:  0.716 | VLoss: 0.0258 | PLoss: -0.0001 | Ent: 0.2078
+[ 97.1%] Iter 9485/9765 | SPS:  2765 | Reward:  0.057 | Value:  0.707 | VLoss: 0.0234 | PLoss: -0.0016 | Ent: 0.2510
+[ 97.1%] Iter 9486/9765 | SPS:  2765 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0298 | PLoss: -0.0013 | Ent: 0.3008
+[ 97.2%] Iter 9487/9765 | SPS:  2765 | Reward:  0.057 | Value:  0.668 | VLoss: 0.0229 | PLoss: -0.0006 | Ent: 0.2825
+[ 97.2%] Iter 9488/9765 | SPS:  2765 | Reward:  0.056 | Value:  0.707 | VLoss: 0.0191 | PLoss: -0.0008 | Ent: 0.2341
+[ 97.2%] Iter 9489/9765 | SPS:  2765 | Reward:  0.055 | Value:  0.705 | VLoss: 0.0293 | PLoss: -0.0003 | Ent: 0.2381
+[ 97.2%] Iter 9490/9765 | SPS:  2765 | Reward:  0.060 | Value:  0.759 | VLoss: 0.0145 | PLoss: -0.0010 | Ent: 0.2209
+[ 97.2%] Iter 9491/9765 | SPS:  2766 | Reward:  0.062 | Value:  0.774 | VLoss: 0.0188 | PLoss: -0.0004 | Ent: 0.2720
+[ 97.2%] Iter 9492/9765 | SPS:  2766 | Reward:  0.055 | Value:  0.799 | VLoss: 0.0131 | PLoss: -0.0004 | Ent: 0.2094
+[ 97.2%] Iter 9493/9765 | SPS:  2766 | Reward:  0.053 | Value:  0.710 | VLoss: 0.0250 | PLoss: -0.0007 | Ent: 0.2493
+[ 97.2%] Iter 9494/9765 | SPS:  2766 | Reward:  0.063 | Value:  0.780 | VLoss: 0.0212 | PLoss: -0.0012 | Ent: 0.2304
+[ 97.2%] Iter 9495/9765 | SPS:  2766 | Reward:  0.060 | Value:  0.728 | VLoss: 0.0239 | PLoss: -0.0009 | Ent: 0.1989
+[ 97.2%] Iter 9496/9765 | SPS:  2766 | Reward:  0.063 | Value:  0.741 | VLoss: 0.0150 | PLoss: -0.0011 | Ent: 0.2061
+[ 97.3%] Iter 9497/9765 | SPS:  2766 | Reward:  0.063 | Value:  0.753 | VLoss: 0.0241 | PLoss: 0.0001 | Ent: 0.1618
+[ 97.3%] Iter 9498/9765 | SPS:  2766 | Reward:  0.062 | Value:  0.747 | VLoss: 0.0151 | PLoss: -0.0003 | Ent: 0.2435
+[ 97.3%] Iter 9499/9765 | SPS:  2766 | Reward:  0.065 | Value:  0.690 | VLoss: 0.0237 | PLoss: -0.0008 | Ent: 0.2384
+[ 97.3%] Iter 9500/9765 | SPS:  2766 | Reward:  0.058 | Value:  0.709 | VLoss: 0.0220 | PLoss: -0.0002 | Ent: 0.2407
+[ 97.3%] Iter 9501/9765 | SPS:  2766 | Reward:  0.058 | Value:  0.734 | VLoss: 0.0179 | PLoss: -0.0015 | Ent: 0.2019
+[ 97.3%] Iter 9502/9765 | SPS:  2766 | Reward:  0.046 | Value:  0.687 | VLoss: 0.0231 | PLoss: -0.0008 | Ent: 0.2719
+[ 97.3%] Iter 9503/9765 | SPS:  2767 | Reward:  0.046 | Value:  0.700 | VLoss: 0.0169 | PLoss: -0.0010 | Ent: 0.2410
+[ 97.3%] Iter 9504/9765 | SPS:  2767 | Reward:  0.064 | Value:  0.705 | VLoss: 0.0251 | PLoss: -0.0035 | Ent: 0.2239
+[ 97.3%] Iter 9505/9765 | SPS:  2767 | Reward:  0.054 | Value:  0.693 | VLoss: 0.0233 | PLoss: -0.0005 | Ent: 0.2671
+[ 97.3%] Iter 9506/9765 | SPS:  2767 | Reward:  0.060 | Value:  0.692 | VLoss: 0.0250 | PLoss: -0.0021 | Ent: 0.2757
+[ 97.4%] Iter 9507/9765 | SPS:  2767 | Reward:  0.062 | Value:  0.698 | VLoss: 0.0225 | PLoss: -0.0005 | Ent: 0.1955
+[ 97.4%] Iter 9508/9765 | SPS:  2767 | Reward:  0.061 | Value:  0.644 | VLoss: 0.0223 | PLoss: -0.0004 | Ent: 0.2361
+[ 97.4%] Iter 9509/9765 | SPS:  2767 | Reward:  0.052 | Value:  0.672 | VLoss: 0.0248 | PLoss: -0.0009 | Ent: 0.2068
+[ 97.4%] Iter 9510/9765 | SPS:  2767 | Reward:  0.056 | Value:  0.721 | VLoss: 0.0211 | PLoss: -0.0003 | Ent: 0.2347
+[ 97.4%] Iter 9511/9765 | SPS:  2767 | Reward:  0.046 | Value:  0.744 | VLoss: 0.0113 | PLoss: -0.0008 | Ent: 0.2390
+[ 97.4%] Iter 9512/9765 | SPS:  2767 | Reward:  0.053 | Value:  0.752 | VLoss: 0.0147 | PLoss: -0.0008 | Ent: 0.2378
+[ 97.4%] Iter 9513/9765 | SPS:  2767 | Reward:  0.064 | Value:  0.706 | VLoss: 0.0279 | PLoss: -0.0011 | Ent: 0.2002
+[ 97.4%] Iter 9514/9765 | SPS:  2767 | Reward:  0.051 | Value:  0.728 | VLoss: 0.0198 | PLoss: -0.0010 | Ent: 0.2165
+[ 97.4%] Iter 9515/9765 | SPS:  2767 | Reward:  0.063 | Value:  0.763 | VLoss: 0.0196 | PLoss: -0.0022 | Ent: 0.2087
+[ 97.5%] Iter 9516/9765 | SPS:  2768 | Reward:  0.062 | Value:  0.729 | VLoss: 0.0255 | PLoss: -0.0001 | Ent: 0.3001
+[ 97.5%] Iter 9517/9765 | SPS:  2768 | Reward:  0.061 | Value:  0.723 | VLoss: 0.0261 | PLoss: -0.0005 | Ent: 0.2029
+[ 97.5%] Iter 9518/9765 | SPS:  2768 | Reward:  0.058 | Value:  0.723 | VLoss: 0.0293 | PLoss: -0.0008 | Ent: 0.2496
+[ 97.5%] Iter 9519/9765 | SPS:  2768 | Reward:  0.059 | Value:  0.753 | VLoss: 0.0180 | PLoss: -0.0004 | Ent: 0.2315
+[ 97.5%] Iter 9520/9765 | SPS:  2768 | Reward:  0.050 | Value:  0.769 | VLoss: 0.0142 | PLoss: -0.0004 | Ent: 0.2012
+[ 97.5%] Iter 9521/9765 | SPS:  2768 | Reward:  0.076 | Value:  0.746 | VLoss: 0.0204 | PLoss: -0.0005 | Ent: 0.2207
+[ 97.5%] Iter 9522/9765 | SPS:  2768 | Reward:  0.072 | Value:  0.707 | VLoss: 0.0231 | PLoss: -0.0013 | Ent: 0.1948
+[ 97.5%] Iter 9523/9765 | SPS:  2768 | Reward:  0.059 | Value:  0.702 | VLoss: 0.0242 | PLoss: -0.0003 | Ent: 0.1912
+[ 97.5%] Iter 9524/9765 | SPS:  2768 | Reward:  0.057 | Value:  0.752 | VLoss: 0.0281 | PLoss: -0.0009 | Ent: 0.2624
+[ 97.5%] Iter 9525/9765 | SPS:  2768 | Reward:  0.061 | Value:  0.687 | VLoss: 0.0190 | PLoss: -0.0011 | Ent: 0.2179
+[ 97.6%] Iter 9526/9765 | SPS:  2768 | Reward:  0.058 | Value:  0.715 | VLoss: 0.0254 | PLoss: -0.0028 | Ent: 0.2387
+[ 97.6%] Iter 9527/9765 | SPS:  2768 | Reward:  0.053 | Value:  0.715 | VLoss: 0.0137 | PLoss: -0.0002 | Ent: 0.2086
+[ 97.6%] Iter 9528/9765 | SPS:  2768 | Reward:  0.058 | Value:  0.655 | VLoss: 0.0153 | PLoss: -0.0009 | Ent: 0.2010
+[ 97.6%] Iter 9529/9765 | SPS:  2769 | Reward:  0.065 | Value:  0.716 | VLoss: 0.0229 | PLoss: -0.0008 | Ent: 0.2469
+[ 97.6%] Iter 9530/9765 | SPS:  2769 | Reward:  0.048 | Value:  0.629 | VLoss: 0.0334 | PLoss: -0.0009 | Ent: 0.2493
+[ 97.6%] Iter 9531/9765 | SPS:  2769 | Reward:  0.057 | Value:  0.757 | VLoss: 0.0145 | PLoss: -0.0013 | Ent: 0.2117
+[ 97.6%] Iter 9532/9765 | SPS:  2769 | Reward:  0.057 | Value:  0.768 | VLoss: 0.0176 | PLoss: -0.0007 | Ent: 0.2041
+[ 97.6%] Iter 9533/9765 | SPS:  2769 | Reward:  0.072 | Value:  0.787 | VLoss: 0.0117 | PLoss: 0.0001 | Ent: 0.2479
+[ 97.6%] Iter 9534/9765 | SPS:  2769 | Reward:  0.061 | Value:  0.780 | VLoss: 0.0190 | PLoss: -0.0006 | Ent: 0.2165
+[ 97.6%] Iter 9535/9765 | SPS:  2769 | Reward:  0.055 | Value:  0.770 | VLoss: 0.0210 | PLoss: -0.0007 | Ent: 0.2480
+[ 97.7%] Iter 9536/9765 | SPS:  2769 | Reward:  0.062 | Value:  0.705 | VLoss: 0.0176 | PLoss: -0.0009 | Ent: 0.2453
+[ 97.7%] Iter 9537/9765 | SPS:  2769 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0207 | PLoss: -0.0003 | Ent: 0.2769
+[ 97.7%] Iter 9538/9765 | SPS:  2769 | Reward:  0.068 | Value:  0.667 | VLoss: 0.0277 | PLoss: -0.0004 | Ent: 0.2445
+[ 97.7%] Iter 9539/9765 | SPS:  2769 | Reward:  0.055 | Value:  0.756 | VLoss: 0.0170 | PLoss: -0.0001 | Ent: 0.2575
+[ 97.7%] Iter 9540/9765 | SPS:  2769 | Reward:  0.052 | Value:  0.705 | VLoss: 0.0170 | PLoss: -0.0011 | Ent: 0.2785
+[ 97.7%] Iter 9541/9765 | SPS:  2770 | Reward:  0.065 | Value:  0.746 | VLoss: 0.0174 | PLoss: -0.0004 | Ent: 0.2178
+[ 97.7%] Iter 9542/9765 | SPS:  2770 | Reward:  0.058 | Value:  0.779 | VLoss: 0.0195 | PLoss: -0.0011 | Ent: 0.2435
+[ 97.7%] Iter 9543/9765 | SPS:  2770 | Reward:  0.065 | Value:  0.808 | VLoss: 0.0161 | PLoss: -0.0003 | Ent: 0.2292
+[ 97.7%] Iter 9544/9765 | SPS:  2770 | Reward:  0.059 | Value:  0.717 | VLoss: 0.0198 | PLoss: -0.0002 | Ent: 0.2348
+[ 97.7%] Iter 9545/9765 | SPS:  2770 | Reward:  0.065 | Value:  0.738 | VLoss: 0.0211 | PLoss: -0.0003 | Ent: 0.2384
+[ 97.8%] Iter 9546/9765 | SPS:  2770 | Reward:  0.078 | Value:  0.754 | VLoss: 0.0216 | PLoss: -0.0011 | Ent: 0.2442
+[ 97.8%] Iter 9547/9765 | SPS:  2770 | Reward:  0.065 | Value:  0.725 | VLoss: 0.0281 | PLoss: 0.0006 | Ent: 0.2207
+[ 97.8%] Iter 9548/9765 | SPS:  2770 | Reward:  0.052 | Value:  0.775 | VLoss: 0.0238 | PLoss: -0.0002 | Ent: 0.2160
+[ 97.8%] Iter 9549/9765 | SPS:  2770 | Reward:  0.059 | Value:  0.733 | VLoss: 0.0182 | PLoss: -0.0009 | Ent: 0.2768
+[ 97.8%] Iter 9550/9765 | SPS:  2770 | Reward:  0.058 | Value:  0.684 | VLoss: 0.0202 | PLoss: -0.0004 | Ent: 0.3305
+[ 97.8%] Iter 9551/9765 | SPS:  2770 | Reward:  0.058 | Value:  0.690 | VLoss: 0.0181 | PLoss: -0.0006 | Ent: 0.2205
+[ 97.8%] Iter 9552/9765 | SPS:  2770 | Reward:  0.057 | Value:  0.715 | VLoss: 0.0225 | PLoss: -0.0011 | Ent: 0.2271
+[ 97.8%] Iter 9553/9765 | SPS:  2770 | Reward:  0.063 | Value:  0.742 | VLoss: 0.0209 | PLoss: -0.0020 | Ent: 0.2338
+[ 97.8%] Iter 9554/9765 | SPS:  2771 | Reward:  0.061 | Value:  0.684 | VLoss: 0.0266 | PLoss: -0.0011 | Ent: 0.2481
+[ 97.8%] Iter 9555/9765 | SPS:  2771 | Reward:  0.066 | Value:  0.734 | VLoss: 0.0157 | PLoss: -0.0006 | Ent: 0.2284
+[ 97.9%] Iter 9556/9765 | SPS:  2771 | Reward:  0.044 | Value:  0.634 | VLoss: 0.0199 | PLoss: -0.0007 | Ent: 0.2651
+[ 97.9%] Iter 9557/9765 | SPS:  2771 | Reward:  0.059 | Value:  0.699 | VLoss: 0.0237 | PLoss: -0.0003 | Ent: 0.1986
+[ 97.9%] Iter 9558/9765 | SPS:  2771 | Reward:  0.063 | Value:  0.723 | VLoss: 0.0227 | PLoss: -0.0008 | Ent: 0.2695
+[ 97.9%] Iter 9559/9765 | SPS:  2771 | Reward:  0.057 | Value:  0.705 | VLoss: 0.0204 | PLoss: -0.0008 | Ent: 0.2296
+[ 97.9%] Iter 9560/9765 | SPS:  2771 | Reward:  0.059 | Value:  0.705 | VLoss: 0.0145 | PLoss: -0.0012 | Ent: 0.2579
+[ 97.9%] Iter 9561/9765 | SPS:  2771 | Reward:  0.055 | Value:  0.675 | VLoss: 0.0201 | PLoss: -0.0007 | Ent: 0.2720
+[ 97.9%] Iter 9562/9765 | SPS:  2771 | Reward:  0.052 | Value:  0.810 | VLoss: 0.0243 | PLoss: -0.0005 | Ent: 0.2409
+[ 97.9%] Iter 9563/9765 | SPS:  2771 | Reward:  0.065 | Value:  0.738 | VLoss: 0.0201 | PLoss: -0.0001 | Ent: 0.2549
+[ 97.9%] Iter 9564/9765 | SPS:  2771 | Reward:  0.066 | Value:  0.720 | VLoss: 0.0216 | PLoss: -0.0007 | Ent: 0.2334
+[ 98.0%] Iter 9565/9765 | SPS:  2771 | Reward:  0.056 | Value:  0.724 | VLoss: 0.0174 | PLoss: -0.0009 | Ent: 0.2588
+[ 98.0%] Iter 9566/9765 | SPS:  2772 | Reward:  0.060 | Value:  0.729 | VLoss: 0.0224 | PLoss: -0.0009 | Ent: 0.2277
+[ 98.0%] Iter 9567/9765 | SPS:  2772 | Reward:  0.065 | Value:  0.729 | VLoss: 0.0233 | PLoss: -0.0005 | Ent: 0.2058
+[ 98.0%] Iter 9568/9765 | SPS:  2772 | Reward:  0.063 | Value:  0.694 | VLoss: 0.0226 | PLoss: 0.0003 | Ent: 0.2316
+[ 98.0%] Iter 9569/9765 | SPS:  2772 | Reward:  0.052 | Value:  0.660 | VLoss: 0.0245 | PLoss: -0.0023 | Ent: 0.2901
+[ 98.0%] Iter 9570/9765 | SPS:  2772 | Reward:  0.068 | Value:  0.723 | VLoss: 0.0279 | PLoss: -0.0005 | Ent: 0.2339
+[ 98.0%] Iter 9571/9765 | SPS:  2772 | Reward:  0.055 | Value:  0.702 | VLoss: 0.0198 | PLoss: -0.0001 | Ent: 0.1939
+[ 98.0%] Iter 9572/9765 | SPS:  2772 | Reward:  0.064 | Value:  0.720 | VLoss: 0.0179 | PLoss: -0.0005 | Ent: 0.2303
+[ 98.0%] Iter 9573/9765 | SPS:  2772 | Reward:  0.074 | Value:  0.772 | VLoss: 0.0267 | PLoss: -0.0010 | Ent: 0.2347
+[ 98.0%] Iter 9574/9765 | SPS:  2772 | Reward:  0.073 | Value:  0.702 | VLoss: 0.0257 | PLoss: -0.0003 | Ent: 0.2221
+[ 98.1%] Iter 9575/9765 | SPS:  2772 | Reward:  0.071 | Value:  0.710 | VLoss: 0.0186 | PLoss: -0.0009 | Ent: 0.2537
+[ 98.1%] Iter 9576/9765 | SPS:  2772 | Reward:  0.064 | Value:  0.758 | VLoss: 0.0235 | PLoss: -0.0004 | Ent: 0.2195
+[ 98.1%] Iter 9577/9765 | SPS:  2772 | Reward:  0.063 | Value:  0.790 | VLoss: 0.0114 | PLoss: -0.0011 | Ent: 0.2130
+[ 98.1%] Iter 9578/9765 | SPS:  2772 | Reward:  0.063 | Value:  0.697 | VLoss: 0.0229 | PLoss: -0.0008 | Ent: 0.2402
+[ 98.1%] Iter 9579/9765 | SPS:  2773 | Reward:  0.057 | Value:  0.726 | VLoss: 0.0162 | PLoss: -0.0002 | Ent: 0.2528
+[ 98.1%] Iter 9580/9765 | SPS:  2773 | Reward:  0.065 | Value:  0.752 | VLoss: 0.0327 | PLoss: -0.0017 | Ent: 0.2496
+[ 98.1%] Iter 9581/9765 | SPS:  2773 | Reward:  0.057 | Value:  0.764 | VLoss: 0.0150 | PLoss: -0.0003 | Ent: 0.1595
+[ 98.1%] Iter 9582/9765 | SPS:  2773 | Reward:  0.062 | Value:  0.713 | VLoss: 0.0259 | PLoss: -0.0005 | Ent: 0.2349
+[ 98.1%] Iter 9583/9765 | SPS:  2773 | Reward:  0.051 | Value:  0.683 | VLoss: 0.0177 | PLoss: -0.0004 | Ent: 0.2491
+[ 98.1%] Iter 9584/9765 | SPS:  2773 | Reward:  0.050 | Value:  0.630 | VLoss: 0.0191 | PLoss: -0.0001 | Ent: 0.2715
+[ 98.2%] Iter 9585/9765 | SPS:  2773 | Reward:  0.040 | Value:  0.684 | VLoss: 0.0245 | PLoss: -0.0009 | Ent: 0.2185
+[ 98.2%] Iter 9586/9765 | SPS:  2773 | Reward:  0.051 | Value:  0.657 | VLoss: 0.0195 | PLoss: -0.0005 | Ent: 0.2272
+[ 98.2%] Iter 9587/9765 | SPS:  2773 | Reward:  0.056 | Value:  0.711 | VLoss: 0.0272 | PLoss: -0.0006 | Ent: 0.2303
+[ 98.2%] Iter 9588/9765 | SPS:  2773 | Reward:  0.059 | Value:  0.657 | VLoss: 0.0241 | PLoss: -0.0011 | Ent: 0.2432
+[ 98.2%] Iter 9589/9765 | SPS:  2773 | Reward:  0.053 | Value:  0.708 | VLoss: 0.0155 | PLoss: -0.0002 | Ent: 0.1563
+[ 98.2%] Iter 9590/9765 | SPS:  2773 | Reward:  0.057 | Value:  0.703 | VLoss: 0.0170 | PLoss: -0.0007 | Ent: 0.2234
+[ 98.2%] Iter 9591/9765 | SPS:  2774 | Reward:  0.058 | Value:  0.714 | VLoss: 0.0217 | PLoss: -0.0006 | Ent: 0.2689
+[ 98.2%] Iter 9592/9765 | SPS:  2774 | Reward:  0.050 | Value:  0.699 | VLoss: 0.0339 | PLoss: -0.0001 | Ent: 0.2235
+[ 98.2%] Iter 9593/9765 | SPS:  2774 | Reward:  0.062 | Value:  0.693 | VLoss: 0.0201 | PLoss: -0.0006 | Ent: 0.2266
+[ 98.2%] Iter 9594/9765 | SPS:  2774 | Reward:  0.066 | Value:  0.692 | VLoss: 0.0163 | PLoss: -0.0004 | Ent: 0.2349
+[ 98.3%] Iter 9595/9765 | SPS:  2774 | Reward:  0.070 | Value:  0.678 | VLoss: 0.0199 | PLoss: -0.0010 | Ent: 0.2696
+[ 98.3%] Iter 9596/9765 | SPS:  2774 | Reward:  0.068 | Value:  0.707 | VLoss: 0.0215 | PLoss: -0.0003 | Ent: 0.2534
+[ 98.3%] Iter 9597/9765 | SPS:  2774 | Reward:  0.057 | Value:  0.688 | VLoss: 0.0243 | PLoss: -0.0008 | Ent: 0.2392
+[ 98.3%] Iter 9598/9765 | SPS:  2774 | Reward:  0.065 | Value:  0.750 | VLoss: 0.0236 | PLoss: -0.0005 | Ent: 0.2522
+[ 98.3%] Iter 9599/9765 | SPS:  2774 | Reward:  0.058 | Value:  0.762 | VLoss: 0.0167 | PLoss: -0.0010 | Ent: 0.1953
+[ 98.3%] Iter 9600/9765 | SPS:  2774 | Reward:  0.063 | Value:  0.731 | VLoss: 0.0291 | PLoss: -0.0013 | Ent: 0.2352
+[ 98.3%] Iter 9601/9765 | SPS:  2774 | Reward:  0.059 | Value:  0.729 | VLoss: 0.0150 | PLoss: -0.0004 | Ent: 0.1784
+[ 98.3%] Iter 9602/9765 | SPS:  2774 | Reward:  0.059 | Value:  0.756 | VLoss: 0.0234 | PLoss: -0.0006 | Ent: 0.2400
+[ 98.3%] Iter 9603/9765 | SPS:  2774 | Reward:  0.048 | Value:  0.662 | VLoss: 0.0169 | PLoss: -0.0004 | Ent: 0.2354
+[ 98.4%] Iter 9604/9765 | SPS:  2775 | Reward:  0.065 | Value:  0.729 | VLoss: 0.0280 | PLoss: -0.0003 | Ent: 0.2293
+[ 98.4%] Iter 9605/9765 | SPS:  2775 | Reward:  0.063 | Value:  0.670 | VLoss: 0.0182 | PLoss: -0.0005 | Ent: 0.3103
+[ 98.4%] Iter 9606/9765 | SPS:  2775 | Reward:  0.045 | Value:  0.675 | VLoss: 0.0122 | PLoss: -0.0000 | Ent: 0.2244
+[ 98.4%] Iter 9607/9765 | SPS:  2775 | Reward:  0.053 | Value:  0.709 | VLoss: 0.0284 | PLoss: -0.0005 | Ent: 0.2495
+[ 98.4%] Iter 9608/9765 | SPS:  2775 | Reward:  0.061 | Value:  0.713 | VLoss: 0.0267 | PLoss: 0.0000 | Ent: 0.2567
+[ 98.4%] Iter 9609/9765 | SPS:  2775 | Reward:  0.036 | Value:  0.645 | VLoss: 0.0220 | PLoss: -0.0009 | Ent: 0.2517
+[ 98.4%] Iter 9610/9765 | SPS:  2775 | Reward:  0.049 | Value:  0.673 | VLoss: 0.0185 | PLoss: 0.0002 | Ent: 0.2112
+[ 98.4%] Iter 9611/9765 | SPS:  2775 | Reward:  0.067 | Value:  0.725 | VLoss: 0.0165 | PLoss: -0.0003 | Ent: 0.2408
+[ 98.4%] Iter 9612/9765 | SPS:  2775 | Reward:  0.067 | Value:  0.726 | VLoss: 0.0212 | PLoss: -0.0008 | Ent: 0.2105
+[ 98.4%] Iter 9613/9765 | SPS:  2775 | Reward:  0.058 | Value:  0.772 | VLoss: 0.0204 | PLoss: -0.0005 | Ent: 0.2026
+[ 98.5%] Iter 9614/9765 | SPS:  2775 | Reward:  0.058 | Value:  0.759 | VLoss: 0.0201 | PLoss: -0.0009 | Ent: 0.2385
+[ 98.5%] Iter 9615/9765 | SPS:  2775 | Reward:  0.047 | Value:  0.689 | VLoss: 0.0133 | PLoss: -0.0010 | Ent: 0.2149
+[ 98.5%] Iter 9616/9765 | SPS:  2775 | Reward:  0.052 | Value:  0.668 | VLoss: 0.0165 | PLoss: -0.0005 | Ent: 0.2872
+[ 98.5%] Iter 9617/9765 | SPS:  2776 | Reward:  0.052 | Value:  0.716 | VLoss: 0.0132 | PLoss: -0.0011 | Ent: 0.2125
+[ 98.5%] Iter 9618/9765 | SPS:  2776 | Reward:  0.060 | Value:  0.708 | VLoss: 0.0223 | PLoss: 0.0001 | Ent: 0.2267
+[ 98.5%] Iter 9619/9765 | SPS:  2776 | Reward:  0.062 | Value:  0.730 | VLoss: 0.0160 | PLoss: -0.0007 | Ent: 0.2243
+[ 98.5%] Iter 9620/9765 | SPS:  2776 | Reward:  0.058 | Value:  0.745 | VLoss: 0.0180 | PLoss: -0.0005 | Ent: 0.2561
+[ 98.5%] Iter 9621/9765 | SPS:  2776 | Reward:  0.062 | Value:  0.615 | VLoss: 0.0228 | PLoss: -0.0006 | Ent: 0.2441
+[ 98.5%] Iter 9622/9765 | SPS:  2776 | Reward:  0.052 | Value:  0.747 | VLoss: 0.0154 | PLoss: -0.0003 | Ent: 0.2033
+[ 98.5%] Iter 9623/9765 | SPS:  2776 | Reward:  0.061 | Value:  0.782 | VLoss: 0.0204 | PLoss: -0.0008 | Ent: 0.2188
+[ 98.6%] Iter 9624/9765 | SPS:  2776 | Reward:  0.059 | Value:  0.742 | VLoss: 0.0240 | PLoss: -0.0010 | Ent: 0.2305
+[ 98.6%] Iter 9625/9765 | SPS:  2776 | Reward:  0.048 | Value:  0.759 | VLoss: 0.0172 | PLoss: -0.0002 | Ent: 0.2008
+[ 98.6%] Iter 9626/9765 | SPS:  2776 | Reward:  0.064 | Value:  0.696 | VLoss: 0.0252 | PLoss: -0.0002 | Ent: 0.1997
+[ 98.6%] Iter 9627/9765 | SPS:  2776 | Reward:  0.052 | Value:  0.737 | VLoss: 0.0185 | PLoss: -0.0002 | Ent: 0.2597
+[ 98.6%] Iter 9628/9765 | SPS:  2776 | Reward:  0.054 | Value:  0.759 | VLoss: 0.0156 | PLoss: -0.0003 | Ent: 0.2588
+[ 98.6%] Iter 9629/9765 | SPS:  2777 | Reward:  0.061 | Value:  0.698 | VLoss: 0.0211 | PLoss: -0.0003 | Ent: 0.2189
+[ 98.6%] Iter 9630/9765 | SPS:  2777 | Reward:  0.061 | Value:  0.699 | VLoss: 0.0218 | PLoss: -0.0004 | Ent: 0.2474
+[ 98.6%] Iter 9631/9765 | SPS:  2777 | Reward:  0.057 | Value:  0.713 | VLoss: 0.0178 | PLoss: -0.0009 | Ent: 0.2351
+[ 98.6%] Iter 9632/9765 | SPS:  2777 | Reward:  0.062 | Value:  0.761 | VLoss: 0.0145 | PLoss: -0.0001 | Ent: 0.2229
+[ 98.6%] Iter 9633/9765 | SPS:  2777 | Reward:  0.053 | Value:  0.749 | VLoss: 0.0157 | PLoss: -0.0005 | Ent: 0.2343
+[ 98.7%] Iter 9634/9765 | SPS:  2777 | Reward:  0.050 | Value:  0.689 | VLoss: 0.0214 | PLoss: -0.0003 | Ent: 0.2755
+[ 98.7%] Iter 9635/9765 | SPS:  2777 | Reward:  0.070 | Value:  0.736 | VLoss: 0.0215 | PLoss: -0.0001 | Ent: 0.2376
+[ 98.7%] Iter 9636/9765 | SPS:  2777 | Reward:  0.057 | Value:  0.696 | VLoss: 0.0308 | PLoss: -0.0005 | Ent: 0.2268
+[ 98.7%] Iter 9637/9765 | SPS:  2777 | Reward:  0.064 | Value:  0.752 | VLoss: 0.0194 | PLoss: -0.0006 | Ent: 0.2283
+[ 98.7%] Iter 9638/9765 | SPS:  2777 | Reward:  0.052 | Value:  0.660 | VLoss: 0.0195 | PLoss: -0.0003 | Ent: 0.2199
+[ 98.7%] Iter 9639/9765 | SPS:  2777 | Reward:  0.048 | Value:  0.693 | VLoss: 0.0162 | PLoss: -0.0001 | Ent: 0.2566
+[ 98.7%] Iter 9640/9765 | SPS:  2777 | Reward:  0.046 | Value:  0.711 | VLoss: 0.0141 | PLoss: -0.0006 | Ent: 0.2182
+[ 98.7%] Iter 9641/9765 | SPS:  2777 | Reward:  0.051 | Value:  0.706 | VLoss: 0.0210 | PLoss: -0.0003 | Ent: 0.2361
+[ 98.7%] Iter 9642/9765 | SPS:  2778 | Reward:  0.061 | Value:  0.660 | VLoss: 0.0163 | PLoss: -0.0001 | Ent: 0.2333
+[ 98.8%] Iter 9643/9765 | SPS:  2778 | Reward:  0.076 | Value:  0.704 | VLoss: 0.0256 | PLoss: -0.0002 | Ent: 0.2373
+[ 98.8%] Iter 9644/9765 | SPS:  2778 | Reward:  0.063 | Value:  0.765 | VLoss: 0.0159 | PLoss: -0.0005 | Ent: 0.2377
+[ 98.8%] Iter 9645/9765 | SPS:  2778 | Reward:  0.052 | Value:  0.684 | VLoss: 0.0309 | PLoss: -0.0002 | Ent: 0.2553
+[ 98.8%] Iter 9646/9765 | SPS:  2778 | Reward:  0.064 | Value:  0.762 | VLoss: 0.0258 | PLoss: -0.0002 | Ent: 0.2679
+[ 98.8%] Iter 9647/9765 | SPS:  2778 | Reward:  0.052 | Value:  0.681 | VLoss: 0.0258 | PLoss: -0.0003 | Ent: 0.2028
+[ 98.8%] Iter 9648/9765 | SPS:  2778 | Reward:  0.048 | Value:  0.687 | VLoss: 0.0203 | PLoss: -0.0013 | Ent: 0.2509
+[ 98.8%] Iter 9649/9765 | SPS:  2778 | Reward:  0.062 | Value:  0.699 | VLoss: 0.0176 | PLoss: -0.0007 | Ent: 0.2307
+[ 98.8%] Iter 9650/9765 | SPS:  2778 | Reward:  0.068 | Value:  0.739 | VLoss: 0.0232 | PLoss: -0.0001 | Ent: 0.1811
+[ 98.8%] Iter 9651/9765 | SPS:  2778 | Reward:  0.062 | Value:  0.691 | VLoss: 0.0181 | PLoss: -0.0002 | Ent: 0.2438
+[ 98.8%] Iter 9652/9765 | SPS:  2778 | Reward:  0.056 | Value:  0.748 | VLoss: 0.0178 | PLoss: -0.0002 | Ent: 0.1943
+[ 98.9%] Iter 9653/9765 | SPS:  2778 | Reward:  0.071 | Value:  0.737 | VLoss: 0.0226 | PLoss: -0.0002 | Ent: 0.2073
+[ 98.9%] Iter 9654/9765 | SPS:  2778 | Reward:  0.050 | Value:  0.710 | VLoss: 0.0216 | PLoss: -0.0000 | Ent: 0.2379
+[ 98.9%] Iter 9655/9765 | SPS:  2779 | Reward:  0.055 | Value:  0.698 | VLoss: 0.0199 | PLoss: -0.0002 | Ent: 0.2158
+[ 98.9%] Iter 9656/9765 | SPS:  2779 | Reward:  0.059 | Value:  0.725 | VLoss: 0.0250 | PLoss: -0.0006 | Ent: 0.2585
+[ 98.9%] Iter 9657/9765 | SPS:  2779 | Reward:  0.055 | Value:  0.779 | VLoss: 0.0148 | PLoss: -0.0012 | Ent: 0.2136
+[ 98.9%] Iter 9658/9765 | SPS:  2779 | Reward:  0.053 | Value:  0.710 | VLoss: 0.0216 | PLoss: -0.0010 | Ent: 0.2669
+[ 98.9%] Iter 9659/9765 | SPS:  2779 | Reward:  0.048 | Value:  0.678 | VLoss: 0.0150 | PLoss: -0.0008 | Ent: 0.2301
+[ 98.9%] Iter 9660/9765 | SPS:  2779 | Reward:  0.053 | Value:  0.656 | VLoss: 0.0255 | PLoss: -0.0003 | Ent: 0.2658
+[ 98.9%] Iter 9661/9765 | SPS:  2779 | Reward:  0.051 | Value:  0.634 | VLoss: 0.0224 | PLoss: -0.0005 | Ent: 0.2382
+[ 98.9%] Iter 9662/9765 | SPS:  2779 | Reward:  0.060 | Value:  0.704 | VLoss: 0.0176 | PLoss: -0.0003 | Ent: 0.1977
+[ 99.0%] Iter 9663/9765 | SPS:  2779 | Reward:  0.052 | Value:  0.704 | VLoss: 0.0220 | PLoss: 0.0002 | Ent: 0.2193
+[ 99.0%] Iter 9664/9765 | SPS:  2779 | Reward:  0.058 | Value:  0.692 | VLoss: 0.0201 | PLoss: 0.0000 | Ent: 0.2606
+[ 99.0%] Iter 9665/9765 | SPS:  2779 | Reward:  0.068 | Value:  0.697 | VLoss: 0.0217 | PLoss: -0.0003 | Ent: 0.2512
+[ 99.0%] Iter 9666/9765 | SPS:  2779 | Reward:  0.073 | Value:  0.714 | VLoss: 0.0205 | PLoss: -0.0001 | Ent: 0.2298
+[ 99.0%] Iter 9667/9765 | SPS:  2780 | Reward:  0.068 | Value:  0.724 | VLoss: 0.0275 | PLoss: -0.0005 | Ent: 0.2439
+[ 99.0%] Iter 9668/9765 | SPS:  2780 | Reward:  0.056 | Value:  0.737 | VLoss: 0.0258 | PLoss: -0.0004 | Ent: 0.2289
+[ 99.0%] Iter 9669/9765 | SPS:  2780 | Reward:  0.062 | Value:  0.710 | VLoss: 0.0211 | PLoss: -0.0003 | Ent: 0.1949
+[ 99.0%] Iter 9670/9765 | SPS:  2780 | Reward:  0.060 | Value:  0.754 | VLoss: 0.0165 | PLoss: -0.0002 | Ent: 0.2306
+[ 99.0%] Iter 9671/9765 | SPS:  2780 | Reward:  0.051 | Value:  0.736 | VLoss: 0.0186 | PLoss: -0.0010 | Ent: 0.2939
+[ 99.0%] Iter 9672/9765 | SPS:  2780 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0247 | PLoss: -0.0002 | Ent: 0.2491
+[ 99.1%] Iter 9673/9765 | SPS:  2780 | Reward:  0.049 | Value:  0.712 | VLoss: 0.0245 | PLoss: -0.0003 | Ent: 0.2251
+[ 99.1%] Iter 9674/9765 | SPS:  2780 | Reward:  0.054 | Value:  0.733 | VLoss: 0.0182 | PLoss: -0.0000 | Ent: 0.1809
+[ 99.1%] Iter 9675/9765 | SPS:  2780 | Reward:  0.056 | Value:  0.651 | VLoss: 0.0202 | PLoss: -0.0007 | Ent: 0.2737
+[ 99.1%] Iter 9676/9765 | SPS:  2780 | Reward:  0.062 | Value:  0.720 | VLoss: 0.0227 | PLoss: -0.0003 | Ent: 0.2360
+[ 99.1%] Iter 9677/9765 | SPS:  2780 | Reward:  0.062 | Value:  0.686 | VLoss: 0.0222 | PLoss: -0.0003 | Ent: 0.2476
+[ 99.1%] Iter 9678/9765 | SPS:  2780 | Reward:  0.054 | Value:  0.676 | VLoss: 0.0227 | PLoss: -0.0002 | Ent: 0.2160
+[ 99.1%] Iter 9679/9765 | SPS:  2780 | Reward:  0.053 | Value:  0.724 | VLoss: 0.0216 | PLoss: -0.0002 | Ent: 0.2029
+[ 99.1%] Iter 9680/9765 | SPS:  2781 | Reward:  0.050 | Value:  0.736 | VLoss: 0.0176 | PLoss: -0.0002 | Ent: 0.2433
+[ 99.1%] Iter 9681/9765 | SPS:  2781 | Reward:  0.060 | Value:  0.743 | VLoss: 0.0167 | PLoss: -0.0002 | Ent: 0.2215
+[ 99.2%] Iter 9682/9765 | SPS:  2781 | Reward:  0.063 | Value:  0.710 | VLoss: 0.0222 | PLoss: -0.0000 | Ent: 0.2242
+[ 99.2%] Iter 9683/9765 | SPS:  2781 | Reward:  0.063 | Value:  0.627 | VLoss: 0.0246 | PLoss: -0.0006 | Ent: 0.2207
+[ 99.2%] Iter 9684/9765 | SPS:  2781 | Reward:  0.045 | Value:  0.735 | VLoss: 0.0190 | PLoss: -0.0004 | Ent: 0.2228
+[ 99.2%] Iter 9685/9765 | SPS:  2781 | Reward:  0.064 | Value:  0.735 | VLoss: 0.0289 | PLoss: -0.0003 | Ent: 0.2175
+[ 99.2%] Iter 9686/9765 | SPS:  2781 | Reward:  0.060 | Value:  0.753 | VLoss: 0.0249 | PLoss: -0.0006 | Ent: 0.2493
+[ 99.2%] Iter 9687/9765 | SPS:  2781 | Reward:  0.062 | Value:  0.681 | VLoss: 0.0249 | PLoss: -0.0003 | Ent: 0.2737
+[ 99.2%] Iter 9688/9765 | SPS:  2781 | Reward:  0.068 | Value:  0.706 | VLoss: 0.0261 | PLoss: -0.0002 | Ent: 0.2333
+[ 99.2%] Iter 9689/9765 | SPS:  2781 | Reward:  0.055 | Value:  0.738 | VLoss: 0.0195 | PLoss: -0.0002 | Ent: 0.2085
+[ 99.2%] Iter 9690/9765 | SPS:  2781 | Reward:  0.065 | Value:  0.743 | VLoss: 0.0195 | PLoss: -0.0003 | Ent: 0.2335
+[ 99.2%] Iter 9691/9765 | SPS:  2781 | Reward:  0.053 | Value:  0.706 | VLoss: 0.0197 | PLoss: -0.0003 | Ent: 0.2725
+[ 99.3%] Iter 9692/9765 | SPS:  2781 | Reward:  0.050 | Value:  0.713 | VLoss: 0.0278 | PLoss: -0.0002 | Ent: 0.2512
+[ 99.3%] Iter 9693/9765 | SPS:  2782 | Reward:  0.068 | Value:  0.740 | VLoss: 0.0229 | PLoss: -0.0006 | Ent: 0.2203
+[ 99.3%] Iter 9694/9765 | SPS:  2782 | Reward:  0.065 | Value:  0.764 | VLoss: 0.0245 | PLoss: -0.0003 | Ent: 0.2348
+[ 99.3%] Iter 9695/9765 | SPS:  2782 | Reward:  0.067 | Value:  0.754 | VLoss: 0.0257 | PLoss: -0.0001 | Ent: 0.2451
+[ 99.3%] Iter 9696/9765 | SPS:  2782 | Reward:  0.058 | Value:  0.686 | VLoss: 0.0231 | PLoss: -0.0004 | Ent: 0.2061
+[ 99.3%] Iter 9697/9765 | SPS:  2782 | Reward:  0.052 | Value:  0.658 | VLoss: 0.0167 | PLoss: -0.0003 | Ent: 0.2295
+[ 99.3%] Iter 9698/9765 | SPS:  2782 | Reward:  0.060 | Value:  0.717 | VLoss: 0.0272 | PLoss: -0.0003 | Ent: 0.2154
+[ 99.3%] Iter 9699/9765 | SPS:  2782 | Reward:  0.072 | Value:  0.802 | VLoss: 0.0113 | PLoss: -0.0003 | Ent: 0.2250
+[ 99.3%] Iter 9700/9765 | SPS:  2782 | Reward:  0.061 | Value:  0.763 | VLoss: 0.0148 | PLoss: -0.0002 | Ent: 0.2695
+[ 99.3%] Iter 9701/9765 | SPS:  2782 | Reward:  0.069 | Value:  0.687 | VLoss: 0.0232 | PLoss: -0.0003 | Ent: 0.2837
+[ 99.4%] Iter 9702/9765 | SPS:  2782 | Reward:  0.051 | Value:  0.723 | VLoss: 0.0164 | PLoss: -0.0001 | Ent: 0.2484
+[ 99.4%] Iter 9703/9765 | SPS:  2782 | Reward:  0.044 | Value:  0.696 | VLoss: 0.0207 | PLoss: -0.0002 | Ent: 0.2018
+[ 99.4%] Iter 9704/9765 | SPS:  2782 | Reward:  0.059 | Value:  0.735 | VLoss: 0.0178 | PLoss: -0.0002 | Ent: 0.1879
+[ 99.4%] Iter 9705/9765 | SPS:  2783 | Reward:  0.048 | Value:  0.699 | VLoss: 0.0383 | PLoss: -0.0003 | Ent: 0.2521
+[ 99.4%] Iter 9706/9765 | SPS:  2783 | Reward:  0.048 | Value:  0.707 | VLoss: 0.0226 | PLoss: -0.0000 | Ent: 0.2153
+[ 99.4%] Iter 9707/9765 | SPS:  2783 | Reward:  0.058 | Value:  0.690 | VLoss: 0.0206 | PLoss: -0.0002 | Ent: 0.2594
+[ 99.4%] Iter 9708/9765 | SPS:  2783 | Reward:  0.069 | Value:  0.759 | VLoss: 0.0197 | PLoss: -0.0002 | Ent: 0.2428
+[ 99.4%] Iter 9709/9765 | SPS:  2783 | Reward:  0.055 | Value:  0.683 | VLoss: 0.0230 | PLoss: -0.0002 | Ent: 0.2288
+[ 99.4%] Iter 9710/9765 | SPS:  2783 | Reward:  0.056 | Value:  0.712 | VLoss: 0.0209 | PLoss: -0.0003 | Ent: 0.2087
+[ 99.4%] Iter 9711/9765 | SPS:  2783 | Reward:  0.069 | Value:  0.725 | VLoss: 0.0210 | PLoss: -0.0001 | Ent: 0.2636
+[ 99.5%] Iter 9712/9765 | SPS:  2783 | Reward:  0.059 | Value:  0.712 | VLoss: 0.0257 | PLoss: -0.0002 | Ent: 0.2694
+[ 99.5%] Iter 9713/9765 | SPS:  2783 | Reward:  0.055 | Value:  0.731 | VLoss: 0.0277 | PLoss: -0.0001 | Ent: 0.2526
+[ 99.5%] Iter 9714/9765 | SPS:  2783 | Reward:  0.062 | Value:  0.753 | VLoss: 0.0232 | PLoss: -0.0003 | Ent: 0.2444
+[ 99.5%] Iter 9715/9765 | SPS:  2783 | Reward:  0.063 | Value:  0.764 | VLoss: 0.0266 | PLoss: -0.0002 | Ent: 0.2171
+[ 99.5%] Iter 9716/9765 | SPS:  2783 | Reward:  0.058 | Value:  0.756 | VLoss: 0.0179 | PLoss: -0.0001 | Ent: 0.2701
+[ 99.5%] Iter 9717/9765 | SPS:  2783 | Reward:  0.062 | Value:  0.769 | VLoss: 0.0146 | PLoss: -0.0002 | Ent: 0.2264
+[ 99.5%] Iter 9718/9765 | SPS:  2784 | Reward:  0.066 | Value:  0.792 | VLoss: 0.0108 | PLoss: 0.0000 | Ent: 0.1914
+[ 99.5%] Iter 9719/9765 | SPS:  2784 | Reward:  0.055 | Value:  0.705 | VLoss: 0.0228 | PLoss: -0.0002 | Ent: 0.2396
+[ 99.5%] Iter 9720/9765 | SPS:  2784 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0191 | PLoss: -0.0001 | Ent: 0.2380
+[ 99.5%] Iter 9721/9765 | SPS:  2784 | Reward:  0.056 | Value:  0.745 | VLoss: 0.0236 | PLoss: -0.0003 | Ent: 0.1932
+[ 99.6%] Iter 9722/9765 | SPS:  2784 | Reward:  0.046 | Value:  0.651 | VLoss: 0.0178 | PLoss: -0.0002 | Ent: 0.2344
+[ 99.6%] Iter 9723/9765 | SPS:  2784 | Reward:  0.052 | Value:  0.697 | VLoss: 0.0348 | PLoss: -0.0003 | Ent: 0.2034
+[ 99.6%] Iter 9724/9765 | SPS:  2784 | Reward:  0.062 | Value:  0.742 | VLoss: 0.0197 | PLoss: -0.0001 | Ent: 0.2140
+[ 99.6%] Iter 9725/9765 | SPS:  2784 | Reward:  0.053 | Value:  0.733 | VLoss: 0.0160 | PLoss: -0.0002 | Ent: 0.2691
+[ 99.6%] Iter 9726/9765 | SPS:  2784 | Reward:  0.072 | Value:  0.741 | VLoss: 0.0210 | PLoss: -0.0001 | Ent: 0.2181
+[ 99.6%] Iter 9727/9765 | SPS:  2784 | Reward:  0.057 | Value:  0.708 | VLoss: 0.0167 | PLoss: -0.0001 | Ent: 0.2357
+[ 99.6%] Iter 9728/9765 | SPS:  2784 | Reward:  0.054 | Value:  0.803 | VLoss: 0.0102 | PLoss: -0.0000 | Ent: 0.2331
+[ 99.6%] Iter 9729/9765 | SPS:  2784 | Reward:  0.057 | Value:  0.705 | VLoss: 0.0190 | PLoss: -0.0001 | Ent: 0.2052
+[ 99.6%] Iter 9730/9765 | SPS:  2785 | Reward:  0.055 | Value:  0.720 | VLoss: 0.0129 | PLoss: -0.0002 | Ent: 0.2330
+[ 99.7%] Iter 9731/9765 | SPS:  2785 | Reward:  0.065 | Value:  0.730 | VLoss: 0.0230 | PLoss: -0.0001 | Ent: 0.2155
+[ 99.7%] Iter 9732/9765 | SPS:  2785 | Reward:  0.074 | Value:  0.732 | VLoss: 0.0176 | PLoss: -0.0000 | Ent: 0.2181
+[ 99.7%] Iter 9733/9765 | SPS:  2785 | Reward:  0.061 | Value:  0.730 | VLoss: 0.0242 | PLoss: -0.0002 | Ent: 0.2298
+[ 99.7%] Iter 9734/9765 | SPS:  2785 | Reward:  0.056 | Value:  0.643 | VLoss: 0.0211 | PLoss: -0.0001 | Ent: 0.2474
+[ 99.7%] Iter 9735/9765 | SPS:  2785 | Reward:  0.071 | Value:  0.718 | VLoss: 0.0258 | PLoss: -0.0001 | Ent: 0.2161
+[ 99.7%] Iter 9736/9765 | SPS:  2785 | Reward:  0.061 | Value:  0.718 | VLoss: 0.0241 | PLoss: -0.0001 | Ent: 0.1978
+[ 99.7%] Iter 9737/9765 | SPS:  2785 | Reward:  0.059 | Value:  0.700 | VLoss: 0.0358 | PLoss: -0.0001 | Ent: 0.2317
+[ 99.7%] Iter 9738/9765 | SPS:  2785 | Reward:  0.070 | Value:  0.710 | VLoss: 0.0339 | PLoss: -0.0001 | Ent: 0.2343
+[ 99.7%] Iter 9739/9765 | SPS:  2785 | Reward:  0.054 | Value:  0.749 | VLoss: 0.0202 | PLoss: -0.0001 | Ent: 0.2692
+[ 99.7%] Iter 9740/9765 | SPS:  2785 | Reward:  0.062 | Value:  0.731 | VLoss: 0.0202 | PLoss: -0.0001 | Ent: 0.2609
+[ 99.8%] Iter 9741/9765 | SPS:  2785 | Reward:  0.051 | Value:  0.757 | VLoss: 0.0184 | PLoss: -0.0000 | Ent: 0.2463
+[ 99.8%] Iter 9742/9765 | SPS:  2785 | Reward:  0.066 | Value:  0.696 | VLoss: 0.0218 | PLoss: -0.0000 | Ent: 0.2917
+[ 99.8%] Iter 9743/9765 | SPS:  2786 | Reward:  0.062 | Value:  0.690 | VLoss: 0.0293 | PLoss: -0.0001 | Ent: 0.2613
+[ 99.8%] Iter 9744/9765 | SPS:  2786 | Reward:  0.058 | Value:  0.747 | VLoss: 0.0208 | PLoss: -0.0000 | Ent: 0.2425
+[ 99.8%] Iter 9745/9765 | SPS:  2786 | Reward:  0.058 | Value:  0.685 | VLoss: 0.0257 | PLoss: -0.0001 | Ent: 0.1983
+[ 99.8%] Iter 9746/9765 | SPS:  2786 | Reward:  0.056 | Value:  0.668 | VLoss: 0.0249 | PLoss: -0.0001 | Ent: 0.3013
+[ 99.8%] Iter 9747/9765 | SPS:  2786 | Reward:  0.050 | Value:  0.743 | VLoss: 0.0200 | PLoss: -0.0002 | Ent: 0.2186
+[ 99.8%] Iter 9748/9765 | SPS:  2786 | Reward:  0.068 | Value:  0.704 | VLoss: 0.0202 | PLoss: -0.0000 | Ent: 0.2603
+[ 99.8%] Iter 9749/9765 | SPS:  2786 | Reward:  0.042 | Value:  0.665 | VLoss: 0.0283 | PLoss: -0.0001 | Ent: 0.2093
+[ 99.8%] Iter 9750/9765 | SPS:  2786 | Reward:  0.058 | Value:  0.762 | VLoss: 0.0210 | PLoss: -0.0001 | Ent: 0.2183
+[ 99.9%] Iter 9751/9765 | SPS:  2786 | Reward:  0.052 | Value:  0.677 | VLoss: 0.0258 | PLoss: -0.0000 | Ent: 0.2561
+[ 99.9%] Iter 9752/9765 | SPS:  2786 | Reward:  0.069 | Value:  0.763 | VLoss: 0.0233 | PLoss: -0.0000 | Ent: 0.2016
+[ 99.9%] Iter 9753/9765 | SPS:  2786 | Reward:  0.065 | Value:  0.696 | VLoss: 0.0202 | PLoss: -0.0000 | Ent: 0.2266
+[ 99.9%] Iter 9754/9765 | SPS:  2786 | Reward:  0.054 | Value:  0.720 | VLoss: 0.0233 | PLoss: -0.0000 | Ent: 0.2624
+[ 99.9%] Iter 9755/9765 | SPS:  2786 | Reward:  0.054 | Value:  0.768 | VLoss: 0.0149 | PLoss: -0.0000 | Ent: 0.2799
+[ 99.9%] Iter 9756/9765 | SPS:  2787 | Reward:  0.062 | Value:  0.636 | VLoss: 0.0149 | PLoss: 0.0000 | Ent: 0.2049
+[ 99.9%] Iter 9757/9765 | SPS:  2787 | Reward:  0.062 | Value:  0.675 | VLoss: 0.0269 | PLoss: 0.0000 | Ent: 0.2060
+[ 99.9%] Iter 9758/9765 | SPS:  2787 | Reward:  0.072 | Value:  0.787 | VLoss: 0.0209 | PLoss: -0.0000 | Ent: 0.2050
+[ 99.9%] Iter 9759/9765 | SPS:  2787 | Reward:  0.046 | Value:  0.718 | VLoss: 0.0224 | PLoss: -0.0001 | Ent: 0.2336
+[ 99.9%] Iter 9760/9765 | SPS:  2787 | Reward:  0.055 | Value:  0.781 | VLoss: 0.0138 | PLoss: -0.0000 | Ent: 0.2113
+[100.0%] Iter 9761/9765 | SPS:  2787 | Reward:  0.052 | Value:  0.715 | VLoss: 0.0173 | PLoss: -0.0000 | Ent: 0.2330
+[100.0%] Iter 9762/9765 | SPS:  2787 | Reward:  0.065 | Value:  0.707 | VLoss: 0.0300 | PLoss: -0.0000 | Ent: 0.2382
+[100.0%] Iter 9763/9765 | SPS:  2787 | Reward:  0.064 | Value:  0.739 | VLoss: 0.0277 | PLoss: -0.0000 | Ent: 0.2081
+[100.0%] Iter 9764/9765 | SPS:  2787 | Reward:  0.069 | Value:  0.711 | VLoss: 0.0248 | PLoss: 0.0000 | Ent: 0.2280
+[100.0%] Iter 9765/9765 | SPS:  2787 | Reward:  0.047 | Value:  0.654 | VLoss: 0.0268 | PLoss: 0.0000 | Ent: 0.2404
diff --git a/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/wandb-metadata.json b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..d890f6e69aa52dbeb80a8be09d4906ea10afa9f1
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2025-11-07T03:38:56.139668Z",
+  "args":  [
+    "--track",
+    "--wandb-project-name",
+    "ragen-bandit"
+  ],
+  "program":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/ppo_frozenlake.py",
+  "codePath":  "cleanrl/ppo_frozenlake.py",
+  "codePathLocal":  "ppo_frozenlake.py",
+  "git":  {
+    "remote":  "https://github.com/vwxyzjn/cleanrl.git",
+    "commit":  "004f8a086a892a2a180f4dd332b90d83a968aa7a"
+  },
+  "email":  "haoyu-wa22@mails.tsinghua.edu.cn",
+  "root":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl",
+  "host":  "pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0",
+  "executable":  "/root/local/miniconda3/envs/ragen/bin/python",
+  "cpu_count":  64,
+  "cpu_count_logical":  128,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "5153960755200",
+      "used":  "31684485120"
+    }
+  },
+  "memory":  {
+    "total":  "2163642122240"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-35e2d43d-4067-82ce-90d4-def9e389bf28"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-af4135e3-88f2-e9ac-518d-502c75a85429"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b7cf0ec6-7c29-1179-dceb-09565da51890"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "qcnee8s40otwmmxavjc9wsiwf1cq85oc"
+}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug-internal.log b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..29255ae48ed7c8d165b7cd5b022e820d45f30f14
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug-internal.log
@@ -0,0 +1,2547 @@
+{"time":"2025-11-07T11:38:56.409015251+08:00","level":"INFO","msg":"stream: starting","core version":"0.22.3"}
+{"time":"2025-11-07T11:38:56.779221075+08:00","level":"INFO","msg":"stream: created new stream","id":"k31oio6c"}
+{"time":"2025-11-07T11:38:56.779452122+08:00","level":"INFO","msg":"handler: started","stream_id":"k31oio6c"}
+{"time":"2025-11-07T11:38:56.782410797+08:00","level":"INFO","msg":"stream: started","id":"k31oio6c"}
+{"time":"2025-11-07T11:38:56.782428601+08:00","level":"INFO","msg":"writer: started","stream_id":"k31oio6c"}
+{"time":"2025-11-07T11:38:56.782430236+08:00","level":"INFO","msg":"sender: started","stream_id":"k31oio6c"}
+{"time":"2025-11-07T11:38:57.177394453+08:00","level":"ERROR","msg":"git repo not found","error":"repository does not exist"}
+{"time":"2025-11-07T11:39:07.313199288+08:00","level":"WARN","msg":"tensorboard: no root directory","error":"timed out after 10s"}
+{"time":"2025-11-07T11:39:07.315364719+08:00","level":"INFO","msg":"tensorboard: tracking new log directory","rootDir":{},"logDir":"LocalOrCloudPath(LocalPath=\"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/FrozenLake__ppo_frozenlake__1__1762486729\")","namespace":""}
+{"time":"2025-11-07T11:39:07.317518283+08:00","level":"INFO","msg":"tensorboard: saving file","fileLocation":"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/FrozenLake__ppo_frozenlake__1__1762486729/events.out.tfevents.1762486737.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.214168.0","runPath":"runs/FrozenLake__ppo_frozenlake__1__1762486729/events.out.tfevents.1762486737.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.214168.0"}
+{"time":"2025-11-07T11:39:07.320277569+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":69}
+{"time":"2025-11-07T11:39:07.320355178+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:39:07.320996348+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":102}
+{"time":"2025-11-07T11:39:07.322047204+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":30}
+{"time":"2025-11-07T11:39:12.326696541+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":172}
+{"time":"2025-11-07T11:39:12.631869617+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":45}
+{"time":"2025-11-07T11:39:22.338061017+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":395}
+{"time":"2025-11-07T11:39:22.33817173+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:39:27.343254296+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":461}
+{"time":"2025-11-07T11:39:27.614602517+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":45}
+{"time":"2025-11-07T11:39:32.348883098+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":546}
+{"time":"2025-11-07T11:39:32.348984765+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:39:32.349436562+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":578}
+{"time":"2025-11-07T11:39:32.350930919+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":10}
+{"time":"2025-11-07T11:39:37.357086324+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":626}
+{"time":"2025-11-07T11:39:37.357181815+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:39:37.35765333+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":658}
+{"time":"2025-11-07T11:39:37.35922207+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:39:42.362226778+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":725}
+{"time":"2025-11-07T11:39:42.594545351+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":359}
+{"time":"2025-11-07T11:39:52.377823845+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1364}
+{"time":"2025-11-07T11:39:52.38462325+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":87}
+{"time":"2025-11-07T11:39:57.384203609+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1508}
+{"time":"2025-11-07T11:39:57.67249197+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":396}
+{"time":"2025-11-07T11:40:02.392442453+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1970}
+{"time":"2025-11-07T11:40:02.392554124+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:02.392967595+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2002}
+{"time":"2025-11-07T11:40:02.397016184+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":189}
+{"time":"2025-11-07T11:40:07.399777157+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2239}
+{"time":"2025-11-07T11:40:07.399868882+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:07.400379737+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2271}
+{"time":"2025-11-07T11:40:07.402852012+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":207}
+{"time":"2025-11-07T11:40:12.406683248+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2525}
+{"time":"2025-11-07T11:40:12.588493126+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":201}
+{"time":"2025-11-07T11:40:17.415520617+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2768}
+{"time":"2025-11-07T11:40:17.415601259+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:17.416094232+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2800}
+{"time":"2025-11-07T11:40:17.417811018+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:40:22.42181448+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2976}
+{"time":"2025-11-07T11:40:22.42330253+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":78}
+{"time":"2025-11-07T11:40:27.427504919+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3097}
+{"time":"2025-11-07T11:40:27.607762263+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":123}
+{"time":"2025-11-07T11:40:32.433141849+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3262}
+{"time":"2025-11-07T11:40:32.433332041+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:32.433702119+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3294}
+{"time":"2025-11-07T11:40:32.435667846+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:40:42.44413146+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3608}
+{"time":"2025-11-07T11:40:42.621851175+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":477}
+{"time":"2025-11-07T11:40:47.454400165+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4162}
+{"time":"2025-11-07T11:40:47.454563648+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:47.454913913+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4194}
+{"time":"2025-11-07T11:40:47.457078442+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":53}
+{"time":"2025-11-07T11:40:52.459268247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4295}
+{"time":"2025-11-07T11:40:52.459388467+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:40:52.459846431+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4327}
+{"time":"2025-11-07T11:40:52.461750012+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":84}
+{"time":"2025-11-07T11:40:57.465284976+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4478}
+{"time":"2025-11-07T11:40:57.607201678+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":591}
+{"time":"2025-11-07T11:41:02.473396926+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5119}
+{"time":"2025-11-07T11:41:02.473520985+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:02.474010499+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5151}
+{"time":"2025-11-07T11:41:02.476880383+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":246}
+{"time":"2025-11-07T11:41:07.48044522+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5447}
+{"time":"2025-11-07T11:41:07.48057485+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:07.48105245+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5479}
+{"time":"2025-11-07T11:41:07.484065202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":246}
+{"time":"2025-11-07T11:41:12.488274702+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5776}
+{"time":"2025-11-07T11:41:12.616056685+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":279}
+{"time":"2025-11-07T11:41:17.496217209+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6102}
+{"time":"2025-11-07T11:41:17.496350355+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:17.496764247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6134}
+{"time":"2025-11-07T11:41:17.499490855+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":208}
+{"time":"2025-11-07T11:41:22.503807416+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6456}
+{"time":"2025-11-07T11:41:22.504036149+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:22.504626098+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6488}
+{"time":"2025-11-07T11:41:22.506491749+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":59}
+{"time":"2025-11-07T11:41:27.510573285+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6592}
+{"time":"2025-11-07T11:41:27.610038946+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":162}
+{"time":"2025-11-07T11:41:37.522811723+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6966}
+{"time":"2025-11-07T11:41:37.522894632+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:37.523397467+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":6998}
+{"time":"2025-11-07T11:41:37.525883848+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":207}
+{"time":"2025-11-07T11:41:42.530317372+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":7265}
+{"time":"2025-11-07T11:41:42.661138356+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":475}
+{"time":"2025-11-07T11:41:47.540235146+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":7852}
+{"time":"2025-11-07T11:41:47.540504683+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:47.541021049+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":7884}
+{"time":"2025-11-07T11:41:47.543720936+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":225}
+{"time":"2025-11-07T11:41:52.549460677+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8272}
+{"time":"2025-11-07T11:41:52.549580223+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:41:52.550054157+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8304}
+{"time":"2025-11-07T11:41:52.551141501+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":92}
+{"time":"2025-11-07T11:41:57.578671241+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8443}
+{"time":"2025-11-07T11:41:57.607702088+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":242}
+{"time":"2025-11-07T11:42:02.585279646+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8766}
+{"time":"2025-11-07T11:42:02.585355385+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:02.586015627+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8798}
+{"time":"2025-11-07T11:42:02.588092957+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":92}
+{"time":"2025-11-07T11:42:07.590627845+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8932}
+{"time":"2025-11-07T11:42:07.590727097+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:07.591834863+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":8965}
+{"time":"2025-11-07T11:42:07.593042819+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:42:12.598089234+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9099}
+{"time":"2025-11-07T11:42:12.612869987+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":203}
+{"time":"2025-11-07T11:42:17.604462127+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9340}
+{"time":"2025-11-07T11:42:17.604587511+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:17.605022865+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9372}
+{"time":"2025-11-07T11:42:17.606829163+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:42:27.616697747+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9514}
+{"time":"2025-11-07T11:42:27.616859743+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:27.617241587+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9546}
+{"time":"2025-11-07T11:42:27.619151718+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":168}
+{"time":"2025-11-07T11:42:32.625179031+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9831}
+{"time":"2025-11-07T11:42:32.625296348+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:32.625722774+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":9863}
+{"time":"2025-11-07T11:42:32.630138571+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":179}
+{"time":"2025-11-07T11:42:37.63241951+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":10084}
+{"time":"2025-11-07T11:42:37.632535337+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:37.633071694+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":10116}
+{"time":"2025-11-07T11:42:37.634934071+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:42:42.638078235+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":10259}
+{"time":"2025-11-07T11:42:42.646223006+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":359}
+{"time":"2025-11-07T11:42:47.645354251+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":10669}
+{"time":"2025-11-07T11:42:47.645522998+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:47.645909667+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":10701}
+{"time":"2025-11-07T11:42:47.649026728+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":286}
+{"time":"2025-11-07T11:42:52.65435262+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11088}
+{"time":"2025-11-07T11:42:52.654457189+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:52.65486145+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11120}
+{"time":"2025-11-07T11:42:52.656167202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":154}
+{"time":"2025-11-07T11:42:57.660286754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11323}
+{"time":"2025-11-07T11:42:57.661015519+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:42:57.662218326+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11496}
+{"time":"2025-11-07T11:42:57.66315429+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":67}
+{"time":"2025-11-07T11:43:02.667594992+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11604}
+{"time":"2025-11-07T11:43:02.667727324+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:02.668145676+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11636}
+{"time":"2025-11-07T11:43:02.669876877+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":91}
+{"time":"2025-11-07T11:43:07.673877091+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11769}
+{"time":"2025-11-07T11:43:07.673970156+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:07.674439967+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11801}
+{"time":"2025-11-07T11:43:07.676545973+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:43:12.701786549+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11935}
+{"time":"2025-11-07T11:43:12.701901933+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:12.702354383+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":11967}
+{"time":"2025-11-07T11:43:12.704159168+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:43:17.724520236+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12109}
+{"time":"2025-11-07T11:43:17.724634425+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:17.725101801+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12141}
+{"time":"2025-11-07T11:43:17.727000233+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":80}
+{"time":"2025-11-07T11:43:22.733103876+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12261}
+{"time":"2025-11-07T11:43:22.733195458+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:22.733625318+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12293}
+{"time":"2025-11-07T11:43:22.735361578+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:43:27.738909371+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12388}
+{"time":"2025-11-07T11:43:27.738998782+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:27.739493354+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12420}
+{"time":"2025-11-07T11:43:27.741187584+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:43:37.750239955+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12768}
+{"time":"2025-11-07T11:43:37.750351374+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:37.750811049+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12800}
+{"time":"2025-11-07T11:43:37.751948588+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":38}
+{"time":"2025-11-07T11:43:42.758235373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12882}
+{"time":"2025-11-07T11:43:42.758381066+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:42.758857465+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":12914}
+{"time":"2025-11-07T11:43:42.761009907+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:43:47.765492778+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13043}
+{"time":"2025-11-07T11:43:47.765673764+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:47.766692827+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13076}
+{"time":"2025-11-07T11:43:47.767687264+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:43:52.771795159+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13242}
+{"time":"2025-11-07T11:43:52.771896107+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:52.772393313+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13274}
+{"time":"2025-11-07T11:43:52.773952844+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":17}
+{"time":"2025-11-07T11:43:57.778861156+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13385}
+{"time":"2025-11-07T11:43:57.778966968+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:43:57.779438878+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13417}
+{"time":"2025-11-07T11:43:57.78020091+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":40}
+{"time":"2025-11-07T11:44:02.784567195+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13500}
+{"time":"2025-11-07T11:44:02.78467103+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:02.785079474+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13532}
+{"time":"2025-11-07T11:44:02.787808838+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:44:07.792416682+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13663}
+{"time":"2025-11-07T11:44:07.792536042+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:07.792954176+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13695}
+{"time":"2025-11-07T11:44:07.7949249+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:44:12.800927213+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":13985}
+{"time":"2025-11-07T11:44:12.80102633+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:17.805630099+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14042}
+{"time":"2025-11-07T11:44:17.805734867+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:17.806179529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14074}
+{"time":"2025-11-07T11:44:17.808661218+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":164}
+{"time":"2025-11-07T11:44:27.824065222+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14491}
+{"time":"2025-11-07T11:44:27.824150075+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:27.824674734+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14523}
+{"time":"2025-11-07T11:44:27.826602931+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":127}
+{"time":"2025-11-07T11:44:32.831281251+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14695}
+{"time":"2025-11-07T11:44:32.831405188+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:32.831830337+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14727}
+{"time":"2025-11-07T11:44:32.834787739+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:44:37.839903384+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14940}
+{"time":"2025-11-07T11:44:37.840021384+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:37.840424921+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":14972}
+{"time":"2025-11-07T11:44:37.842622733+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:44:42.849801952+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15221}
+{"time":"2025-11-07T11:44:42.849920595+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:42.85047043+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15253}
+{"time":"2025-11-07T11:44:42.851270807+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":55}
+{"time":"2025-11-07T11:44:52.862368816+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15598}
+{"time":"2025-11-07T11:44:52.862509422+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:52.862926965+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15630}
+{"time":"2025-11-07T11:44:52.864765704+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:44:57.871920739+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15857}
+{"time":"2025-11-07T11:44:57.87202618+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:44:57.872475971+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":15889}
+{"time":"2025-11-07T11:44:57.875967942+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":77}
+{"time":"2025-11-07T11:45:02.876968135+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16011}
+{"time":"2025-11-07T11:45:02.877128172+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:02.877540803+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16043}
+{"time":"2025-11-07T11:45:02.880386652+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:45:07.884045898+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16256}
+{"time":"2025-11-07T11:45:07.884170445+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:07.884614273+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16288}
+{"time":"2025-11-07T11:45:07.886683747+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:45:12.891600076+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16530}
+{"time":"2025-11-07T11:45:12.89169361+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:12.892498149+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16562}
+{"time":"2025-11-07T11:45:12.894704461+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":62}
+{"time":"2025-11-07T11:45:17.899683889+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16676}
+{"time":"2025-11-07T11:45:17.899816791+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:17.901043785+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16709}
+{"time":"2025-11-07T11:45:17.902366846+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":120}
+{"time":"2025-11-07T11:45:22.908893062+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16873}
+{"time":"2025-11-07T11:45:22.909034362+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:22.909438564+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":16905}
+{"time":"2025-11-07T11:45:22.911896535+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:45:32.922413664+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17286}
+{"time":"2025-11-07T11:45:32.925578932+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":201}
+{"time":"2025-11-07T11:45:37.928785471+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17531}
+{"time":"2025-11-07T11:45:37.92890613+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:37.929281337+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17563}
+{"time":"2025-11-07T11:45:37.931480986+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:45:42.937122793+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17763}
+{"time":"2025-11-07T11:45:42.937251269+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:42.937711051+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17795}
+{"time":"2025-11-07T11:45:42.938896118+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":104}
+{"time":"2025-11-07T11:45:47.942910509+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17943}
+{"time":"2025-11-07T11:45:47.942982195+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:47.943490828+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":17975}
+{"time":"2025-11-07T11:45:47.947244993+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:45:52.949420506+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18147}
+{"time":"2025-11-07T11:45:52.949514985+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:52.950004019+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18179}
+{"time":"2025-11-07T11:45:52.952600367+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":130}
+{"time":"2025-11-07T11:45:57.955431662+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18373}
+{"time":"2025-11-07T11:45:57.955549968+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:45:57.956075035+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18405}
+{"time":"2025-11-07T11:45:57.958080082+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":111}
+{"time":"2025-11-07T11:46:02.962215582+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18560}
+{"time":"2025-11-07T11:46:02.962352006+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:02.962782357+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18592}
+{"time":"2025-11-07T11:46:02.965017435+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:46:12.975383856+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":18976}
+{"time":"2025-11-07T11:46:12.975473374+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:12.976563599+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19009}
+{"time":"2025-11-07T11:46:12.977770567+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":124}
+{"time":"2025-11-07T11:46:18.002776802+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19177}
+{"time":"2025-11-07T11:46:18.002870522+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:18.003291163+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19209}
+{"time":"2025-11-07T11:46:18.005674831+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:46:23.011463805+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19382}
+{"time":"2025-11-07T11:46:23.011571063+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:23.012006876+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19414}
+{"time":"2025-11-07T11:46:23.013868457+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:46:28.017526624+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19591}
+{"time":"2025-11-07T11:46:28.017591563+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:28.01811408+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19623}
+{"time":"2025-11-07T11:46:28.019987083+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":127}
+{"time":"2025-11-07T11:46:33.023987343+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19849}
+{"time":"2025-11-07T11:46:33.024119534+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:33.024516909+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":19881}
+{"time":"2025-11-07T11:46:33.025783512+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":115}
+{"time":"2025-11-07T11:46:38.030363393+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20124}
+{"time":"2025-11-07T11:46:38.030422395+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:38.030944024+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20156}
+{"time":"2025-11-07T11:46:38.035101543+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":291}
+{"time":"2025-11-07T11:46:43.040158502+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20590}
+{"time":"2025-11-07T11:46:43.040264828+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:43.040949214+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20622}
+{"time":"2025-11-07T11:46:43.041658477+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":32}
+{"time":"2025-11-07T11:46:48.047209314+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20694}
+{"time":"2025-11-07T11:46:48.047335694+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:48.047724259+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20726}
+{"time":"2025-11-07T11:46:48.04929921+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:46:53.055641125+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20939}
+{"time":"2025-11-07T11:46:53.056229089+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:53.056655773+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":20971}
+{"time":"2025-11-07T11:46:53.058495037+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":216}
+{"time":"2025-11-07T11:46:58.063629875+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21255}
+{"time":"2025-11-07T11:46:58.063725043+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:46:58.064309083+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21287}
+{"time":"2025-11-07T11:46:58.067136483+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":230}
+{"time":"2025-11-07T11:47:03.071972518+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21575}
+{"time":"2025-11-07T11:47:03.072329463+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:03.073282129+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21608}
+{"time":"2025-11-07T11:47:03.075240773+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":196}
+{"time":"2025-11-07T11:47:08.08023138+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21927}
+{"time":"2025-11-07T11:47:08.080332319+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:08.080806537+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":21959}
+{"time":"2025-11-07T11:47:08.081846433+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":91}
+{"time":"2025-11-07T11:47:13.087274352+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22097}
+{"time":"2025-11-07T11:47:13.087393376+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:13.088418234+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22130}
+{"time":"2025-11-07T11:47:13.091504268+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":168}
+{"time":"2025-11-07T11:47:18.095261572+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22339}
+{"time":"2025-11-07T11:47:18.095400398+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:18.096416555+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22372}
+{"time":"2025-11-07T11:47:18.097794709+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:47:23.105291621+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22561}
+{"time":"2025-11-07T11:47:23.105406739+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:23.105878408+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22593}
+{"time":"2025-11-07T11:47:23.107545761+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":33}
+{"time":"2025-11-07T11:47:28.111027079+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22778}
+{"time":"2025-11-07T11:47:28.111117686+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:38.137781178+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":22998}
+{"time":"2025-11-07T11:47:38.13789674+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:38.138382103+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":23030}
+{"time":"2025-11-07T11:47:38.140592247+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:47:43.144192278+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":23164}
+{"time":"2025-11-07T11:47:43.144331039+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:43.144802687+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":23196}
+{"time":"2025-11-07T11:47:43.145965834+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":90}
+{"time":"2025-11-07T11:47:53.159004462+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":23694}
+{"time":"2025-11-07T11:47:53.159111791+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:53.159563409+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":23726}
+{"time":"2025-11-07T11:47:53.160848947+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":134}
+{"time":"2025-11-07T11:47:58.170361195+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24052}
+{"time":"2025-11-07T11:47:58.170454105+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:47:58.170865929+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24084}
+{"time":"2025-11-07T11:47:58.171831454+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":65}
+{"time":"2025-11-07T11:48:03.175988921+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24227}
+{"time":"2025-11-07T11:48:03.176063544+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:03.176539079+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24259}
+{"time":"2025-11-07T11:48:03.177332865+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":54}
+{"time":"2025-11-07T11:48:08.184184489+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24387}
+{"time":"2025-11-07T11:48:08.184592316+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:08.185568697+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24420}
+{"time":"2025-11-07T11:48:08.188091842+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":262}
+{"time":"2025-11-07T11:48:13.193424885+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24798}
+{"time":"2025-11-07T11:48:13.19353719+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:13.193903479+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":24830}
+{"time":"2025-11-07T11:48:13.195488539+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":141}
+{"time":"2025-11-07T11:48:18.200349182+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25022}
+{"time":"2025-11-07T11:48:18.20045368+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:18.20093742+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25054}
+{"time":"2025-11-07T11:48:18.205370732+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":204}
+{"time":"2025-11-07T11:48:23.208316506+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25314}
+{"time":"2025-11-07T11:48:23.208463306+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:23.208880729+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25346}
+{"time":"2025-11-07T11:48:23.210855293+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":117}
+{"time":"2025-11-07T11:48:33.222558065+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25782}
+{"time":"2025-11-07T11:48:33.222694197+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:33.223090867+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25814}
+{"time":"2025-11-07T11:48:33.22366645+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":20}
+{"time":"2025-11-07T11:48:38.22685519+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25874}
+{"time":"2025-11-07T11:48:38.226999234+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:38.227463101+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":25906}
+{"time":"2025-11-07T11:48:38.230044906+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:48:43.232042208+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26002}
+{"time":"2025-11-07T11:48:43.232181602+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:43.232677446+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26034}
+{"time":"2025-11-07T11:48:43.23376266+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:48:48.238554663+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26179}
+{"time":"2025-11-07T11:48:48.238705973+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:48.239323632+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26211}
+{"time":"2025-11-07T11:48:48.242769481+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":281}
+{"time":"2025-11-07T11:48:53.248385514+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26593}
+{"time":"2025-11-07T11:48:53.24846217+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:53.248983046+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26625}
+{"time":"2025-11-07T11:48:53.251116134+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":31}
+{"time":"2025-11-07T11:48:58.256937334+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26846}
+{"time":"2025-11-07T11:48:58.257009959+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:48:58.257422658+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":26878}
+{"time":"2025-11-07T11:48:58.258560903+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":108}
+{"time":"2025-11-07T11:49:03.26432435+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27084}
+{"time":"2025-11-07T11:49:03.264500377+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:03.264905827+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27116}
+{"time":"2025-11-07T11:49:03.267260194+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":239}
+{"time":"2025-11-07T11:49:08.270665668+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27482}
+{"time":"2025-11-07T11:49:08.270781926+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:08.271299502+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27514}
+{"time":"2025-11-07T11:49:08.27381606+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:49:13.278779995+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27879}
+{"time":"2025-11-07T11:49:13.278836748+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:13.2793335+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27911}
+{"time":"2025-11-07T11:49:13.280049999+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":20}
+{"time":"2025-11-07T11:49:18.285269942+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":27972}
+{"time":"2025-11-07T11:49:18.285453656+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:18.285928766+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28004}
+{"time":"2025-11-07T11:49:18.289921616+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":91}
+{"time":"2025-11-07T11:49:23.290701389+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28140}
+{"time":"2025-11-07T11:49:23.290841452+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:23.291379964+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28172}
+{"time":"2025-11-07T11:49:23.294277256+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:49:33.305475644+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28632}
+{"time":"2025-11-07T11:49:33.305546996+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:33.30598751+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28664}
+{"time":"2025-11-07T11:49:33.306511722+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":7}
+{"time":"2025-11-07T11:49:38.314349547+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":28972}
+{"time":"2025-11-07T11:49:38.315423559+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":27}
+{"time":"2025-11-07T11:49:43.319667199+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29053}
+{"time":"2025-11-07T11:49:43.319724192+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:43.320773258+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29086}
+{"time":"2025-11-07T11:49:43.321780874+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":79}
+{"time":"2025-11-07T11:49:48.325453712+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29568}
+{"time":"2025-11-07T11:49:48.325547621+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:48.32596133+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29600}
+{"time":"2025-11-07T11:49:48.327417796+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":139}
+{"time":"2025-11-07T11:49:53.331943251+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29787}
+{"time":"2025-11-07T11:49:53.33249561+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:53.332604071+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":29819}
+{"time":"2025-11-07T11:49:53.335861382+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":207}
+{"time":"2025-11-07T11:49:58.339813789+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30084}
+{"time":"2025-11-07T11:49:58.33994278+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:49:58.340395036+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30116}
+{"time":"2025-11-07T11:49:58.343429417+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":240}
+{"time":"2025-11-07T11:50:03.347873973+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30424}
+{"time":"2025-11-07T11:50:03.347986349+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:03.34843142+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30456}
+{"time":"2025-11-07T11:50:03.352095501+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":228}
+{"time":"2025-11-07T11:50:08.356569123+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30727}
+{"time":"2025-11-07T11:50:08.356670691+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:08.357101421+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30759}
+{"time":"2025-11-07T11:50:08.359069562+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":130}
+{"time":"2025-11-07T11:50:13.36273032+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30939}
+{"time":"2025-11-07T11:50:13.362806631+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:13.363347451+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":30971}
+{"time":"2025-11-07T11:50:13.364817996+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":125}
+{"time":"2025-11-07T11:50:18.370334354+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31144}
+{"time":"2025-11-07T11:50:18.370451983+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:18.371506829+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31177}
+{"time":"2025-11-07T11:50:18.373162903+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":124}
+{"time":"2025-11-07T11:50:23.377822992+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31363}
+{"time":"2025-11-07T11:50:23.378932101+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31395}
+{"time":"2025-11-07T11:50:23.378236537+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:23.380953653+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":70}
+{"time":"2025-11-07T11:50:28.385018583+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31590}
+{"time":"2025-11-07T11:50:28.385122605+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:28.385570266+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31622}
+{"time":"2025-11-07T11:50:28.387017158+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":9}
+{"time":"2025-11-07T11:50:33.39211735+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31743}
+{"time":"2025-11-07T11:50:33.39221626+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:33.392928434+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31775}
+{"time":"2025-11-07T11:50:33.394140284+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":102}
+{"time":"2025-11-07T11:50:38.398600536+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31933}
+{"time":"2025-11-07T11:50:38.398696157+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:43.405005013+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":31999}
+{"time":"2025-11-07T11:50:43.407472593+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":44}
+{"time":"2025-11-07T11:50:48.410849042+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32083}
+{"time":"2025-11-07T11:50:48.4110039+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:48.411385841+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32115}
+{"time":"2025-11-07T11:50:48.413225833+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:50:58.424581105+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32288}
+{"time":"2025-11-07T11:50:58.424666557+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:50:58.425071693+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32320}
+{"time":"2025-11-07T11:50:58.426069733+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:51:18.452686806+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32620}
+{"time":"2025-11-07T11:51:18.452754436+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:51:18.453245055+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32652}
+{"time":"2025-11-07T11:51:18.45611169+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:51:23.458016688+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32742}
+{"time":"2025-11-07T11:51:23.458214599+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:51:23.458627947+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":32774}
+{"time":"2025-11-07T11:51:23.460127937+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":11}
+{"time":"2025-11-07T11:51:48.508996748+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33200}
+{"time":"2025-11-07T11:51:48.509130891+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:51:48.509604177+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33232}
+{"time":"2025-11-07T11:51:48.510509234+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":8}
+{"time":"2025-11-07T11:51:53.514021579+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33290}
+{"time":"2025-11-07T11:51:53.514143533+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:03.527763868+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33443}
+{"time":"2025-11-07T11:52:03.527927322+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:03.528270726+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33475}
+{"time":"2025-11-07T11:52:03.529163424+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":13}
+{"time":"2025-11-07T11:52:13.54454995+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33609}
+{"time":"2025-11-07T11:52:13.544666798+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:13.545659887+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33642}
+{"time":"2025-11-07T11:52:13.546609337+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:52:18.550058474+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33763}
+{"time":"2025-11-07T11:52:18.55013781+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:23.555655712+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33815}
+{"time":"2025-11-07T11:52:23.555743858+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:23.55619846+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33847}
+{"time":"2025-11-07T11:52:23.557959761+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:52:33.568047217+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":33981}
+{"time":"2025-11-07T11:52:33.568177421+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:33.568618148+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34013}
+{"time":"2025-11-07T11:52:33.569753997+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:52:38.574669714+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34067}
+{"time":"2025-11-07T11:52:38.574776557+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:38.575194797+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34099}
+{"time":"2025-11-07T11:52:38.576912068+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":8}
+{"time":"2025-11-07T11:52:43.581609541+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34149}
+{"time":"2025-11-07T11:52:43.581712993+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:43.582152868+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34181}
+{"time":"2025-11-07T11:52:43.582862064+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":10}
+{"time":"2025-11-07T11:52:48.587218724+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34231}
+{"time":"2025-11-07T11:52:48.587325062+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:52:48.587800529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34263}
+{"time":"2025-11-07T11:52:48.589583567+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:52:58.616666209+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34470}
+{"time":"2025-11-07T11:52:58.616815837+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:03.625429893+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34548}
+{"time":"2025-11-07T11:53:03.625549889+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:13.640430585+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34878}
+{"time":"2025-11-07T11:53:13.640579425+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:13.640954207+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":34910}
+{"time":"2025-11-07T11:53:13.644883376+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":392}
+{"time":"2025-11-07T11:53:18.668552896+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":35352}
+{"time":"2025-11-07T11:53:18.668679577+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:18.66973734+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":35385}
+{"time":"2025-11-07T11:53:18.671606922+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":204}
+{"time":"2025-11-07T11:53:23.677199512+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":35687}
+{"time":"2025-11-07T11:53:23.67728983+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:23.677775926+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":35719}
+{"time":"2025-11-07T11:53:23.679416808+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":186}
+{"time":"2025-11-07T11:53:28.686762405+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36003}
+{"time":"2025-11-07T11:53:28.686893597+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:28.687325261+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36035}
+{"time":"2025-11-07T11:53:28.690336571+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:53:33.697765209+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36185}
+{"time":"2025-11-07T11:53:33.69788075+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:33.698630218+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36217}
+{"time":"2025-11-07T11:53:33.699673648+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":71}
+{"time":"2025-11-07T11:53:38.704769891+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36334}
+{"time":"2025-11-07T11:53:38.704869209+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:38.705264577+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36366}
+{"time":"2025-11-07T11:53:38.707883748+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":168}
+{"time":"2025-11-07T11:53:43.712732408+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36574}
+{"time":"2025-11-07T11:53:43.712865202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:43.713215641+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36606}
+{"time":"2025-11-07T11:53:43.714757542+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:53:48.71990907+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36692}
+{"time":"2025-11-07T11:53:48.720019764+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:58.730951607+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36839}
+{"time":"2025-11-07T11:53:58.731067292+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:53:58.7315243+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36871}
+{"time":"2025-11-07T11:53:58.733168341+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":36}
+{"time":"2025-11-07T11:54:03.735618487+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36946}
+{"time":"2025-11-07T11:54:03.735734583+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:54:03.736151704+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":36978}
+{"time":"2025-11-07T11:54:03.737706961+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":11}
+{"time":"2025-11-07T11:54:13.748675+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37113}
+{"time":"2025-11-07T11:54:13.748744809+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:54:13.749859048+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37146}
+{"time":"2025-11-07T11:54:13.750567066+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":9}
+{"time":"2025-11-07T11:54:28.772445365+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37408}
+{"time":"2025-11-07T11:54:28.772601661+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:54:28.773006853+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37440}
+{"time":"2025-11-07T11:54:28.773746336+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":4}
+{"time":"2025-11-07T11:54:43.790056466+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37654}
+{"time":"2025-11-07T11:54:43.79016807+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:54:43.790632993+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37686}
+{"time":"2025-11-07T11:54:43.791249576+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":6}
+{"time":"2025-11-07T11:54:53.804765613+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37830}
+{"time":"2025-11-07T11:54:53.804853199+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:54:53.805327632+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":37862}
+{"time":"2025-11-07T11:54:53.810860951+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":363}
+{"time":"2025-11-07T11:54:58.814269118+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":38380}
+{"time":"2025-11-07T11:54:58.814331068+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:03.820397803+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":38437}
+{"time":"2025-11-07T11:55:03.820569742+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:03.821054698+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":38469}
+{"time":"2025-11-07T11:55:03.823558493+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":127}
+{"time":"2025-11-07T11:55:08.829186723+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":38653}
+{"time":"2025-11-07T11:55:08.829284727+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:08.83017729+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":38686}
+{"time":"2025-11-07T11:55:08.83458366+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":325}
+{"time":"2025-11-07T11:55:13.853323634+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39096}
+{"time":"2025-11-07T11:55:13.853457973+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:13.853900685+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39128}
+{"time":"2025-11-07T11:55:13.856651254+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":208}
+{"time":"2025-11-07T11:55:18.860479932+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39383}
+{"time":"2025-11-07T11:55:18.860635886+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:18.861053302+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39415}
+{"time":"2025-11-07T11:55:18.863768296+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":208}
+{"time":"2025-11-07T11:55:23.870187681+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39727}
+{"time":"2025-11-07T11:55:23.870246384+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:23.870803881+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39759}
+{"time":"2025-11-07T11:55:23.871966428+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":110}
+{"time":"2025-11-07T11:55:28.877354373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39914}
+{"time":"2025-11-07T11:55:28.877521084+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:28.878472765+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":39947}
+{"time":"2025-11-07T11:55:28.879867441+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:55:33.883531444+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40188}
+{"time":"2025-11-07T11:55:33.883650386+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:33.884654691+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40221}
+{"time":"2025-11-07T11:55:33.885519468+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":19}
+{"time":"2025-11-07T11:55:38.890677439+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40284}
+{"time":"2025-11-07T11:55:38.890803123+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:38.891310682+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40316}
+{"time":"2025-11-07T11:55:38.893275+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:55:43.899156669+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40485}
+{"time":"2025-11-07T11:55:43.899273184+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:43.899662279+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40517}
+{"time":"2025-11-07T11:55:43.900369746+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:55:48.904334263+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40573}
+{"time":"2025-11-07T11:55:48.904431827+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:48.90517039+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40605}
+{"time":"2025-11-07T11:55:48.90741307+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:55:53.911250247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40841}
+{"time":"2025-11-07T11:55:53.911334643+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:53.911845565+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":40873}
+{"time":"2025-11-07T11:55:53.914549878+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":230}
+{"time":"2025-11-07T11:55:58.918651404+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41151}
+{"time":"2025-11-07T11:55:58.918736976+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:55:58.919198558+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41183}
+{"time":"2025-11-07T11:55:58.921529659+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":86}
+{"time":"2025-11-07T11:56:03.924839438+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41458}
+{"time":"2025-11-07T11:56:03.924941659+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:03.925535229+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41490}
+{"time":"2025-11-07T11:56:03.926761531+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":66}
+{"time":"2025-11-07T11:56:08.932924853+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41767}
+{"time":"2025-11-07T11:56:08.933033534+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:08.93507659+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":41864}
+{"time":"2025-11-07T11:56:08.93623962+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":102}
+{"time":"2025-11-07T11:56:18.947908947+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":42341}
+{"time":"2025-11-07T11:56:18.948021904+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:18.948497405+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":42373}
+{"time":"2025-11-07T11:56:18.95044537+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:56:28.976847669+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":42830}
+{"time":"2025-11-07T11:56:28.97694912+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:28.977418568+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":42862}
+{"time":"2025-11-07T11:56:28.978605715+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":51}
+{"time":"2025-11-07T11:56:39.06960205+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43074}
+{"time":"2025-11-07T11:56:39.069673604+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:39.070240373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43106}
+{"time":"2025-11-07T11:56:39.07134157+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:56:54.089830813+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43419}
+{"time":"2025-11-07T11:56:54.089938957+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:54.090469247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43451}
+{"time":"2025-11-07T11:56:54.092288774+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":79}
+{"time":"2025-11-07T11:56:59.096542676+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43581}
+{"time":"2025-11-07T11:56:59.096698365+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:56:59.097086284+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43613}
+{"time":"2025-11-07T11:56:59.100114302+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":247}
+{"time":"2025-11-07T11:57:04.103734509+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43901}
+{"time":"2025-11-07T11:57:04.103938991+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:04.10422249+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":43933}
+{"time":"2025-11-07T11:57:04.107721724+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":91}
+{"time":"2025-11-07T11:57:09.11459218+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44131}
+{"time":"2025-11-07T11:57:09.114672972+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:09.115250738+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44163}
+{"time":"2025-11-07T11:57:09.118048545+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":148}
+{"time":"2025-11-07T11:57:14.121899576+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44366}
+{"time":"2025-11-07T11:57:14.122105364+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:14.122476232+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44398}
+{"time":"2025-11-07T11:57:14.128207286+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":325}
+{"time":"2025-11-07T11:57:19.133900447+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44846}
+{"time":"2025-11-07T11:57:19.133988012+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:19.134837407+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":44879}
+{"time":"2025-11-07T11:57:19.136228279+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":131}
+{"time":"2025-11-07T11:57:24.141371865+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45057}
+{"time":"2025-11-07T11:57:24.141508714+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:24.14191876+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45089}
+{"time":"2025-11-07T11:57:24.144190153+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":167}
+{"time":"2025-11-07T11:57:29.150287393+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45423}
+{"time":"2025-11-07T11:57:29.150408118+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:29.150887034+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45456}
+{"time":"2025-11-07T11:57:29.152829171+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:57:39.164345419+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45680}
+{"time":"2025-11-07T11:57:39.164538992+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:39.164928943+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":45712}
+{"time":"2025-11-07T11:57:39.168742186+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":327}
+{"time":"2025-11-07T11:57:44.196614684+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":46157}
+{"time":"2025-11-07T11:57:44.196688775+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:44.197190152+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":46189}
+{"time":"2025-11-07T11:57:44.198997729+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":55}
+{"time":"2025-11-07T11:57:54.21112788+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":46608}
+{"time":"2025-11-07T11:57:54.211206421+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:54.211721383+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":46640}
+{"time":"2025-11-07T11:57:54.214073105+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":260}
+{"time":"2025-11-07T11:57:59.218196436+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":46984}
+{"time":"2025-11-07T11:57:59.218380235+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:57:59.228591716+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47016}
+{"time":"2025-11-07T11:57:59.231256449+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":184}
+{"time":"2025-11-07T11:58:04.243285085+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47276}
+{"time":"2025-11-07T11:58:04.243460234+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:04.2444614+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47309}
+{"time":"2025-11-07T11:58:04.246001664+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":167}
+{"time":"2025-11-07T11:58:09.25143005+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47518}
+{"time":"2025-11-07T11:58:09.251567869+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:09.252741592+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47551}
+{"time":"2025-11-07T11:58:09.255204288+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:58:14.258054923+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47687}
+{"time":"2025-11-07T11:58:14.258155351+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:14.258600844+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":47719}
+{"time":"2025-11-07T11:58:14.261035223+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":169}
+{"time":"2025-11-07T11:58:29.283345195+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48092}
+{"time":"2025-11-07T11:58:29.283451569+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:29.283876021+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48124}
+{"time":"2025-11-07T11:58:29.284929577+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2025-11-07T11:58:34.292672342+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48217}
+{"time":"2025-11-07T11:58:34.296363436+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":206}
+{"time":"2025-11-07T11:58:39.301723106+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48472}
+{"time":"2025-11-07T11:58:39.301877206+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:39.302264966+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48504}
+{"time":"2025-11-07T11:58:39.30549257+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":247}
+{"time":"2025-11-07T11:58:49.439349235+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":48934}
+{"time":"2025-11-07T11:58:49.444765064+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":434}
+{"time":"2025-11-07T11:58:54.452100215+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":49684}
+{"time":"2025-11-07T11:58:54.45217833+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:54.452655114+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":49716}
+{"time":"2025-11-07T11:58:54.453230878+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":21}
+{"time":"2025-11-07T11:58:59.460668682+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":49849}
+{"time":"2025-11-07T11:58:59.460759786+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:58:59.46139993+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":49881}
+{"time":"2025-11-07T11:58:59.463041548+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":189}
+{"time":"2025-11-07T11:59:04.488582509+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50155}
+{"time":"2025-11-07T11:59:04.488733989+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:04.489167606+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50187}
+{"time":"2025-11-07T11:59:04.491687913+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":208}
+{"time":"2025-11-07T11:59:09.495422862+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50443}
+{"time":"2025-11-07T11:59:09.499421952+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":239}
+{"time":"2025-11-07T11:59:14.504203958+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50728}
+{"time":"2025-11-07T11:59:14.504289441+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:14.504775273+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50760}
+{"time":"2025-11-07T11:59:14.507047541+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":129}
+{"time":"2025-11-07T11:59:19.51158445+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":50971}
+{"time":"2025-11-07T11:59:19.51167864+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:19.512184679+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51003}
+{"time":"2025-11-07T11:59:19.51379756+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":50}
+{"time":"2025-11-07T11:59:24.518735651+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51137}
+{"time":"2025-11-07T11:59:24.518949588+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:24.51993778+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51170}
+{"time":"2025-11-07T11:59:24.520542645+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":47}
+{"time":"2025-11-07T11:59:29.524471723+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51283}
+{"time":"2025-11-07T11:59:29.524607851+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:29.525032282+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51315}
+{"time":"2025-11-07T11:59:29.526923176+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":68}
+{"time":"2025-11-07T11:59:34.530836908+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51523}
+{"time":"2025-11-07T11:59:34.530927214+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:39.535859565+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51589}
+{"time":"2025-11-07T11:59:39.535993914+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:39.536897448+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51622}
+{"time":"2025-11-07T11:59:39.538218136+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T11:59:44.543004684+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51889}
+{"time":"2025-11-07T11:59:44.543066841+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:44.543539011+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":51921}
+{"time":"2025-11-07T11:59:44.548490021+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":448}
+{"time":"2025-11-07T11:59:49.550207221+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":52438}
+{"time":"2025-11-07T11:59:49.550404407+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:49.550851771+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":52470}
+{"time":"2025-11-07T11:59:49.557660384+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T11:59:54.560099881+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":53286}
+{"time":"2025-11-07T11:59:54.560270755+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:54.560750835+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":53318}
+{"time":"2025-11-07T11:59:54.566347296+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":527}
+{"time":"2025-11-07T11:59:59.57015269+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":53970}
+{"time":"2025-11-07T11:59:59.570272028+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T11:59:59.570834928+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":54002}
+{"time":"2025-11-07T11:59:59.57674833+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":556}
+{"time":"2025-11-07T12:00:04.593104129+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":54820}
+{"time":"2025-11-07T12:00:04.593241262+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:04.593626283+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":54852}
+{"time":"2025-11-07T12:00:04.598310907+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":430}
+{"time":"2025-11-07T12:00:09.603702363+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":55352}
+{"time":"2025-11-07T12:00:09.603872209+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:09.604269929+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":55384}
+{"time":"2025-11-07T12:00:09.610431465+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:00:14.615090322+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":56335}
+{"time":"2025-11-07T12:00:14.615696794+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:14.616118157+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":56369}
+{"time":"2025-11-07T12:00:14.619567146+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":391}
+{"time":"2025-11-07T12:00:19.62391205+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":56838}
+{"time":"2025-11-07T12:00:19.624041428+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:19.624579143+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":56870}
+{"time":"2025-11-07T12:00:19.631004754+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":628}
+{"time":"2025-11-07T12:00:24.634807911+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":57595}
+{"time":"2025-11-07T12:00:24.634918543+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:24.635407408+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":57627}
+{"time":"2025-11-07T12:00:24.641307812+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":609}
+{"time":"2025-11-07T12:00:29.645595571+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":58442}
+{"time":"2025-11-07T12:00:29.650922153+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":551}
+{"time":"2025-11-07T12:00:34.66912916+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":59395}
+{"time":"2025-11-07T12:00:34.669323641+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:34.66989581+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":59427}
+{"time":"2025-11-07T12:00:34.672135732+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":287}
+{"time":"2025-11-07T12:00:39.676163209+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":59784}
+{"time":"2025-11-07T12:00:39.676249515+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:39.676746591+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":59816}
+{"time":"2025-11-07T12:00:39.683596763+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:00:44.686577562+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":60521}
+{"time":"2025-11-07T12:00:44.686734155+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:44.687107116+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":60553}
+{"time":"2025-11-07T12:00:44.693349032+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":598}
+{"time":"2025-11-07T12:00:49.696920366+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":61220}
+{"time":"2025-11-07T12:00:49.697144288+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:49.697394968+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":61252}
+{"time":"2025-11-07T12:00:49.703310712+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":639}
+{"time":"2025-11-07T12:00:54.721072098+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":62019}
+{"time":"2025-11-07T12:00:54.721151063+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:54.721599232+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":62051}
+{"time":"2025-11-07T12:00:54.726936689+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":576}
+{"time":"2025-11-07T12:00:59.731456794+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":62849}
+{"time":"2025-11-07T12:00:59.731543292+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:00:59.731948305+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":62881}
+{"time":"2025-11-07T12:00:59.736944014+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":486}
+{"time":"2025-11-07T12:01:04.740192699+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":63684}
+{"time":"2025-11-07T12:01:04.740271825+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:04.740825253+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":63716}
+{"time":"2025-11-07T12:01:04.745502713+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":389}
+{"time":"2025-11-07T12:01:09.748185932+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":64224}
+{"time":"2025-11-07T12:01:09.748287292+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:09.748815725+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":64256}
+{"time":"2025-11-07T12:01:09.754871848+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":587}
+{"time":"2025-11-07T12:01:14.757757069+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":64921}
+{"time":"2025-11-07T12:01:14.757925127+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:14.758274647+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":64953}
+{"time":"2025-11-07T12:01:14.764800993+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:01:19.770558717+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":65665}
+{"time":"2025-11-07T12:01:19.770626337+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:19.771022878+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":65697}
+{"time":"2025-11-07T12:01:19.777596373+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":583}
+{"time":"2025-11-07T12:01:24.780963472+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":66384}
+{"time":"2025-11-07T12:01:24.781100776+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:24.781506694+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":66416}
+{"time":"2025-11-07T12:01:24.786933709+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":602}
+{"time":"2025-11-07T12:01:29.793987271+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":67214}
+{"time":"2025-11-07T12:01:29.794089769+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:29.794479165+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":67246}
+{"time":"2025-11-07T12:01:29.797961516+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":512}
+{"time":"2025-11-07T12:01:34.8014957+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":67876}
+{"time":"2025-11-07T12:01:34.801606159+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:34.80301878+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":67909}
+{"time":"2025-11-07T12:01:34.807739758+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":610}
+{"time":"2025-11-07T12:01:39.847522092+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":68973}
+{"time":"2025-11-07T12:01:39.847642332+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:39.848031864+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":69005}
+{"time":"2025-11-07T12:01:39.850027242+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":270}
+{"time":"2025-11-07T12:01:44.854048363+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":69351}
+{"time":"2025-11-07T12:01:44.854172864+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:44.854658153+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":69383}
+{"time":"2025-11-07T12:01:44.859078495+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":632}
+{"time":"2025-11-07T12:01:49.865302089+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":70272}
+{"time":"2025-11-07T12:01:49.865368535+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:49.865833438+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":70304}
+{"time":"2025-11-07T12:01:49.868590338+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":408}
+{"time":"2025-11-07T12:01:54.872515072+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":70786}
+{"time":"2025-11-07T12:01:54.872629161+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:54.87306208+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":70818}
+{"time":"2025-11-07T12:01:54.879215935+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":673}
+{"time":"2025-11-07T12:01:59.881269363+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":71563}
+{"time":"2025-11-07T12:01:59.881355344+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:01:59.881892464+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":71595}
+{"time":"2025-11-07T12:01:59.887771856+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:02:04.897599303+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":72632}
+{"time":"2025-11-07T12:02:04.897702718+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:04.898623316+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":72665}
+{"time":"2025-11-07T12:02:04.901557+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":304}
+{"time":"2025-11-07T12:02:09.906427432+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":73039}
+{"time":"2025-11-07T12:02:09.906513549+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:09.907015287+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":73071}
+{"time":"2025-11-07T12:02:09.913731631+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:02:14.918367745+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":73899}
+{"time":"2025-11-07T12:02:14.918448356+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:14.918857956+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":73931}
+{"time":"2025-11-07T12:02:14.923853588+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":516}
+{"time":"2025-11-07T12:02:19.927502126+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":74592}
+{"time":"2025-11-07T12:02:19.927634277+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:19.928160088+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":74624}
+{"time":"2025-11-07T12:02:19.934516534+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":575}
+{"time":"2025-11-07T12:02:24.94168514+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":75448}
+{"time":"2025-11-07T12:02:24.941845838+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:24.942240926+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":75480}
+{"time":"2025-11-07T12:02:24.947147104+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":443}
+{"time":"2025-11-07T12:02:29.962932575+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":76009}
+{"time":"2025-11-07T12:02:29.963041063+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:29.963623874+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":76041}
+{"time":"2025-11-07T12:02:29.96933309+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":622}
+{"time":"2025-11-07T12:02:34.972857763+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":76751}
+{"time":"2025-11-07T12:02:34.972955388+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:34.973424791+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":76783}
+{"time":"2025-11-07T12:02:34.979223091+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":618}
+{"time":"2025-11-07T12:02:39.9836319+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":77471}
+{"time":"2025-11-07T12:02:39.983727785+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:39.985866222+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":77505}
+{"time":"2025-11-07T12:02:39.985950815+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:39.986005818+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":77537}
+{"time":"2025-11-07T12:02:39.990224168+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":602}
+{"time":"2025-11-07T12:02:44.997162985+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":78571}
+{"time":"2025-11-07T12:02:44.997378223+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:44.997903401+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":78603}
+{"time":"2025-11-07T12:02:45.000117513+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":276}
+{"time":"2025-11-07T12:02:50.004011035+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":79009}
+{"time":"2025-11-07T12:02:50.00412343+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:50.005062529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":79042}
+{"time":"2025-11-07T12:02:50.010709984+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":575}
+{"time":"2025-11-07T12:02:55.023713679+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":79686}
+{"time":"2025-11-07T12:02:55.023929931+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:02:55.024274007+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":79718}
+{"time":"2025-11-07T12:02:55.030830027+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:03:00.035158677+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":80521}
+{"time":"2025-11-07T12:03:00.035626874+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:00.036336964+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":80554}
+{"time":"2025-11-07T12:03:00.039939219+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":541}
+{"time":"2025-11-07T12:03:05.044624789+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":81167}
+{"time":"2025-11-07T12:03:05.044710529+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:05.045107027+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":81199}
+{"time":"2025-11-07T12:03:05.052166034+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":663}
+{"time":"2025-11-07T12:03:10.064164099+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":82041}
+{"time":"2025-11-07T12:03:10.069747313+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":571}
+{"time":"2025-11-07T12:03:15.073828393+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":82682}
+{"time":"2025-11-07T12:03:15.073989666+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:15.074449743+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":82714}
+{"time":"2025-11-07T12:03:15.080562871+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":597}
+{"time":"2025-11-07T12:03:20.081921009+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":83380}
+{"time":"2025-11-07T12:03:20.082080961+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:20.082406755+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":83412}
+{"time":"2025-11-07T12:03:20.089152791+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:03:25.091770304+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":84118}
+{"time":"2025-11-07T12:03:25.091875512+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:25.092345912+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":84150}
+{"time":"2025-11-07T12:03:25.099284739+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:03:30.103835504+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":84858}
+{"time":"2025-11-07T12:03:30.103984209+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:30.104522605+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":84890}
+{"time":"2025-11-07T12:03:30.111103695+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:03:35.116303566+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":85613}
+{"time":"2025-11-07T12:03:35.116398787+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:35.116906329+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":85645}
+{"time":"2025-11-07T12:03:35.123630824+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":596}
+{"time":"2025-11-07T12:03:40.132710063+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":86329}
+{"time":"2025-11-07T12:03:40.132800093+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:40.133370133+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":86361}
+{"time":"2025-11-07T12:03:40.139649017+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":642}
+{"time":"2025-11-07T12:03:45.145403736+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":87160}
+{"time":"2025-11-07T12:03:45.14963107+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":583}
+{"time":"2025-11-07T12:03:50.153720757+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":87814}
+{"time":"2025-11-07T12:03:50.153818579+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:50.155051556+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":87847}
+{"time":"2025-11-07T12:03:50.160867443+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":634}
+{"time":"2025-11-07T12:03:55.164816099+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":88550}
+{"time":"2025-11-07T12:03:55.164941954+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:03:55.16549362+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":88582}
+{"time":"2025-11-07T12:03:55.172246605+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:04:00.17652714+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":89289}
+{"time":"2025-11-07T12:04:00.176612738+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:00.177093546+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":89321}
+{"time":"2025-11-07T12:04:00.183797416+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":602}
+{"time":"2025-11-07T12:04:05.19466506+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":90034}
+{"time":"2025-11-07T12:04:05.194737861+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:05.195274937+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":90066}
+{"time":"2025-11-07T12:04:05.201554296+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":590}
+{"time":"2025-11-07T12:04:10.205097019+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":90799}
+{"time":"2025-11-07T12:04:10.205217309+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:10.205717379+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":90831}
+{"time":"2025-11-07T12:04:10.21153049+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":563}
+{"time":"2025-11-07T12:04:15.218935913+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":91618}
+{"time":"2025-11-07T12:04:15.223504913+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":516}
+{"time":"2025-11-07T12:04:20.22794348+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":92204}
+{"time":"2025-11-07T12:04:20.22803011+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:20.22911269+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":92237}
+{"time":"2025-11-07T12:04:20.235091661+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":611}
+{"time":"2025-11-07T12:04:25.251330314+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":92946}
+{"time":"2025-11-07T12:04:25.251435197+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:25.251988715+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":92978}
+{"time":"2025-11-07T12:04:25.258306849+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":632}
+{"time":"2025-11-07T12:04:30.26204445+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":93682}
+{"time":"2025-11-07T12:04:30.26212778+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:30.262693474+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":93714}
+{"time":"2025-11-07T12:04:30.26865079+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:04:35.274505411+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":94418}
+{"time":"2025-11-07T12:04:35.274587369+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:35.27505082+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":94450}
+{"time":"2025-11-07T12:04:35.281178262+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":597}
+{"time":"2025-11-07T12:04:40.285210312+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":95146}
+{"time":"2025-11-07T12:04:40.285283412+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:40.285843+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":95178}
+{"time":"2025-11-07T12:04:40.29152296+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":607}
+{"time":"2025-11-07T12:04:45.297365518+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":95990}
+{"time":"2025-11-07T12:04:45.297902031+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":96022}
+{"time":"2025-11-07T12:04:45.297479966+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:45.302024944+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":503}
+{"time":"2025-11-07T12:04:50.306479071+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":96685}
+{"time":"2025-11-07T12:04:50.306961393+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:50.307692182+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":96718}
+{"time":"2025-11-07T12:04:50.31153293+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":545}
+{"time":"2025-11-07T12:04:55.31838011+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":97455}
+{"time":"2025-11-07T12:04:55.318709509+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:04:55.319326965+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":97487}
+{"time":"2025-11-07T12:04:55.322407782+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":473}
+{"time":"2025-11-07T12:05:00.326688254+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":98033}
+{"time":"2025-11-07T12:05:00.326875002+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:00.327311638+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":98065}
+{"time":"2025-11-07T12:05:00.333134269+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:05:05.338834224+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":99111}
+{"time":"2025-11-07T12:05:05.338951062+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:05.339380327+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":99143}
+{"time":"2025-11-07T12:05:05.343456634+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":295}
+{"time":"2025-11-07T12:05:10.3471648+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":99527}
+{"time":"2025-11-07T12:05:10.347228587+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:10.347738578+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":99559}
+{"time":"2025-11-07T12:05:10.353369312+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":617}
+{"time":"2025-11-07T12:05:15.358224262+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":100248}
+{"time":"2025-11-07T12:05:15.358324243+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:15.359419221+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":100281}
+{"time":"2025-11-07T12:05:15.36566223+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:05:20.369469677+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":100985}
+{"time":"2025-11-07T12:05:20.369648474+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:20.369967841+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":101017}
+{"time":"2025-11-07T12:05:20.37704147+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:05:25.381153724+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":101738}
+{"time":"2025-11-07T12:05:25.381227946+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:25.38183915+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":101770}
+{"time":"2025-11-07T12:05:25.388792936+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":622}
+{"time":"2025-11-07T12:05:30.392055613+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":102482}
+{"time":"2025-11-07T12:05:30.398092914+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":609}
+{"time":"2025-11-07T12:05:35.410694342+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":103254}
+{"time":"2025-11-07T12:05:35.410761466+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:35.411306965+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":103286}
+{"time":"2025-11-07T12:05:35.416259478+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":502}
+{"time":"2025-11-07T12:05:40.422232553+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":103858}
+{"time":"2025-11-07T12:05:40.422375155+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:40.422804333+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":103890}
+{"time":"2025-11-07T12:05:40.429631009+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:05:45.434315928+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":104599}
+{"time":"2025-11-07T12:05:45.440100687+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":667}
+{"time":"2025-11-07T12:05:50.445596526+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":105382}
+{"time":"2025-11-07T12:05:50.445978923+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:50.446424569+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":105414}
+{"time":"2025-11-07T12:05:50.450769826+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":551}
+{"time":"2025-11-07T12:05:55.476930421+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":106157}
+{"time":"2025-11-07T12:05:55.47715251+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:05:55.477767585+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":106189}
+{"time":"2025-11-07T12:05:55.482191002+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":512}
+{"time":"2025-11-07T12:06:00.499705956+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":106773}
+{"time":"2025-11-07T12:06:00.499829002+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:00.50030791+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":106805}
+{"time":"2025-11-07T12:06:00.505851981+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:06:05.510651042+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":107601}
+{"time":"2025-11-07T12:06:05.510725738+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:05.511212966+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":107633}
+{"time":"2025-11-07T12:06:05.516481423+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":546}
+{"time":"2025-11-07T12:06:10.520155226+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":108285}
+{"time":"2025-11-07T12:06:10.520320388+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:10.521423614+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":108318}
+{"time":"2025-11-07T12:06:10.526834215+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":599}
+{"time":"2025-11-07T12:06:15.537212362+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":109261}
+{"time":"2025-11-07T12:06:15.537409967+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:15.537873541+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":109293}
+{"time":"2025-11-07T12:06:15.540360295+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":323}
+{"time":"2025-11-07T12:06:20.545678234+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":109759}
+{"time":"2025-11-07T12:06:20.545746962+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:20.54620117+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":109791}
+{"time":"2025-11-07T12:06:20.551942612+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":563}
+{"time":"2025-11-07T12:06:25.557065371+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":110478}
+{"time":"2025-11-07T12:06:25.557142597+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:25.557646334+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":110510}
+{"time":"2025-11-07T12:06:25.563597952+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":582}
+{"time":"2025-11-07T12:06:30.567604545+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":111164}
+{"time":"2025-11-07T12:06:30.567722178+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:30.568188776+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":111196}
+{"time":"2025-11-07T12:06:30.574536141+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:06:35.581022729+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":112241}
+{"time":"2025-11-07T12:06:35.581111087+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:35.581704141+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":112273}
+{"time":"2025-11-07T12:06:35.585445387+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":297}
+{"time":"2025-11-07T12:06:40.590930137+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":112648}
+{"time":"2025-11-07T12:06:40.591010099+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:40.591499745+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":112680}
+{"time":"2025-11-07T12:06:40.597871804+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":587}
+{"time":"2025-11-07T12:06:45.601994428+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":113392}
+{"time":"2025-11-07T12:06:45.602112668+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:45.602545983+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":113424}
+{"time":"2025-11-07T12:06:45.608170699+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":583}
+{"time":"2025-11-07T12:06:50.613008151+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":114241}
+{"time":"2025-11-07T12:06:50.6175219+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":504}
+{"time":"2025-11-07T12:06:55.622223146+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":114835}
+{"time":"2025-11-07T12:06:55.622295782+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:06:55.622763342+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":114867}
+{"time":"2025-11-07T12:06:55.628328024+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":616}
+{"time":"2025-11-07T12:07:00.635004754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":115568}
+{"time":"2025-11-07T12:07:00.635089202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:00.635467224+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":115600}
+{"time":"2025-11-07T12:07:00.639626731+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":623}
+{"time":"2025-11-07T12:07:05.645893243+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":116348}
+{"time":"2025-11-07T12:07:05.646508095+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:05.647033619+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":116380}
+{"time":"2025-11-07T12:07:05.650831302+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":581}
+{"time":"2025-11-07T12:07:10.656572268+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":117249}
+{"time":"2025-11-07T12:07:10.656623641+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:10.657116288+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":117281}
+{"time":"2025-11-07T12:07:10.659897383+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":418}
+{"time":"2025-11-07T12:07:15.666338742+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":118026}
+{"time":"2025-11-07T12:07:15.666416928+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:15.666822531+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":118058}
+{"time":"2025-11-07T12:07:15.669354543+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":381}
+{"time":"2025-11-07T12:07:20.673984379+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":118508}
+{"time":"2025-11-07T12:07:20.674179684+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:20.674565401+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":118540}
+{"time":"2025-11-07T12:07:20.68044676+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:07:25.683473226+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":119248}
+{"time":"2025-11-07T12:07:25.683661085+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:25.683972836+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":119280}
+{"time":"2025-11-07T12:07:25.690646552+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:07:30.693694502+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":120027}
+{"time":"2025-11-07T12:07:30.693866066+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:30.694183102+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":120059}
+{"time":"2025-11-07T12:07:30.700921526+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:07:35.704417165+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":120772}
+{"time":"2025-11-07T12:07:35.704536046+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:35.705064852+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":120804}
+{"time":"2025-11-07T12:07:35.710917165+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:07:40.715277931+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":121562}
+{"time":"2025-11-07T12:07:40.715355191+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:40.71582293+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":121594}
+{"time":"2025-11-07T12:07:40.721477666+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":578}
+{"time":"2025-11-07T12:07:45.728086536+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":122489}
+{"time":"2025-11-07T12:07:45.728203284+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:45.728760372+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":122522}
+{"time":"2025-11-07T12:07:45.73131296+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":390}
+{"time":"2025-11-07T12:07:50.735810958+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":122982}
+{"time":"2025-11-07T12:07:50.735933217+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:50.736287608+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":123014}
+{"time":"2025-11-07T12:07:50.742685334+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:07:55.746427821+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":123719}
+{"time":"2025-11-07T12:07:55.746550618+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:07:55.746931531+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":123751}
+{"time":"2025-11-07T12:07:55.753412877+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:08:00.757709859+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":124459}
+{"time":"2025-11-07T12:08:00.757846355+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:00.758315446+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":124491}
+{"time":"2025-11-07T12:08:00.764275178+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:08:05.768128418+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":125206}
+{"time":"2025-11-07T12:08:05.768264213+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:05.768613704+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":125238}
+{"time":"2025-11-07T12:08:05.774437189+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":628}
+{"time":"2025-11-07T12:08:10.778729452+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":125937}
+{"time":"2025-11-07T12:08:10.778847853+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:10.779330404+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":125969}
+{"time":"2025-11-07T12:08:10.784937142+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:08:15.789862704+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":126754}
+{"time":"2025-11-07T12:08:15.790001494+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:15.790384472+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":126786}
+{"time":"2025-11-07T12:08:15.794990059+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":558}
+{"time":"2025-11-07T12:08:20.800029133+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":127579}
+{"time":"2025-11-07T12:08:20.800103472+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:20.800558792+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":127611}
+{"time":"2025-11-07T12:08:20.805149459+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":471}
+{"time":"2025-11-07T12:08:25.809465877+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":128160}
+{"time":"2025-11-07T12:08:25.809609706+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:25.810133936+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":128192}
+{"time":"2025-11-07T12:08:25.816480375+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":628}
+{"time":"2025-11-07T12:08:30.821777581+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":129078}
+{"time":"2025-11-07T12:08:30.822727438+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:30.823480274+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":129111}
+{"time":"2025-11-07T12:08:30.827247687+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":449}
+{"time":"2025-11-07T12:08:35.831575521+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":129630}
+{"time":"2025-11-07T12:08:35.831734328+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:35.832248029+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":129662}
+{"time":"2025-11-07T12:08:35.838500721+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":618}
+{"time":"2025-11-07T12:08:40.850652159+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":130414}
+{"time":"2025-11-07T12:08:40.850763666+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:40.851232227+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":130446}
+{"time":"2025-11-07T12:08:40.857235009+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":549}
+{"time":"2025-11-07T12:08:45.890672047+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":131172}
+{"time":"2025-11-07T12:08:45.890739062+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:45.891301727+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":131204}
+{"time":"2025-11-07T12:08:45.897026095+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":531}
+{"time":"2025-11-07T12:08:50.900824317+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":131804}
+{"time":"2025-11-07T12:08:50.900935086+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:50.901786133+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":131837}
+{"time":"2025-11-07T12:08:50.908185913+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:08:55.913626669+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":132613}
+{"time":"2025-11-07T12:08:55.913706386+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:08:55.914666671+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":132646}
+{"time":"2025-11-07T12:08:55.920065203+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":606}
+{"time":"2025-11-07T12:09:00.926018089+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":133485}
+{"time":"2025-11-07T12:09:00.926084297+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:00.926576601+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":133517}
+{"time":"2025-11-07T12:09:00.930001323+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":516}
+{"time":"2025-11-07T12:09:05.934126886+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":134103}
+{"time":"2025-11-07T12:09:05.93425924+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:05.93495447+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":134135}
+{"time":"2025-11-07T12:09:05.940604948+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:09:10.945866376+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":134841}
+{"time":"2025-11-07T12:09:10.951388665+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":668}
+{"time":"2025-11-07T12:09:15.956293572+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":135593}
+{"time":"2025-11-07T12:09:15.956357518+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:15.95692133+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":135625}
+{"time":"2025-11-07T12:09:15.963671606+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":624}
+{"time":"2025-11-07T12:09:20.96630516+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":136319}
+{"time":"2025-11-07T12:09:20.966433722+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:20.966815716+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":136351}
+{"time":"2025-11-07T12:09:20.973253438+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:09:25.977846595+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":137085}
+{"time":"2025-11-07T12:09:25.97792875+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:25.978361487+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":137117}
+{"time":"2025-11-07T12:09:25.984009744+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":608}
+{"time":"2025-11-07T12:09:30.98926635+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":137831}
+{"time":"2025-11-07T12:09:30.989412665+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:30.989795789+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":137863}
+{"time":"2025-11-07T12:09:30.995893463+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":602}
+{"time":"2025-11-07T12:09:36.001428188+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":138636}
+{"time":"2025-11-07T12:09:36.001491257+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:36.001922097+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":138668}
+{"time":"2025-11-07T12:09:36.006757615+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":543}
+{"time":"2025-11-07T12:09:41.01501854+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":139290}
+{"time":"2025-11-07T12:09:41.015244877+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:41.015520907+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":139322}
+{"time":"2025-11-07T12:09:41.022014603+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":619}
+{"time":"2025-11-07T12:09:46.028371842+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":140025}
+{"time":"2025-11-07T12:09:46.028440858+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:46.028905068+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":140057}
+{"time":"2025-11-07T12:09:46.035302629+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":624}
+{"time":"2025-11-07T12:09:51.041333579+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":140750}
+{"time":"2025-11-07T12:09:51.041445519+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:51.041814973+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":140782}
+{"time":"2025-11-07T12:09:51.048369495+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:09:56.052092992+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":141489}
+{"time":"2025-11-07T12:09:56.052198996+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:09:56.052631073+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":141521}
+{"time":"2025-11-07T12:09:56.058828066+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:10:01.066940724+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":142627}
+{"time":"2025-11-07T12:10:01.067032422+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:01.067536762+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":142659}
+{"time":"2025-11-07T12:10:01.069670726+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":238}
+{"time":"2025-11-07T12:10:06.074060035+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":142966}
+{"time":"2025-11-07T12:10:06.07415349+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:06.075603094+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":142999}
+{"time":"2025-11-07T12:10:06.080257795+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:10:11.087662207+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":143847}
+{"time":"2025-11-07T12:10:11.087796088+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:11.088179292+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":143879}
+{"time":"2025-11-07T12:10:11.092297146+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":494}
+{"time":"2025-11-07T12:10:16.097031341+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":144445}
+{"time":"2025-11-07T12:10:16.097167235+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:16.097542264+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":144477}
+{"time":"2025-11-07T12:10:16.104140968+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:10:21.107131312+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":145182}
+{"time":"2025-11-07T12:10:21.107293041+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:21.107706511+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":145214}
+{"time":"2025-11-07T12:10:21.114160748+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:10:26.117424569+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":145940}
+{"time":"2025-11-07T12:10:26.117547702+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:26.117980241+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":145972}
+{"time":"2025-11-07T12:10:26.123938627+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":617}
+{"time":"2025-11-07T12:10:31.128274333+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":146679}
+{"time":"2025-11-07T12:10:31.133970194+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":650}
+{"time":"2025-11-07T12:10:36.139315858+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":147433}
+{"time":"2025-11-07T12:10:36.139384202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:36.139796668+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":147465}
+{"time":"2025-11-07T12:10:36.145155085+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":602}
+{"time":"2025-11-07T12:10:41.151635515+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":148394}
+{"time":"2025-11-07T12:10:41.151810949+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:41.152418423+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":148426}
+{"time":"2025-11-07T12:10:41.155081319+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":379}
+{"time":"2025-11-07T12:10:46.160363548+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":148895}
+{"time":"2025-11-07T12:10:46.160432338+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:46.162143079+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":148928}
+{"time":"2025-11-07T12:10:46.166739079+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":617}
+{"time":"2025-11-07T12:10:51.171805207+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":149620}
+{"time":"2025-11-07T12:10:51.171887554+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:51.172321805+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":149652}
+{"time":"2025-11-07T12:10:51.177890832+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":590}
+{"time":"2025-11-07T12:10:56.181546523+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":150312}
+{"time":"2025-11-07T12:10:56.181701963+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:10:56.182086659+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":150344}
+{"time":"2025-11-07T12:10:56.188281379+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:11:01.193807423+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":151114}
+{"time":"2025-11-07T12:11:01.193896395+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:01.1942855+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":151146}
+{"time":"2025-11-07T12:11:01.199910769+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":574}
+{"time":"2025-11-07T12:11:06.204399594+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":151959}
+{"time":"2025-11-07T12:11:06.204488925+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:06.204942484+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":151991}
+{"time":"2025-11-07T12:11:06.209805232+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":467}
+{"time":"2025-11-07T12:11:11.213379662+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":152530}
+{"time":"2025-11-07T12:11:11.213506113+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:11.213837814+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":152562}
+{"time":"2025-11-07T12:11:11.220622802+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:11:16.22561455+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":153308}
+{"time":"2025-11-07T12:11:16.225793289+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:16.226869886+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":153341}
+{"time":"2025-11-07T12:11:16.232554045+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:11:21.237394616+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":154173}
+{"time":"2025-11-07T12:11:21.237479119+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:21.237893042+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":154205}
+{"time":"2025-11-07T12:11:21.242838232+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":510}
+{"time":"2025-11-07T12:11:26.247621932+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":154784}
+{"time":"2025-11-07T12:11:26.247833254+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:26.248119165+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":154816}
+{"time":"2025-11-07T12:11:26.254473672+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:11:31.259346881+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":155580}
+{"time":"2025-11-07T12:11:31.259411158+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:31.260834797+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":155613}
+{"time":"2025-11-07T12:11:31.264926334+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":580}
+{"time":"2025-11-07T12:11:36.270232015+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":156263}
+{"time":"2025-11-07T12:11:36.270331081+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:36.270761793+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":156295}
+{"time":"2025-11-07T12:11:36.276703887+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:11:41.281108107+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":157068}
+{"time":"2025-11-07T12:11:41.286202206+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":601}
+{"time":"2025-11-07T12:11:46.291274153+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":157816}
+{"time":"2025-11-07T12:11:46.291363772+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:46.291876144+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":157848}
+{"time":"2025-11-07T12:11:46.296929017+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":561}
+{"time":"2025-11-07T12:11:51.299228582+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":158493}
+{"time":"2025-11-07T12:11:51.29932694+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:51.300427382+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":158526}
+{"time":"2025-11-07T12:11:51.305764627+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":621}
+{"time":"2025-11-07T12:11:56.309596226+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":159223}
+{"time":"2025-11-07T12:11:56.309664672+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:11:56.310190061+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":159255}
+{"time":"2025-11-07T12:11:56.317064253+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:12:01.320187407+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":159956}
+{"time":"2025-11-07T12:12:01.320350384+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:01.320842035+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":159988}
+{"time":"2025-11-07T12:12:01.326710927+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":616}
+{"time":"2025-11-07T12:12:06.338905108+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":160694}
+{"time":"2025-11-07T12:12:06.339037661+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:06.339397079+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":160726}
+{"time":"2025-11-07T12:12:06.345485421+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:12:11.349853526+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":161432}
+{"time":"2025-11-07T12:12:11.349978763+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:11.350466401+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":161464}
+{"time":"2025-11-07T12:12:11.356415772+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:12:16.360762047+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":162190}
+{"time":"2025-11-07T12:12:16.360851899+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:16.361415839+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":162222}
+{"time":"2025-11-07T12:12:16.366945306+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":619}
+{"time":"2025-11-07T12:12:21.372903987+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":163045}
+{"time":"2025-11-07T12:12:21.377632398+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":534}
+{"time":"2025-11-07T12:12:26.385824422+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":164185}
+{"time":"2025-11-07T12:12:26.385898155+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:26.386404548+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":164217}
+{"time":"2025-11-07T12:12:26.387541626+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":100}
+{"time":"2025-11-07T12:12:31.393127909+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":164397}
+{"time":"2025-11-07T12:12:31.393322699+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:31.393733123+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":164429}
+{"time":"2025-11-07T12:12:31.400151432+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":628}
+{"time":"2025-11-07T12:12:36.404116336+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":165126}
+{"time":"2025-11-07T12:12:36.404275235+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:36.404608105+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":165158}
+{"time":"2025-11-07T12:12:36.410787965+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:12:41.415684509+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":165865}
+{"time":"2025-11-07T12:12:41.415775815+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:41.41617699+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":165897}
+{"time":"2025-11-07T12:12:41.422396133+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:12:46.426706578+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":166611}
+{"time":"2025-11-07T12:12:46.426818769+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:46.427437933+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":166643}
+{"time":"2025-11-07T12:12:46.432375281+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:12:51.441388255+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":167644}
+{"time":"2025-11-07T12:12:51.441451359+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:51.441921696+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":167676}
+{"time":"2025-11-07T12:12:51.444143449+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":335}
+{"time":"2025-11-07T12:12:56.452775555+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":168198}
+{"time":"2025-11-07T12:12:56.453009103+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:12:56.453599947+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":168230}
+{"time":"2025-11-07T12:12:56.45677602+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":542}
+{"time":"2025-11-07T12:13:01.469932619+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":169166}
+{"time":"2025-11-07T12:13:01.469983665+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:01.470578879+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":169198}
+{"time":"2025-11-07T12:13:01.474870032+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":332}
+{"time":"2025-11-07T12:13:06.480051952+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":169888}
+{"time":"2025-11-07T12:13:06.480171013+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:06.480807401+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":169921}
+{"time":"2025-11-07T12:13:06.484879244+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":347}
+{"time":"2025-11-07T12:13:11.49140753+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":170368}
+{"time":"2025-11-07T12:13:11.49155581+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:11.491998671+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":170400}
+{"time":"2025-11-07T12:13:11.498462769+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":565}
+{"time":"2025-11-07T12:13:16.502074522+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":171036}
+{"time":"2025-11-07T12:13:16.50221303+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:16.50266592+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":171068}
+{"time":"2025-11-07T12:13:16.509344346+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:13:21.514201035+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":171857}
+{"time":"2025-11-07T12:13:21.514320974+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:21.514773865+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":171889}
+{"time":"2025-11-07T12:13:21.520648636+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":586}
+{"time":"2025-11-07T12:13:26.528854046+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":172683}
+{"time":"2025-11-07T12:13:26.528930931+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:26.529356779+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":172715}
+{"time":"2025-11-07T12:13:26.53470552+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":507}
+{"time":"2025-11-07T12:13:31.540010767+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":173313}
+{"time":"2025-11-07T12:13:31.54099059+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:31.541396837+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":173346}
+{"time":"2025-11-07T12:13:31.545845515+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":616}
+{"time":"2025-11-07T12:13:36.554544267+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":174444}
+{"time":"2025-11-07T12:13:36.554615335+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:36.555103627+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":174476}
+{"time":"2025-11-07T12:13:36.557109413+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":224}
+{"time":"2025-11-07T12:13:41.561704819+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":174770}
+{"time":"2025-11-07T12:13:41.561807305+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:41.562309287+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":174802}
+{"time":"2025-11-07T12:13:41.56908221+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:13:46.573125977+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":175528}
+{"time":"2025-11-07T12:13:46.573223958+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:46.573685464+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":175560}
+{"time":"2025-11-07T12:13:46.580050831+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":618}
+{"time":"2025-11-07T12:13:51.584214608+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":176363}
+{"time":"2025-11-07T12:13:51.584325227+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:51.585541038+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":176396}
+{"time":"2025-11-07T12:13:51.590924242+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":520}
+{"time":"2025-11-07T12:13:56.595364359+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":177169}
+{"time":"2025-11-07T12:13:56.595433782+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:13:56.596778148+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":177202}
+{"time":"2025-11-07T12:13:56.601573687+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":452}
+{"time":"2025-11-07T12:14:01.60719753+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":177725}
+{"time":"2025-11-07T12:14:01.607324809+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:01.607675632+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":177757}
+{"time":"2025-11-07T12:14:01.614040358+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:14:06.618885404+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":178510}
+{"time":"2025-11-07T12:14:06.618983667+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:06.619534883+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":178542}
+{"time":"2025-11-07T12:14:06.625074609+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":590}
+{"time":"2025-11-07T12:14:11.629103241+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":179201}
+{"time":"2025-11-07T12:14:11.629258851+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:11.62974324+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":179233}
+{"time":"2025-11-07T12:14:11.636392834+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:14:16.641911073+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":179989}
+{"time":"2025-11-07T12:14:16.647082228+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":621}
+{"time":"2025-11-07T12:14:21.652094127+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":180711}
+{"time":"2025-11-07T12:14:21.652308585+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:21.652848191+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":180743}
+{"time":"2025-11-07T12:14:21.657399341+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":605}
+{"time":"2025-11-07T12:14:26.664696247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":181450}
+{"time":"2025-11-07T12:14:26.665199558+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:26.665847143+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":181482}
+{"time":"2025-11-07T12:14:26.670542381+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":577}
+{"time":"2025-11-07T12:14:31.685887374+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":182160}
+{"time":"2025-11-07T12:14:31.68606968+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:31.686430371+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":182192}
+{"time":"2025-11-07T12:14:31.692781533+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":593}
+{"time":"2025-11-07T12:14:36.698419544+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":182983}
+{"time":"2025-11-07T12:14:36.698540929+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:36.699003053+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":183015}
+{"time":"2025-11-07T12:14:36.704145636+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":508}
+{"time":"2025-11-07T12:14:41.709443728+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":183592}
+{"time":"2025-11-07T12:14:41.709566917+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:41.709959139+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":183624}
+{"time":"2025-11-07T12:14:41.716680226+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:14:46.721215916+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":184332}
+{"time":"2025-11-07T12:14:46.721310725+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:46.721837095+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":184364}
+{"time":"2025-11-07T12:14:46.728560442+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:14:51.738601303+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":185075}
+{"time":"2025-11-07T12:14:51.738701922+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:51.739752617+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":185108}
+{"time":"2025-11-07T12:14:51.746211608+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:14:56.774157268+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":185935}
+{"time":"2025-11-07T12:14:56.774257745+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:14:56.774813823+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":185967}
+{"time":"2025-11-07T12:14:56.78104925+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":539}
+{"time":"2025-11-07T12:15:01.789800816+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":186669}
+{"time":"2025-11-07T12:15:01.794927162+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":589}
+{"time":"2025-11-07T12:15:06.800345982+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":187326}
+{"time":"2025-11-07T12:15:06.800451573+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:06.800849394+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":187358}
+{"time":"2025-11-07T12:15:06.807115766+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":597}
+{"time":"2025-11-07T12:15:11.811881554+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":188025}
+{"time":"2025-11-07T12:15:11.811958934+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:11.81313582+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":188058}
+{"time":"2025-11-07T12:15:11.819122548+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:15:16.822009238+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":188763}
+{"time":"2025-11-07T12:15:16.822118691+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:16.822581257+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":188795}
+{"time":"2025-11-07T12:15:16.828839374+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":597}
+{"time":"2025-11-07T12:15:21.83056754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":189470}
+{"time":"2025-11-07T12:15:21.830656575+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:21.831076336+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":189502}
+{"time":"2025-11-07T12:15:21.837745218+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":628}
+{"time":"2025-11-07T12:15:26.840049845+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":190242}
+{"time":"2025-11-07T12:15:26.840131239+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:26.840654697+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":190274}
+{"time":"2025-11-07T12:15:26.847093382+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":594}
+{"time":"2025-11-07T12:15:31.849629929+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":190939}
+{"time":"2025-11-07T12:15:31.849714295+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:31.850305526+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":190971}
+{"time":"2025-11-07T12:15:31.856156894+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:15:36.861255654+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":191963}
+{"time":"2025-11-07T12:15:36.861319473+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:36.861806982+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":191995}
+{"time":"2025-11-07T12:15:36.866115461+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":351}
+{"time":"2025-11-07T12:15:41.870207647+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":192416}
+{"time":"2025-11-07T12:15:41.870286084+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:41.870693093+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":192448}
+{"time":"2025-11-07T12:15:41.877013038+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:15:46.878859752+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":193188}
+{"time":"2025-11-07T12:15:46.878963387+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:46.879989865+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":193221}
+{"time":"2025-11-07T12:15:46.88568925+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":605}
+{"time":"2025-11-07T12:15:51.903926342+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":193928}
+{"time":"2025-11-07T12:15:51.90403965+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:51.904418592+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":193960}
+{"time":"2025-11-07T12:15:51.910852379+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":643}
+{"time":"2025-11-07T12:15:56.913152101+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":194707}
+{"time":"2025-11-07T12:15:56.913205738+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:15:56.913664748+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":194739}
+{"time":"2025-11-07T12:15:56.919586008+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":561}
+{"time":"2025-11-07T12:16:01.922361618+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":195371}
+{"time":"2025-11-07T12:16:01.92246383+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:01.92367923+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":195404}
+{"time":"2025-11-07T12:16:01.928872489+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":638}
+{"time":"2025-11-07T12:16:06.937111138+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":196129}
+{"time":"2025-11-07T12:16:06.937246454+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:06.937754141+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":196161}
+{"time":"2025-11-07T12:16:06.943483735+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":634}
+{"time":"2025-11-07T12:16:11.957550546+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":197189}
+{"time":"2025-11-07T12:16:11.95764532+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:11.958138473+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":197221}
+{"time":"2025-11-07T12:16:11.960501766+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":336}
+{"time":"2025-11-07T12:16:16.964195335+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":197651}
+{"time":"2025-11-07T12:16:16.964270208+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:16.964701491+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":197683}
+{"time":"2025-11-07T12:16:16.969648424+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":614}
+{"time":"2025-11-07T12:16:21.972343333+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":198369}
+{"time":"2025-11-07T12:16:21.972421203+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:21.973697593+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":198402}
+{"time":"2025-11-07T12:16:21.979795158+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:16:26.985401501+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":199145}
+{"time":"2025-11-07T12:16:26.985517853+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:26.986059567+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":199177}
+{"time":"2025-11-07T12:16:26.992680487+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:16:31.997310686+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":200029}
+{"time":"2025-11-07T12:16:31.99781982+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:31.998318734+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":200061}
+{"time":"2025-11-07T12:16:32.001961391+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":490}
+{"time":"2025-11-07T12:16:37.006239228+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":200623}
+{"time":"2025-11-07T12:16:37.006375395+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:37.006888297+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":200655}
+{"time":"2025-11-07T12:16:37.01347668+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:16:42.016196346+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":201459}
+{"time":"2025-11-07T12:16:42.016320474+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:42.017199+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":201492}
+{"time":"2025-11-07T12:16:42.022920388+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":579}
+{"time":"2025-11-07T12:16:47.026308775+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":202158}
+{"time":"2025-11-07T12:16:47.026767531+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:47.027476555+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":202191}
+{"time":"2025-11-07T12:16:47.03193319+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":620}
+{"time":"2025-11-07T12:16:52.038419138+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":203512}
+{"time":"2025-11-07T12:16:52.038512184+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:52.038930581+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":203544}
+{"time":"2025-11-07T12:16:52.039477084+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":5}
+{"time":"2025-11-07T12:16:57.046940348+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":204040}
+{"time":"2025-11-07T12:16:57.047005814+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:16:57.047430068+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":204072}
+{"time":"2025-11-07T12:16:57.049153858+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":217}
+{"time":"2025-11-07T12:17:02.066895219+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":204412}
+{"time":"2025-11-07T12:17:02.067005452+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:02.067402184+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":204444}
+{"time":"2025-11-07T12:17:02.072700337+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":583}
+{"time":"2025-11-07T12:17:07.075362167+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":205098}
+{"time":"2025-11-07T12:17:07.075459719+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:07.076559333+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":205131}
+{"time":"2025-11-07T12:17:07.082663871+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":666}
+{"time":"2025-11-07T12:17:12.091920759+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":205981}
+{"time":"2025-11-07T12:17:12.092058748+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:12.092488022+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":206013}
+{"time":"2025-11-07T12:17:12.097840051+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":531}
+{"time":"2025-11-07T12:17:17.102902185+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":206815}
+{"time":"2025-11-07T12:17:17.102993736+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:17.103423819+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":206847}
+{"time":"2025-11-07T12:17:17.106650278+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":437}
+{"time":"2025-11-07T12:17:22.11375137+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":207855}
+{"time":"2025-11-07T12:17:22.113835777+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:22.114267567+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":207887}
+{"time":"2025-11-07T12:17:22.115569678+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":135}
+{"time":"2025-11-07T12:17:27.119786958+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":208091}
+{"time":"2025-11-07T12:17:27.119901844+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:27.120241645+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":208123}
+{"time":"2025-11-07T12:17:27.126596863+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:17:32.131088212+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":208875}
+{"time":"2025-11-07T12:17:32.136528354+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":625}
+{"time":"2025-11-07T12:17:37.142021672+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":209636}
+{"time":"2025-11-07T12:17:37.142260586+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:37.143008728+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":209669}
+{"time":"2025-11-07T12:17:37.146443038+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":569}
+{"time":"2025-11-07T12:17:42.150863892+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":210307}
+{"time":"2025-11-07T12:17:42.157115521+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":669}
+{"time":"2025-11-07T12:17:47.163170189+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":211391}
+{"time":"2025-11-07T12:17:47.163253596+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:47.163713083+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":211423}
+{"time":"2025-11-07T12:17:47.167577007+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":293}
+{"time":"2025-11-07T12:17:52.171860515+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":211822}
+{"time":"2025-11-07T12:17:52.171959567+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:52.172510068+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":211854}
+{"time":"2025-11-07T12:17:52.178967307+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":600}
+{"time":"2025-11-07T12:17:57.183034204+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":212525}
+{"time":"2025-11-07T12:17:57.18312767+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:17:57.183675522+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":212557}
+{"time":"2025-11-07T12:17:57.190909228+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:18:02.19419595+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":213363}
+{"time":"2025-11-07T12:18:02.194592424+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:02.19521098+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":213395}
+{"time":"2025-11-07T12:18:02.199558528+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":537}
+{"time":"2025-11-07T12:18:07.202264833+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":214001}
+{"time":"2025-11-07T12:18:07.20235053+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:07.202953385+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":214033}
+{"time":"2025-11-07T12:18:07.20938115+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:18:12.213812954+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":214917}
+{"time":"2025-11-07T12:18:12.217770414+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":493}
+{"time":"2025-11-07T12:18:17.22295247+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":215543}
+{"time":"2025-11-07T12:18:17.223023053+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:17.223466713+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":215575}
+{"time":"2025-11-07T12:18:17.227274493+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":573}
+{"time":"2025-11-07T12:18:22.232998782+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":216370}
+{"time":"2025-11-07T12:18:22.233071179+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:22.23349042+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":216402}
+{"time":"2025-11-07T12:18:22.237126616+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":484}
+{"time":"2025-11-07T12:18:27.240946474+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":216968}
+{"time":"2025-11-07T12:18:27.241077065+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:27.24166313+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":217000}
+{"time":"2025-11-07T12:18:27.248447779+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":626}
+{"time":"2025-11-07T12:18:32.250954499+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":217695}
+{"time":"2025-11-07T12:18:32.251146637+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:32.251498462+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":217727}
+{"time":"2025-11-07T12:18:32.257627815+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:18:37.260649953+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":218474}
+{"time":"2025-11-07T12:18:37.260710487+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:37.261320692+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":218506}
+{"time":"2025-11-07T12:18:37.267323486+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":596}
+{"time":"2025-11-07T12:18:42.270962029+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":219172}
+{"time":"2025-11-07T12:18:42.271158886+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:42.271566828+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":219204}
+{"time":"2025-11-07T12:18:42.277943736+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:18:47.281859732+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":219914}
+{"time":"2025-11-07T12:18:47.281966458+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:47.282438748+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":219946}
+{"time":"2025-11-07T12:18:47.28861947+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":663}
+{"time":"2025-11-07T12:18:52.299149123+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":220717}
+{"time":"2025-11-07T12:18:52.299256827+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:52.299728714+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":220749}
+{"time":"2025-11-07T12:18:52.304830626+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":571}
+{"time":"2025-11-07T12:18:57.323069808+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":221556}
+{"time":"2025-11-07T12:18:57.323172075+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:18:57.323626393+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":221588}
+{"time":"2025-11-07T12:18:57.328714525+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":493}
+{"time":"2025-11-07T12:19:02.337833987+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":222199}
+{"time":"2025-11-07T12:19:02.33835876+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:02.339208362+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":222232}
+{"time":"2025-11-07T12:19:02.34317076+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":605}
+{"time":"2025-11-07T12:19:07.347029072+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":222907}
+{"time":"2025-11-07T12:19:07.347129132+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:07.348462517+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":222940}
+{"time":"2025-11-07T12:19:07.353658671+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:19:12.356573303+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":223646}
+{"time":"2025-11-07T12:19:12.605061917+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:19:17.365634735+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":224407}
+{"time":"2025-11-07T12:19:17.365705947+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:17.366286144+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":224439}
+{"time":"2025-11-07T12:19:17.372368018+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":614}
+{"time":"2025-11-07T12:19:22.374472399+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":225123}
+{"time":"2025-11-07T12:19:22.374569448+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:22.374952377+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":225155}
+{"time":"2025-11-07T12:19:22.380994694+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":621}
+{"time":"2025-11-07T12:19:27.389610238+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":225851}
+{"time":"2025-11-07T12:19:27.632118966+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":641}
+{"time":"2025-11-07T12:19:32.398112836+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":226664}
+{"time":"2025-11-07T12:19:32.398198666+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:32.398710433+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":226696}
+{"time":"2025-11-07T12:19:32.404271203+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":573}
+{"time":"2025-11-07T12:19:37.407374305+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":227343}
+{"time":"2025-11-07T12:19:37.407450334+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:37.408018586+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":227375}
+{"time":"2025-11-07T12:19:37.414848382+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":632}
+{"time":"2025-11-07T12:19:42.417881298+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":228078}
+{"time":"2025-11-07T12:19:42.661424884+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:19:47.428870653+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":228884}
+{"time":"2025-11-07T12:19:47.429005526+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:47.42955187+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":228916}
+{"time":"2025-11-07T12:19:47.434657569+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":569}
+{"time":"2025-11-07T12:19:52.441164042+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":229856}
+{"time":"2025-11-07T12:19:52.441264385+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:19:52.441693126+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":229888}
+{"time":"2025-11-07T12:19:52.444484372+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":376}
+{"time":"2025-11-07T12:19:57.447866135+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":230335}
+{"time":"2025-11-07T12:19:57.604425036+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:20:02.460127059+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":231341}
+{"time":"2025-11-07T12:20:02.460188921+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:02.460710529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":231373}
+{"time":"2025-11-07T12:20:02.463650985+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":369}
+{"time":"2025-11-07T12:20:07.468963053+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":231864}
+{"time":"2025-11-07T12:20:07.469064778+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:07.469418371+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":231896}
+{"time":"2025-11-07T12:20:07.474698356+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":584}
+{"time":"2025-11-07T12:20:12.47936696+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":232551}
+{"time":"2025-11-07T12:20:12.686578682+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":710}
+{"time":"2025-11-07T12:20:17.491492541+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":233330}
+{"time":"2025-11-07T12:20:17.491615838+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:17.492555652+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":233363}
+{"time":"2025-11-07T12:20:17.498244327+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:20:22.499806659+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":234068}
+{"time":"2025-11-07T12:20:22.499960539+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:22.500368522+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":234100}
+{"time":"2025-11-07T12:20:22.506630842+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:20:27.508910445+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":234806}
+{"time":"2025-11-07T12:20:27.691682256+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:20:32.517945262+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":235546}
+{"time":"2025-11-07T12:20:32.518066432+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:32.518473659+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":235578}
+{"time":"2025-11-07T12:20:32.525032195+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:20:37.529162455+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":236441}
+{"time":"2025-11-07T12:20:37.529241521+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:37.53035628+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":236474}
+{"time":"2025-11-07T12:20:37.535080066+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":479}
+{"time":"2025-11-07T12:20:42.540938609+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":237024}
+{"time":"2025-11-07T12:20:42.589840621+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:20:47.551121764+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":237793}
+{"time":"2025-11-07T12:20:47.551212445+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:47.551844228+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":237825}
+{"time":"2025-11-07T12:20:47.55619786+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":606}
+{"time":"2025-11-07T12:20:52.561912399+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":238644}
+{"time":"2025-11-07T12:20:52.562364684+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:20:52.563174948+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":238677}
+{"time":"2025-11-07T12:20:52.56653534+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":492}
+{"time":"2025-11-07T12:20:57.570639651+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":239238}
+{"time":"2025-11-07T12:20:57.662586281+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:21:02.583459039+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":240318}
+{"time":"2025-11-07T12:21:02.583534526+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:02.584248774+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":240351}
+{"time":"2025-11-07T12:21:02.586364298+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":296}
+{"time":"2025-11-07T12:21:07.589954221+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":240717}
+{"time":"2025-11-07T12:21:07.59615609+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":668}
+{"time":"2025-11-07T12:21:12.600031018+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":241466}
+{"time":"2025-11-07T12:21:12.600423715+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:12.601004639+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":241498}
+{"time":"2025-11-07T12:21:12.605396715+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":627}
+{"time":"2025-11-07T12:21:17.609090514+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":242329}
+{"time":"2025-11-07T12:21:17.609369059+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:17.609681291+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":242361}
+{"time":"2025-11-07T12:21:17.614799874+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":502}
+{"time":"2025-11-07T12:21:22.618009454+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":243036}
+{"time":"2025-11-07T12:21:22.61810359+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:22.618554267+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":243068}
+{"time":"2025-11-07T12:21:22.623765803+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":533}
+{"time":"2025-11-07T12:21:27.62586002+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":243670}
+{"time":"2025-11-07T12:21:27.633437491+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:21:32.639092373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":244411}
+{"time":"2025-11-07T12:21:32.639326791+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:32.639767074+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":244443}
+{"time":"2025-11-07T12:21:32.646053116+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:21:37.648239687+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":245148}
+{"time":"2025-11-07T12:21:37.648368974+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:37.648843063+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":245180}
+{"time":"2025-11-07T12:21:37.655287393+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:21:42.660985505+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":245886}
+{"time":"2025-11-07T12:21:42.691366334+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":703}
+{"time":"2025-11-07T12:21:47.679689402+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":246844}
+{"time":"2025-11-07T12:21:47.679786252+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:47.680268484+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":246876}
+{"time":"2025-11-07T12:21:47.683582605+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":460}
+{"time":"2025-11-07T12:21:52.68650846+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":247424}
+{"time":"2025-11-07T12:21:52.686601639+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:52.687140974+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":247456}
+{"time":"2025-11-07T12:21:52.693158372+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":618}
+{"time":"2025-11-07T12:21:57.696043731+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":248145}
+{"time":"2025-11-07T12:21:57.69616204+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:21:57.696784192+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":248177}
+{"time":"2025-11-07T12:21:57.70256706+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:22:02.709509879+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":249091}
+{"time":"2025-11-07T12:22:02.709616264+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:02.710030179+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":249123}
+{"time":"2025-11-07T12:22:02.713029927+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":429}
+{"time":"2025-11-07T12:22:07.718468517+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":249869}
+{"time":"2025-11-07T12:22:07.718568001+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:07.719068535+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":249901}
+{"time":"2025-11-07T12:22:07.722238152+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":430}
+{"time":"2025-11-07T12:22:12.725637102+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":250403}
+{"time":"2025-11-07T12:22:12.725719806+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:12.72669548+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":250436}
+{"time":"2025-11-07T12:22:12.732381202+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:22:17.734840874+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":251143}
+{"time":"2025-11-07T12:22:17.734932444+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:17.735467057+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":251175}
+{"time":"2025-11-07T12:22:17.742256692+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":654}
+{"time":"2025-11-07T12:22:22.751181413+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":251922}
+{"time":"2025-11-07T12:22:22.751308817+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:22.751720937+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":251954}
+{"time":"2025-11-07T12:22:22.75832661+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":634}
+{"time":"2025-11-07T12:22:27.760422981+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":252664}
+{"time":"2025-11-07T12:22:27.760546155+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:27.761091668+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":252696}
+{"time":"2025-11-07T12:22:27.767437984+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":673}
+{"time":"2025-11-07T12:22:32.770036586+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":253473}
+{"time":"2025-11-07T12:22:32.776475272+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":634}
+{"time":"2025-11-07T12:22:37.77889839+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":254180}
+{"time":"2025-11-07T12:22:37.779070551+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:37.780463317+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":254213}
+{"time":"2025-11-07T12:22:37.785177704+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":664}
+{"time":"2025-11-07T12:22:42.792936031+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":255000}
+{"time":"2025-11-07T12:22:42.793035532+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:42.793449417+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":255032}
+{"time":"2025-11-07T12:22:42.797896977+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":594}
+{"time":"2025-11-07T12:22:47.800560555+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":255695}
+{"time":"2025-11-07T12:22:47.800718965+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:47.801129587+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":255727}
+{"time":"2025-11-07T12:22:47.80737772+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:22:52.809808136+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":256434}
+{"time":"2025-11-07T12:22:52.80989763+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:52.816609029+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":256467}
+{"time":"2025-11-07T12:22:52.822604839+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:22:57.824638467+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":257250}
+{"time":"2025-11-07T12:22:57.824712933+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:22:57.825165671+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":257282}
+{"time":"2025-11-07T12:22:57.83078212+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":601}
+{"time":"2025-11-07T12:23:07.843434997+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":258701}
+{"time":"2025-11-07T12:23:07.843539638+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:07.844478524+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":258734}
+{"time":"2025-11-07T12:23:07.849972876+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":625}
+{"time":"2025-11-07T12:23:12.853032854+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":259430}
+{"time":"2025-11-07T12:23:12.853147416+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:12.853648811+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":259462}
+{"time":"2025-11-07T12:23:12.859954776+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:23:17.864346798+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":260327}
+{"time":"2025-11-07T12:23:17.864537275+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:17.865092963+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":260359}
+{"time":"2025-11-07T12:23:17.868381541+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":519}
+{"time":"2025-11-07T12:23:22.871633639+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":260947}
+{"time":"2025-11-07T12:23:22.871781879+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:22.872099953+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":260979}
+{"time":"2025-11-07T12:23:22.878281315+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:23:27.881853464+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":261694}
+{"time":"2025-11-07T12:23:27.882125076+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:27.883084746+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":261727}
+{"time":"2025-11-07T12:23:27.887667985+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":670}
+{"time":"2025-11-07T12:23:32.890666131+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":262472}
+{"time":"2025-11-07T12:23:32.89073209+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:32.891325943+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":262504}
+{"time":"2025-11-07T12:23:32.897803514+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":631}
+{"time":"2025-11-07T12:23:37.899636412+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":263329}
+{"time":"2025-11-07T12:23:37.899709931+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:37.900128715+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":263361}
+{"time":"2025-11-07T12:23:37.905889307+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":512}
+{"time":"2025-11-07T12:23:42.907919778+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":264116}
+{"time":"2025-11-07T12:23:42.912917137+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":497}
+{"time":"2025-11-07T12:23:47.917009897+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":264760}
+{"time":"2025-11-07T12:23:47.917088008+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:47.917529539+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":264792}
+{"time":"2025-11-07T12:23:47.923310656+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":600}
+{"time":"2025-11-07T12:23:52.928038479+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":265579}
+{"time":"2025-11-07T12:23:52.928099208+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:52.928556183+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":265611}
+{"time":"2025-11-07T12:23:52.934014055+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":519}
+{"time":"2025-11-07T12:23:57.937600195+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":266201}
+{"time":"2025-11-07T12:23:57.937772183+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:23:57.938670496+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":266234}
+{"time":"2025-11-07T12:23:57.944474032+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:24:02.94873173+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":267178}
+{"time":"2025-11-07T12:24:02.95300591+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":432}
+{"time":"2025-11-07T12:24:07.977529224+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":267733}
+{"time":"2025-11-07T12:24:07.977634634+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:07.978038564+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":267765}
+{"time":"2025-11-07T12:24:07.984840485+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":622}
+{"time":"2025-11-07T12:24:12.989791659+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":268678}
+{"time":"2025-11-07T12:24:12.989877979+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:12.990368738+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":268710}
+{"time":"2025-11-07T12:24:12.993343944+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":417}
+{"time":"2025-11-07T12:24:17.996928985+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":269197}
+{"time":"2025-11-07T12:24:17.997017818+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:17.997448687+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":269229}
+{"time":"2025-11-07T12:24:18.003921034+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:24:23.006287529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":269934}
+{"time":"2025-11-07T12:24:23.00641718+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:23.006830454+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":269966}
+{"time":"2025-11-07T12:24:23.012998353+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:24:28.018591048+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":270723}
+{"time":"2025-11-07T12:24:28.019068765+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:28.01967051+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":270755}
+{"time":"2025-11-07T12:24:28.023445262+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":561}
+{"time":"2025-11-07T12:24:33.035902945+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":271513}
+{"time":"2025-11-07T12:24:33.041386611+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":568}
+{"time":"2025-11-07T12:24:38.04485085+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":272150}
+{"time":"2025-11-07T12:24:38.044996955+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:38.045413378+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":272182}
+{"time":"2025-11-07T12:24:38.052414602+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:24:43.055019934+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":272893}
+{"time":"2025-11-07T12:24:43.055148544+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:43.055465148+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":272925}
+{"time":"2025-11-07T12:24:43.062202052+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:24:48.063409157+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":273674}
+{"time":"2025-11-07T12:24:48.063524302+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:48.06392645+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":273706}
+{"time":"2025-11-07T12:24:48.070774183+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":632}
+{"time":"2025-11-07T12:24:53.073284432+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":274440}
+{"time":"2025-11-07T12:24:53.073369203+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:24:53.073795672+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":274472}
+{"time":"2025-11-07T12:24:53.080298295+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":604}
+{"time":"2025-11-07T12:24:58.084403263+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":275190}
+{"time":"2025-11-07T12:24:58.090692946+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":652}
+{"time":"2025-11-07T12:25:03.095539268+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":276017}
+{"time":"2025-11-07T12:25:03.101259807+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":578}
+{"time":"2025-11-07T12:25:08.103411228+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":276666}
+{"time":"2025-11-07T12:25:08.103575069+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:08.103996666+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":276698}
+{"time":"2025-11-07T12:25:08.110465718+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:25:13.114038191+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":277446}
+{"time":"2025-11-07T12:25:13.114188316+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:13.114515388+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":277478}
+{"time":"2025-11-07T12:25:13.118765732+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:25:18.123955923+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":278273}
+{"time":"2025-11-07T12:25:18.124028964+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:18.12451575+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":278305}
+{"time":"2025-11-07T12:25:18.130380133+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":547}
+{"time":"2025-11-07T12:25:23.132304784+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":278924}
+{"time":"2025-11-07T12:25:23.13243992+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:23.133493314+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":278957}
+{"time":"2025-11-07T12:25:23.139324494+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":674}
+{"time":"2025-11-07T12:25:28.143115427+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":279733}
+{"time":"2025-11-07T12:25:28.143203185+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:28.143840266+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":279765}
+{"time":"2025-11-07T12:25:28.148879671+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":606}
+{"time":"2025-11-07T12:25:33.15238911+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":280440}
+{"time":"2025-11-07T12:25:33.15249277+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:33.153350701+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":280473}
+{"time":"2025-11-07T12:25:33.159191523+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:25:38.162527101+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":281335}
+{"time":"2025-11-07T12:25:38.162593071+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:38.163031484+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":281367}
+{"time":"2025-11-07T12:25:38.168129201+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":480}
+{"time":"2025-11-07T12:25:43.177632411+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":282008}
+{"time":"2025-11-07T12:25:43.183182402+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":579}
+{"time":"2025-11-07T12:25:48.186991061+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":282658}
+{"time":"2025-11-07T12:25:48.187093219+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:48.187436078+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":282690}
+{"time":"2025-11-07T12:25:48.193928369+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:25:53.198447894+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":283435}
+{"time":"2025-11-07T12:25:53.198566948+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:53.199106003+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":283467}
+{"time":"2025-11-07T12:25:53.205115742+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:25:58.208411379+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":284348}
+{"time":"2025-11-07T12:25:58.208494461+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:25:58.20933227+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":284381}
+{"time":"2025-11-07T12:25:58.212295583+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":463}
+{"time":"2025-11-07T12:26:03.216337415+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":284913}
+{"time":"2025-11-07T12:26:03.216495199+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:03.216914107+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":284945}
+{"time":"2025-11-07T12:26:03.222534445+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:26:08.225640197+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":285838}
+{"time":"2025-11-07T12:26:08.225722883+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:08.226122927+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":285870}
+{"time":"2025-11-07T12:26:08.230751927+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":450}
+{"time":"2025-11-07T12:26:13.23386155+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":286392}
+{"time":"2025-11-07T12:26:13.233945663+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:13.235081472+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":286425}
+{"time":"2025-11-07T12:26:13.241404442+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:26:18.244634453+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":287253}
+{"time":"2025-11-07T12:26:18.244701004+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:18.245189658+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":287285}
+{"time":"2025-11-07T12:26:18.250842353+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":554}
+{"time":"2025-11-07T12:26:23.253507532+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":287909}
+{"time":"2025-11-07T12:26:23.25362872+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:23.254133167+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":287941}
+{"time":"2025-11-07T12:26:23.260865839+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:26:28.263791736+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":288648}
+{"time":"2025-11-07T12:26:28.263952242+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:28.264284482+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":288680}
+{"time":"2025-11-07T12:26:28.270592174+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:26:33.273341203+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":289401}
+{"time":"2025-11-07T12:26:33.273440186+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:33.273986029+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":289433}
+{"time":"2025-11-07T12:26:33.280998309+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":663}
+{"time":"2025-11-07T12:26:38.283561226+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":290165}
+{"time":"2025-11-07T12:26:38.28372198+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:38.284214202+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":290197}
+{"time":"2025-11-07T12:26:38.2908668+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:26:43.293346715+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":290907}
+{"time":"2025-11-07T12:26:43.293514061+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:43.294597447+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":290940}
+{"time":"2025-11-07T12:26:43.299718056+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:26:48.302950662+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":291687}
+{"time":"2025-11-07T12:26:48.303050838+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:48.304322662+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":291720}
+{"time":"2025-11-07T12:26:48.309898722+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":633}
+{"time":"2025-11-07T12:26:53.312879718+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":292430}
+{"time":"2025-11-07T12:26:53.312957654+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:53.313399754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":292462}
+{"time":"2025-11-07T12:26:53.31943847+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":670}
+{"time":"2025-11-07T12:26:58.323475883+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":293485}
+{"time":"2025-11-07T12:26:58.323559686+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:26:58.324015613+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":293517}
+{"time":"2025-11-07T12:26:58.326243956+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":355}
+{"time":"2025-11-07T12:27:03.334318989+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":294246}
+{"time":"2025-11-07T12:27:03.334493841+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:03.33485754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":294278}
+{"time":"2025-11-07T12:27:03.339296273+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":332}
+{"time":"2025-11-07T12:27:08.341829186+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":294712}
+{"time":"2025-11-07T12:27:08.341936221+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:08.342899444+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":294745}
+{"time":"2025-11-07T12:27:08.348878731+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":644}
+{"time":"2025-11-07T12:27:13.351084625+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":295463}
+{"time":"2025-11-07T12:27:13.351217174+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:13.352189581+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":295496}
+{"time":"2025-11-07T12:27:13.358860168+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":674}
+{"time":"2025-11-07T12:27:18.36329634+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":296240}
+{"time":"2025-11-07T12:27:18.363394097+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:18.36473393+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":296273}
+{"time":"2025-11-07T12:27:18.370539716+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:27:23.3740153+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":296981}
+{"time":"2025-11-07T12:27:23.37412215+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:23.37474824+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":297013}
+{"time":"2025-11-07T12:27:23.381739207+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":674}
+{"time":"2025-11-07T12:27:28.384081445+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":297759}
+{"time":"2025-11-07T12:27:28.384170427+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:28.384731437+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":297791}
+{"time":"2025-11-07T12:27:28.39099832+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:27:33.394183054+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":298538}
+{"time":"2025-11-07T12:27:33.394253307+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:33.394871727+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":298570}
+{"time":"2025-11-07T12:27:33.401157236+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":595}
+{"time":"2025-11-07T12:27:38.405250028+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":299292}
+{"time":"2025-11-07T12:27:38.405321229+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:38.405933524+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":299324}
+{"time":"2025-11-07T12:27:38.412932574+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":620}
+{"time":"2025-11-07T12:27:43.415704513+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":300015}
+{"time":"2025-11-07T12:27:43.415848704+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:43.417297849+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":300048}
+{"time":"2025-11-07T12:27:43.422233594+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:27:48.425981702+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":300790}
+{"time":"2025-11-07T12:27:48.426362961+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:48.427274912+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":300823}
+{"time":"2025-11-07T12:27:48.431663219+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":599}
+{"time":"2025-11-07T12:27:53.436913593+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":301622}
+{"time":"2025-11-07T12:27:53.43701588+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:53.437406968+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":301654}
+{"time":"2025-11-07T12:27:53.440799226+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":547}
+{"time":"2025-11-07T12:27:58.444431401+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":302289}
+{"time":"2025-11-07T12:27:58.444589013+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:27:58.445018359+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":302321}
+{"time":"2025-11-07T12:27:58.451015829+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":620}
+{"time":"2025-11-07T12:28:03.454946694+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":303222}
+{"time":"2025-11-07T12:28:03.455032905+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:03.455596898+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":303254}
+{"time":"2025-11-07T12:28:03.460655017+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":442}
+{"time":"2025-11-07T12:28:08.471560347+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":303821}
+{"time":"2025-11-07T12:28:08.471692438+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:08.472081488+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":303853}
+{"time":"2025-11-07T12:28:08.477556197+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":605}
+{"time":"2025-11-07T12:28:13.481747019+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":304529}
+{"time":"2025-11-07T12:28:13.48185685+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:13.482911865+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":304562}
+{"time":"2025-11-07T12:28:13.488465947+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:28:18.492121966+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":305270}
+{"time":"2025-11-07T12:28:18.492230706+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:18.492729019+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":305302}
+{"time":"2025-11-07T12:28:18.499521383+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:28:23.50189367+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":306047}
+{"time":"2025-11-07T12:28:23.501966109+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:23.502649431+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":306079}
+{"time":"2025-11-07T12:28:23.509621665+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:28:28.512430301+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":306805}
+{"time":"2025-11-07T12:28:28.512519069+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:28.512993915+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":306837}
+{"time":"2025-11-07T12:28:28.517600419+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":618}
+{"time":"2025-11-07T12:28:33.521313451+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":307577}
+{"time":"2025-11-07T12:28:33.52138139+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:33.52194542+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":307609}
+{"time":"2025-11-07T12:28:33.528157108+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":625}
+{"time":"2025-11-07T12:28:38.531009572+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":308394}
+{"time":"2025-11-07T12:28:38.531091317+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:38.531562331+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":308426}
+{"time":"2025-11-07T12:28:38.537095549+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":546}
+{"time":"2025-11-07T12:28:43.538985538+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":309043}
+{"time":"2025-11-07T12:28:43.539096112+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:43.539594049+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":309075}
+{"time":"2025-11-07T12:28:43.54633796+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:28:48.549275371+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":309788}
+{"time":"2025-11-07T12:28:48.549343041+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:48.549917867+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":309820}
+{"time":"2025-11-07T12:28:48.556003428+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:28:53.559680751+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":310562}
+{"time":"2025-11-07T12:28:53.559806085+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:53.560267747+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":310594}
+{"time":"2025-11-07T12:28:53.565962832+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:28:58.572220181+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":311492}
+{"time":"2025-11-07T12:28:58.572341358+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:28:58.572704969+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":311524}
+{"time":"2025-11-07T12:28:58.575753167+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":486}
+{"time":"2025-11-07T12:29:03.583822198+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":312099}
+{"time":"2025-11-07T12:29:03.583918828+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:03.584333547+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":312131}
+{"time":"2025-11-07T12:29:03.590238554+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":617}
+{"time":"2025-11-07T12:29:08.593670584+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":312821}
+{"time":"2025-11-07T12:29:08.593829908+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:08.594720124+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":312854}
+{"time":"2025-11-07T12:29:08.600648246+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":673}
+{"time":"2025-11-07T12:29:13.604780762+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":313644}
+{"time":"2025-11-07T12:29:13.604959446+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:13.605302833+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":313676}
+{"time":"2025-11-07T12:29:13.611048958+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":632}
+{"time":"2025-11-07T12:29:18.617495935+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":314378}
+{"time":"2025-11-07T12:29:18.617590105+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:18.61877295+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":314411}
+{"time":"2025-11-07T12:29:18.6242934+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:29:23.626062466+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":315117}
+{"time":"2025-11-07T12:29:23.626152622+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:23.627308788+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":315150}
+{"time":"2025-11-07T12:29:23.633625941+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:29:28.639309032+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":315980}
+{"time":"2025-11-07T12:29:28.639453052+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:28.639797269+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":316012}
+{"time":"2025-11-07T12:29:28.643659952+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":553}
+{"time":"2025-11-07T12:29:33.647069276+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":316636}
+{"time":"2025-11-07T12:29:33.647174074+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:33.647688425+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":316668}
+{"time":"2025-11-07T12:29:33.654107258+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:29:38.657815216+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":317442}
+{"time":"2025-11-07T12:29:38.657957974+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:38.658428152+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":317474}
+{"time":"2025-11-07T12:29:38.663676079+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":608}
+{"time":"2025-11-07T12:29:43.668702112+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":318154}
+{"time":"2025-11-07T12:29:43.668808862+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:43.669326963+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":318186}
+{"time":"2025-11-07T12:29:43.67598157+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:29:48.678258885+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":319143}
+{"time":"2025-11-07T12:29:48.678326324+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:48.678826693+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":319175}
+{"time":"2025-11-07T12:29:48.683745918+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":426}
+{"time":"2025-11-07T12:29:53.687650779+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":319722}
+{"time":"2025-11-07T12:29:53.68774229+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:53.688315127+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":319754}
+{"time":"2025-11-07T12:29:53.694546717+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":585}
+{"time":"2025-11-07T12:29:58.696969162+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":320412}
+{"time":"2025-11-07T12:29:58.697067326+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:29:58.697612574+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":320444}
+{"time":"2025-11-07T12:29:58.70385278+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:30:03.707165346+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":321189}
+{"time":"2025-11-07T12:30:03.707296755+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:03.707597117+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":321221}
+{"time":"2025-11-07T12:30:03.714031023+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:30:08.717240257+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":321950}
+{"time":"2025-11-07T12:30:08.717386832+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:08.717745307+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":321982}
+{"time":"2025-11-07T12:30:08.723638235+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":614}
+{"time":"2025-11-07T12:30:13.72657206+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":322726}
+{"time":"2025-11-07T12:30:13.726675212+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:13.727030373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":322758}
+{"time":"2025-11-07T12:30:13.732649777+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":619}
+{"time":"2025-11-07T12:30:18.73857981+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":323677}
+{"time":"2025-11-07T12:30:18.743087187+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":438}
+{"time":"2025-11-07T12:30:23.748038075+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":324362}
+{"time":"2025-11-07T12:30:23.748119944+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:23.748688653+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":324394}
+{"time":"2025-11-07T12:30:23.753505974+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":500}
+{"time":"2025-11-07T12:30:28.758663306+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":325017}
+{"time":"2025-11-07T12:30:28.758911599+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:28.759392205+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":325049}
+{"time":"2025-11-07T12:30:28.763329596+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":585}
+{"time":"2025-11-07T12:30:33.766994373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":325703}
+{"time":"2025-11-07T12:30:33.767140868+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:33.767471564+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":325735}
+{"time":"2025-11-07T12:30:33.773978772+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":639}
+{"time":"2025-11-07T12:30:38.79361725+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":326476}
+{"time":"2025-11-07T12:30:38.793761743+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:38.795913642+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":326508}
+{"time":"2025-11-07T12:30:38.802959868+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":643}
+{"time":"2025-11-07T12:30:43.806785576+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":327232}
+{"time":"2025-11-07T12:30:43.812854073+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":700}
+{"time":"2025-11-07T12:30:48.818005696+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":328107}
+{"time":"2025-11-07T12:30:48.818159742+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:48.818517764+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":328139}
+{"time":"2025-11-07T12:30:48.821838417+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":531}
+{"time":"2025-11-07T12:30:53.826071644+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":328747}
+{"time":"2025-11-07T12:30:53.831512128+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":661}
+{"time":"2025-11-07T12:30:58.834563749+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":329479}
+{"time":"2025-11-07T12:30:58.834662198+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:30:58.835110064+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":329511}
+{"time":"2025-11-07T12:30:58.841778151+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:31:03.844077755+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":330220}
+{"time":"2025-11-07T12:31:03.844192833+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:03.845215647+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":330253}
+{"time":"2025-11-07T12:31:03.85212981+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":674}
+{"time":"2025-11-07T12:31:08.856852456+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":331225}
+{"time":"2025-11-07T12:31:08.85693832+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:08.857379373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":331257}
+{"time":"2025-11-07T12:31:08.862083714+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":408}
+{"time":"2025-11-07T12:31:13.865471264+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":331747}
+{"time":"2025-11-07T12:31:13.865610994+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:13.866074184+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":331779}
+{"time":"2025-11-07T12:31:13.872601505+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":667}
+{"time":"2025-11-07T12:31:18.87530258+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":332515}
+{"time":"2025-11-07T12:31:18.875425401+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:18.875815244+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":332547}
+{"time":"2025-11-07T12:31:18.882857019+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:31:23.886378797+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":333335}
+{"time":"2025-11-07T12:31:23.886449315+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:23.886888949+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":333367}
+{"time":"2025-11-07T12:31:23.892450926+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":555}
+{"time":"2025-11-07T12:31:28.89669466+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":334049}
+{"time":"2025-11-07T12:31:28.896817728+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:28.897171242+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":334081}
+{"time":"2025-11-07T12:31:28.90311457+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":622}
+{"time":"2025-11-07T12:31:33.909070134+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":334908}
+{"time":"2025-11-07T12:31:33.909409707+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:33.909943505+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":334940}
+{"time":"2025-11-07T12:31:33.913353139+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":501}
+{"time":"2025-11-07T12:31:38.917278355+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":335512}
+{"time":"2025-11-07T12:31:38.917406757+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:38.918443324+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":335545}
+{"time":"2025-11-07T12:31:38.92441552+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:31:43.92798623+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":336363}
+{"time":"2025-11-07T12:31:43.928069816+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:43.928479775+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":336395}
+{"time":"2025-11-07T12:31:43.933514867+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":565}
+{"time":"2025-11-07T12:31:48.936070196+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":337030}
+{"time":"2025-11-07T12:31:48.93622009+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:48.936654827+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":337062}
+{"time":"2025-11-07T12:31:48.943372743+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:31:53.946964876+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":337943}
+{"time":"2025-11-07T12:31:53.947083985+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:53.947537633+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":337975}
+{"time":"2025-11-07T12:31:53.952611157+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":502}
+{"time":"2025-11-07T12:31:58.955817379+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":338613}
+{"time":"2025-11-07T12:31:58.955889099+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:31:58.95698083+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":338646}
+{"time":"2025-11-07T12:31:58.96285186+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":571}
+{"time":"2025-11-07T12:32:03.966347632+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":339286}
+{"time":"2025-11-07T12:32:03.966439291+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:03.966917564+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":339318}
+{"time":"2025-11-07T12:32:03.973309165+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:32:08.977058696+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":340215}
+{"time":"2025-11-07T12:32:08.977213044+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:08.977565917+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":340247}
+{"time":"2025-11-07T12:32:08.98270658+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":487}
+{"time":"2025-11-07T12:32:13.991487709+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":340969}
+{"time":"2025-11-07T12:32:13.991593141+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:13.991961002+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":341001}
+{"time":"2025-11-07T12:32:13.995115077+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":473}
+{"time":"2025-11-07T12:32:18.998973387+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":341622}
+{"time":"2025-11-07T12:32:18.999066964+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:18.999505388+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":341654}
+{"time":"2025-11-07T12:32:19.005169103+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":558}
+{"time":"2025-11-07T12:32:24.008783529+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":342283}
+{"time":"2025-11-07T12:32:24.008956106+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:24.009385251+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":342315}
+{"time":"2025-11-07T12:32:24.015414028+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:32:29.021790842+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":343547}
+{"time":"2025-11-07T12:32:29.021866558+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:29.022284032+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":343579}
+{"time":"2025-11-07T12:32:29.023808486+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":152}
+{"time":"2025-11-07T12:32:34.028927619+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":343948}
+{"time":"2025-11-07T12:32:34.028993037+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:34.02937415+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":343980}
+{"time":"2025-11-07T12:32:34.034503922+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":489}
+{"time":"2025-11-07T12:32:39.039202754+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":344912}
+{"time":"2025-11-07T12:32:39.039267062+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:39.039704278+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":344944}
+{"time":"2025-11-07T12:32:39.043479578+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":304}
+{"time":"2025-11-07T12:32:44.048186126+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":345411}
+{"time":"2025-11-07T12:32:44.048290028+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:44.048621058+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":345443}
+{"time":"2025-11-07T12:32:44.054097124+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":586}
+{"time":"2025-11-07T12:32:49.058172572+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":346267}
+{"time":"2025-11-07T12:32:49.058271606+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:49.058666922+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":346299}
+{"time":"2025-11-07T12:32:49.063505288+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":468}
+{"time":"2025-11-07T12:32:54.067730375+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":346916}
+{"time":"2025-11-07T12:32:54.067856592+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:54.068247592+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":346948}
+{"time":"2025-11-07T12:32:54.073979911+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":557}
+{"time":"2025-11-07T12:32:59.077332414+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":347607}
+{"time":"2025-11-07T12:32:59.077450505+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:32:59.078625477+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":347640}
+{"time":"2025-11-07T12:32:59.084554988+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":646}
+{"time":"2025-11-07T12:33:04.088180198+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":348402}
+{"time":"2025-11-07T12:33:04.088257429+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:04.089225979+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":348435}
+{"time":"2025-11-07T12:33:04.094747127+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":589}
+{"time":"2025-11-07T12:33:09.099884874+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":349240}
+{"time":"2025-11-07T12:33:09.100253621+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:09.10042869+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":349272}
+{"time":"2025-11-07T12:33:09.10467572+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":490}
+{"time":"2025-11-07T12:33:14.107815731+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":349835}
+{"time":"2025-11-07T12:33:14.107925663+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:14.108383968+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":349867}
+{"time":"2025-11-07T12:33:14.115013107+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:33:19.116919631+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":350645}
+{"time":"2025-11-07T12:33:19.117003503+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:19.117456282+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":350677}
+{"time":"2025-11-07T12:33:19.123772901+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":604}
+{"time":"2025-11-07T12:33:24.126408097+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":351466}
+{"time":"2025-11-07T12:33:24.126518775+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:24.127076524+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":351498}
+{"time":"2025-11-07T12:33:24.133812831+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":562}
+{"time":"2025-11-07T12:33:29.136823599+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":352131}
+{"time":"2025-11-07T12:33:29.137029058+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:29.137520262+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":352163}
+{"time":"2025-11-07T12:33:29.144275223+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:33:34.147944076+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":352870}
+{"time":"2025-11-07T12:33:34.148075934+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:34.148593115+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":352902}
+{"time":"2025-11-07T12:33:34.15516856+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:33:39.160588665+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":353842}
+{"time":"2025-11-07T12:33:39.165411263+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":434}
+{"time":"2025-11-07T12:33:44.169503664+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":354379}
+{"time":"2025-11-07T12:33:44.169614117+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:44.17004469+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":354411}
+{"time":"2025-11-07T12:33:44.175488399+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":605}
+{"time":"2025-11-07T12:33:49.178968872+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":355092}
+{"time":"2025-11-07T12:33:49.17911926+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:49.179929932+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":355125}
+{"time":"2025-11-07T12:33:49.185711907+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":670}
+{"time":"2025-11-07T12:33:54.191228289+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":356066}
+{"time":"2025-11-07T12:33:54.191946199+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":356098}
+{"time":"2025-11-07T12:33:54.19139736+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:54.195114374+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":435}
+{"time":"2025-11-07T12:33:59.200896014+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":356722}
+{"time":"2025-11-07T12:33:59.200990421+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:33:59.201368448+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":356754}
+{"time":"2025-11-07T12:33:59.204781671+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":519}
+{"time":"2025-11-07T12:34:04.209180543+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":357382}
+{"time":"2025-11-07T12:34:04.209321055+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:04.209838331+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":357414}
+{"time":"2025-11-07T12:34:04.215835056+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":638}
+{"time":"2025-11-07T12:34:09.219439064+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":358121}
+{"time":"2025-11-07T12:34:09.219572667+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:09.219944202+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":358153}
+{"time":"2025-11-07T12:34:09.225981689+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":639}
+{"time":"2025-11-07T12:34:14.252982095+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":358903}
+{"time":"2025-11-07T12:34:14.253097002+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:14.253657551+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":358935}
+{"time":"2025-11-07T12:34:14.258392293+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:34:19.263206539+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":359685}
+{"time":"2025-11-07T12:34:19.263309816+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:19.263855364+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":359717}
+{"time":"2025-11-07T12:34:19.270152179+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":592}
+{"time":"2025-11-07T12:34:24.273580246+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":360426}
+{"time":"2025-11-07T12:34:24.273687903+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:24.274147001+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":360458}
+{"time":"2025-11-07T12:34:24.280508949+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":630}
+{"time":"2025-11-07T12:34:29.286243368+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":361161}
+{"time":"2025-11-07T12:34:29.286415182+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:29.28674683+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":361193}
+{"time":"2025-11-07T12:34:29.292880942+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:34:34.313415876+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":361929}
+{"time":"2025-11-07T12:34:34.313564768+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:34.313941372+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":361961}
+{"time":"2025-11-07T12:34:34.319934472+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":646}
+{"time":"2025-11-07T12:34:39.326455099+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":362978}
+{"time":"2025-11-07T12:34:39.326553225+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:39.326983284+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":363010}
+{"time":"2025-11-07T12:34:39.32924643+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":335}
+{"time":"2025-11-07T12:34:44.336179802+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":363503}
+{"time":"2025-11-07T12:34:44.343799974+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":623}
+{"time":"2025-11-07T12:34:49.351520361+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":364488}
+{"time":"2025-11-07T12:34:49.351606904+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:49.352106455+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":364520}
+{"time":"2025-11-07T12:34:49.354550511+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":344}
+{"time":"2025-11-07T12:34:54.360029559+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":365062}
+{"time":"2025-11-07T12:34:54.360122406+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:54.361059808+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":365095}
+{"time":"2025-11-07T12:34:54.365499996+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":548}
+{"time":"2025-11-07T12:34:59.370641379+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":365777}
+{"time":"2025-11-07T12:34:59.370728715+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:34:59.371117067+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":365809}
+{"time":"2025-11-07T12:34:59.376078054+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":574}
+{"time":"2025-11-07T12:35:04.380501609+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":366666}
+{"time":"2025-11-07T12:35:04.380581188+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:04.380986318+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":366698}
+{"time":"2025-11-07T12:35:04.385736063+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":423}
+{"time":"2025-11-07T12:35:09.38954812+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":367314}
+{"time":"2025-11-07T12:35:09.389641619+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:09.390080053+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":367346}
+{"time":"2025-11-07T12:35:09.395539729+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":554}
+{"time":"2025-11-07T12:35:14.399438357+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":367972}
+{"time":"2025-11-07T12:35:14.399600277+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:14.39999693+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":368004}
+{"time":"2025-11-07T12:35:14.407148002+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":650}
+{"time":"2025-11-07T12:35:19.429073925+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":368771}
+{"time":"2025-11-07T12:35:19.429168916+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:19.429582538+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":368803}
+{"time":"2025-11-07T12:35:19.436001802+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":616}
+{"time":"2025-11-07T12:35:24.439641878+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":369488}
+{"time":"2025-11-07T12:35:24.439771236+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:24.440124621+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":369520}
+{"time":"2025-11-07T12:35:24.447103539+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:35:29.450284322+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":370229}
+{"time":"2025-11-07T12:35:29.450447166+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:29.450843661+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":370261}
+{"time":"2025-11-07T12:35:29.45699645+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:35:34.461146805+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":370967}
+{"time":"2025-11-07T12:35:34.461254854+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:34.461685168+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":370999}
+{"time":"2025-11-07T12:35:34.468117477+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:35:39.472017228+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":371714}
+{"time":"2025-11-07T12:35:39.472220925+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:39.472715781+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":371746}
+{"time":"2025-11-07T12:35:39.478957968+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":668}
+{"time":"2025-11-07T12:35:44.483082271+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":372485}
+{"time":"2025-11-07T12:35:44.489046215+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":669}
+{"time":"2025-11-07T12:35:49.494743557+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":373324}
+{"time":"2025-11-07T12:35:49.494891101+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:49.496083295+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":373378}
+{"time":"2025-11-07T12:35:49.499495257+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":514}
+{"time":"2025-11-07T12:35:54.503806657+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":373963}
+{"time":"2025-11-07T12:35:54.503972614+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:54.504997083+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":373996}
+{"time":"2025-11-07T12:35:54.511034086+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:35:59.515043097+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":374870}
+{"time":"2025-11-07T12:35:59.515144688+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:35:59.51555606+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":374902}
+{"time":"2025-11-07T12:35:59.520828957+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":550}
+{"time":"2025-11-07T12:36:04.526075128+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":375994}
+{"time":"2025-11-07T12:36:04.52614469+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:04.527162619+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":376027}
+{"time":"2025-11-07T12:36:04.5299766+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":204}
+{"time":"2025-11-07T12:36:09.535079317+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":376300}
+{"time":"2025-11-07T12:36:09.535205473+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:09.536223712+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":376333}
+{"time":"2025-11-07T12:36:09.542173213+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":636}
+{"time":"2025-11-07T12:36:14.547654383+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":377251}
+{"time":"2025-11-07T12:36:14.548268067+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:14.548592568+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":377283}
+{"time":"2025-11-07T12:36:14.551861604+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":467}
+{"time":"2025-11-07T12:36:19.557284201+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":377896}
+{"time":"2025-11-07T12:36:19.557366187+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:19.557852973+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":377928}
+{"time":"2025-11-07T12:36:19.563348796+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":601}
+{"time":"2025-11-07T12:36:24.567514917+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":378598}
+{"time":"2025-11-07T12:36:24.567672172+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:24.567994921+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":378630}
+{"time":"2025-11-07T12:36:24.575016897+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":637}
+{"time":"2025-11-07T12:36:29.578748846+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":379345}
+{"time":"2025-11-07T12:36:29.578838691+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:29.579308917+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":379377}
+{"time":"2025-11-07T12:36:29.586253625+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":671}
+{"time":"2025-11-07T12:36:34.589488121+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":380155}
+{"time":"2025-11-07T12:36:34.589558014+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:34.590480059+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":380188}
+{"time":"2025-11-07T12:36:34.596163409+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":598}
+{"time":"2025-11-07T12:36:39.602022056+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":381094}
+{"time":"2025-11-07T12:36:39.602227149+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:39.602849141+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":381126}
+{"time":"2025-11-07T12:36:39.606016676+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":439}
+{"time":"2025-11-07T12:36:44.610887404+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":381649}
+{"time":"2025-11-07T12:36:44.616530746+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":697}
+{"time":"2025-11-07T12:36:49.622182527+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":382466}
+{"time":"2025-11-07T12:36:49.622266536+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:49.622801584+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":382498}
+{"time":"2025-11-07T12:36:49.627645895+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":586}
+{"time":"2025-11-07T12:36:54.631408653+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":383200}
+{"time":"2025-11-07T12:36:54.63153355+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:54.632083852+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":383232}
+{"time":"2025-11-07T12:36:54.638494616+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":631}
+{"time":"2025-11-07T12:36:59.641090341+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":383936}
+{"time":"2025-11-07T12:36:59.641300242+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:36:59.641517627+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":383968}
+{"time":"2025-11-07T12:36:59.648417795+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":676}
+{"time":"2025-11-07T12:37:04.652146374+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":384750}
+{"time":"2025-11-07T12:37:04.65227299+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:04.652691474+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":384782}
+{"time":"2025-11-07T12:37:04.659097526+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":600}
+{"time":"2025-11-07T12:37:09.664988466+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":385787}
+{"time":"2025-11-07T12:37:09.668639949+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":374}
+{"time":"2025-11-07T12:37:14.672258046+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":386245}
+{"time":"2025-11-07T12:37:14.672400115+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:14.672927596+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":386277}
+{"time":"2025-11-07T12:37:14.67907601+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":626}
+{"time":"2025-11-07T12:37:19.692108789+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":386987}
+{"time":"2025-11-07T12:37:19.692294771+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:19.692796057+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":387019}
+{"time":"2025-11-07T12:37:19.699386275+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":661}
+{"time":"2025-11-07T12:37:24.705212344+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":387915}
+{"time":"2025-11-07T12:37:24.705397657+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:24.706400631+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":387948}
+{"time":"2025-11-07T12:37:24.709710064+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":511}
+{"time":"2025-11-07T12:37:29.713317894+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":388561}
+{"time":"2025-11-07T12:37:29.713397873+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:29.713910156+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":388593}
+{"time":"2025-11-07T12:37:29.718141835+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":606}
+{"time":"2025-11-07T12:37:34.722087377+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":389272}
+{"time":"2025-11-07T12:37:34.722306674+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:34.722582479+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":389304}
+{"time":"2025-11-07T12:37:34.729352947+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":674}
+{"time":"2025-11-07T12:37:39.732706464+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":390117}
+{"time":"2025-11-07T12:37:39.732786191+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:39.733362439+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":390149}
+{"time":"2025-11-07T12:37:39.738958567+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":608}
+{"time":"2025-11-07T12:37:44.747035914+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":390931}
+{"time":"2025-11-07T12:37:44.747108634+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:44.7475733+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":390963}
+{"time":"2025-11-07T12:37:44.753049694+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":534}
+{"time":"2025-11-07T12:37:49.75837369+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":391725}
+{"time":"2025-11-07T12:37:49.758508797+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:49.758887097+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":391757}
+{"time":"2025-11-07T12:37:49.764225775+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":519}
+{"time":"2025-11-07T12:37:54.770780885+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":392675}
+{"time":"2025-11-07T12:37:54.77387722+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":380}
+{"time":"2025-11-07T12:37:59.778632083+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":393147}
+{"time":"2025-11-07T12:37:59.778758386+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:37:59.779097594+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":393179}
+{"time":"2025-11-07T12:37:59.784674807+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":616}
+{"time":"2025-11-07T12:38:04.787795537+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":393869}
+{"time":"2025-11-07T12:38:04.787906785+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:04.788990426+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":393902}
+{"time":"2025-11-07T12:38:04.7953259+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":672}
+{"time":"2025-11-07T12:38:09.799197273+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":394717}
+{"time":"2025-11-07T12:38:09.799263569+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:09.799693145+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":394749}
+{"time":"2025-11-07T12:38:09.80536919+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":604}
+{"time":"2025-11-07T12:38:14.810739566+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":395680}
+{"time":"2025-11-07T12:38:14.81081083+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:14.811626084+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":395713}
+{"time":"2025-11-07T12:38:14.811732589+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:14.811799506+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":395745}
+{"time":"2025-11-07T12:38:14.815689993+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":383}
+{"time":"2025-11-07T12:38:19.821237152+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":396204}
+{"time":"2025-11-07T12:38:19.821349254+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:19.822391853+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":396237}
+{"time":"2025-11-07T12:38:19.828299124+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":635}
+{"time":"2025-11-07T12:38:24.831673626+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":397193}
+{"time":"2025-11-07T12:38:24.831749252+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:24.832282373+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":397225}
+{"time":"2025-11-07T12:38:24.837416365+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":426}
+{"time":"2025-11-07T12:38:29.841814515+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":397724}
+{"time":"2025-11-07T12:38:29.841949046+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:29.84277576+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":397757}
+{"time":"2025-11-07T12:38:29.848503418+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":675}
+{"time":"2025-11-07T12:38:34.855618547+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":398678}
+{"time":"2025-11-07T12:38:34.855699393+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:34.856140359+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":398710}
+{"time":"2025-11-07T12:38:34.859588039+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":501}
+{"time":"2025-11-07T12:38:39.865317473+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":399293}
+{"time":"2025-11-07T12:38:39.871146664+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":685}
+{"time":"2025-11-07T12:38:44.878800509+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":400137}
+{"time":"2025-11-07T12:38:44.878886935+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:38:44.879408558+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":400169}
+{"time":"2025-11-07T12:38:44.884862992+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":561}
+{"time":"2025-11-07T12:38:45.543431116+08:00","level":"INFO","msg":"stream: closing","id":"k31oio6c"}
+{"time":"2025-11-07T12:38:45.54656445+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":400769}
+{"time":"2025-11-07T12:38:45.547626739+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":84}
+{"time":"2025-11-07T12:38:49.064707374+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2025-11-07T12:38:49.399801242+08:00","level":"INFO","msg":"handler: closed","stream_id":"k31oio6c"}
+{"time":"2025-11-07T12:38:49.40214424+08:00","level":"INFO","msg":"sender: closed","stream_id":"k31oio6c"}
+{"time":"2025-11-07T12:38:49.402246481+08:00","level":"INFO","msg":"stream: closed","id":"k31oio6c"}
diff --git a/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug.log b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..da88c49425fb5cfedcb9e140bd4de0a0d34dc582
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug.log
@@ -0,0 +1,24 @@
+2025-11-07 11:38:56,186 INFO    MainThread:214168 [wandb_setup.py:_flush():81] Current SDK version is 0.22.3
+2025-11-07 11:38:56,186 INFO    MainThread:214168 [wandb_setup.py:_flush():81] Configure stats pid to 214168
+2025-11-07 11:38:56,186 INFO    MainThread:214168 [wandb_setup.py:_flush():81] Loading settings from /root/.config/wandb/settings
+2025-11-07 11:38:56,186 INFO    MainThread:214168 [wandb_setup.py:_flush():81] Loading settings from /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/settings
+2025-11-07 11:38:56,186 INFO    MainThread:214168 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2025-11-07 11:38:56,187 INFO    MainThread:214168 [wandb_init.py:setup_run_log_directory():706] Logging user logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug.log
+2025-11-07 11:38:56,187 INFO    MainThread:214168 [wandb_init.py:setup_run_log_directory():707] Logging internal logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_113856-k31oio6c/logs/debug-internal.log
+2025-11-07 11:38:56,188 INFO    MainThread:214168 [wandb_init.py:init():833] calling init triggers
+2025-11-07 11:38:56,188 INFO    MainThread:214168 [wandb_init.py:init():838] wandb.init called with sweep_config: {}
+config: {'exp_name': 'ppo_frozenlake', 'seed': 1, 'torch_deterministic': True, 'cuda': True, 'track': True, 'wandb_project_name': 'ragen-bandit', 'wandb_entity': None, 'capture_video': False, 'env_id': 'FrozenLake', 'total_timesteps': 10000000, 'learning_rate': 0.00025, 'num_envs': 8, 'num_steps': 128, 'anneal_lr': True, 'gamma': 0.99, 'gae_lambda': 0.95, 'num_minibatches': 4, 'update_epochs': 4, 'norm_adv': True, 'clip_coef': 0.2, 'clip_vloss': True, 'ent_coef': 0.01, 'vf_coef': 0.5, 'max_grad_norm': 0.5, 'target_kl': None, 'grid_size': 4, 'is_slippery': True, 'batch_size': 1024, 'minibatch_size': 256, 'num_iterations': 9765, '_wandb': {'code_path': 'code/cleanrl/ppo_frozenlake.py'}}
+2025-11-07 11:38:56,188 INFO    MainThread:214168 [wandb_init.py:init():881] starting backend
+2025-11-07 11:38:56,395 INFO    MainThread:214168 [wandb_init.py:init():884] sending inform_init request
+2025-11-07 11:38:56,406 INFO    MainThread:214168 [wandb_init.py:init():892] backend started and connected
+2025-11-07 11:38:56,408 INFO    MainThread:214168 [wandb_init.py:init():962] updated telemetry
+2025-11-07 11:38:56,438 INFO    MainThread:214168 [wandb_init.py:init():986] communicating run to backend with 90.0 second timeout
+2025-11-07 11:38:57,166 INFO    MainThread:214168 [wandb_init.py:init():1033] starting run threads in backend
+2025-11-07 11:38:57,309 INFO    MainThread:214168 [wandb_run.py:_console_start():2506] atexit reg
+2025-11-07 11:38:57,310 INFO    MainThread:214168 [wandb_run.py:_redirect():2354] redirect: wrap_raw
+2025-11-07 11:38:57,310 INFO    MainThread:214168 [wandb_run.py:_redirect():2423] Wrapping output streams.
+2025-11-07 11:38:57,310 INFO    MainThread:214168 [wandb_run.py:_redirect():2446] Redirects installed.
+2025-11-07 11:38:57,312 INFO    MainThread:214168 [wandb_init.py:init():1073] run started, returning control to user process
+2025-11-07 11:38:57,312 INFO    MainThread:214168 [wandb_run.py:_tensorboard_callback():1598] tensorboard callback: runs/FrozenLake__ppo_frozenlake__1__1762486729, True
+2025-11-07 12:38:45,543 INFO    wandb-AsyncioManager-main:214168 [service_client.py:_forward_responses():80] Reached EOF.
+2025-11-07 12:38:45,543 INFO    wandb-AsyncioManager-main:214168 [mailbox.py:close():137] Closing mailbox, abandoning 1 handles.
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/code/cleanrl/dqn_bandit.py b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/code/cleanrl/dqn_bandit.py
new file mode 100644
index 0000000000000000000000000000000000000000..1762efada08f2b5a1a0aa8b78b2d8051353e312a
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/code/cleanrl/dqn_bandit.py
@@ -0,0 +1,274 @@
+# DQN implementation for RAGEN Bandit Environment
+# Adapted from dqn_atari.py for single-step bandit problem
+import os
+import random
+import sys
+import time
+from dataclasses import dataclass
+
+# Add parent directory to path to import the wrapper
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+from ragen.env.bandit.env import BanditEnv
+from ragen.env.bandit.config import BanditEnvConfig
+from ragen_wrappers import BanditWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+
+    # Algorithm specific arguments
+    env_id: str = "Bandit"
+    """the id of the environment"""
+    total_timesteps: int = 100000
+    """total timesteps of the experiments"""
+    learning_rate: float = 1e-3
+    """the learning rate of the optimizer"""
+    num_envs: int = 1
+    """the number of parallel game environments (DQN typically uses 1)"""
+    buffer_size: int = 10000
+    """the replay memory buffer size"""
+    gamma: float = 0.0
+    """the discount factor gamma (0 for bandit since it's single-step)"""
+    tau: float = 1.0
+    """the target network update rate"""
+    target_network_frequency: int = 500
+    """the timesteps it takes to update the target network"""
+    batch_size: int = 32
+    """the batch size of sample from the reply memory"""
+    start_e: float = 1.0
+    """the starting epsilon for exploration"""
+    end_e: float = 0.05
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.5
+    """the fraction of `total-timesteps` it takes from start-e to go end-e"""
+    learning_starts: int = 1000
+    """timestep to start learning"""
+    train_frequency: int = 1
+    """the frequency of training"""
+    steps_per_episode: int = 10
+    """number of steps per episode for multi-step bandit"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, steps_per_episode):
+    def thunk():
+        config = BanditEnvConfig()
+        env = BanditEnv(config)
+        env = BanditWrapper(env, steps_per_episode=steps_per_episode)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    """
+    Q-Network for Bandit environment.
+    Input: [arm0_pulls, arm0_avg_reward, arm1_pulls, arm1_avg_reward]
+    Output: Q-values for each arm
+    """
+    def __init__(self, env):
+        super().__init__()
+        obs_shape = np.array(env.single_observation_space.shape).prod()
+        self.network = nn.Sequential(
+            nn.Linear(obs_shape, 128),
+            nn.ReLU(),
+            nn.Linear(128, 128),
+            nn.ReLU(),
+            nn.Linear(128, env.single_action_space.n),
+        )
+
+    def forward(self, x):
+        return self.network(x)
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    assert args.num_envs == 1, "vectorized envs are not supported at the moment"
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed + i, args.steps_per_episode) 
+         for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.Adam(q_network.parameters(), lr=args.learning_rate)
+    target_network = QNetwork(envs).to(device)
+    target_network.load_state_dict(q_network.state_dict())
+
+    # Use simple replay buffer (no special memory optimization needed for bandit)
+    from cleanrl_utils.buffers import ReplayBuffer
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        optimize_memory_usage=False,
+        handle_timeout_termination=False,
+    )
+    
+    start_time = time.time()
+    
+    # Track episode returns
+    episode_returns = []
+    recent_episode_returns = []
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = q_network(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if info and "episode" in info:
+                    episode_return = info['episode']['r']
+                    episode_length = info['episode']['l']
+                    episode_returns.append(episode_return)
+                    recent_episode_returns.append(episode_return)
+                    if len(recent_episode_returns) > 10:
+                        recent_episode_returns.pop(0)
+                    writer.add_scalar("charts/episodic_return", episode_return, global_step)
+                    writer.add_scalar("charts/episodic_length", episode_length, global_step)
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            if global_step % args.train_frequency == 0:
+                data = rb.sample(args.batch_size)
+                with torch.no_grad():
+                    target_max, _ = target_network(data.next_observations).max(dim=1)
+                    td_target = data.rewards.flatten() + args.gamma * target_max * (1 - data.dones.flatten())
+                old_val = q_network(data.observations).gather(1, data.actions).squeeze()
+                loss = F.mse_loss(td_target, old_val)
+
+                if global_step % 100 == 0:
+                    writer.add_scalar("losses/td_loss", loss, global_step)
+                    writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+                    writer.add_scalar("charts/epsilon", epsilon, global_step)
+                    
+                    # Console output with key metrics
+                    sps = int(global_step / (time.time() - start_time))
+                    progress = 100 * global_step / args.total_timesteps
+                    avg_episode_return = np.mean(recent_episode_returns) if recent_episode_returns else 0.0
+                    
+                    print(f"[{progress:5.1f}%] Step {global_step:6d}/{args.total_timesteps} | "
+                          f"SPS: {sps:5d} | "
+                          f"EpRet: {avg_episode_return:7.3f} | "
+                          f"Loss: {loss.item():.4f} | "
+                          f"Q-val: {old_val.mean().item():.4f} | "
+                          f"Eps: {epsilon:.3f}")
+                    
+                    writer.add_scalar("charts/SPS", sps, global_step)
+                    if recent_episode_returns:
+                        writer.add_scalar("charts/avg_episodic_return", avg_episode_return, global_step)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                optimizer.step()
+
+            # update target network
+            if global_step % args.target_network_frequency == 0:
+                for target_network_param, q_network_param in zip(target_network.parameters(), q_network.parameters()):
+                    target_network_param.data.copy_(
+                        args.tau * q_network_param.data + (1.0 - args.tau) * target_network_param.data
+                    )
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        torch.save(q_network.state_dict(), model_path)
+        print(f"model saved to {model_path}")
+
+    envs.close()
+    writer.close()
+    
+    print("\n" + "="*60)
+    print("Training Complete!")
+    print("="*60)
+    if episode_returns:
+        print(f"Final Average Episode Return (last 10): {np.mean(recent_episode_returns):.3f}")
+        print(f"Overall Average Episode Return: {np.mean(episode_returns):.3f}")
+        print(f"Best Episode Return: {max(episode_returns):.3f}")
+    print("="*60)
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b906123a360341d422bfb4e66a9fc3edcb561d64
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/config.yaml
@@ -0,0 +1,145 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/dqn_bandit.py
+        e:
+            tkb6lq3xdk95jdrk8itil4tq7t5n4mli:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-bandit
+                codePath: cleanrl/dqn_bandit.py
+                codePathLocal: dqn_bandit.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "31715360768"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/dqn_bandit.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T04:52:10.930319Z"
+                writerId: tkb6lq3xdk95jdrk8itil4tq7t5n4mli
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+batch_size:
+    value: 32
+buffer_size:
+    value: 10000
+capture_video:
+    value: false
+cuda:
+    value: true
+end_e:
+    value: 0.05
+env_id:
+    value: Bandit
+exp_name:
+    value: dqn_bandit
+exploration_fraction:
+    value: 0.5
+gamma:
+    value: 0
+learning_rate:
+    value: 0.001
+learning_starts:
+    value: 1000
+num_envs:
+    value: 1
+save_model:
+    value: false
+seed:
+    value: 1
+start_e:
+    value: 1
+steps_per_episode:
+    value: 10
+target_network_frequency:
+    value: 500
+tau:
+    value: 1
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 100000
+track:
+    value: true
+train_frequency:
+    value: 1
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-bandit
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/wandb-summary.json b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..379fe911e23a828fc54c3b9736a610509e9c45f0
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/files/wandb-summary.json
@@ -0,0 +1 @@
+{"_wandb":{"runtime":10},"_runtime":10}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug-internal.log b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..2551cb5d0100780312cb54f0333b97d71f6c983d
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug-internal.log
@@ -0,0 +1,15 @@
+{"time":"2025-11-07T12:52:11.203166597+08:00","level":"INFO","msg":"stream: starting","core version":"0.22.3"}
+{"time":"2025-11-07T12:52:11.577529761+08:00","level":"INFO","msg":"stream: created new stream","id":"py6fnjml"}
+{"time":"2025-11-07T12:52:11.5778486+08:00","level":"INFO","msg":"handler: started","stream_id":"py6fnjml"}
+{"time":"2025-11-07T12:52:11.581014824+08:00","level":"INFO","msg":"stream: started","id":"py6fnjml"}
+{"time":"2025-11-07T12:52:11.581034736+08:00","level":"INFO","msg":"writer: started","stream_id":"py6fnjml"}
+{"time":"2025-11-07T12:52:11.581039208+08:00","level":"INFO","msg":"sender: started","stream_id":"py6fnjml"}
+{"time":"2025-11-07T12:52:11.932469757+08:00","level":"ERROR","msg":"git repo not found","error":"repository does not exist"}
+{"time":"2025-11-07T12:52:12.181621426+08:00","level":"INFO","msg":"stream: closing","id":"py6fnjml"}
+{"time":"2025-11-07T12:52:22.072619394+08:00","level":"WARN","msg":"tensorboard: no root directory","error":"timed out after 10s"}
+{"time":"2025-11-07T12:52:22.075996307+08:00","level":"INFO","msg":"tensorboard: tracking new log directory","rootDir":{},"logDir":"LocalOrCloudPath(LocalPath=\"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491124\")","namespace":""}
+{"time":"2025-11-07T12:52:22.07870238+08:00","level":"INFO","msg":"tensorboard: saving file","fileLocation":"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491124/events.out.tfevents.1762491132.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.233243.0","runPath":"runs/Bandit__dqn_bandit__1__1762491124/events.out.tfevents.1762491132.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.233243.0"}
+{"time":"2025-11-07T12:52:23.615791082+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2025-11-07T12:52:23.935404884+08:00","level":"INFO","msg":"handler: closed","stream_id":"py6fnjml"}
+{"time":"2025-11-07T12:52:23.937646561+08:00","level":"INFO","msg":"sender: closed","stream_id":"py6fnjml"}
+{"time":"2025-11-07T12:52:23.937747171+08:00","level":"INFO","msg":"stream: closed","id":"py6fnjml"}
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug.log b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..e2feec3120a8f2d49cd412892ce9567146b46306
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug.log
@@ -0,0 +1,24 @@
+2025-11-07 12:52:10,970 INFO    MainThread:233243 [wandb_setup.py:_flush():81] Current SDK version is 0.22.3
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_setup.py:_flush():81] Configure stats pid to 233243
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_setup.py:_flush():81] Loading settings from /root/.config/wandb/settings
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_setup.py:_flush():81] Loading settings from /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/settings
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_init.py:setup_run_log_directory():706] Logging user logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug.log
+2025-11-07 12:52:10,971 INFO    MainThread:233243 [wandb_init.py:setup_run_log_directory():707] Logging internal logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125210-py6fnjml/logs/debug-internal.log
+2025-11-07 12:52:10,973 INFO    MainThread:233243 [wandb_init.py:init():833] calling init triggers
+2025-11-07 12:52:10,973 INFO    MainThread:233243 [wandb_init.py:init():838] wandb.init called with sweep_config: {}
+config: {'exp_name': 'dqn_bandit', 'seed': 1, 'torch_deterministic': True, 'cuda': True, 'track': True, 'wandb_project_name': 'ragen-bandit', 'wandb_entity': None, 'capture_video': False, 'save_model': False, 'env_id': 'Bandit', 'total_timesteps': 100000, 'learning_rate': 0.001, 'num_envs': 1, 'buffer_size': 10000, 'gamma': 0.0, 'tau': 1.0, 'target_network_frequency': 500, 'batch_size': 32, 'start_e': 1.0, 'end_e': 0.05, 'exploration_fraction': 0.5, 'learning_starts': 1000, 'train_frequency': 1, 'steps_per_episode': 10, '_wandb': {'code_path': 'code/cleanrl/dqn_bandit.py'}}
+2025-11-07 12:52:10,973 INFO    MainThread:233243 [wandb_init.py:init():881] starting backend
+2025-11-07 12:52:11,189 INFO    MainThread:233243 [wandb_init.py:init():884] sending inform_init request
+2025-11-07 12:52:11,200 INFO    MainThread:233243 [wandb_init.py:init():892] backend started and connected
+2025-11-07 12:52:11,203 INFO    MainThread:233243 [wandb_init.py:init():962] updated telemetry
+2025-11-07 12:52:11,235 INFO    MainThread:233243 [wandb_init.py:init():986] communicating run to backend with 90.0 second timeout
+2025-11-07 12:52:11,919 INFO    MainThread:233243 [wandb_init.py:init():1033] starting run threads in backend
+2025-11-07 12:52:12,067 INFO    MainThread:233243 [wandb_run.py:_console_start():2506] atexit reg
+2025-11-07 12:52:12,067 INFO    MainThread:233243 [wandb_run.py:_redirect():2354] redirect: wrap_raw
+2025-11-07 12:52:12,067 INFO    MainThread:233243 [wandb_run.py:_redirect():2423] Wrapping output streams.
+2025-11-07 12:52:12,067 INFO    MainThread:233243 [wandb_run.py:_redirect():2446] Redirects installed.
+2025-11-07 12:52:12,070 INFO    MainThread:233243 [wandb_init.py:init():1073] run started, returning control to user process
+2025-11-07 12:52:12,071 INFO    MainThread:233243 [wandb_run.py:_tensorboard_callback():1598] tensorboard callback: runs/Bandit__dqn_bandit__1__1762491124, True
+2025-11-07 12:52:12,181 INFO    wandb-AsyncioManager-main:233243 [service_client.py:_forward_responses():80] Reached EOF.
+2025-11-07 12:52:12,181 INFO    wandb-AsyncioManager-main:233243 [mailbox.py:close():137] Closing mailbox, abandoning 2 handles.
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/code/cleanrl/dqn_bandit.py b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/code/cleanrl/dqn_bandit.py
new file mode 100644
index 0000000000000000000000000000000000000000..1762efada08f2b5a1a0aa8b78b2d8051353e312a
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/code/cleanrl/dqn_bandit.py
@@ -0,0 +1,274 @@
+# DQN implementation for RAGEN Bandit Environment
+# Adapted from dqn_atari.py for single-step bandit problem
+import os
+import random
+import sys
+import time
+from dataclasses import dataclass
+
+# Add parent directory to path to import the wrapper
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+from ragen.env.bandit.env import BanditEnv
+from ragen.env.bandit.config import BanditEnvConfig
+from ragen_wrappers import BanditWrapper
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+
+    # Algorithm specific arguments
+    env_id: str = "Bandit"
+    """the id of the environment"""
+    total_timesteps: int = 100000
+    """total timesteps of the experiments"""
+    learning_rate: float = 1e-3
+    """the learning rate of the optimizer"""
+    num_envs: int = 1
+    """the number of parallel game environments (DQN typically uses 1)"""
+    buffer_size: int = 10000
+    """the replay memory buffer size"""
+    gamma: float = 0.0
+    """the discount factor gamma (0 for bandit since it's single-step)"""
+    tau: float = 1.0
+    """the target network update rate"""
+    target_network_frequency: int = 500
+    """the timesteps it takes to update the target network"""
+    batch_size: int = 32
+    """the batch size of sample from the reply memory"""
+    start_e: float = 1.0
+    """the starting epsilon for exploration"""
+    end_e: float = 0.05
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.5
+    """the fraction of `total-timesteps` it takes from start-e to go end-e"""
+    learning_starts: int = 1000
+    """timestep to start learning"""
+    train_frequency: int = 1
+    """the frequency of training"""
+    steps_per_episode: int = 10
+    """number of steps per episode for multi-step bandit"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, steps_per_episode):
+    def thunk():
+        config = BanditEnvConfig()
+        env = BanditEnv(config)
+        env = BanditWrapper(env, steps_per_episode=steps_per_episode)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    """
+    Q-Network for Bandit environment.
+    Input: [arm0_pulls, arm0_avg_reward, arm1_pulls, arm1_avg_reward]
+    Output: Q-values for each arm
+    """
+    def __init__(self, env):
+        super().__init__()
+        obs_shape = np.array(env.single_observation_space.shape).prod()
+        self.network = nn.Sequential(
+            nn.Linear(obs_shape, 128),
+            nn.ReLU(),
+            nn.Linear(128, 128),
+            nn.ReLU(),
+            nn.Linear(128, env.single_action_space.n),
+        )
+
+    def forward(self, x):
+        return self.network(x)
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    assert args.num_envs == 1, "vectorized envs are not supported at the moment"
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed + i, args.steps_per_episode) 
+         for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.Adam(q_network.parameters(), lr=args.learning_rate)
+    target_network = QNetwork(envs).to(device)
+    target_network.load_state_dict(q_network.state_dict())
+
+    # Use simple replay buffer (no special memory optimization needed for bandit)
+    from cleanrl_utils.buffers import ReplayBuffer
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        optimize_memory_usage=False,
+        handle_timeout_termination=False,
+    )
+    
+    start_time = time.time()
+    
+    # Track episode returns
+    episode_returns = []
+    recent_episode_returns = []
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = q_network(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if info and "episode" in info:
+                    episode_return = info['episode']['r']
+                    episode_length = info['episode']['l']
+                    episode_returns.append(episode_return)
+                    recent_episode_returns.append(episode_return)
+                    if len(recent_episode_returns) > 10:
+                        recent_episode_returns.pop(0)
+                    writer.add_scalar("charts/episodic_return", episode_return, global_step)
+                    writer.add_scalar("charts/episodic_length", episode_length, global_step)
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            if global_step % args.train_frequency == 0:
+                data = rb.sample(args.batch_size)
+                with torch.no_grad():
+                    target_max, _ = target_network(data.next_observations).max(dim=1)
+                    td_target = data.rewards.flatten() + args.gamma * target_max * (1 - data.dones.flatten())
+                old_val = q_network(data.observations).gather(1, data.actions).squeeze()
+                loss = F.mse_loss(td_target, old_val)
+
+                if global_step % 100 == 0:
+                    writer.add_scalar("losses/td_loss", loss, global_step)
+                    writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+                    writer.add_scalar("charts/epsilon", epsilon, global_step)
+                    
+                    # Console output with key metrics
+                    sps = int(global_step / (time.time() - start_time))
+                    progress = 100 * global_step / args.total_timesteps
+                    avg_episode_return = np.mean(recent_episode_returns) if recent_episode_returns else 0.0
+                    
+                    print(f"[{progress:5.1f}%] Step {global_step:6d}/{args.total_timesteps} | "
+                          f"SPS: {sps:5d} | "
+                          f"EpRet: {avg_episode_return:7.3f} | "
+                          f"Loss: {loss.item():.4f} | "
+                          f"Q-val: {old_val.mean().item():.4f} | "
+                          f"Eps: {epsilon:.3f}")
+                    
+                    writer.add_scalar("charts/SPS", sps, global_step)
+                    if recent_episode_returns:
+                        writer.add_scalar("charts/avg_episodic_return", avg_episode_return, global_step)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                optimizer.step()
+
+            # update target network
+            if global_step % args.target_network_frequency == 0:
+                for target_network_param, q_network_param in zip(target_network.parameters(), q_network.parameters()):
+                    target_network_param.data.copy_(
+                        args.tau * q_network_param.data + (1.0 - args.tau) * target_network_param.data
+                    )
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        torch.save(q_network.state_dict(), model_path)
+        print(f"model saved to {model_path}")
+
+    envs.close()
+    writer.close()
+    
+    print("\n" + "="*60)
+    print("Training Complete!")
+    print("="*60)
+    if episode_returns:
+        print(f"Final Average Episode Return (last 10): {np.mean(recent_episode_returns):.3f}")
+        print(f"Overall Average Episode Return: {np.mean(episode_returns):.3f}")
+        print(f"Best Episode Return: {max(episode_returns):.3f}")
+    print("="*60)
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..4f5c0d4f182ff26c78ab2c4b72e0a7cb53dd8279
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/config.yaml
@@ -0,0 +1,145 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/dqn_bandit.py
+        e:
+            gslyp7kt2rm94medkhny5d0o16ypm0nc:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-bandit
+                codePath: cleanrl/dqn_bandit.py
+                codePathLocal: dqn_bandit.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "31715966976"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/dqn_bandit.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T04:53:07.905986Z"
+                writerId: gslyp7kt2rm94medkhny5d0o16ypm0nc
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+batch_size:
+    value: 32
+buffer_size:
+    value: 10000
+capture_video:
+    value: false
+cuda:
+    value: true
+end_e:
+    value: 0.05
+env_id:
+    value: Bandit
+exp_name:
+    value: dqn_bandit
+exploration_fraction:
+    value: 0.5
+gamma:
+    value: 0
+learning_rate:
+    value: 0.001
+learning_starts:
+    value: 1000
+num_envs:
+    value: 1
+save_model:
+    value: false
+seed:
+    value: 1
+start_e:
+    value: 1
+steps_per_episode:
+    value: 10
+target_network_frequency:
+    value: 500
+tau:
+    value: 1
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 100000
+track:
+    value: true
+train_frequency:
+    value: 1
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-bandit
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/requirements.txt b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..252d801b74f9b4745c1e1681ec0a02a5dca88347
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/files/requirements.txt
@@ -0,0 +1,304 @@
+setuptools==80.9.0
+wheel==0.45.1
+pip==25.2
+zipp==3.23.0
+verl==0.2.0.dev0
+ragen==0.1
+triton==3.2.0
+nvidia-cusparselt-cu12==0.6.2
+mpmath==1.3.0
+typing_extensions==4.15.0
+sympy==1.13.1
+nvidia-nvtx-cu12==12.4.127
+nvidia-nvjitlink-cu12==12.4.127
+nvidia-nccl-cu12==2.21.5
+nvidia-curand-cu12==10.3.5.147
+nvidia-cufft-cu12==11.2.1.3
+nvidia-cuda-runtime-cu12==12.4.127
+nvidia-cuda-nvrtc-cu12==12.4.127
+nvidia-cuda-cupti-cu12==12.4.127
+nvidia-cublas-cu12==12.4.5.8
+networkx==3.5
+MarkupSafe==2.1.5
+fsspec==2025.9.0
+filelock==3.19.1
+nvidia-cusparse-cu12==12.3.1.170
+nvidia-cudnn-cu12==9.1.0.70
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.6.1.9
+torch==2.6.0+cu124
+einops==0.8.1
+flash_attn==2.7.4.post1
+pytz==2025.2
+pyperclip==1.11.0
+pylatexenc==2.10
+pyjnius==1.7.0
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+gym-notices==0.1.0
+flatbuffers==25.9.23
+fastrlock==0.8.3
+Farama-Notifications==0.0.4
+cymem==2.0.11
+antlr4-python3-runtime==4.9.3
+xxhash==3.6.0
+wrapt==2.0.0
+Werkzeug==3.1.3
+websockets==15.0.1
+wcwidth==0.2.14
+wasabi==1.1.3
+uvloop==0.22.1
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.2
+traitlets==5.14.3
+tqdm==4.67.1
+threadpoolctl==3.6.0
+tabulate==0.9.0
+spacy-loggers==1.0.5
+spacy-legacy==3.0.12
+soupsieve==2.8
+sniffio==1.3.1
+smmap==5.0.2
+six==1.17.0
+shellingham==1.5.4
+sentencepiece==0.2.1
+safetensors==0.6.2
+rpds-py==0.28.0
+rignore==0.7.6
+regex==2025.11.3
+RapidFuzz==3.14.3
+pyzmq==27.1.0
+PyYAML==6.0.3
+pytokens==0.3.0
+python-multipart==0.0.20
+python-json-logger==4.0.0
+python-dotenv==1.2.1
+PySocks==1.7.1
+pyparsing==3.2.5
+PyJWT==2.10.1
+Pygments==2.19.2
+pygame==2.6.1
+pydantic_core==2.41.5
+pycparser==2.23
+pycountry==24.6.1
+pybind11==3.0.1
+pyarrow==22.0.0
+psutil==7.1.3
+protobuf==6.33.0
+propcache==0.4.1
+prometheus_client==0.23.1
+platformdirs==4.5.0
+pillow==11.3.0
+pexpect==4.9.0
+pathvalidate==3.3.1
+pathspec==0.12.1
+pathable==0.4.4
+partial-json-parser==0.2.1.1.post6
+parso==0.8.5
+packaging==25.0
+orjson==3.11.4
+numpy==1.26.4
+ninja==1.13.0
+nest-asyncio==1.6.0
+mypy_extensions==1.1.0
+murmurhash==1.0.13
+multidict==6.7.0
+msgspec==0.19.0
+msgpack==1.1.2
+more-itertools==10.8.0
+mdurl==0.1.2
+marisa-trie==1.3.1
+llvmlite==0.43.0
+llguidance==0.7.30
+lark==1.2.2
+kiwisolver==1.4.9
+joblib==1.5.2
+jiter==0.11.1
+jeepney==0.9.0
+jaraco.context==6.0.1
+itsdangerous==2.2.0
+interegular==0.3.3
+idna==3.11
+humanfriendly==10.0
+httpx-sse==0.4.3
+httptools==0.7.1
+html2text==2025.4.15
+hf-xet==1.2.0
+h11==0.16.0
+frozenlist==1.8.0
+fonttools==4.60.1
+executing==2.2.1
+exceptiongroup==1.3.0
+eval_type_backport==0.2.2
+docutils==0.22.3
+docstring_parser==0.17.0
+dnspython==2.8.0
+distro==1.9.0
+diskcache==5.6.3
+dill==0.4.0
+decorator==5.2.1
+debugpy==1.8.17
+Cython==3.2.0
+cycler==0.12.1
+codetiming==1.4.0
+cloudpickle==3.1.2
+cloudpathlib==0.23.0
+click==8.2.1
+charset-normalizer==3.4.4
+certifi==2025.10.5
+catalogue==2.0.10
+cachetools==6.2.1
+blinker==1.9.0
+blake3==1.0.8
+beartype==0.22.5
+attrs==25.4.0
+asttokens==3.0.0
+astor==0.8.1
+annotated-types==0.7.0
+annotated-doc==0.0.3
+airportsdata==20250909
+aiohappyeyeballs==2.6.1
+yarl==1.22.0
+uvicorn==0.38.0
+typer-slim==0.20.0
+thefuzz==0.22.1
+stack-data==0.6.3
+srsly==2.5.1
+smart_open==7.4.4
+sentry-sdk==2.43.0
+scipy==1.16.3
+requests==2.32.5
+referencing==0.36.2
+rank-bm25==0.2.2
+python-dateutil==2.9.0.post0
+pydantic==2.12.4
+py-key-value-shared==0.2.8
+prompt_toolkit==3.0.52
+preshed==3.0.10
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+numba==0.60.0
+nltk==3.9.2
+multiprocess==0.70.18
+matplotlib-inline==0.2.1
+markdown-it-py==4.0.0
+language_data==1.3.0
+jedi==0.19.2
+jaraco.functools==4.3.0
+jaraco.classes==3.4.0
+ipython_pygments_lexers==1.1.1
+importlib_metadata==8.7.0
+ImageIO==2.37.2
+httpcore==1.0.9
+gymnasium==1.2.2
+gym==0.26.2
+gitdb==4.0.12
+gguf==0.10.0
+Flask==3.1.2
+faiss-cpu==1.12.0
+email-validator==2.3.0
+depyf==0.18.0
+cupy-cuda12x==13.6.0
+contourpy==1.3.3
+coloredlogs==15.0.1
+cffi==2.0.0
+blis==1.3.0
+black==25.9.0
+beautifulsoup4==4.14.2
+anyio==4.11.0
+aiosignal==1.4.0
+watchfiles==1.1.1
+tiktoken==0.12.0
+starlette==0.49.3
+sse-starlette==3.0.3
+scikit-learn==1.7.2
+rich==14.2.0
+pydantic-settings==2.11.0
+pydantic-extra-types==2.10.6
+py-key-value-aio==0.2.8
+pandas==2.3.3
+openapi-pydantic==0.5.1
+onnxruntime==1.23.2
+matplotlib==3.10.7
+lm-format-enforcer==0.10.12
+langcodes==3.5.0
+jsonschema-specifications==2025.9.1
+jsonschema-path==0.3.4
+ipython==9.7.0
+hydra-core==1.3.2
+huggingface-hub==0.36.0
+httpx==0.28.1
+gym-sokoban==0.0.6
+GitPython==3.1.45
+cryptography==46.0.3
+confection==0.1.5
+cleantext==1.1.4
+aiohttp==3.13.2
+xformers==0.0.29.post2
+weasel==0.4.2
+wandb==0.22.3
+typer==0.19.2
+torchvision==0.21.0
+torchdata==0.11.0
+torchaudio==2.6.0
+tokenizers==0.22.1
+thinc==8.3.7
+tensordict==0.8.3
+SecretStorage==3.4.0
+rich-toolkit==0.15.1
+rich-rst==1.3.2
+prometheus-fastapi-instrumentator==7.1.0
+openai==2.7.1
+jsonschema==4.25.1
+gdown==5.2.0
+fastapi==0.121.0
+Authlib==1.6.5
+anthropic==0.72.0
+accelerate==1.11.0
+transformers==4.57.1
+together==1.5.30
+spacy==3.8.7
+ray==2.51.1
+outlines_core==0.1.26
+mistral_common==1.8.5
+mcp==1.20.0
+keyring==25.6.0
+fastapi-cloud-cli==0.3.1
+fastapi-cli==0.0.14
+datasets==4.4.1
+cyclopts==4.2.1
+xgrammar==0.1.16
+peft==0.17.1
+outlines==0.1.11
+compressed-tensors==0.9.2
+fastmcp==2.13.0.2
+vllm==0.8.2
+pyserini==1.3.0
+typeguard==4.4.4
+shtab==1.7.2
+tyro==0.9.35
+tensorboard-data-server==0.7.2
+Markdown==3.10
+grpcio==1.76.0
+absl-py==2.3.1
+tensorboard==2.20.0
+ragen==0.1
+verl==0.2.0.dev0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.0.0
+inflect==7.3.1
+jaraco.collections==5.1.0
+jaraco.context==5.3.0
+jaraco.functools==4.0.1
+jaraco.text==3.12.1
+more-itertools==10.3.0
+packaging==24.2
+platformdirs==4.2.2
+tomli==2.0.1
+typeguard==4.3.0
+typing_extensions==4.12.2
+wheel==0.45.1
+zipp==3.19.2
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug-internal.log b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..812247db9f7c2ce1fc26b4a6074049d571349c52
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug-internal.log
@@ -0,0 +1,86 @@
+{"time":"2025-11-07T12:53:08.172031891+08:00","level":"INFO","msg":"stream: starting","core version":"0.22.3"}
+{"time":"2025-11-07T12:53:08.533993776+08:00","level":"INFO","msg":"stream: created new stream","id":"gt2g963v"}
+{"time":"2025-11-07T12:53:08.534293491+08:00","level":"INFO","msg":"handler: started","stream_id":"gt2g963v"}
+{"time":"2025-11-07T12:53:08.537890578+08:00","level":"INFO","msg":"stream: started","id":"gt2g963v"}
+{"time":"2025-11-07T12:53:08.537903912+08:00","level":"INFO","msg":"writer: started","stream_id":"gt2g963v"}
+{"time":"2025-11-07T12:53:08.537910822+08:00","level":"INFO","msg":"sender: started","stream_id":"gt2g963v"}
+{"time":"2025-11-07T12:53:08.960706405+08:00","level":"ERROR","msg":"git repo not found","error":"repository does not exist"}
+{"time":"2025-11-07T12:53:19.096437914+08:00","level":"WARN","msg":"tensorboard: no root directory","error":"timed out after 10s"}
+{"time":"2025-11-07T12:53:19.097330577+08:00","level":"INFO","msg":"tensorboard: tracking new log directory","rootDir":{},"logDir":"LocalOrCloudPath(LocalPath=\"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491180\")","namespace":""}
+{"time":"2025-11-07T12:53:19.098968164+08:00","level":"INFO","msg":"tensorboard: saving file","fileLocation":"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491180/events.out.tfevents.1762491189.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.233508.0","runPath":"runs/Bandit__dqn_bandit__1__1762491180/events.out.tfevents.1762491189.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.233508.0"}
+{"time":"2025-11-07T12:53:19.101717897+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":128}
+{"time":"2025-11-07T12:53:19.105065758+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":385}
+{"time":"2025-11-07T12:53:24.10924085+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":585}
+{"time":"2025-11-07T12:53:24.440936183+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":185}
+{"time":"2025-11-07T12:53:34.122016617+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1082}
+{"time":"2025-11-07T12:53:34.122109163+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:53:34.12251808+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1114}
+{"time":"2025-11-07T12:53:34.124134613+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":102}
+{"time":"2025-11-07T12:53:39.12743858+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1269}
+{"time":"2025-11-07T12:53:39.365468821+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":75}
+{"time":"2025-11-07T12:53:44.134465237+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1417}
+{"time":"2025-11-07T12:53:44.134608021+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:53:44.134749151+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1449}
+{"time":"2025-11-07T12:53:44.135688519+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":77}
+{"time":"2025-11-07T12:53:49.140296383+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1568}
+{"time":"2025-11-07T12:53:49.140414396+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:53:54.145000563+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1627}
+{"time":"2025-11-07T12:53:54.343755292+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":45}
+{"time":"2025-11-07T12:53:59.150032219+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1782}
+{"time":"2025-11-07T12:53:59.150118931+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:53:59.150258994+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":1814}
+{"time":"2025-11-07T12:53:59.154124665+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":232}
+{"time":"2025-11-07T12:54:04.156942707+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2100}
+{"time":"2025-11-07T12:54:04.157033991+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:04.157520981+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2133}
+{"time":"2025-11-07T12:54:04.159020017+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":11}
+{"time":"2025-11-07T12:54:09.162162541+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2207}
+{"time":"2025-11-07T12:54:09.415849773+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":139}
+{"time":"2025-11-07T12:54:14.168178299+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2535}
+{"time":"2025-11-07T12:54:14.168300295+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:14.168401838+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2567}
+{"time":"2025-11-07T12:54:14.172136564+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":265}
+{"time":"2025-11-07T12:54:19.179495936+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2950}
+{"time":"2025-11-07T12:54:19.179564083+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:19.180121589+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":2982}
+{"time":"2025-11-07T12:54:19.182154713+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":110}
+{"time":"2025-11-07T12:54:24.187115668+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3167}
+{"time":"2025-11-07T12:54:24.334547491+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":185}
+{"time":"2025-11-07T12:54:29.192892466+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3419}
+{"time":"2025-11-07T12:54:29.193146867+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:29.193764349+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3451}
+{"time":"2025-11-07T12:54:29.195578107+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":111}
+{"time":"2025-11-07T12:54:34.200135408+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3637}
+{"time":"2025-11-07T12:54:34.200246529+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:34.200654397+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3669}
+{"time":"2025-11-07T12:54:34.20139884+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":47}
+{"time":"2025-11-07T12:54:39.208643304+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3771}
+{"time":"2025-11-07T12:54:39.405178302+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":89}
+{"time":"2025-11-07T12:54:44.21363748+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3920}
+{"time":"2025-11-07T12:54:44.213745535+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:44.213868926+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":3952}
+{"time":"2025-11-07T12:54:44.214946343+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":4}
+{"time":"2025-11-07T12:54:49.218617242+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4024}
+{"time":"2025-11-07T12:54:49.218728334+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:54:49.219163025+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4056}
+{"time":"2025-11-07T12:54:49.22116412+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":68}
+{"time":"2025-11-07T12:54:54.223888538+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4165}
+{"time":"2025-11-07T12:54:54.384123987+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":3}
+{"time":"2025-11-07T12:55:09.239847321+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4307}
+{"time":"2025-11-07T12:55:09.409316986+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":3}
+{"time":"2025-11-07T12:55:24.256736392+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4449}
+{"time":"2025-11-07T12:55:24.404054363+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":3}
+{"time":"2025-11-07T12:55:34.268166686+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4562}
+{"time":"2025-11-07T12:55:34.26828271+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:55:34.270280693+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4595}
+{"time":"2025-11-07T12:55:34.272120407+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":137}
+{"time":"2025-11-07T12:55:39.276872315+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":4817}
+{"time":"2025-11-07T12:55:39.409576893+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":281}
+{"time":"2025-11-07T12:55:44.283456263+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5161}
+{"time":"2025-11-07T12:55:44.283600607+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":1}
+{"time":"2025-11-07T12:55:44.284151663+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5194}
+{"time":"2025-11-07T12:55:44.2875506+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":114}
+{"time":"2025-11-07T12:55:47.105188628+08:00","level":"INFO","msg":"stream: closing","id":"gt2g963v"}
+{"time":"2025-11-07T12:55:47.108071616+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":5362}
+{"time":"2025-11-07T12:55:47.109245835+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":50}
+{"time":"2025-11-07T12:55:48.193707174+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug.log b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug.log
new file mode 100644
index 0000000000000000000000000000000000000000..cc04c5d0fe4e7d881bea876cc7386331bd9bf43e
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug.log
@@ -0,0 +1,375 @@
+2025-11-07 12:53:07,949 INFO    MainThread:233508 [wandb_setup.py:_flush():81] Current SDK version is 0.22.3
+2025-11-07 12:53:07,949 INFO    MainThread:233508 [wandb_setup.py:_flush():81] Configure stats pid to 233508
+2025-11-07 12:53:07,949 INFO    MainThread:233508 [wandb_setup.py:_flush():81] Loading settings from /root/.config/wandb/settings
+2025-11-07 12:53:07,949 INFO    MainThread:233508 [wandb_setup.py:_flush():81] Loading settings from /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/settings
+2025-11-07 12:53:07,949 INFO    MainThread:233508 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2025-11-07 12:53:07,950 INFO    MainThread:233508 [wandb_init.py:setup_run_log_directory():706] Logging user logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug.log
+2025-11-07 12:53:07,950 INFO    MainThread:233508 [wandb_init.py:setup_run_log_directory():707] Logging internal logs to /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/wandb/run-20251107_125307-gt2g963v/logs/debug-internal.log
+2025-11-07 12:53:07,951 INFO    MainThread:233508 [wandb_init.py:init():833] calling init triggers
+2025-11-07 12:53:07,951 INFO    MainThread:233508 [wandb_init.py:init():838] wandb.init called with sweep_config: {}
+config: {'exp_name': 'dqn_bandit', 'seed': 1, 'torch_deterministic': True, 'cuda': True, 'track': True, 'wandb_project_name': 'ragen-bandit', 'wandb_entity': None, 'capture_video': False, 'save_model': False, 'env_id': 'Bandit', 'total_timesteps': 100000, 'learning_rate': 0.001, 'num_envs': 1, 'buffer_size': 10000, 'gamma': 0.0, 'tau': 1.0, 'target_network_frequency': 500, 'batch_size': 32, 'start_e': 1.0, 'end_e': 0.05, 'exploration_fraction': 0.5, 'learning_starts': 1000, 'train_frequency': 1, 'steps_per_episode': 10, '_wandb': {'code_path': 'code/cleanrl/dqn_bandit.py'}}
+2025-11-07 12:53:07,951 INFO    MainThread:233508 [wandb_init.py:init():881] starting backend
+2025-11-07 12:53:08,158 INFO    MainThread:233508 [wandb_init.py:init():884] sending inform_init request
+2025-11-07 12:53:08,168 INFO    MainThread:233508 [wandb_init.py:init():892] backend started and connected
+2025-11-07 12:53:08,170 INFO    MainThread:233508 [wandb_init.py:init():962] updated telemetry
+2025-11-07 12:53:08,201 INFO    MainThread:233508 [wandb_init.py:init():986] communicating run to backend with 90.0 second timeout
+2025-11-07 12:53:08,947 INFO    MainThread:233508 [wandb_init.py:init():1033] starting run threads in backend
+2025-11-07 12:53:09,092 INFO    MainThread:233508 [wandb_run.py:_console_start():2506] atexit reg
+2025-11-07 12:53:09,092 INFO    MainThread:233508 [wandb_run.py:_redirect():2354] redirect: wrap_raw
+2025-11-07 12:53:09,092 INFO    MainThread:233508 [wandb_run.py:_redirect():2423] Wrapping output streams.
+2025-11-07 12:53:09,092 INFO    MainThread:233508 [wandb_run.py:_redirect():2446] Redirects installed.
+2025-11-07 12:53:09,095 INFO    MainThread:233508 [wandb_init.py:init():1073] run started, returning control to user process
+2025-11-07 12:53:09,095 INFO    MainThread:233508 [wandb_run.py:_tensorboard_callback():1598] tensorboard callback: runs/Bandit__dqn_bandit__1__1762491180, True
+2025-11-07 12:55:47,105 INFO    wandb-AsyncioManager-main:233508 [service_client.py:_forward_responses():80] Reached EOF.
+2025-11-07 12:55:47,105 INFO    wandb-AsyncioManager-main:233508 [mailbox.py:close():137] Closing mailbox, abandoning 1 handles.
+2025-11-07 12:55:47,387 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,389 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,389 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,391 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,392 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,392 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,397 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,397 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,401 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,402 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,402 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,403 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,403 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,404 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,404 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,404 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,404 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,405 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,405 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,406 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,406 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,407 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,407 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,407 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,408 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,408 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
+2025-11-07 12:55:47,410 ERROR   wandb-AsyncioManager-main:233508 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
+Traceback (most recent call last):
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
+    await fn()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
+    await self._send_server_request(request)
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
+    await self._writer.drain()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 392, in drain
+    await self._protocol._drain_helper()
+  File "/root/local/miniconda3/envs/ragen/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
+    raise ConnectionResetError('Connection lost')
+ConnectionResetError: Connection lost
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/code/cleanrl/dqn_bandit.py b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/code/cleanrl/dqn_bandit.py
new file mode 100644
index 0000000000000000000000000000000000000000..3c6be3a159b622c05d7fd8d4f8e9555cb139a162
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/code/cleanrl/dqn_bandit.py
@@ -0,0 +1,344 @@
+# DQN implementation for RAGEN Bandit Environment
+# Adapted from dqn_atari.py for single-step bandit problem
+import os
+import random
+import sys
+import time
+from dataclasses import dataclass
+
+# Add parent directory to path to import the wrapper
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import gymnasium as gym
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.optim as optim
+import tyro
+from torch.utils.tensorboard import SummaryWriter
+
+from ragen.env.bandit.env import BanditEnv
+from ragen.env.bandit.config import BanditEnvConfig
+
+
+# Define BanditWrapper locally to avoid dependency issues
+class BanditWrapper(gym.Wrapper):
+    """
+    Wrapper for RAGEN Bandit environment to make it compatible with DQN.
+    Converts single-step Bandit to multi-step environment.
+    
+    State representation: [arm0_count, arm0_avg_reward, arm1_count, arm1_avg_reward]
+    This allows the agent to learn which arm is better based on historical performance.
+    """
+    def __init__(self, env, steps_per_episode=10):
+        super().__init__(env)
+        # State: [arm0_pulls, arm0_avg_reward, arm1_pulls, arm1_avg_reward]
+        self.observation_space = gym.spaces.Box(low=0, high=1000, shape=(4,), dtype=np.float32)
+        self.action_space = gym.spaces.Discrete(2)
+        self.steps_per_episode = steps_per_episode
+        self.current_step = 0
+        self._arm_counts = [0, 0]
+        self._arm_rewards = [0.0, 0.0]
+        
+    def reset(self, **kwargs):
+        # Filter out 'options' parameter that gymnasium passes but RAGEN doesn't support
+        seed = kwargs.get('seed', None)
+        mode = kwargs.get('mode', None)
+        self.env.reset(seed=seed, mode=mode)
+        
+        # Reset step counter but keep arm statistics for learning
+        self.current_step = 0
+        
+        # Return current state based on accumulated history
+        state = self._get_state()
+        return state, {}
+    
+    def _get_state(self):
+        """Get current state representation."""
+        return np.array([
+            self._arm_counts[0],
+            self._arm_rewards[0] / max(1, self._arm_counts[0]),
+            self._arm_counts[1],
+            self._arm_rewards[1] / max(1, self._arm_counts[1]),
+        ], dtype=np.float32)
+    
+    def step(self, action):
+        # Map action from 0,1 to 1,2 (RAGEN uses 1-indexed actions)
+        ragen_action = action + 1
+        obs, reward, done, info = self.env.step(ragen_action)
+        
+        # Update statistics for the chosen arm
+        self._arm_counts[action] += 1
+        self._arm_rewards[action] += reward
+        self.current_step += 1
+        
+        # Episode ends after N steps (not after single step)
+        terminated = (self.current_step >= self.steps_per_episode)
+        truncated = False
+        
+        # Debug: print when episode ends
+        if terminated and self._arm_counts[0] + self._arm_counts[1] < 100:
+            print(f"DEBUG: Episode ended at step {self.current_step}, arm_counts: {self._arm_counts}")
+        
+        # Reset the underlying Bandit env but keep our statistics
+        if not terminated:
+            self.env.reset()
+        
+        state = self._get_state()
+        return state, reward, terminated, truncated, info
+
+
+@dataclass
+class Args:
+    exp_name: str = os.path.basename(__file__)[: -len(".py")]
+    """the name of this experiment"""
+    seed: int = 1
+    """seed of the experiment"""
+    torch_deterministic: bool = True
+    """if toggled, `torch.backends.cudnn.deterministic=False`"""
+    cuda: bool = True
+    """if toggled, cuda will be enabled by default"""
+    track: bool = False
+    """if toggled, this experiment will be tracked with Weights and Biases"""
+    wandb_project_name: str = "cleanRL"
+    """the wandb's project name"""
+    wandb_entity: str = None
+    """the entity (team) of wandb's project"""
+    capture_video: bool = False
+    """whether to capture videos of the agent performances (check out `videos` folder)"""
+    save_model: bool = False
+    """whether to save model into the `runs/{run_name}` folder"""
+
+    # Algorithm specific arguments
+    env_id: str = "Bandit"
+    """the id of the environment"""
+    total_timesteps: int = 100000
+    """total timesteps of the experiments"""
+    learning_rate: float = 1e-3
+    """the learning rate of the optimizer"""
+    num_envs: int = 1
+    """the number of parallel game environments (DQN typically uses 1)"""
+    buffer_size: int = 10000
+    """the replay memory buffer size"""
+    gamma: float = 0.99
+    """the discount factor gamma (0.99 for multi-step bandit)"""
+    tau: float = 1.0
+    """the target network update rate"""
+    target_network_frequency: int = 500
+    """the timesteps it takes to update the target network"""
+    batch_size: int = 32
+    """the batch size of sample from the reply memory"""
+    start_e: float = 1.0
+    """the starting epsilon for exploration"""
+    end_e: float = 0.05
+    """the ending epsilon for exploration"""
+    exploration_fraction: float = 0.5
+    """the fraction of `total-timesteps` it takes from start-e to go end-e"""
+    learning_starts: int = 1000
+    """timestep to start learning"""
+    train_frequency: int = 1
+    """the frequency of training"""
+    steps_per_episode: int = 10
+    """number of steps per episode for multi-step bandit"""
+
+
+def make_env(env_id, idx, capture_video, run_name, seed, steps_per_episode):
+    def thunk():
+        config = BanditEnvConfig()
+        env = BanditEnv(config)
+        env = BanditWrapper(env, steps_per_episode=steps_per_episode)
+        env = gym.wrappers.RecordEpisodeStatistics(env)
+        if capture_video and idx == 0:
+            env = gym.wrappers.RecordVideo(env, f"videos/{run_name}")
+        return env
+    return thunk
+
+
+# ALGO LOGIC: initialize agent here:
+class QNetwork(nn.Module):
+    """
+    Q-Network for Bandit environment.
+    Input: [arm0_pulls, arm0_avg_reward, arm1_pulls, arm1_avg_reward]
+    Output: Q-values for each arm
+    """
+    def __init__(self, env):
+        super().__init__()
+        obs_shape = np.array(env.single_observation_space.shape).prod()
+        self.network = nn.Sequential(
+            nn.Linear(obs_shape, 128),
+            nn.ReLU(),
+            nn.Linear(128, 128),
+            nn.ReLU(),
+            nn.Linear(128, env.single_action_space.n),
+        )
+
+    def forward(self, x):
+        return self.network(x)
+
+
+def linear_schedule(start_e: float, end_e: float, duration: int, t: int):
+    slope = (end_e - start_e) / duration
+    return max(slope * t + start_e, end_e)
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    assert args.num_envs == 1, "vectorized envs are not supported at the moment"
+    run_name = f"{args.env_id}__{args.exp_name}__{args.seed}__{int(time.time())}"
+    
+    if args.track:
+        import wandb
+
+        wandb.init(
+            project=args.wandb_project_name,
+            entity=args.wandb_entity,
+            sync_tensorboard=True,
+            config=vars(args),
+            name=run_name,
+            monitor_gym=True,
+            save_code=True,
+        )
+    writer = SummaryWriter(f"runs/{run_name}")
+    writer.add_text(
+        "hyperparameters",
+        "|param|value|\n|-|-|\n%s" % ("\n".join([f"|{key}|{value}|" for key, value in vars(args).items()])),
+    )
+
+    # TRY NOT TO MODIFY: seeding
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    device = torch.device("cuda" if torch.cuda.is_available() and args.cuda else "cpu")
+
+    # env setup
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, i, args.capture_video, run_name, args.seed + i, args.steps_per_episode) 
+         for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    q_network = QNetwork(envs).to(device)
+    optimizer = optim.Adam(q_network.parameters(), lr=args.learning_rate)
+    target_network = QNetwork(envs).to(device)
+    target_network.load_state_dict(q_network.state_dict())
+
+    # Use simple replay buffer (no special memory optimization needed for bandit)
+    from cleanrl_utils.buffers import ReplayBuffer
+    rb = ReplayBuffer(
+        args.buffer_size,
+        envs.single_observation_space,
+        envs.single_action_space,
+        device,
+        optimize_memory_usage=False,
+        handle_timeout_termination=False,
+    )
+    
+    start_time = time.time()
+    
+    # Track episode returns
+    episode_returns = []
+    recent_episode_returns = []
+
+    # TRY NOT TO MODIFY: start the game
+    obs, _ = envs.reset(seed=args.seed)
+    for global_step in range(args.total_timesteps):
+        # ALGO LOGIC: put action logic here
+        epsilon = linear_schedule(args.start_e, args.end_e, args.exploration_fraction * args.total_timesteps, global_step)
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = q_network(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+
+        # TRY NOT TO MODIFY: execute the game and log data.
+        next_obs, rewards, terminations, truncations, infos = envs.step(actions)
+
+        # TRY NOT TO MODIFY: record rewards for plotting purposes
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if info and "episode" in info:
+                    episode_return = info['episode']['r']
+                    episode_length = info['episode']['l']
+                    episode_returns.append(episode_return)
+                    recent_episode_returns.append(episode_return)
+                    if len(recent_episode_returns) > 10:
+                        recent_episode_returns.pop(0)
+                    writer.add_scalar("charts/episodic_return", episode_return, global_step)
+                    writer.add_scalar("charts/episodic_length", episode_length, global_step)
+                    
+                    # Debug: print episode completion
+                    if global_step % 1000 == 0:
+                        print(f"  → Episode completed! Return: {episode_return:.3f}, Length: {episode_length}")
+
+        # TRY NOT TO MODIFY: save data to reply buffer; handle `final_observation`
+        real_next_obs = next_obs.copy()
+        for idx, trunc in enumerate(truncations):
+            if trunc:
+                real_next_obs[idx] = infos["final_observation"][idx]
+        rb.add(obs, real_next_obs, actions, rewards, terminations, infos)
+
+        # TRY NOT TO MODIFY: CRUCIAL step easy to overlook
+        obs = next_obs
+
+        # ALGO LOGIC: training.
+        if global_step > args.learning_starts:
+            if global_step % args.train_frequency == 0:
+                data = rb.sample(args.batch_size)
+                with torch.no_grad():
+                    target_max, _ = target_network(data.next_observations).max(dim=1)
+                    td_target = data.rewards.flatten() + args.gamma * target_max * (1 - data.dones.flatten())
+                old_val = q_network(data.observations).gather(1, data.actions).squeeze()
+                loss = F.mse_loss(td_target, old_val)
+
+                if global_step % 100 == 0:
+                    writer.add_scalar("losses/td_loss", loss, global_step)
+                    writer.add_scalar("losses/q_values", old_val.mean().item(), global_step)
+                    writer.add_scalar("charts/epsilon", epsilon, global_step)
+                    
+                    # Console output with key metrics
+                    sps = int(global_step / (time.time() - start_time))
+                    progress = 100 * global_step / args.total_timesteps
+                    avg_episode_return = np.mean(recent_episode_returns) if recent_episode_returns else 0.0
+                    
+                    print(f"[{progress:5.1f}%] Step {global_step:6d}/{args.total_timesteps} | "
+                          f"SPS: {sps:5d} | "
+                          f"EpRet: {avg_episode_return:7.3f} | "
+                          f"Loss: {loss.item():.4f} | "
+                          f"Q-val: {old_val.mean().item():.4f} | "
+                          f"Eps: {epsilon:.3f}")
+                    
+                    writer.add_scalar("charts/SPS", sps, global_step)
+                    if recent_episode_returns:
+                        writer.add_scalar("charts/avg_episodic_return", avg_episode_return, global_step)
+
+                # optimize the model
+                optimizer.zero_grad()
+                loss.backward()
+                optimizer.step()
+
+            # update target network
+            if global_step % args.target_network_frequency == 0:
+                for target_network_param, q_network_param in zip(target_network.parameters(), q_network.parameters()):
+                    target_network_param.data.copy_(
+                        args.tau * q_network_param.data + (1.0 - args.tau) * target_network_param.data
+                    )
+
+    if args.save_model:
+        model_path = f"runs/{run_name}/{args.exp_name}.cleanrl_model"
+        torch.save(q_network.state_dict(), model_path)
+        print(f"model saved to {model_path}")
+
+    envs.close()
+    writer.close()
+    
+    print("\n" + "="*60)
+    print("Training Complete!")
+    print("="*60)
+    if episode_returns:
+        print(f"Final Average Episode Return (last 10): {np.mean(recent_episode_returns):.3f}")
+        print(f"Overall Average Episode Return: {np.mean(episode_returns):.3f}")
+        print(f"Best Episode Return: {max(episode_returns):.3f}")
+    print("="*60)
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/config.yaml b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/config.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..cb8e1868a138ba023573a5cf0470df19a9430876
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/config.yaml
@@ -0,0 +1,145 @@
+_wandb:
+    value:
+        cli_version: 0.22.3
+        code_path: code/cleanrl/dqn_bandit.py
+        e:
+            jbqdicudyaaspbdzq2z1959r4qyohhsf:
+                args:
+                    - --track
+                    - --wandb-project-name
+                    - ragen-bandit
+                codePath: cleanrl/dqn_bandit.py
+                codePathLocal: dqn_bandit.py
+                cpu_count: 64
+                cpu_count_logical: 128
+                cudaVersion: "12.4"
+                disk:
+                    /:
+                        total: "5153960755200"
+                        used: "31724965888"
+                email: haoyu-wa22@mails.tsinghua.edu.cn
+                executable: /root/local/miniconda3/envs/ragen/bin/python
+                git:
+                    commit: 004f8a086a892a2a180f4dd332b90d83a968aa7a
+                    remote: https://github.com/vwxyzjn/cleanrl.git
+                gpu: NVIDIA H100 80GB HBM3
+                gpu_count: 8
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-35e2d43d-4067-82ce-90d4-def9e389bf28
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-af4135e3-88f2-e9ac-518d-502c75a85429
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "85520809984"
+                      name: NVIDIA H100 80GB HBM3
+                      uuid: GPU-b7cf0ec6-7c29-1179-dceb-09565da51890
+                host: pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0
+                memory:
+                    total: "2163642122240"
+                os: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35
+                program: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl/dqn_bandit.py
+                python: CPython 3.12.12
+                root: /mnt/general/wanghy/RAGEN/cleanrl/cleanrl
+                startedAt: "2025-11-07T04:56:47.687990Z"
+                writerId: jbqdicudyaaspbdzq2z1959r4qyohhsf
+        m: []
+        python_version: 3.12.12
+        t:
+            "1":
+                - 1
+                - 49
+                - 51
+                - 105
+            "2":
+                - 1
+                - 49
+                - 51
+                - 105
+            "3":
+                - 13
+                - 16
+                - 35
+            "4": 3.12.12
+            "5": 0.22.3
+            "12": 0.22.3
+            "13": linux-x86_64
+batch_size:
+    value: 32
+buffer_size:
+    value: 10000
+capture_video:
+    value: false
+cuda:
+    value: true
+end_e:
+    value: 0.05
+env_id:
+    value: Bandit
+exp_name:
+    value: dqn_bandit
+exploration_fraction:
+    value: 0.5
+gamma:
+    value: 0.99
+learning_rate:
+    value: 0.001
+learning_starts:
+    value: 1000
+num_envs:
+    value: 1
+save_model:
+    value: false
+seed:
+    value: 1
+start_e:
+    value: 1
+steps_per_episode:
+    value: 10
+target_network_frequency:
+    value: 500
+tau:
+    value: 1
+torch_deterministic:
+    value: true
+total_timesteps:
+    value: 100000
+track:
+    value: true
+train_frequency:
+    value: 1
+wandb_entity:
+    value: null
+wandb_project_name:
+    value: ragen-bandit
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/output.log b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/output.log
new file mode 100644
index 0000000000000000000000000000000000000000..4b24f2815d527b592d0d9c96101b70e405c2e17f
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/output.log
@@ -0,0 +1,49 @@
+DEBUG: Episode ended at step 10, arm_counts: [6, 4]
+DEBUG: Episode ended at step 10, arm_counts: [9, 11]
+DEBUG: Episode ended at step 10, arm_counts: [12, 18]
+DEBUG: Episode ended at step 10, arm_counts: [20, 20]
+DEBUG: Episode ended at step 10, arm_counts: [25, 25]
+DEBUG: Episode ended at step 10, arm_counts: [33, 27]
+DEBUG: Episode ended at step 10, arm_counts: [37, 33]
+DEBUG: Episode ended at step 10, arm_counts: [41, 39]
+DEBUG: Episode ended at step 10, arm_counts: [45, 45]
+[  1.1%] Step   1100/100000 | SPS:  2447 | EpRet:   0.000 | Loss: 0.5871 | Q-val: -1.5165 | Eps: 0.979
+[  1.2%] Step   1200/100000 | SPS:  2143 | EpRet:   0.000 | Loss: 2.2378 | Q-val: -1.1003 | Eps: 0.977
+[  1.3%] Step   1300/100000 | SPS:  1942 | EpRet:   0.000 | Loss: 2.8699 | Q-val: -0.4089 | Eps: 0.975
+[  1.4%] Step   1400/100000 | SPS:  1794 | EpRet:   0.000 | Loss: 0.4940 | Q-val: -1.8337 | Eps: 0.973
+[  1.5%] Step   1500/100000 | SPS:  1652 | EpRet:   0.000 | Loss: 7.6582 | Q-val: -3.1953 | Eps: 0.972
+[  1.6%] Step   1600/100000 | SPS:  1266 | EpRet:   0.000 | Loss: 2.2227 | Q-val: -2.6707 | Eps: 0.970
+[  1.7%] Step   1700/100000 | SPS:  1113 | EpRet:   0.000 | Loss: 27.5951 | Q-val: -5.2100 | Eps: 0.968
+[  1.8%] Step   1800/100000 | SPS:   774 | EpRet:   0.000 | Loss: 2.7080 | Q-val: -2.7679 | Eps: 0.966
+[  1.9%] Step   1900/100000 | SPS:   685 | EpRet:   0.000 | Loss: 11.6545 | Q-val: -5.6619 | Eps: 0.964
+[  2.0%] Step   2000/100000 | SPS:   688 | EpRet:   0.000 | Loss: 2.5694 | Q-val: -4.9087 | Eps: 0.962
+[  2.1%] Step   2100/100000 | SPS:   696 | EpRet:   0.000 | Loss: 1.4418 | Q-val: -3.4358 | Eps: 0.960
+[  2.2%] Step   2200/100000 | SPS:   704 | EpRet:   0.000 | Loss: 3.4676 | Q-val: -3.2345 | Eps: 0.958
+[  2.3%] Step   2300/100000 | SPS:   711 | EpRet:   0.000 | Loss: 4.6980 | Q-val: -4.6280 | Eps: 0.956
+[  2.4%] Step   2400/100000 | SPS:   718 | EpRet:   0.000 | Loss: 4.4856 | Q-val: -3.7042 | Eps: 0.954
+[  2.5%] Step   2500/100000 | SPS:   724 | EpRet:   0.000 | Loss: 16.9031 | Q-val: -1.1285 | Eps: 0.953
+[  2.6%] Step   2600/100000 | SPS:   730 | EpRet:   0.000 | Loss: 0.8887 | Q-val: 0.1586 | Eps: 0.951
+[  2.7%] Step   2700/100000 | SPS:   736 | EpRet:   0.000 | Loss: 0.4255 | Q-val: -0.1371 | Eps: 0.949
+[  2.8%] Step   2800/100000 | SPS:   741 | EpRet:   0.000 | Loss: 0.1293 | Q-val: -0.6650 | Eps: 0.947
+[  2.9%] Step   2900/100000 | SPS:   745 | EpRet:   0.000 | Loss: 1.3904 | Q-val: -1.6370 | Eps: 0.945
+[  3.0%] Step   3000/100000 | SPS:   749 | EpRet:   0.000 | Loss: 0.3428 | Q-val: -0.7361 | Eps: 0.943
+[  3.1%] Step   3100/100000 | SPS:   753 | EpRet:   0.000 | Loss: 0.2346 | Q-val: 0.5201 | Eps: 0.941
+[  3.2%] Step   3200/100000 | SPS:   756 | EpRet:   0.000 | Loss: 0.7521 | Q-val: -0.1650 | Eps: 0.939
+[  3.3%] Step   3300/100000 | SPS:   729 | EpRet:   0.000 | Loss: 0.4594 | Q-val: 0.0121 | Eps: 0.937
+[  3.4%] Step   3400/100000 | SPS:   698 | EpRet:   0.000 | Loss: 0.5089 | Q-val: 0.7252 | Eps: 0.935
+[  3.5%] Step   3500/100000 | SPS:   667 | EpRet:   0.000 | Loss: 2.0388 | Q-val: 0.8035 | Eps: 0.933
+[  3.6%] Step   3600/100000 | SPS:   619 | EpRet:   0.000 | Loss: 1.6456 | Q-val: 2.8996 | Eps: 0.932
+[  3.7%] Step   3700/100000 | SPS:   564 | EpRet:   0.000 | Loss: 0.3920 | Q-val: 2.0191 | Eps: 0.930
+[  3.8%] Step   3800/100000 | SPS:   517 | EpRet:   0.000 | Loss: 1.0114 | Q-val: 1.1074 | Eps: 0.928
+[  3.9%] Step   3900/100000 | SPS:   477 | EpRet:   0.000 | Loss: 1.5410 | Q-val: 1.7671 | Eps: 0.926
+[  4.0%] Step   4000/100000 | SPS:   448 | EpRet:   0.000 | Loss: 0.4592 | Q-val: 1.6611 | Eps: 0.924
+[  4.1%] Step   4100/100000 | SPS:   419 | EpRet:   0.000 | Loss: 0.5028 | Q-val: 1.8597 | Eps: 0.922
+[  4.2%] Step   4200/100000 | SPS:   399 | EpRet:   0.000 | Loss: 1.5164 | Q-val: 1.6604 | Eps: 0.920
+[  4.3%] Step   4300/100000 | SPS:   376 | EpRet:   0.000 | Loss: 2.9306 | Q-val: 1.3875 | Eps: 0.918
+[  4.4%] Step   4400/100000 | SPS:   363 | EpRet:   0.000 | Loss: 0.6109 | Q-val: 2.2573 | Eps: 0.916
+[  4.5%] Step   4500/100000 | SPS:   345 | EpRet:   0.000 | Loss: 1.7421 | Q-val: 1.6360 | Eps: 0.914
+Traceback (most recent call last):
+  File "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/dqn_bandit.py", line 253, in <module>
+    q_values = q_network(torch.Tensor(obs).to(device))
+                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+KeyboardInterrupt
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/requirements.txt b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..252d801b74f9b4745c1e1681ec0a02a5dca88347
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/requirements.txt
@@ -0,0 +1,304 @@
+setuptools==80.9.0
+wheel==0.45.1
+pip==25.2
+zipp==3.23.0
+verl==0.2.0.dev0
+ragen==0.1
+triton==3.2.0
+nvidia-cusparselt-cu12==0.6.2
+mpmath==1.3.0
+typing_extensions==4.15.0
+sympy==1.13.1
+nvidia-nvtx-cu12==12.4.127
+nvidia-nvjitlink-cu12==12.4.127
+nvidia-nccl-cu12==2.21.5
+nvidia-curand-cu12==10.3.5.147
+nvidia-cufft-cu12==11.2.1.3
+nvidia-cuda-runtime-cu12==12.4.127
+nvidia-cuda-nvrtc-cu12==12.4.127
+nvidia-cuda-cupti-cu12==12.4.127
+nvidia-cublas-cu12==12.4.5.8
+networkx==3.5
+MarkupSafe==2.1.5
+fsspec==2025.9.0
+filelock==3.19.1
+nvidia-cusparse-cu12==12.3.1.170
+nvidia-cudnn-cu12==9.1.0.70
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.6.1.9
+torch==2.6.0+cu124
+einops==0.8.1
+flash_attn==2.7.4.post1
+pytz==2025.2
+pyperclip==1.11.0
+pylatexenc==2.10
+pyjnius==1.7.0
+py-cpuinfo==9.0.0
+pure_eval==0.2.3
+ptyprocess==0.7.0
+gym-notices==0.1.0
+flatbuffers==25.9.23
+fastrlock==0.8.3
+Farama-Notifications==0.0.4
+cymem==2.0.11
+antlr4-python3-runtime==4.9.3
+xxhash==3.6.0
+wrapt==2.0.0
+Werkzeug==3.1.3
+websockets==15.0.1
+wcwidth==0.2.14
+wasabi==1.1.3
+uvloop==0.22.1
+urllib3==2.5.0
+tzdata==2025.2
+typing-inspection==0.4.2
+traitlets==5.14.3
+tqdm==4.67.1
+threadpoolctl==3.6.0
+tabulate==0.9.0
+spacy-loggers==1.0.5
+spacy-legacy==3.0.12
+soupsieve==2.8
+sniffio==1.3.1
+smmap==5.0.2
+six==1.17.0
+shellingham==1.5.4
+sentencepiece==0.2.1
+safetensors==0.6.2
+rpds-py==0.28.0
+rignore==0.7.6
+regex==2025.11.3
+RapidFuzz==3.14.3
+pyzmq==27.1.0
+PyYAML==6.0.3
+pytokens==0.3.0
+python-multipart==0.0.20
+python-json-logger==4.0.0
+python-dotenv==1.2.1
+PySocks==1.7.1
+pyparsing==3.2.5
+PyJWT==2.10.1
+Pygments==2.19.2
+pygame==2.6.1
+pydantic_core==2.41.5
+pycparser==2.23
+pycountry==24.6.1
+pybind11==3.0.1
+pyarrow==22.0.0
+psutil==7.1.3
+protobuf==6.33.0
+propcache==0.4.1
+prometheus_client==0.23.1
+platformdirs==4.5.0
+pillow==11.3.0
+pexpect==4.9.0
+pathvalidate==3.3.1
+pathspec==0.12.1
+pathable==0.4.4
+partial-json-parser==0.2.1.1.post6
+parso==0.8.5
+packaging==25.0
+orjson==3.11.4
+numpy==1.26.4
+ninja==1.13.0
+nest-asyncio==1.6.0
+mypy_extensions==1.1.0
+murmurhash==1.0.13
+multidict==6.7.0
+msgspec==0.19.0
+msgpack==1.1.2
+more-itertools==10.8.0
+mdurl==0.1.2
+marisa-trie==1.3.1
+llvmlite==0.43.0
+llguidance==0.7.30
+lark==1.2.2
+kiwisolver==1.4.9
+joblib==1.5.2
+jiter==0.11.1
+jeepney==0.9.0
+jaraco.context==6.0.1
+itsdangerous==2.2.0
+interegular==0.3.3
+idna==3.11
+humanfriendly==10.0
+httpx-sse==0.4.3
+httptools==0.7.1
+html2text==2025.4.15
+hf-xet==1.2.0
+h11==0.16.0
+frozenlist==1.8.0
+fonttools==4.60.1
+executing==2.2.1
+exceptiongroup==1.3.0
+eval_type_backport==0.2.2
+docutils==0.22.3
+docstring_parser==0.17.0
+dnspython==2.8.0
+distro==1.9.0
+diskcache==5.6.3
+dill==0.4.0
+decorator==5.2.1
+debugpy==1.8.17
+Cython==3.2.0
+cycler==0.12.1
+codetiming==1.4.0
+cloudpickle==3.1.2
+cloudpathlib==0.23.0
+click==8.2.1
+charset-normalizer==3.4.4
+certifi==2025.10.5
+catalogue==2.0.10
+cachetools==6.2.1
+blinker==1.9.0
+blake3==1.0.8
+beartype==0.22.5
+attrs==25.4.0
+asttokens==3.0.0
+astor==0.8.1
+annotated-types==0.7.0
+annotated-doc==0.0.3
+airportsdata==20250909
+aiohappyeyeballs==2.6.1
+yarl==1.22.0
+uvicorn==0.38.0
+typer-slim==0.20.0
+thefuzz==0.22.1
+stack-data==0.6.3
+srsly==2.5.1
+smart_open==7.4.4
+sentry-sdk==2.43.0
+scipy==1.16.3
+requests==2.32.5
+referencing==0.36.2
+rank-bm25==0.2.2
+python-dateutil==2.9.0.post0
+pydantic==2.12.4
+py-key-value-shared==0.2.8
+prompt_toolkit==3.0.52
+preshed==3.0.10
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+numba==0.60.0
+nltk==3.9.2
+multiprocess==0.70.18
+matplotlib-inline==0.2.1
+markdown-it-py==4.0.0
+language_data==1.3.0
+jedi==0.19.2
+jaraco.functools==4.3.0
+jaraco.classes==3.4.0
+ipython_pygments_lexers==1.1.1
+importlib_metadata==8.7.0
+ImageIO==2.37.2
+httpcore==1.0.9
+gymnasium==1.2.2
+gym==0.26.2
+gitdb==4.0.12
+gguf==0.10.0
+Flask==3.1.2
+faiss-cpu==1.12.0
+email-validator==2.3.0
+depyf==0.18.0
+cupy-cuda12x==13.6.0
+contourpy==1.3.3
+coloredlogs==15.0.1
+cffi==2.0.0
+blis==1.3.0
+black==25.9.0
+beautifulsoup4==4.14.2
+anyio==4.11.0
+aiosignal==1.4.0
+watchfiles==1.1.1
+tiktoken==0.12.0
+starlette==0.49.3
+sse-starlette==3.0.3
+scikit-learn==1.7.2
+rich==14.2.0
+pydantic-settings==2.11.0
+pydantic-extra-types==2.10.6
+py-key-value-aio==0.2.8
+pandas==2.3.3
+openapi-pydantic==0.5.1
+onnxruntime==1.23.2
+matplotlib==3.10.7
+lm-format-enforcer==0.10.12
+langcodes==3.5.0
+jsonschema-specifications==2025.9.1
+jsonschema-path==0.3.4
+ipython==9.7.0
+hydra-core==1.3.2
+huggingface-hub==0.36.0
+httpx==0.28.1
+gym-sokoban==0.0.6
+GitPython==3.1.45
+cryptography==46.0.3
+confection==0.1.5
+cleantext==1.1.4
+aiohttp==3.13.2
+xformers==0.0.29.post2
+weasel==0.4.2
+wandb==0.22.3
+typer==0.19.2
+torchvision==0.21.0
+torchdata==0.11.0
+torchaudio==2.6.0
+tokenizers==0.22.1
+thinc==8.3.7
+tensordict==0.8.3
+SecretStorage==3.4.0
+rich-toolkit==0.15.1
+rich-rst==1.3.2
+prometheus-fastapi-instrumentator==7.1.0
+openai==2.7.1
+jsonschema==4.25.1
+gdown==5.2.0
+fastapi==0.121.0
+Authlib==1.6.5
+anthropic==0.72.0
+accelerate==1.11.0
+transformers==4.57.1
+together==1.5.30
+spacy==3.8.7
+ray==2.51.1
+outlines_core==0.1.26
+mistral_common==1.8.5
+mcp==1.20.0
+keyring==25.6.0
+fastapi-cloud-cli==0.3.1
+fastapi-cli==0.0.14
+datasets==4.4.1
+cyclopts==4.2.1
+xgrammar==0.1.16
+peft==0.17.1
+outlines==0.1.11
+compressed-tensors==0.9.2
+fastmcp==2.13.0.2
+vllm==0.8.2
+pyserini==1.3.0
+typeguard==4.4.4
+shtab==1.7.2
+tyro==0.9.35
+tensorboard-data-server==0.7.2
+Markdown==3.10
+grpcio==1.76.0
+absl-py==2.3.1
+tensorboard==2.20.0
+ragen==0.1
+verl==0.2.0.dev0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+importlib_metadata==8.0.0
+inflect==7.3.1
+jaraco.collections==5.1.0
+jaraco.context==5.3.0
+jaraco.functools==4.0.1
+jaraco.text==3.12.1
+more-itertools==10.3.0
+packaging==24.2
+platformdirs==4.2.2
+tomli==2.0.1
+typeguard==4.3.0
+typing_extensions==4.12.2
+wheel==0.45.1
+zipp==3.19.2
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-metadata.json b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-metadata.json
new file mode 100644
index 0000000000000000000000000000000000000000..6e77c56819547d5aa3dc9be03249749e9180ac8e
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-metadata.json
@@ -0,0 +1,94 @@
+{
+  "os":  "Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.35",
+  "python":  "CPython 3.12.12",
+  "startedAt":  "2025-11-07T04:56:47.687990Z",
+  "args":  [
+    "--track",
+    "--wandb-project-name",
+    "ragen-bandit"
+  ],
+  "program":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/dqn_bandit.py",
+  "codePath":  "cleanrl/dqn_bandit.py",
+  "codePathLocal":  "dqn_bandit.py",
+  "git":  {
+    "remote":  "https://github.com/vwxyzjn/cleanrl.git",
+    "commit":  "004f8a086a892a2a180f4dd332b90d83a968aa7a"
+  },
+  "email":  "haoyu-wa22@mails.tsinghua.edu.cn",
+  "root":  "/mnt/general/wanghy/RAGEN/cleanrl/cleanrl",
+  "host":  "pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0",
+  "executable":  "/root/local/miniconda3/envs/ragen/bin/python",
+  "cpu_count":  64,
+  "cpu_count_logical":  128,
+  "gpu":  "NVIDIA H100 80GB HBM3",
+  "gpu_count":  8,
+  "disk":  {
+    "/":  {
+      "total":  "5153960755200",
+      "used":  "31724965888"
+    }
+  },
+  "memory":  {
+    "total":  "2163642122240"
+  },
+  "gpu_nvidia":  [
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-35e2d43d-4067-82ce-90d4-def9e389bf28"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-af4135e3-88f2-e9ac-518d-502c75a85429"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-d7fdeeba-fe9b-ec03-d9f7-6724fe4266b5"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-ccc4f668-3882-5a8e-2c07-c5cd08f6f666"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-7b73c0cf-d3d5-e10c-7176-a43be1e41001"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-81b58d94-5d1f-8ec2-f9d2-fd56172ed177"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-03e8bc66-3b44-6794-49fd-5392fbdda6d1"
+    },
+    {
+      "name":  "NVIDIA H100 80GB HBM3",
+      "memoryTotal":  "85520809984",
+      "cudaCores":  16896,
+      "architecture":  "Hopper",
+      "uuid":  "GPU-b7cf0ec6-7c29-1179-dceb-09565da51890"
+    }
+  ],
+  "cudaVersion":  "12.4",
+  "writerId":  "jbqdicudyaaspbdzq2z1959r4qyohhsf"
+}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-summary.json b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-summary.json
new file mode 100644
index 0000000000000000000000000000000000000000..d32c8e2339c92b780094efd1ddb2bfe62be153b0
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/files/wandb-summary.json
@@ -0,0 +1 @@
+{"losses/q_values":1.63602,"_runtime":14,"global_step":4500,"_step":139,"losses/td_loss":1.7421228,"_wandb":{"runtime":14},"charts/epsilon":0.9145,"_timestamp":1.7624914231757627e+09,"charts/SPS":345}
\ No newline at end of file
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/logs/debug-internal.log b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/logs/debug-internal.log
new file mode 100644
index 0000000000000000000000000000000000000000..a61ba666dd77c3c9e5ab5f378fed38b993c06108
--- /dev/null
+++ b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/logs/debug-internal.log
@@ -0,0 +1,15 @@
+{"time":"2025-11-07T12:56:47.957309906+08:00","level":"INFO","msg":"stream: starting","core version":"0.22.3"}
+{"time":"2025-11-07T12:56:48.33142385+08:00","level":"INFO","msg":"stream: created new stream","id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.331674194+08:00","level":"INFO","msg":"handler: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335368381+08:00","level":"INFO","msg":"stream: started","id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335386015+08:00","level":"INFO","msg":"writer: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.335389832+08:00","level":"INFO","msg":"sender: started","stream_id":"g1edw7ov"}
+{"time":"2025-11-07T12:56:48.726770104+08:00","level":"ERROR","msg":"git repo not found","error":"repository does not exist"}
+{"time":"2025-11-07T12:56:58.861202459+08:00","level":"WARN","msg":"tensorboard: no root directory","error":"timed out after 10s"}
+{"time":"2025-11-07T12:56:58.862839009+08:00","level":"INFO","msg":"tensorboard: tracking new log directory","rootDir":{},"logDir":"LocalOrCloudPath(LocalPath=\"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491400\")","namespace":""}
+{"time":"2025-11-07T12:56:58.864923146+08:00","level":"INFO","msg":"tensorboard: saving file","fileLocation":"/mnt/general/wanghy/RAGEN/cleanrl/cleanrl/runs/Bandit__dqn_bandit__1__1762491400/events.out.tfevents.1762491408.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.234009.0","runPath":"runs/Bandit__dqn_bandit__1__1762491400/events.out.tfevents.1762491408.pt-4d654cf4576f4d23ad3d3919f12932fe-worker-0.234009.0"}
+{"time":"2025-11-07T12:56:58.868040818+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":122}
+{"time":"2025-11-07T12:56:58.871369044+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":311}
+{"time":"2025-11-07T12:57:03.576733669+08:00","level":"INFO","msg":"stream: closing","id":"g1edw7ov"}
+{"time":"2025-11-07T12:57:03.579454439+08:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":506}
+{"time":"2025-11-07T12:57:03.581505913+08:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":20}
diff --git a/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/run-g1edw7ov.wandb b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/run-g1edw7ov.wandb
new file mode 100644
index 0000000000000000000000000000000000000000..2b5373e0e2a608d6d5a167503ee192409bd27cee
Binary files /dev/null and b/cleanrl/cleanrl/wandb/run-20251107_125647-g1edw7ov/run-g1edw7ov.wandb differ
diff --git a/cleanrl/cleanrl_utils/__init__.py b/cleanrl/cleanrl_utils/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/cleanrl/cleanrl_utils/add_header.py b/cleanrl/cleanrl_utils/add_header.py
new file mode 100644
index 0000000000000000000000000000000000000000..f81a73c266202da854376bbddb014f3972155bd6
--- /dev/null
+++ b/cleanrl/cleanrl_utils/add_header.py
@@ -0,0 +1,28 @@
+import os
+
+
+def add_header(dirname: str):
+    """
+    Add a header string with documentation link
+    to each file in the directory `dirname`.
+    """
+
+    for filename in os.listdir(dirname):
+        if filename.endswith(".py"):
+            with open(os.path.join(dirname, filename)) as f:
+                lines = f.readlines()
+
+            # hacky bit
+            exp_name = filename.split(".")[0]
+            algo_name = exp_name.split("_")[0]
+            header_string = f"# docs and experiment results can be found at https://docs.cleanrl.dev/rl-algorithms/{algo_name}/#{exp_name}py"
+
+            if not lines[0].startswith(header_string):
+                print(f"adding headers for {filename}")
+                lines.insert(0, header_string + "\n")
+                with open(os.path.join(dirname, filename), "w") as f:
+                    f.writelines(lines)
+
+
+if __name__ == "__main__":
+    add_header("cleanrl")
diff --git a/cleanrl/cleanrl_utils/atari_wrappers.py b/cleanrl/cleanrl_utils/atari_wrappers.py
new file mode 100644
index 0000000000000000000000000000000000000000..238690af0b0f2beb17cd59a1440eff1f81de90cb
--- /dev/null
+++ b/cleanrl/cleanrl_utils/atari_wrappers.py
@@ -0,0 +1,325 @@
+# Copyright notice
+#
+# This file contains code adapted from stable-baselines3
+# (https://github.com/DLR-RM/stable-baselines3/blob/master/stable_baselines3/common/atari_wrappers.py)
+# licensed under the MIT License.
+#
+# Copyright (c) 2019-2023 Antonin Raffin, Ashley Hill, Anssi Kanervisto,
+# Maximilian Ernestus, Rinu Boney, Pavan Goli, and other contributors
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+from __future__ import annotations
+
+from typing import SupportsFloat
+
+import gymnasium as gym
+import numpy as np
+from gymnasium import spaces
+
+try:
+    import cv2
+
+    cv2.ocl.setUseOpenCL(False)
+except ImportError:
+    cv2 = None  # type: ignore[assignment]
+
+
+class StickyActionEnv(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Sticky action.
+
+    Paper: https://arxiv.org/abs/1709.06009
+    Official implementation: https://github.com/mgbellemare/Arcade-Learning-Environment
+
+    :param env: Environment to wrap
+    :param action_repeat_probability: Probability of repeating the last action
+    """
+
+    def __init__(self, env: gym.Env, action_repeat_probability: float) -> None:
+        super().__init__(env)
+        self.action_repeat_probability = action_repeat_probability
+        assert env.unwrapped.get_action_meanings()[0] == "NOOP"  # type: ignore[attr-defined]
+
+    def reset(self, **kwargs):
+        self._sticky_action = 0  # NOOP
+        return self.env.reset(**kwargs)
+
+    def step(self, action: int):
+        if self.np_random.random() >= self.action_repeat_probability:
+            self._sticky_action = action
+        return self.env.step(self._sticky_action)
+
+
+class NoopResetEnv(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Sample initial states by taking random number of no-ops on reset.
+    No-op is assumed to be action 0.
+
+    :param env: Environment to wrap
+    :param noop_max: Maximum value of no-ops to run
+    """
+
+    def __init__(self, env: gym.Env, noop_max: int = 30) -> None:
+        super().__init__(env)
+        self.noop_max = noop_max
+        self.override_num_noops = None
+        self.noop_action = 0
+        assert env.unwrapped.get_action_meanings()[0] == "NOOP"  # type: ignore[attr-defined]
+
+    def reset(self, **kwargs):
+        self.env.reset(**kwargs)
+        if self.override_num_noops is not None:
+            noops = self.override_num_noops
+        else:
+            noops = self.unwrapped.np_random.integers(1, self.noop_max + 1)
+        assert noops > 0
+        obs = np.zeros(0)
+        info: dict = {}
+        for _ in range(noops):
+            obs, _, terminated, truncated, info = self.env.step(self.noop_action)
+            if terminated or truncated:
+                obs, info = self.env.reset(**kwargs)
+        return obs, info
+
+
+class FireResetEnv(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Take action on reset for environments that are fixed until firing.
+
+    :param env: Environment to wrap
+    """
+
+    def __init__(self, env: gym.Env) -> None:
+        super().__init__(env)
+        assert env.unwrapped.get_action_meanings()[1] == "FIRE"  # type: ignore[attr-defined]
+        assert len(env.unwrapped.get_action_meanings()) >= 3  # type: ignore[attr-defined]
+
+    def reset(self, **kwargs):
+        self.env.reset(**kwargs)
+        obs, _, terminated, truncated, _ = self.env.step(1)
+        if terminated or truncated:
+            self.env.reset(**kwargs)
+        obs, _, terminated, truncated, _ = self.env.step(2)
+        if terminated or truncated:
+            self.env.reset(**kwargs)
+        return obs, {}
+
+
+class EpisodicLifeEnv(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Make end-of-life == end-of-episode, but only reset on true game over.
+    Done by DeepMind for the DQN and co. since it helps value estimation.
+
+    :param env: Environment to wrap
+    """
+
+    def __init__(self, env: gym.Env) -> None:
+        super().__init__(env)
+        self.lives = 0
+        self.was_real_done = True
+
+    def step(self, action: int):
+        obs, reward, terminated, truncated, info = self.env.step(action)
+        self.was_real_done = terminated or truncated
+        # check current lives, make loss of life terminal,
+        # then update lives to handle bonus lives
+        lives = self.env.unwrapped.ale.lives()  # type: ignore[attr-defined]
+        if 0 < lives < self.lives:
+            # for Qbert sometimes we stay in lives == 0 condition for a few frames
+            # so its important to keep lives > 0, so that we only reset once
+            # the environment advertises done.
+            terminated = True
+        self.lives = lives
+        return obs, reward, terminated, truncated, info
+
+    def reset(self, **kwargs):
+        """
+        Calls the Gym environment reset, only when lives are exhausted.
+        This way all states are still reachable even though lives are episodic,
+        and the learner need not know about any of this behind-the-scenes.
+
+        :param kwargs: Extra keywords passed to env.reset() call
+        :return: the first observation of the environment
+        """
+        if self.was_real_done:
+            obs, info = self.env.reset(**kwargs)
+        else:
+            # no-op step to advance from terminal/lost life state
+            obs, _, terminated, truncated, info = self.env.step(0)
+
+            # The no-op step can lead to a game over, so we need to check it again
+            # to see if we should reset the environment and avoid the
+            # monitor.py `RuntimeError: Tried to step environment that needs reset`
+            if terminated or truncated:
+                obs, info = self.env.reset(**kwargs)
+        self.lives = self.env.unwrapped.ale.lives()  # type: ignore[attr-defined]
+        return obs, info
+
+
+class MaxAndSkipEnv(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Return only every ``skip``-th frame (frameskipping)
+    and return the max between the two last frames.
+
+    :param env: Environment to wrap
+    :param skip: Number of ``skip``-th frame
+        The same action will be taken ``skip`` times.
+    """
+
+    def __init__(self, env: gym.Env, skip: int = 4) -> None:
+        super().__init__(env)
+        # most recent raw observations (for max pooling across time steps)
+        assert env.observation_space.dtype is not None, "No dtype specified for the observation space"
+        assert env.observation_space.shape is not None, "No shape defined for the observation space"
+        self._obs_buffer = np.zeros((2, *env.observation_space.shape), dtype=env.observation_space.dtype)
+        self._skip = skip
+
+    def step(self, action: int):
+        """
+        Step the environment with the given action
+        Repeat action, sum reward, and max over last observations.
+
+        :param action: the action
+        :return: observation, reward, terminated, truncated, information
+        """
+        total_reward = 0.0
+        terminated = truncated = False
+        for i in range(self._skip):
+            obs, reward, terminated, truncated, info = self.env.step(action)
+            done = terminated or truncated
+            if i == self._skip - 2:
+                self._obs_buffer[0] = obs
+            if i == self._skip - 1:
+                self._obs_buffer[1] = obs
+            total_reward += float(reward)
+            if done:
+                break
+        # Note that the observation on the done=True frame
+        # doesn't matter
+        max_frame = self._obs_buffer.max(axis=0)
+
+        return max_frame, total_reward, terminated, truncated, info
+
+
+class ClipRewardEnv(gym.RewardWrapper):
+    """
+    Clip the reward to {+1, 0, -1} by its sign.
+
+    :param env: Environment to wrap
+    """
+
+    def __init__(self, env: gym.Env) -> None:
+        super().__init__(env)
+
+    def reward(self, reward: SupportsFloat) -> float:
+        """
+        Bin reward to {+1, 0, -1} by its sign.
+
+        :param reward:
+        :return:
+        """
+        return np.sign(float(reward))
+
+
+class WarpFrame(gym.ObservationWrapper[np.ndarray, int, np.ndarray]):
+    """
+    Convert to grayscale and warp frames to 84x84 (default)
+    as done in the Nature paper and later work.
+
+    :param env: Environment to wrap
+    :param width: New frame width
+    :param height: New frame height
+    """
+
+    def __init__(self, env: gym.Env, width: int = 84, height: int = 84) -> None:
+        super().__init__(env)
+        self.width = width
+        self.height = height
+        assert isinstance(env.observation_space, spaces.Box), f"Expected Box space, got {env.observation_space}"
+
+        self.observation_space = spaces.Box(
+            low=0,
+            high=255,
+            shape=(self.height, self.width, 1),
+            dtype=env.observation_space.dtype,  # type: ignore[arg-type]
+        )
+
+    def observation(self, frame: np.ndarray) -> np.ndarray:
+        """
+        returns the current observation from a frame
+
+        :param frame: environment frame
+        :return: the observation
+        """
+        assert cv2 is not None, "OpenCV is not installed, you can do `pip install opencv-python`"
+        frame = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
+        frame = cv2.resize(frame, (self.width, self.height), interpolation=cv2.INTER_AREA)
+        return frame[:, :, None]
+
+
+class AtariWrapper(gym.Wrapper[np.ndarray, int, np.ndarray, int]):
+    """
+    Atari 2600 preprocessings
+
+    Specifically:
+
+    * Noop reset: obtain initial state by taking random number of no-ops on reset.
+    * Frame skipping: 4 by default
+    * Max-pooling: most recent two observations
+    * Termination signal when a life is lost.
+    * Resize to a square image: 84x84 by default
+    * Grayscale observation
+    * Clip reward to {-1, 0, 1}
+    * Sticky actions: disabled by default
+
+    See https://danieltakeshi.github.io/2016/11/25/frame-skipping-and-preprocessing-for-deep-q-networks-on-atari-2600-games/
+    for a visual explanation.
+
+    .. warning::
+        Use this wrapper only with Atari v4 without frame skip: ``env_id = "*NoFrameskip-v4"``.
+
+    :param env: Environment to wrap
+    :param noop_max: Max number of no-ops
+    :param frame_skip: Frequency at which the agent experiences the game.
+        This correspond to repeating the action ``frame_skip`` times.
+    :param screen_size: Resize Atari frame
+    :param terminal_on_life_loss: If True, then step() returns done=True whenever a life is lost.
+    :param clip_reward: If True (default), the reward is clip to {-1, 0, 1} depending on its sign.
+    :param action_repeat_probability: Probability of repeating the last action
+    """
+
+    def __init__(
+        self,
+        env: gym.Env,
+        noop_max: int = 30,
+        frame_skip: int = 4,
+        screen_size: int = 84,
+        terminal_on_life_loss: bool = True,
+        clip_reward: bool = True,
+        action_repeat_probability: float = 0.0,
+    ) -> None:
+        if action_repeat_probability > 0.0:
+            env = StickyActionEnv(env, action_repeat_probability)
+        if noop_max > 0:
+            env = NoopResetEnv(env, noop_max=noop_max)
+        # frame_skip=1 is the same as no frame-skip (action repeat)
+        if frame_skip > 1:
+            env = MaxAndSkipEnv(env, skip=frame_skip)
+        if terminal_on_life_loss:
+            env = EpisodicLifeEnv(env)
+        if "FIRE" in env.unwrapped.get_action_meanings():  # type: ignore[attr-defined]
+            env = FireResetEnv(env)
+        env = WarpFrame(env, width=screen_size, height=screen_size)
+        if clip_reward:
+            env = ClipRewardEnv(env)
+
+        super().__init__(env)
diff --git a/cleanrl/cleanrl_utils/benchmark.py b/cleanrl/cleanrl_utils/benchmark.py
new file mode 100644
index 0000000000000000000000000000000000000000..042a223f7bb29659496aeb0575af53271f3657eb
--- /dev/null
+++ b/cleanrl/cleanrl_utils/benchmark.py
@@ -0,0 +1,152 @@
+import math
+import os
+import shlex
+import subprocess
+import uuid
+from dataclasses import dataclass
+from typing import List, Optional
+
+import requests
+import tyro
+
+
+@dataclass
+class Args:
+    env_ids: List[str]
+    """the ids of the environment to compare"""
+    command: str
+    """the command to run"""
+    num_seeds: int = 3
+    """the number of random seeds"""
+    start_seed: int = 1
+    """the number of the starting seed"""
+    workers: int = 0
+    """the number of workers to run benchmark experimenets"""
+    auto_tag: bool = True
+    """if toggled, the runs will be tagged with git tags, commit, and pull request number if possible"""
+    slurm_template_path: Optional[str] = None
+    """the path to the slurm template file (see docs for more details)"""
+    slurm_gpus_per_task: Optional[int] = None
+    """the number of gpus per task to use for slurm jobs"""
+    slurm_total_cpus: Optional[int] = None
+    """the number of gpus per task to use for slurm jobs"""
+    slurm_ntasks: Optional[int] = None
+    """the number of tasks to use for slurm jobs"""
+    slurm_nodes: Optional[int] = None
+    """the number of nodes to use for slurm jobs"""
+
+
+def run_experiment(command: str):
+    command_list = shlex.split(command)
+    print(f"running {command}")
+
+    # Use subprocess.PIPE to capture the output
+    fd = subprocess.Popen(command_list, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+    output, errors = fd.communicate()
+
+    return_code = fd.returncode
+    assert return_code == 0, f"Command failed with error: {errors.decode('utf-8')}"
+
+    # Convert bytes to string and strip leading/trailing whitespaces
+    return output.decode("utf-8").strip()
+
+
+def autotag() -> str:
+    wandb_tag = ""
+    print("autotag feature is enabled")
+    git_tag = ""
+    try:
+        git_tag = subprocess.check_output(["git", "describe", "--tags"]).decode("ascii").strip()
+        print(f"identified git tag: {git_tag}")
+    except subprocess.CalledProcessError as e:
+        print(e)
+    if len(git_tag) == 0:
+        try:
+            count = int(subprocess.check_output(["git", "rev-list", "--count", "HEAD"]).decode("ascii").strip())
+            hash = subprocess.check_output(["git", "rev-parse", "--short", "HEAD"]).decode("ascii").strip()
+            git_tag = f"no-tag-{count}-g{hash}"
+            print(f"identified git tag: {git_tag}")
+        except subprocess.CalledProcessError as e:
+            print(e)
+    wandb_tag = git_tag
+
+    git_commit = subprocess.check_output(["git", "rev-parse", "--verify", "HEAD"]).decode("ascii").strip()
+    try:
+        # try finding the pull request number on github
+        prs = requests.get(f"https://api.github.com/search/issues?q=repo:vwxyzjn/cleanrl+is:pr+{git_commit}")
+        if prs.status_code == 200:
+            prs = prs.json()
+            if len(prs["items"]) > 0:
+                pr = prs["items"][0]
+                pr_number = pr["number"]
+                wandb_tag += f",pr-{pr_number}"
+        print(f"identified github pull request: {pr_number}")
+    except Exception as e:
+        print(e)
+
+    return wandb_tag
+
+
+if __name__ == "__main__":
+    args = tyro.cli(Args)
+    if args.auto_tag:
+        existing_wandb_tag = os.environ.get("WANDB_TAGS", "")
+        wandb_tag = autotag()
+        if len(wandb_tag) > 0:
+            if len(existing_wandb_tag) > 0:
+                os.environ["WANDB_TAGS"] = ",".join([existing_wandb_tag, wandb_tag])
+            else:
+                os.environ["WANDB_TAGS"] = wandb_tag
+    print("WANDB_TAGS: ", os.environ.get("WANDB_TAGS", ""))
+    commands = []
+    for seed in range(0, args.num_seeds):
+        for env_id in args.env_ids:
+            commands += [" ".join([args.command, "--env-id", env_id, "--seed", str(args.start_seed + seed)])]
+
+    print("======= commands to run:")
+    for command in commands:
+        print(command)
+
+    if args.workers > 0 and args.slurm_template_path is None:
+        from concurrent.futures import ThreadPoolExecutor
+
+        executor = ThreadPoolExecutor(max_workers=args.workers, thread_name_prefix="cleanrl-benchmark-worker-")
+        for command in commands:
+            executor.submit(run_experiment, command)
+        executor.shutdown(wait=True)
+    else:
+        print("not running the experiments because --workers is set to 0; just printing the commands to run")
+
+    # SLURM logic
+    if args.slurm_template_path is not None:
+        if not os.path.exists("slurm"):
+            os.makedirs("slurm")
+        if not os.path.exists("slurm/logs"):
+            os.makedirs("slurm/logs")
+        print("======= slurm commands to run:")
+        with open(args.slurm_template_path) as f:
+            slurm_template = f.read()
+        slurm_template = slurm_template.replace("{{array}}", f"0-{len(commands) - 1}%{args.workers}")
+        slurm_template = slurm_template.replace("{{env_ids}}", f"({' '.join(args.env_ids)})")
+        slurm_template = slurm_template.replace(
+            "{{seeds}}",
+            f"({' '.join([str(args.start_seed + int(seed)) for seed in range(args.num_seeds)])})",
+        )
+        slurm_template = slurm_template.replace("{{len_seeds}}", f"{args.num_seeds}")
+        slurm_template = slurm_template.replace("{{command}}", args.command)
+        slurm_template = slurm_template.replace("{{gpus_per_task}}", f"{args.slurm_gpus_per_task}")
+        total_gpus = args.slurm_gpus_per_task * args.slurm_ntasks
+        slurm_cpus_per_gpu = math.ceil(args.slurm_total_cpus / total_gpus)
+        slurm_template = slurm_template.replace("{{cpus_per_gpu}}", f"{slurm_cpus_per_gpu}")
+        slurm_template = slurm_template.replace("{{ntasks}}", f"{args.slurm_ntasks}")
+        if args.slurm_nodes is not None:
+            slurm_template = slurm_template.replace("{{nodes}}", f"#SBATCH --nodes={args.slurm_nodes}")
+        else:
+            slurm_template = slurm_template.replace("{{nodes}}", "")
+        filename = str(uuid.uuid4())
+        open(os.path.join("slurm", f"{filename}.slurm"), "w").write(slurm_template)
+        slurm_path = os.path.join("slurm", f"{filename}.slurm")
+        print(f"saving command in {slurm_path}")
+        if args.workers > 0:
+            job_id = run_experiment(f"sbatch --parsable {slurm_path}")
+            print(f"Job ID: {job_id}")
diff --git a/cleanrl/cleanrl_utils/buffers.py b/cleanrl/cleanrl_utils/buffers.py
new file mode 100644
index 0000000000000000000000000000000000000000..544faaf478831e528fd90b1ea02005be9befcdea
--- /dev/null
+++ b/cleanrl/cleanrl_utils/buffers.py
@@ -0,0 +1,610 @@
+# Copyright notice
+#
+# This file contains code adapted from stable-baselines3
+# (https://github.com/DLR-RM/stable-baselines3/blob/master/stable_baselines3/common/buffers.py)
+# licensed under the MIT License.
+#
+# Copyright (c) 2019-2023 Antonin Raffin, Ashley Hill, Anssi Kanervisto,
+# Maximilian Ernestus, Rinu Boney, Pavan Goli, and other contributors
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+from __future__ import annotations
+
+import warnings
+from abc import ABC, abstractmethod
+from collections.abc import Generator
+from typing import Any, NamedTuple
+
+import numpy as np
+import torch as th
+from gymnasium import spaces
+
+try:
+    # Check memory used by replay buffer when possible
+    import psutil
+except ImportError:
+    psutil = None
+
+
+__all__ = [
+    "BaseBuffer",
+    "RolloutBuffer",
+    "ReplayBuffer",
+    "RolloutBufferSamples",
+    "ReplayBufferSamples",
+]
+
+
+class RolloutBufferSamples(NamedTuple):
+    observations: th.Tensor
+    actions: th.Tensor
+    old_values: th.Tensor
+    old_log_prob: th.Tensor
+    advantages: th.Tensor
+    returns: th.Tensor
+
+
+class ReplayBufferSamples(NamedTuple):
+    observations: th.Tensor
+    actions: th.Tensor
+    next_observations: th.Tensor
+    dones: th.Tensor
+    rewards: th.Tensor
+
+
+def get_action_dim(action_space: spaces.Space) -> int:
+    """
+    Get the dimension of the action space.
+
+    :param action_space:
+    :return:
+    """
+    if isinstance(action_space, spaces.Box):
+        return int(np.prod(action_space.shape))
+    elif isinstance(action_space, spaces.Discrete):
+        # Action is an int
+        return 1
+    elif isinstance(action_space, spaces.MultiDiscrete):
+        # Number of discrete actions
+        return int(len(action_space.nvec))
+    elif isinstance(action_space, spaces.MultiBinary):
+        # Number of binary actions
+        assert isinstance(
+            action_space.n, int
+        ), f"Multi-dimensional MultiBinary({action_space.n}) action space is not supported. You can flatten it instead."
+        return int(action_space.n)
+    else:
+        raise NotImplementedError(f"{action_space} action space is not supported")
+
+
+def get_obs_shape(
+    observation_space: spaces.Space,
+) -> tuple[int, ...] | dict[str, tuple[int, ...]]:
+    """
+    Get the shape of the observation (useful for the buffers).
+
+    :param observation_space:
+    :return:
+    """
+    if isinstance(observation_space, spaces.Box):
+        return observation_space.shape
+    elif isinstance(observation_space, spaces.Discrete):
+        # Observation is an int
+        return (1,)
+    elif isinstance(observation_space, spaces.MultiDiscrete):
+        # Number of discrete features
+        return (int(len(observation_space.nvec)),)
+    elif isinstance(observation_space, spaces.MultiBinary):
+        # Number of binary features
+        return observation_space.shape
+    elif isinstance(observation_space, spaces.Dict):
+        return {key: get_obs_shape(subspace) for (key, subspace) in observation_space.spaces.items()}  # type: ignore[misc]
+
+    else:
+        raise NotImplementedError(f"{observation_space} observation space is not supported")
+
+
+def get_device(device: th.device | str = "auto") -> th.device:
+    """
+    Retrieve PyTorch device.
+    It checks that the requested device is available first.
+    For now, it supports only cpu and cuda.
+    By default, it tries to use the gpu.
+
+    :param device: One for 'auto', 'cuda', 'cpu'
+    :return: Supported Pytorch device
+    """
+    # Cuda by default
+    if device == "auto":
+        device = "cuda"
+    # Force conversion to th.device
+    device = th.device(device)
+
+    # Cuda not available
+    if device.type == th.device("cuda").type and not th.cuda.is_available():
+        return th.device("cpu")
+
+    return device
+
+
+class BaseBuffer(ABC):
+    """
+    Base class that represent a buffer (rollout or replay)
+
+    :param buffer_size: Max number of element in the buffer
+    :param observation_space: Observation space
+    :param action_space: Action space
+    :param device: PyTorch device
+        to which the values will be converted
+    :param n_envs: Number of parallel environments
+    """
+
+    observation_space: spaces.Space
+    obs_shape: tuple[int, ...]
+
+    def __init__(
+        self,
+        buffer_size: int,
+        observation_space: spaces.Space,
+        action_space: spaces.Space,
+        device: th.device | str = "auto",
+        n_envs: int = 1,
+    ):
+        super().__init__()
+        self.buffer_size = buffer_size
+        self.observation_space = observation_space
+        self.action_space = action_space
+        self.obs_shape = get_obs_shape(observation_space)  # type: ignore[assignment]
+
+        self.action_dim = get_action_dim(action_space)
+        self.pos = 0
+        self.full = False
+        self.device = get_device(device)
+        self.n_envs = n_envs
+
+    @staticmethod
+    def swap_and_flatten(arr: np.ndarray) -> np.ndarray:
+        """
+        Swap and then flatten axes 0 (buffer_size) and 1 (n_envs)
+        to convert shape from [n_steps, n_envs, ...] (when ... is the shape of the features)
+        to [n_steps * n_envs, ...] (which maintain the order)
+
+        :param arr:
+        :return:
+        """
+        shape = arr.shape
+        if len(shape) < 3:
+            shape = (*shape, 1)
+        return arr.swapaxes(0, 1).reshape(shape[0] * shape[1], *shape[2:])
+
+    def size(self) -> int:
+        """
+        :return: The current size of the buffer
+        """
+        if self.full:
+            return self.buffer_size
+        return self.pos
+
+    def add(self, *args, **kwargs) -> None:
+        """
+        Add elements to the buffer.
+        """
+        raise NotImplementedError()
+
+    def extend(self, *args, **kwargs) -> None:
+        """
+        Add a new batch of transitions to the buffer
+        """
+        # Do a for loop along the batch axis
+        for data in zip(*args):
+            self.add(*data)
+
+    def reset(self) -> None:
+        """
+        Reset the buffer.
+        """
+        self.pos = 0
+        self.full = False
+
+    def sample(self, batch_size: int):
+        """
+        :param batch_size: Number of element to sample
+        :return:
+        """
+        upper_bound = self.buffer_size if self.full else self.pos
+        batch_inds = np.random.randint(0, upper_bound, size=batch_size)
+        return self._get_samples(batch_inds)
+
+    @abstractmethod
+    def _get_samples(self, batch_inds: np.ndarray) -> ReplayBufferSamples | RolloutBufferSamples:
+        """
+        :param batch_inds:
+        :return:
+        """
+        raise NotImplementedError()
+
+    def to_torch(self, array: np.ndarray, copy: bool = True) -> th.Tensor:
+        """
+        Convert a numpy array to a PyTorch tensor.
+        Note: it copies the data by default
+
+        :param array:
+        :param copy: Whether to copy or not the data (may be useful to avoid changing things
+            by reference). This argument is inoperative if the device is not the CPU.
+        :return:
+        """
+        if copy:
+            return th.tensor(array, device=self.device)
+        return th.as_tensor(array, device=self.device)
+
+
+class ReplayBuffer(BaseBuffer):
+    """
+    Replay buffer used in off-policy algorithms like SAC/TD3.
+
+    :param buffer_size: Max number of element in the buffer
+    :param observation_space: Observation space
+    :param action_space: Action space
+    :param device: PyTorch device
+    :param n_envs: Number of parallel environments
+    :param optimize_memory_usage: Enable a memory efficient variant
+        of the replay buffer which reduces by almost a factor two the memory used,
+        at a cost of more complexity.
+        See https://github.com/DLR-RM/stable-baselines3/issues/37#issuecomment-637501195
+        and https://github.com/DLR-RM/stable-baselines3/pull/28#issuecomment-637559274
+        Cannot be used in combination with handle_timeout_termination.
+    :param handle_timeout_termination: Handle timeout termination (due to timelimit)
+        separately and treat the task as infinite horizon task.
+        https://github.com/DLR-RM/stable-baselines3/issues/284
+    """
+
+    observations: np.ndarray
+    next_observations: np.ndarray
+    actions: np.ndarray
+    rewards: np.ndarray
+    dones: np.ndarray
+    timeouts: np.ndarray
+
+    def __init__(
+        self,
+        buffer_size: int,
+        observation_space: spaces.Space,
+        action_space: spaces.Space,
+        device: th.device | str = "auto",
+        n_envs: int = 1,
+        optimize_memory_usage: bool = False,
+        handle_timeout_termination: bool = True,
+    ):
+        super().__init__(buffer_size, observation_space, action_space, device, n_envs=n_envs)
+
+        # Adjust buffer size
+        self.buffer_size = max(buffer_size // n_envs, 1)
+
+        # Check that the replay buffer can fit into the memory
+        if psutil is not None:
+            mem_available = psutil.virtual_memory().available
+
+        # there is a bug if both optimize_memory_usage and handle_timeout_termination are true
+        # see https://github.com/DLR-RM/stable-baselines3/issues/934
+        if optimize_memory_usage and handle_timeout_termination:
+            raise ValueError(
+                "ReplayBuffer does not support optimize_memory_usage = True "
+                "and handle_timeout_termination = True simultaneously."
+            )
+        self.optimize_memory_usage = optimize_memory_usage
+
+        self.observations = np.zeros((self.buffer_size, self.n_envs, *self.obs_shape), dtype=observation_space.dtype)
+
+        if not optimize_memory_usage:
+            # When optimizing memory, `observations` contains also the next observation
+            self.next_observations = np.zeros((self.buffer_size, self.n_envs, *self.obs_shape), dtype=observation_space.dtype)
+
+        self.actions = np.zeros(
+            (self.buffer_size, self.n_envs, self.action_dim), dtype=self._maybe_cast_dtype(action_space.dtype)
+        )
+
+        self.rewards = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.dones = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        # Handle timeouts termination properly if needed
+        # see https://github.com/DLR-RM/stable-baselines3/issues/284
+        self.handle_timeout_termination = handle_timeout_termination
+        self.timeouts = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+
+        if psutil is not None:
+            total_memory_usage: float = (
+                self.observations.nbytes + self.actions.nbytes + self.rewards.nbytes + self.dones.nbytes
+            )
+
+            if not optimize_memory_usage:
+                total_memory_usage += self.next_observations.nbytes
+
+            if total_memory_usage > mem_available:
+                # Convert to GB
+                total_memory_usage /= 1e9
+                mem_available /= 1e9
+                warnings.warn(
+                    "This system does not have apparently enough memory to store the complete "
+                    f"replay buffer {total_memory_usage:.2f}GB > {mem_available:.2f}GB"
+                )
+
+    def add(
+        self,
+        obs: np.ndarray,
+        next_obs: np.ndarray,
+        action: np.ndarray,
+        reward: np.ndarray,
+        done: np.ndarray,
+        infos: list[dict[str, Any]],
+    ) -> None:
+        # Reshape needed when using multiple envs with discrete observations
+        # as numpy cannot broadcast (n_discrete,) to (n_discrete, 1)
+        if isinstance(self.observation_space, spaces.Discrete):
+            obs = obs.reshape((self.n_envs, *self.obs_shape))
+            next_obs = next_obs.reshape((self.n_envs, *self.obs_shape))
+
+        # Reshape to handle multi-dim and discrete action spaces, see GH #970 #1392
+        action = action.reshape((self.n_envs, self.action_dim))
+
+        # Copy to avoid modification by reference
+        self.observations[self.pos] = np.array(obs)
+
+        if self.optimize_memory_usage:
+            self.observations[(self.pos + 1) % self.buffer_size] = np.array(next_obs)
+        else:
+            self.next_observations[self.pos] = np.array(next_obs)
+
+        self.actions[self.pos] = np.array(action)
+        self.rewards[self.pos] = np.array(reward)
+        self.dones[self.pos] = np.array(done)
+
+        if self.handle_timeout_termination:
+            self.timeouts[self.pos] = np.array([info.get("TimeLimit.truncated", False) for info in infos])
+
+        self.pos += 1
+        if self.pos == self.buffer_size:
+            self.full = True
+            self.pos = 0
+
+    def sample(self, batch_size: int) -> ReplayBufferSamples:
+        """
+        Sample elements from the replay buffer.
+        Custom sampling when using memory efficient variant,
+        as we should not sample the element with index `self.pos`
+        See https://github.com/DLR-RM/stable-baselines3/pull/28#issuecomment-637559274
+
+        :param batch_size: Number of element to sample
+        :return:
+        """
+        if not self.optimize_memory_usage:
+            return super().sample(batch_size=batch_size)
+        # Do not sample the element with index `self.pos` as the transitions is invalid
+        # (we use only one array to store `obs` and `next_obs`)
+        if self.full:
+            batch_inds = (np.random.randint(1, self.buffer_size, size=batch_size) + self.pos) % self.buffer_size
+        else:
+            batch_inds = np.random.randint(0, self.pos, size=batch_size)
+        return self._get_samples(batch_inds)
+
+    def _get_samples(self, batch_inds: np.ndarray) -> ReplayBufferSamples:
+        # Sample randomly the env idx
+        env_indices = np.random.randint(0, high=self.n_envs, size=(len(batch_inds),))
+
+        if self.optimize_memory_usage:
+            next_obs = self.observations[(batch_inds + 1) % self.buffer_size, env_indices, :]
+        else:
+            next_obs = self.next_observations[batch_inds, env_indices, :]
+
+        data = (
+            self.observations[batch_inds, env_indices, :],
+            self.actions[batch_inds, env_indices, :],
+            next_obs,
+            # Only use dones that are not due to timeouts
+            # deactivated by default (timeouts is initialized as an array of False)
+            (self.dones[batch_inds, env_indices] * (1 - self.timeouts[batch_inds, env_indices])).reshape(-1, 1),
+            self.rewards[batch_inds, env_indices].reshape(-1, 1),
+        )
+        return ReplayBufferSamples(*tuple(map(self.to_torch, data)))
+
+    @staticmethod
+    def _maybe_cast_dtype(dtype: np.typing.DTypeLike) -> np.typing.DTypeLike:
+        """
+        Cast `np.float64` action datatype to `np.float32`,
+        keep the others dtype unchanged.
+        See GH#1572 for more information.
+
+        :param dtype: The original action space dtype
+        :return: ``np.float32`` if the dtype was float64,
+            the original dtype otherwise.
+        """
+        if dtype == np.float64:
+            return np.float32
+        return dtype
+
+
+class RolloutBuffer(BaseBuffer):
+    """
+    Rollout buffer used in on-policy algorithms like A2C/PPO.
+    It corresponds to ``buffer_size`` transitions collected
+    using the current policy.
+    This experience will be discarded after the policy update.
+    In order to use PPO objective, we also store the current value of each state
+    and the log probability of each taken action.
+
+    The term rollout here refers to the model-free notion and should not
+    be used with the concept of rollout used in model-based RL or planning.
+    Hence, it is only involved in policy and value function training but not action selection.
+
+    :param buffer_size: Max number of element in the buffer
+    :param observation_space: Observation space
+    :param action_space: Action space
+    :param device: PyTorch device
+    :param gae_lambda: Factor for trade-off of bias vs variance for Generalized Advantage Estimator
+        Equivalent to classic advantage when set to 1.
+    :param gamma: Discount factor
+    :param n_envs: Number of parallel environments
+    """
+
+    observations: np.ndarray
+    actions: np.ndarray
+    rewards: np.ndarray
+    advantages: np.ndarray
+    returns: np.ndarray
+    episode_starts: np.ndarray
+    log_probs: np.ndarray
+    values: np.ndarray
+
+    def __init__(
+        self,
+        buffer_size: int,
+        observation_space: spaces.Space,
+        action_space: spaces.Space,
+        device: th.device | str = "auto",
+        gae_lambda: float = 1,
+        gamma: float = 0.99,
+        n_envs: int = 1,
+    ):
+        super().__init__(buffer_size, observation_space, action_space, device, n_envs=n_envs)
+        self.gae_lambda = gae_lambda
+        self.gamma = gamma
+        self.generator_ready = False
+        self.reset()
+
+    def reset(self) -> None:
+        self.observations = np.zeros((self.buffer_size, self.n_envs, *self.obs_shape), dtype=np.float32)
+        self.actions = np.zeros((self.buffer_size, self.n_envs, self.action_dim), dtype=np.float32)
+        self.rewards = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.returns = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.episode_starts = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.values = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.log_probs = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.advantages = np.zeros((self.buffer_size, self.n_envs), dtype=np.float32)
+        self.generator_ready = False
+        super().reset()
+
+    def compute_returns_and_advantage(self, last_values: th.Tensor, dones: np.ndarray) -> None:
+        """
+        Post-processing step: compute the lambda-return (TD(lambda) estimate)
+        and GAE(lambda) advantage.
+
+        Uses Generalized Advantage Estimation (https://arxiv.org/abs/1506.02438)
+        to compute the advantage. To obtain Monte-Carlo advantage estimate (A(s) = R - V(S))
+        where R is the sum of discounted reward with value bootstrap
+        (because we don't always have full episode), set ``gae_lambda=1.0`` during initialization.
+
+        The TD(lambda) estimator has also two special cases:
+        - TD(1) is Monte-Carlo estimate (sum of discounted rewards)
+        - TD(0) is one-step estimate with bootstrapping (r_t + gamma * v(s_{t+1}))
+
+        For more information, see discussion in https://github.com/DLR-RM/stable-baselines3/pull/375.
+
+        :param last_values: state value estimation for the last step (one for each env)
+        :param dones: if the last step was a terminal step (one bool for each env).
+        """
+        # Convert to numpy
+        last_values = last_values.clone().cpu().numpy().flatten()  # type: ignore[assignment]
+
+        last_gae_lam = 0
+        for step in reversed(range(self.buffer_size)):
+            if step == self.buffer_size - 1:
+                next_non_terminal = 1.0 - dones.astype(np.float32)
+                next_values = last_values
+            else:
+                next_non_terminal = 1.0 - self.episode_starts[step + 1]
+                next_values = self.values[step + 1]
+            delta = self.rewards[step] + self.gamma * next_values * next_non_terminal - self.values[step]
+            last_gae_lam = delta + self.gamma * self.gae_lambda * next_non_terminal * last_gae_lam
+            self.advantages[step] = last_gae_lam
+        # TD(lambda) estimator, see Github PR #375 or "Telescoping in TD(lambda)"
+        # in David Silver Lecture 4: https://www.youtube.com/watch?v=PnHCvfgC_ZA
+        self.returns = self.advantages + self.values
+
+    def add(
+        self,
+        obs: np.ndarray,
+        action: np.ndarray,
+        reward: np.ndarray,
+        episode_start: np.ndarray,
+        value: th.Tensor,
+        log_prob: th.Tensor,
+    ) -> None:
+        """
+        :param obs: Observation
+        :param action: Action
+        :param reward:
+        :param episode_start: Start of episode signal.
+        :param value: estimated value of the current state
+            following the current policy.
+        :param log_prob: log probability of the action
+            following the current policy.
+        """
+        if len(log_prob.shape) == 0:
+            # Reshape 0-d tensor to avoid error
+            log_prob = log_prob.reshape(-1, 1)
+
+        # Reshape needed when using multiple envs with discrete observations
+        # as numpy cannot broadcast (n_discrete,) to (n_discrete, 1)
+        if isinstance(self.observation_space, spaces.Discrete):
+            obs = obs.reshape((self.n_envs, *self.obs_shape))
+
+        # Reshape to handle multi-dim and discrete action spaces, see GH #970 #1392
+        action = action.reshape((self.n_envs, self.action_dim))
+
+        self.observations[self.pos] = np.array(obs)
+        self.actions[self.pos] = np.array(action)
+        self.rewards[self.pos] = np.array(reward)
+        self.episode_starts[self.pos] = np.array(episode_start)
+        self.values[self.pos] = value.clone().cpu().numpy().flatten()
+        self.log_probs[self.pos] = log_prob.clone().cpu().numpy()
+        self.pos += 1
+        if self.pos == self.buffer_size:
+            self.full = True
+
+    def get(self, batch_size: int | None = None) -> Generator[RolloutBufferSamples]:
+        assert self.full, ""
+        indices = np.random.permutation(self.buffer_size * self.n_envs)
+        # Prepare the data
+        if not self.generator_ready:
+            _tensor_names = [
+                "observations",
+                "actions",
+                "values",
+                "log_probs",
+                "advantages",
+                "returns",
+            ]
+
+            for tensor in _tensor_names:
+                self.__dict__[tensor] = self.swap_and_flatten(self.__dict__[tensor])
+            self.generator_ready = True
+
+        # Return everything, don't create minibatches
+        if batch_size is None:
+            batch_size = self.buffer_size * self.n_envs
+
+        start_idx = 0
+        while start_idx < self.buffer_size * self.n_envs:
+            yield self._get_samples(indices[start_idx : start_idx + batch_size])
+            start_idx += batch_size
+
+    def _get_samples(
+        self,
+        batch_inds: np.ndarray,
+    ) -> RolloutBufferSamples:
+        data = (
+            self.observations[batch_inds],
+            self.actions[batch_inds],
+            self.values[batch_inds].flatten(),
+            self.log_probs[batch_inds].flatten(),
+            self.advantages[batch_inds].flatten(),
+            self.returns[batch_inds].flatten(),
+        )
+        return RolloutBufferSamples(*tuple(map(self.to_torch, data)))
diff --git a/cleanrl/cleanrl_utils/docker_build.py b/cleanrl/cleanrl_utils/docker_build.py
new file mode 100644
index 0000000000000000000000000000000000000000..35e20f59a430431ad931d073f0cb1c424d512dcf
--- /dev/null
+++ b/cleanrl/cleanrl_utils/docker_build.py
@@ -0,0 +1,12 @@
+import argparse
+import subprocess
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--tag", type=str, default="cleanrl:latest", help="the name of this experiment")
+args = parser.parse_args()
+
+subprocess.run(
+    f"docker build -t {args.tag} .",
+    shell=True,
+    check=True,
+)
diff --git a/cleanrl/cleanrl_utils/docker_queue.py b/cleanrl/cleanrl_utils/docker_queue.py
new file mode 100644
index 0000000000000000000000000000000000000000..3d75b8618559ff21ebcef30d733c03d121999e5e
--- /dev/null
+++ b/cleanrl/cleanrl_utils/docker_queue.py
@@ -0,0 +1,84 @@
+"""
+See https://github.com/docker/docker-py/issues/2395
+At the moment, nvidia-container-toolkit still includes nvidia-container-runtime. So, you can still add nvidia-container-runtime as a runtime in /etc/docker/daemon.json:
+
+{
+  "runtimes": {
+    "nvidia": {
+      "path": "nvidia-container-runtime",
+      "runtimeArgs": []
+    }
+  }
+}
+Then restart the docker service (sudo systemctl restart docker) and use runtime="nvidia" in docker-py as before.
+"""
+
+import argparse
+import shlex
+import time
+
+import docker
+
+parser = argparse.ArgumentParser(description="CleanRL Docker Submission")
+# Common arguments
+parser.add_argument("--exp-script", type=str, default="test1.sh", help="the file name of this experiment")
+# parser.add_argument('--cuda', type=lambda x:bool(strtobool(x)), default=True, nargs='?', const=True,
+#                     help='if toggled, cuda will not be enabled by default')
+parser.add_argument("--num-vcpus", type=int, default=16, help="total number of vcpus used in the host machine")
+parser.add_argument("--frequency", type=int, default=1, help="the number of seconds to check container update status")
+args = parser.parse_args()
+
+client = docker.from_env()
+
+# c = client.containers.run("ubuntu:latest", "echo hello world", detach=True)
+
+with open(args.exp_script) as f:
+    lines = f.readlines()
+
+tasks = []
+for line in lines:
+    line.replace("\n", "")
+    line_split = shlex.split(line)
+    for idx, item in enumerate(line_split):
+        if item == "-e":
+            break
+    env_vars = line_split[idx + 1 : idx + 2]
+    image = line_split[idx + 2]
+    commands = line_split[idx + 3 :]
+    tasks += [[image, env_vars, commands]]
+
+running_containers = []
+vcpus = list(range(args.num_vcpus))
+while len(tasks) != 0:
+    time.sleep(args.frequency)
+
+    # update running_containers
+    new_running_containers = []
+    for item in running_containers:
+        c = item[0]
+        c.reload()
+        if c.status != "exited":
+            new_running_containers += [item]
+        else:
+            print(f"✅ task on vcpu {item[1]} has finished")
+            vcpus += [item[1]]
+    running_containers = new_running_containers
+
+    if len(vcpus) != 0:
+        task = tasks.pop()
+        vcpu = vcpus.pop()
+        # if args.cuda:
+        #     c = client.containers.run(
+        #         image=task[0],
+        #         environment=task[1],
+        #         command=task[2],
+        #         runtime="nvidia",
+        #         cpuset_cpus=str(vcpu),
+        #         detach=True)
+        #     running_containers += [[c, vcpu]]
+        # else:
+        c = client.containers.run(image=task[0], environment=task[1], command=task[2], cpuset_cpus=str(vcpu), detach=True)
+        running_containers += [[c, vcpu]]
+        print("========================")
+        print(f"remaining tasks={len(tasks)}, running containers={len(running_containers)}")
+        print(f"running on vcpu {vcpu}", task)
diff --git a/cleanrl/cleanrl_utils/enjoy.py b/cleanrl/cleanrl_utils/enjoy.py
new file mode 100644
index 0000000000000000000000000000000000000000..afc8696691d9d11b9345c0ab4fb056c722c13d4e
--- /dev/null
+++ b/cleanrl/cleanrl_utils/enjoy.py
@@ -0,0 +1,43 @@
+import argparse
+
+from huggingface_hub import hf_hub_download
+
+from cleanrl_utils.evals import MODELS
+
+
+def parse_args():
+    # fmt: off
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--exp-name", type=str, default="dqn_atari",
+        help="the name of this experiment (e.g., ppo, dqn_atari)")
+    parser.add_argument("--seed", type=int, default=1,
+        help="seed of the experiment")
+    parser.add_argument("--hf-entity", type=str, default="cleanrl",
+        help="the user or org name of the model repository from the Hugging Face Hub")
+    parser.add_argument("--hf-repository", type=str, default="",
+        help="the huggingface repo (e.g., cleanrl/BreakoutNoFrameskip-v4-dqn_atari-seed1)")
+    parser.add_argument("--env-id", type=str, default="BreakoutNoFrameskip-v4",
+        help="the id of the environment")
+    parser.add_argument("--eval-episodes", type=int, default=10,
+        help="the number of evaluation episodes")
+    args = parser.parse_args()
+    # fmt: on
+    return args
+
+
+if __name__ == "__main__":
+    args = parse_args()
+    Model, make_env, evaluate = MODELS[args.exp_name]()
+    if not args.hf_repository:
+        args.hf_repository = f"{args.hf_entity}/{args.env_id}-{args.exp_name}-seed{args.seed}"
+    print(f"loading saved models from {args.hf_repository}...")
+    model_path = hf_hub_download(repo_id=args.hf_repository, filename=f"{args.exp_name}.cleanrl_model")
+    evaluate(
+        model_path,
+        make_env,
+        args.env_id,
+        eval_episodes=args.eval_episodes,
+        run_name=f"eval",
+        Model=Model,
+        capture_video=args.capture_video,
+    )
diff --git a/cleanrl/cleanrl_utils/evals/c51_eval.py b/cleanrl/cleanrl_utils/evals/c51_eval.py
new file mode 100644
index 0000000000000000000000000000000000000000..fc1e41e2256cdc54922c301bbf4ab9fb6a3df830
--- /dev/null
+++ b/cleanrl/cleanrl_utils/evals/c51_eval.py
@@ -0,0 +1,64 @@
+import random
+from argparse import Namespace
+from typing import Callable
+
+import gymnasium as gym
+import numpy as np
+import torch
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: torch.nn.Module,
+    device: torch.device = torch.device("cpu"),
+    epsilon: float = 0.05,
+    capture_video: bool = True,
+):
+    envs = gym.vector.SyncVectorEnv([make_env(env_id, 0, 0, capture_video, run_name)])
+    model_data = torch.load(model_path, map_location="cpu")
+    args = Namespace(**model_data["args"])
+    model = Model(envs, n_atoms=args.n_atoms, v_min=args.v_min, v_max=args.v_max)
+    model.load_state_dict(model_data["model_weights"])
+    model = model.to(device)
+    model.eval()
+
+    obs, _ = envs.reset()
+    episodic_returns = []
+    while len(episodic_returns) < eval_episodes:
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            actions, _ = model.get_action(torch.Tensor(obs).to(device))
+            actions = actions.cpu().numpy()
+        next_obs, _, _, _, infos = envs.step(actions)
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if "episode" not in info:
+                    continue
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={info['episode']['r']}")
+                episodic_returns += [info["episode"]["r"]]
+        obs = next_obs
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.c51 import QNetwork, make_env
+
+    model_path = hf_hub_download(repo_id="cleanrl/CartPole-v1-c51-seed1", filename="c51.cleanrl_model")
+    evaluate(
+        model_path,
+        make_env,
+        "CartPole-v1",
+        eval_episodes=10,
+        run_name=f"eval",
+        Model=QNetwork,
+        device="cpu",
+        capture_video=False,
+    )
diff --git a/cleanrl/cleanrl_utils/evals/c51_jax_eval.py b/cleanrl/cleanrl_utils/evals/c51_jax_eval.py
new file mode 100644
index 0000000000000000000000000000000000000000..e1cde8c75d4d5b91375816b30bc2cddac521dca3
--- /dev/null
+++ b/cleanrl/cleanrl_utils/evals/c51_jax_eval.py
@@ -0,0 +1,71 @@
+import random
+from argparse import Namespace
+from typing import Callable
+
+import flax
+import flax.linen as nn
+import gymnasium as gym
+import jax
+import jax.numpy as jnp
+import numpy as np
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: nn.Module,
+    epsilon: float = 0.05,
+    capture_video: bool = True,
+    seed=1,
+):
+    envs = gym.vector.SyncVectorEnv([make_env(env_id, 0, 0, capture_video, run_name)])
+    obs, _ = envs.reset()
+    model_data = None
+    with open(model_path, "rb") as f:
+        model_data = flax.serialization.from_bytes(model_data, f.read())
+    args = Namespace(**model_data["args"])
+    model = Model(action_dim=envs.single_action_space.n, n_atoms=args.n_atoms)
+    # q_key = jax.random.PRNGKey(seed)
+    params = model_data["model_weights"]
+    model.apply = jax.jit(model.apply)
+    atoms = jnp.asarray(np.linspace(args.v_min, args.v_max, num=args.n_atoms))
+
+    episodic_returns = []
+    while len(episodic_returns) < eval_episodes:
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            pmfs = model.apply(params, obs)
+            q_vals = (pmfs * atoms).sum(axis=-1)
+            actions = q_vals.argmax(axis=-1)
+            actions = jax.device_get(actions)
+        next_obs, _, _, _, infos = envs.step(actions)
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if "episode" not in info:
+                    continue
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={info['episode']['r']}")
+                episodic_returns += [info["episode"]["r"]]
+        obs = next_obs
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.c51_jax import QNetwork, make_env
+
+    model_path = hf_hub_download(repo_id="cleanrl/CartPole-v1-c51_jax-seed1", filename="c51_jax.cleanrl_model")
+    evaluate(
+        model_path,
+        make_env,
+        "CartPole-v1",
+        eval_episodes=10,
+        run_name=f"eval",
+        Model=QNetwork,
+        capture_video=False,
+    )
diff --git a/cleanrl/cleanrl_utils/evals/dqn_eval.py b/cleanrl/cleanrl_utils/evals/dqn_eval.py
new file mode 100644
index 0000000000000000000000000000000000000000..af637e39d9b51192f9fe3f519baea56c2af10b1e
--- /dev/null
+++ b/cleanrl/cleanrl_utils/evals/dqn_eval.py
@@ -0,0 +1,60 @@
+import random
+from typing import Callable
+
+import gymnasium as gym
+import numpy as np
+import torch
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: torch.nn.Module,
+    device: torch.device = torch.device("cpu"),
+    epsilon: float = 0.05,
+    capture_video: bool = True,
+):
+    envs = gym.vector.SyncVectorEnv([make_env(env_id, 0, 0, capture_video, run_name)])
+    model = Model(envs).to(device)
+    model.load_state_dict(torch.load(model_path, map_location=device))
+    model.eval()
+
+    obs, _ = envs.reset()
+    episodic_returns = []
+    while len(episodic_returns) < eval_episodes:
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = model(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+        next_obs, _, _, _, infos = envs.step(actions)
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if "episode" not in info:
+                    continue
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={info['episode']['r']}")
+                episodic_returns += [info["episode"]["r"]]
+        obs = next_obs
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.dqn import QNetwork, make_env
+
+    model_path = hf_hub_download(repo_id="cleanrl/CartPole-v1-dqn-seed1", filename="q_network.pth")
+    evaluate(
+        model_path,
+        make_env,
+        "CartPole-v1",
+        eval_episodes=10,
+        run_name=f"eval",
+        Model=QNetwork,
+        device="cpu",
+        capture_video=False,
+    )
diff --git a/cleanrl/cleanrl_utils/evals/dqn_jax_eval.py b/cleanrl/cleanrl_utils/evals/dqn_jax_eval.py
new file mode 100644
index 0000000000000000000000000000000000000000..baa858e4c9b23e7dda2e19973abe80ecdce4c293
--- /dev/null
+++ b/cleanrl/cleanrl_utils/evals/dqn_jax_eval.py
@@ -0,0 +1,65 @@
+import random
+from typing import Callable
+
+import flax
+import flax.linen as nn
+import gymnasium as gym
+import jax
+import numpy as np
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: nn.Module,
+    epsilon: float = 0.05,
+    capture_video: bool = True,
+    seed=1,
+):
+    envs = gym.vector.SyncVectorEnv([make_env(env_id, 0, 0, capture_video, run_name)])
+    obs, _ = envs.reset()
+    model = Model(action_dim=envs.single_action_space.n)
+    q_key = jax.random.PRNGKey(seed)
+    params = model.init(q_key, obs)
+    with open(model_path, "rb") as f:
+        params = flax.serialization.from_bytes(params, f.read())
+    model.apply = jax.jit(model.apply)
+
+    episodic_returns = []
+    while len(episodic_returns) < eval_episodes:
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = model.apply(params, obs)
+            actions = q_values.argmax(axis=-1)
+            actions = jax.device_get(actions)
+        next_obs, _, _, _, infos = envs.step(actions)
+        if "final_info" in infos:
+            for info in infos["final_info"]:
+                if "episode" not in info:
+                    continue
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={info['episode']['r']}")
+                episodic_returns += [info["episode"]["r"]]
+        obs = next_obs
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.dqn_jax import QNetwork, make_env
+
+    model_path = hf_hub_download(repo_id="vwxyzjn/CartPole-v1-dqn_jax-seed1", filename="dqn_jax.cleanrl_model")
+    evaluate(
+        model_path,
+        make_env,
+        "CartPole-v1",
+        eval_episodes=10,
+        run_name=f"eval",
+        Model=QNetwork,
+        capture_video=False,
+    )
diff --git a/cleanrl/cleanrl_utils/evals/ppo_envpool_jax_eval.py b/cleanrl/cleanrl_utils/evals/ppo_envpool_jax_eval.py
new file mode 100644
index 0000000000000000000000000000000000000000..7fe1b7de4ff4bd16dc5cb1a2be225093e6d9bc0c
--- /dev/null
+++ b/cleanrl/cleanrl_utils/evals/ppo_envpool_jax_eval.py
@@ -0,0 +1,104 @@
+import os
+from typing import Callable
+
+import cv2
+import flax
+import flax.linen as nn
+import jax
+import jax.numpy as jnp
+import numpy as np
+from moviepy.video.io.ImageSequenceClip import ImageSequenceClip
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: nn.Module,
+    capture_video: bool = True,
+    seed=1,
+):
+    envs = make_env(env_id, seed, num_envs=1)()
+    Network, Actor, Critic = Model
+    next_obs = envs.reset()
+    network = Network()
+    actor = Actor(action_dim=envs.single_action_space.n)
+    critic = Critic()
+    key = jax.random.PRNGKey(seed)
+    key, network_key, actor_key, critic_key = jax.random.split(key, 4)
+    network_params = network.init(network_key, np.array([envs.single_observation_space.sample()]))
+    actor_params = actor.init(actor_key, network.apply(network_params, np.array([envs.single_observation_space.sample()])))
+    critic_params = critic.init(critic_key, network.apply(network_params, np.array([envs.single_observation_space.sample()])))
+    # note: critic_params is not used in this script
+    with open(model_path, "rb") as f:
+        (args, (network_params, actor_params, critic_params)) = flax.serialization.from_bytes(
+            (None, (network_params, actor_params, critic_params)), f.read()
+        )
+
+    @jax.jit
+    def get_action_and_value(
+        network_params: flax.core.FrozenDict,
+        actor_params: flax.core.FrozenDict,
+        next_obs: np.ndarray,
+        key: jax.random.PRNGKey,
+    ):
+        hidden = network.apply(network_params, next_obs)
+        logits = actor.apply(actor_params, hidden)
+        # sample action: Gumbel-softmax trick
+        # see https://stats.stackexchange.com/questions/359442/sampling-from-a-categorical-distribution
+        key, subkey = jax.random.split(key)
+        u = jax.random.uniform(subkey, shape=logits.shape)
+        action = jnp.argmax(logits - jnp.log(-jnp.log(u)), axis=1)
+        return action, key
+
+    # a simple non-vectorized version
+
+    episodic_returns = []
+    for episode in range(eval_episodes):
+        episodic_return = 0
+        next_obs = envs.reset()
+        terminated = False
+
+        if capture_video:
+            recorded_frames = []
+            # conversion from grayscale into rgb
+            recorded_frames.append(cv2.cvtColor(next_obs[0][-1], cv2.COLOR_GRAY2RGB))
+        while not terminated:
+            actions, key = get_action_and_value(network_params, actor_params, next_obs, key)
+            next_obs, _, _, infos = envs.step(np.array(actions))
+            episodic_return += infos["reward"][0]
+            terminated = sum(infos["terminated"]) == 1
+
+            if capture_video and episode == 0:
+                recorded_frames.append(cv2.cvtColor(next_obs[0][-1], cv2.COLOR_GRAY2RGB))
+
+            if terminated:
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={episodic_return}")
+                episodic_returns.append(episodic_return)
+                if capture_video and episode == 0:
+                    clip = ImageSequenceClip(recorded_frames, fps=24)
+                    os.makedirs(f"videos/{run_name}", exist_ok=True)
+                    clip.write_videofile(f"videos/{run_name}/{episode}.mp4", logger="bar")
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.ppo_atari_envpool_xla_jax_scan import Actor, Critic, Network, make_env
+
+    model_path = hf_hub_download(
+        repo_id="vwxyzjn/Pong-v5-ppo_atari_envpool_xla_jax_scan-seed1", filename="ppo_atari_envpool_xla_jax_scan.cleanrl_model"
+    )
+    evaluate(
+        model_path,
+        make_env,
+        "Pong-v5",
+        eval_episodes=10,
+        run_name=f"eval",
+        Model=(Network, Actor, Critic),
+        capture_video=False,
+    )
diff --git a/cleanrl/cleanrl_utils/huggingface.py b/cleanrl/cleanrl_utils/huggingface.py
new file mode 100644
index 0000000000000000000000000000000000000000..f3fe83e3fb5c309a1cfcde9fa6c558a6fa884539
--- /dev/null
+++ b/cleanrl/cleanrl_utils/huggingface.py
@@ -0,0 +1,145 @@
+import argparse
+import sys
+from pathlib import Path
+from pprint import pformat
+from typing import List
+
+import numpy as np
+from tenacity import retry, stop_after_attempt, wait_fixed
+
+HUGGINGFACE_VIDEO_PREVIEW_FILE_NAME = "replay.mp4"
+HUGGINGFACE_README_FILE_NAME = "README.md"
+
+
+@retry(stop=stop_after_attempt(10), wait=wait_fixed(3))
+def push_to_hub(
+    args: argparse.Namespace,
+    episodic_returns: List,
+    repo_id: str,
+    algo_name: str,
+    folder_path: str,
+    video_folder_path: str = "",
+    revision: str = "main",
+    create_pr: bool = False,
+    private: bool = False,
+):
+    # Step 1: lazy import and create / read a huggingface repo
+    from huggingface_hub import CommitOperationAdd, CommitOperationDelete, HfApi
+    from huggingface_hub.repocard import metadata_eval_result, metadata_save
+
+    api = HfApi()
+    repo_url = api.create_repo(
+        repo_id=repo_id,
+        exist_ok=True,
+        private=private,
+    )
+    # parse the default entity
+    entity, repo = repo_url.split("/")[-2:]
+    repo_id = f"{entity}/{repo}"
+
+    # Step 2: clean up data
+    # delete previous tfevents and mp4 files
+    operations = [
+        CommitOperationDelete(path_in_repo=file)
+        for file in api.list_repo_files(repo_id=repo_id)
+        if ".tfevents" in file or file.endswith(".mp4")
+    ]
+
+    # Step 3: Generate the model card
+    algorithm_variant_filename = sys.argv[0].split("/")[-1]
+    model_card = f"""
+# (CleanRL) **{algo_name}** Agent Playing **{args.env_id}**
+
+This is a trained model of a {algo_name} agent playing {args.env_id}.
+The model was trained by using [CleanRL](https://github.com/vwxyzjn/cleanrl) and the most up-to-date training code can be
+found [here](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/{args.exp_name}.py).
+
+## Get Started
+
+To use this model, please install the `cleanrl` package with the following command:
+
+```
+pip install "cleanrl[{args.exp_name}]"
+python -m cleanrl_utils.enjoy --exp-name {args.exp_name} --env-id {args.env_id}
+```
+
+Please refer to the [documentation](https://docs.cleanrl.dev/get-started/zoo/) for more detail.
+
+
+## Command to reproduce the training
+
+```bash
+curl -OL https://huggingface.co/{repo_id}/raw/main/{algorithm_variant_filename}
+curl -OL https://huggingface.co/{repo_id}/raw/main/pyproject.toml
+curl -OL https://huggingface.co/{repo_id}/raw/main/poetry.lock
+poetry install --all-extras
+python {algorithm_variant_filename} {" ".join(sys.argv[1:])}
+```
+
+# Hyperparameters
+```python
+{pformat(vars(args))}
+```
+    """
+    readme_path = Path(folder_path) / HUGGINGFACE_README_FILE_NAME
+    readme = model_card
+
+    # metadata
+    metadata = {}
+    metadata["tags"] = [
+        args.env_id,
+        "deep-reinforcement-learning",
+        "reinforcement-learning",
+        "custom-implementation",
+    ]
+    metadata["library_name"] = "cleanrl"
+    eval = metadata_eval_result(
+        model_pretty_name=algo_name,
+        task_pretty_name="reinforcement-learning",
+        task_id="reinforcement-learning",
+        metrics_pretty_name="mean_reward",
+        metrics_id="mean_reward",
+        metrics_value=f"{np.average(episodic_returns):.2f} +/- {np.std(episodic_returns):.2f}",
+        dataset_pretty_name=args.env_id,
+        dataset_id=args.env_id,
+    )
+    metadata = {**metadata, **eval}
+
+    with open(readme_path, "w", encoding="utf-8") as f:
+        f.write(readme)
+    metadata_save(readme_path, metadata)
+
+    # fetch mp4 files
+    if video_folder_path:
+        # Push all video files
+        video_files = list(Path(video_folder_path).glob("*.mp4"))
+        operations += [CommitOperationAdd(path_or_fileobj=str(file), path_in_repo=str(file)) for file in video_files]
+        # Push latest one in root directory
+        latest_file = max(video_files, key=lambda file: int("".join(filter(str.isdigit, file.stem))))
+        operations.append(
+            CommitOperationAdd(path_or_fileobj=str(latest_file), path_in_repo=HUGGINGFACE_VIDEO_PREVIEW_FILE_NAME)
+        )
+
+    # fetch folder files
+    operations += [
+        CommitOperationAdd(path_or_fileobj=str(item), path_in_repo=str(item.relative_to(folder_path)))
+        for item in Path(folder_path).glob("*")
+    ]
+
+    # fetch source code
+    operations.append(CommitOperationAdd(path_or_fileobj=sys.argv[0], path_in_repo=sys.argv[0].split("/")[-1]))
+
+    # upload poetry files at the root of the repository
+    git_root = Path(__file__).parent.parent
+    operations.append(CommitOperationAdd(path_or_fileobj=str(git_root / "pyproject.toml"), path_in_repo="pyproject.toml"))
+    operations.append(CommitOperationAdd(path_or_fileobj=str(git_root / "poetry.lock"), path_in_repo="poetry.lock"))
+
+    api.create_commit(
+        repo_id=repo_id,
+        operations=operations,
+        commit_message="pushing model",
+        revision=revision,
+        create_pr=create_pr,
+    )
+    print(f"Model pushed to {repo_url}")
+    return repo_url
diff --git a/cleanrl/cleanrl_utils/paper_plot.py b/cleanrl/cleanrl_utils/paper_plot.py
new file mode 100644
index 0000000000000000000000000000000000000000..6acfae97b03da76b080bd19467d90fba08b5241a
--- /dev/null
+++ b/cleanrl/cleanrl_utils/paper_plot.py
@@ -0,0 +1,308 @@
+import argparse
+import os
+import pickle
+from os import path
+
+import matplotlib as mpl
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import wandb
+
+sns.set_style("whitegrid")
+mpl.rcParams["text.usetex"] = True
+mpl.rcParams["text.latex.preamble"] = r"\usepackage{amsmath}"  # for \text command
+
+parser = argparse.ArgumentParser(description="CleanRL Plots")
+# Common arguments
+parser.add_argument(
+    "--wandb-project", type=str, default="cleanrl/cleanrl.benchmark", help="the name of wandb project (e.g. cleanrl/cleanrl)"
+)
+parser.add_argument(
+    "--feature-of-interest", type=str, default="charts/episodic_return", help="which feature to be plotted on the y-axis"
+)
+parser.add_argument("--hyper-params-tuned", nargs="+", default=[], help="the hyper parameters tuned")
+# parser.add_argument('--scan-history', type=lambda x:bool(strtobool(x)), default=False, nargs='?', const=True,
+#                     help='if toggled, cuda will not be enabled by default')
+parser.add_argument("--interested-exp-names", nargs="+", default=[], help="the hyper parameters tuned")
+parser.add_argument("--samples", type=int, default=500, help="the sampled point of the run")
+parser.add_argument("--smooth-weight", type=float, default=0.95, help="the weight parameter of the exponential moving average")
+parser.add_argument(
+    "--last-n-episodes",
+    type=int,
+    default=10,
+    help="for analysis only; the last n episodes from which the mean of the feature of interest is calculated",
+)
+parser.add_argument("--num-points-x-axis", type=int, default=500, help="the number of points in the x-axis")
+parser.add_argument("--font-size", type=int, default=18, help="the font size of the plots")
+parser.add_argument("--x-label", type=str, default="Time Steps", help="the label of x-axis")
+parser.add_argument("--y-label", type=str, default="Episodic Return", help="the label of y-axis")
+parser.add_argument("--y-lim-bottom", type=float, default=0.0, help="the bottom limit for the y-axis")
+parser.add_argument("--output-format", type=str, default="pdf", help="either `pdf`, `png`, or `svg`")
+args = parser.parse_args()
+api = wandb.Api()
+
+# hacks
+env_dict = {
+    # 'MicrortsAttackShapedReward-v1': 'MicrortsAttackHRL-v1',
+    # 'MicrortsProduceCombatUnitsShapedReward-v1':  'MicrortsProduceCombatUnitHRL-v1',
+    # 'MicrortsRandomEnemyShapedReward3-v1': 'MicrortsRandomEnemyHRL3-v1',
+}
+exp_convert_dict = {
+    "ppo_atari_visual": "PPO",
+    "dqn_atari_visual": "DQN",
+    "apex_dqn_atari_visual": "Ape-X DQN",
+    "c51_atari_visual": "C51",
+    # 'rnd_ppo_gamma_0.999_nocliploss_lr_1e-4_128envs_entcoef_0.001_stickyaction': "PPO RND",
+    "ddpg_continuous_action": "DDPG",
+    # 'dqn': 'DQN',
+    # 'ppg_procgen_fast': 'PPG',
+    # 'ppg_procgen_impala_cnn': 'PPG-IMPALA-CNN',
+    # 'ppo': "PPO",
+    # 'ppo_car_racing': "PPO",
+    "ppo_continuous_action": "PPO",
+    # 'ppo_procgen_fast': "PPO",
+    # "ppo_procgen_impala_cnn": "PPO-IMPALA-CNN",
+    "td3_continuous_action": "TD3",
+}
+
+# args.feature_of_interest = 'charts/episodic_return'
+feature_name = args.feature_of_interest.replace("/", "_")
+if not os.path.exists(feature_name):
+    os.makedirs(feature_name)
+    with open(f"{feature_name}/cache.pkl", "wb") as handle:
+        pickle.dump([[], [], [], {}, [], set()], handle, protocol=pickle.HIGHEST_PROTOCOL)
+with open(f"{feature_name}/cache.pkl", "rb") as handle:
+    summary_list, config_list, name_list, envs, exp_names, ids = pickle.load(handle)
+
+# Change oreilly-class/cifar to <entity/project-name>
+runs = api.runs(args.wandb_project)
+data = []
+for idx, run in enumerate(runs):
+    if run.id not in ids:
+        ids.add(run.id)
+        if args.feature_of_interest in run.summary:
+            metrics_dataframe = run.history(keys=[args.feature_of_interest, "global_step"], samples=args.samples)
+            exp_name = run.config["exp_name"]
+            for param in args.hyper_params_tuned:
+                if param in run.config:
+                    exp_name += "-" + param + "-" + str(run.config[param]) + "-"
+
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "algo", exp_name)
+            exp_names += [exp_name]
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "seed", run.config["seed"])
+
+            data += [metrics_dataframe]
+            if run.config["env_id"] not in envs:
+                envs[run.config["env_id"]] = [metrics_dataframe]
+                envs[run.config["env_id"] + "total_timesteps"] = run.config["total_timesteps"]
+            else:
+                envs[run.config["env_id"]] += [metrics_dataframe]
+
+            # run.summary are the output key/values like accuracy.  We call ._json_dict to omit large files
+            summary_list.append(run.summary._json_dict)
+
+            # run.config is the input metrics.  We remove special values that start with _.
+            config_list.append({k: v for k, v in run.config.items() if not k.startswith("_")})
+
+            # run.name is the name of the run.
+            name_list.append(run.name)
+
+
+summary_df = pd.DataFrame.from_records(summary_list)
+config_df = pd.DataFrame.from_records(config_list)
+name_df = pd.DataFrame({"name": name_list})
+all_df = pd.concat([name_df, config_df, summary_df], axis=1)
+# data = pd.concat(data, ignore_index=True)
+with open(f"{feature_name}/cache.pkl", "wb") as handle:
+    pickle.dump([summary_list, config_list, name_list, envs, exp_names, ids], handle, protocol=pickle.HIGHEST_PROTOCOL)
+print("data loaded")
+
+
+# https://stackoverflow.com/questions/42281844/what-is-the-mathematics-behind-the-smoothing-parameter-in-tensorboards-scalar#_=_
+def smooth(scalars, weight):  # Weight between 0 and 1
+    last = scalars[0]  # First value in the plot (first timestep)
+    smoothed = list()
+    for point in scalars:
+        smoothed_val = last * weight + (1 - weight) * point  # Calculate smoothed value
+        smoothed.append(smoothed_val)  # Save it
+        last = smoothed_val  # Anchor the last smoothed value
+
+    return smoothed
+
+
+# smoothing
+for env in envs:
+    if not env.endswith("total_timesteps"):
+        for idx, metrics_dataframe in enumerate(envs[env]):
+            envs[env][idx] = metrics_dataframe.dropna(subset=[args.feature_of_interest])
+#             envs[env][idx][args.feature_of_interest] = smooth(metrics_dataframe[args.feature_of_interest], 0.85)
+
+sns.set(style="darkgrid")
+
+
+def get_df_for_env(env_id):
+    env_total_timesteps = envs[env_id + "total_timesteps"]
+    env_increment = env_total_timesteps / 500
+    envs_same_x_axis = []
+    for sampled_run in envs[env_id]:
+        df = pd.DataFrame(columns=sampled_run.columns)
+        x_axis = [i * env_increment for i in range(500 - 2)]
+        current_row = 0
+        for timestep in x_axis:
+            while sampled_run.iloc[current_row]["global_step"] < timestep:
+                current_row += 1
+                if current_row > len(sampled_run) - 2:
+                    break
+            if current_row > len(sampled_run) - 2:
+                break
+            temp_row = sampled_run.iloc[current_row].copy()
+            temp_row["global_step"] = timestep
+            df = df.append(temp_row)
+
+        envs_same_x_axis += [df]
+    return pd.concat(envs_same_x_axis, ignore_index=True)
+
+
+def export_legend(ax, filename="legend.pdf"):
+    try:
+        # import matplotlib as mpl
+        # mpl.rcParams['text.usetex'] = True
+        # mpl.rcParams['text.latex.preamble'] = [r'\usepackage{amsmath}'] #for \text command
+        fig2 = plt.figure()
+        ax2 = fig2.add_subplot()
+        ax2.axis("off")
+        handles, labels = ax.get_legend_handles_labels()
+
+        legend = ax2.legend(
+            handles=handles, labels=labels, frameon=False, loc="lower center", ncol=6, fontsize=20, handlelength=1
+        )
+        for text in legend.get_texts():
+            if text.get_text() in exp_convert_dict:
+                text.set_text(exp_convert_dict[text.get_text()])
+            text.set_text(text.get_text().replace("_", "-"))
+        for line in legend.get_lines():
+            line.set_linewidth(4.0)
+        fig = legend.figure
+        fig.canvas.draw()
+
+        bbox = legend.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
+        fig.savefig(filename, dpi="figure", bbox_inches=bbox)
+        fig.clf()
+    except:
+        print(f"export legend failed: {filename}")
+
+
+if not os.path.exists(f"{feature_name}/data"):
+    os.makedirs(f"{feature_name}/data")
+if not os.path.exists(f"{feature_name}/plots"):
+    os.makedirs(f"{feature_name}/plots")
+if not os.path.exists(f"{feature_name}/legends"):
+    os.makedirs(f"{feature_name}/legends")
+
+
+interested_exp_names = sorted(list(exp_convert_dict.keys()))  # ['ppo_continuous_action', 'ppo_atari_visual']
+palette = sns.color_palette(n_colors=len(set(exp_convert_dict.values())))
+palette_dict = dict(zip(set(exp_convert_dict.values()), palette))
+current_palette_dict = dict(zip(interested_exp_names, [palette_dict[exp_convert_dict[k]] for k in interested_exp_names]))
+if args.interested_exp_names:
+    interested_exp_names = args.interested_exp_names
+print(interested_exp_names)
+# raise
+# print(current_palette_dict)
+
+legend_df = pd.DataFrame()
+
+# hack
+algos_in_legend = []
+
+if args.font_size:
+    plt.rc("axes", titlesize=args.font_size)  # fontsize of the axes title
+    plt.rc("axes", labelsize=args.font_size)  # fontsize of the x and y labels
+    plt.rc("xtick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("ytick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("legend", fontsize=args.font_size)  # legend fontsize
+
+stats = {item: [] for item in ["env_id", "exp_name", args.feature_of_interest]}
+# uncommenet the following to generate all figures
+for env in set(all_df["env_id"]):
+    if not path.exists(f"{feature_name}/data/{env}.pkl"):
+        with open(f"{feature_name}/data/{env}.pkl", "wb") as handle:
+            data = get_df_for_env(env)
+            data["seed"] = data["seed"].astype(float)
+            data[args.feature_of_interest] = data[args.feature_of_interest].astype(float)
+            pickle.dump(data, handle, protocol=pickle.HIGHEST_PROTOCOL)
+    else:
+        with open(f"{feature_name}/data/{env}.pkl", "rb") as handle:
+            data = pickle.load(handle)
+            print(f"{env}'s data loaded")
+
+    def _smooth(df):
+        df[args.feature_of_interest] = smooth(list(df[args.feature_of_interest]), args.smooth_weight)
+        return df
+
+    plot_data = data.groupby(["seed", "algo"]).apply(_smooth).loc[data["algo"].isin(interested_exp_names)]
+    if len(plot_data) == 0:
+        continue
+    ax = sns.lineplot(
+        data=plot_data, x="global_step", y=args.feature_of_interest, hue="algo", ci="sd", palette=current_palette_dict
+    )
+    ax.ticklabel_format(style="sci", scilimits=(0, 0), axis="x")
+    ax.set(xlabel=args.x_label, ylabel=args.y_label)
+    ax.legend().remove()
+    if args.y_lim_bottom:
+        plt.ylim(bottom=args.y_lim_bottom)
+    plt.title(env)
+    plt.tight_layout()
+    plt.savefig(f"{feature_name}/plots/{env}.{args.output_format}")
+    plt.clf()
+
+    env_algos = data["algo"].unique()
+    for algo in env_algos:
+        algo_data = data.loc[data["algo"].isin([algo])]
+        last_n_episodes_global_step = sorted(algo_data["global_step"].unique())[-args.last_n_episodes]
+        last_n_episodes_features = (
+            algo_data[algo_data["global_step"] > last_n_episodes_global_step]
+            .groupby(["seed"])
+            .mean()[args.feature_of_interest]
+        )
+
+        for item in last_n_episodes_features:
+            stats[args.feature_of_interest] += [item]
+            if algo in exp_convert_dict:
+                stats["exp_name"] += [exp_convert_dict[algo]]
+            else:
+                stats["exp_name"] += [algo]
+            stats["env_id"] += [env]
+
+    # export legend
+    # legend_df = pd.DataFrame()
+    # legend_df = legend_df.append(plot_data)
+    # legend_df = legend_df.reset_index()
+    # ax = sns.lineplot(data=legend_df, x="global_step", y=args.feature_of_interest, hue="algo", ci='sd', palette=current_palette_dict)
+    # ax.set(xlabel=args.x_label, ylabel=args.y_label)
+    # ax.legend().remove()
+    # export_legend(ax, f"{feature_name}/legends/{env}.{args.output_format}")
+    # plt.clf()
+
+    # hack
+    algo_in_legend = exp_convert_dict[plot_data["algo"].iloc[0]]
+    if algo_in_legend not in algos_in_legend:
+        legend_df = legend_df.append(plot_data.iloc[:5])
+        algos_in_legend += [algo_in_legend]
+
+legend_df = legend_df.reset_index()
+ax = sns.lineplot(
+    data=legend_df, x="global_step", y=args.feature_of_interest, hue="algo", ci="sd", palette=current_palette_dict
+)
+ax.set(xlabel=args.x_label, ylabel=args.y_label)
+ax.legend().remove()
+export_legend(ax, f"{feature_name}/legend.{args.output_format}")
+plt.clf()
+
+
+# analysis
+stats_df = pd.DataFrame(stats)
+g = stats_df.groupby(["env_id", "exp_name"]).agg(lambda x: f"{np.mean(x):.2f} ± {np.std(x):.2f}")
+print(g.reset_index().pivot("exp_name", "env_id", args.feature_of_interest).to_latex().replace("±", r"$\pm$"))
diff --git a/cleanrl/cleanrl_utils/plot.py b/cleanrl/cleanrl_utils/plot.py
new file mode 100644
index 0000000000000000000000000000000000000000..0ebcb150b4df2e07819d940d100f9581ef15452d
--- /dev/null
+++ b/cleanrl/cleanrl_utils/plot.py
@@ -0,0 +1,290 @@
+import argparse
+import os
+import pickle
+from os import path
+
+import matplotlib as mpl
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import wandb
+
+mpl.rcParams["text.usetex"] = True
+mpl.rcParams["text.latex.preamble"] = r"\usepackage{amsmath}"  # for \text command
+
+
+parser = argparse.ArgumentParser(description="CleanRL Plots")
+# Common arguments
+parser.add_argument(
+    "--wandb-project", type=str, default="cleanrl/cleanrl.benchmark", help="the name of wandb project (e.g. cleanrl/cleanrl)"
+)
+parser.add_argument(
+    "--feature-of-interest", type=str, default="charts/episodic_return", help="which feature to be plotted on the y-axis"
+)
+parser.add_argument("--hyper-params-tuned", nargs="+", default=[], help="the hyper parameters tuned")
+# parser.add_argument('--scan-history', type=lambda x:bool(strtobool(x)), default=False, nargs='?', const=True,
+#                     help='if toggled, cuda will not be enabled by default')
+parser.add_argument("--interested-exp-names", nargs="+", default=[], help="the hyper parameters tuned")
+parser.add_argument("--samples", type=int, default=500, help="the sampled point of the run")
+parser.add_argument("--smooth-weight", type=float, default=0.90, help="the weight parameter of the exponential moving average")
+parser.add_argument(
+    "--last-n-episodes",
+    type=int,
+    default=10,
+    help="for analysis only; the last n episodes from which the mean of the feature of interest is calculated",
+)
+parser.add_argument("--num-points-x-axis", type=int, default=500, help="the number of points in the x-axis")
+parser.add_argument("--font-size", type=int, default=18, help="the font size of the plots")
+parser.add_argument("--x-label", type=str, default="Time Steps", help="the label of x-axis")
+parser.add_argument("--y-label", type=str, default="Episodic Return", help="the label of y-axis")
+parser.add_argument("--y-lim-bottom", type=float, default=0.0, help="the bottom limit for the y-axis")
+parser.add_argument("--output-format", type=str, default="pdf", help="either `pdf`, `png`, or `svg`")
+args = parser.parse_args()
+api = wandb.Api()
+
+# hacks
+env_dict = {
+    # 'MicrortsAttackShapedReward-v1': 'MicrortsAttackHRL-v1',
+    # 'MicrortsProduceCombatUnitsShapedReward-v1':  'MicrortsProduceCombatUnitHRL-v1',
+    # 'MicrortsRandomEnemyShapedReward3-v1': 'MicrortsRandomEnemyHRL3-v1',
+}
+exp_convert_dict = {
+    "ppo_atari_visual": "PPO",
+    "dqn_atari_visual": "DQN",
+    "apex_dqn_atari_visual": "Ape-X DQN",
+    "c51_atari_visual": "C51",
+    # 'ppo_no_mask-0': 'Invalid action penalty, $r_{\\text{invalid}}=0$',
+    # 'ppo_no_mask--0.1': 'Invalid action penalty, $r_{\\text{invalid}}=-0.1$',
+    # 'ppo_no_mask--0.01': 'Invalid action penalty, $r_{\\text{invalid}}=-0.01$',
+    # 'ppo_no_mask--1': 'Invalid action penalty, $r_{\\text{invalid}}=-1$',
+    # 'ppo-maskrm': 'Masking removed',
+    # 'ppo_no_adj': 'Naive invalid action masking',
+}
+
+# args.feature_of_interest = 'charts/episodic_return'
+feature_name = args.feature_of_interest.replace("/", "_")
+if not os.path.exists(feature_name):
+    os.makedirs(feature_name)
+    with open(f"{feature_name}/cache.pkl", "wb") as handle:
+        pickle.dump([[], [], [], {}, [], set()], handle, protocol=pickle.HIGHEST_PROTOCOL)
+with open(f"{feature_name}/cache.pkl", "rb") as handle:
+    summary_list, config_list, name_list, envs, exp_names, ids = pickle.load(handle)
+
+# Change oreilly-class/cifar to <entity/project-name>
+runs = api.runs(args.wandb_project)
+data = []
+for idx, run in enumerate(runs):
+    if run.id not in ids:
+        ids.add(run.id)
+        if args.feature_of_interest in run.summary:
+            metrics_dataframe = run.history(keys=[args.feature_of_interest, "global_step"], samples=args.samples)
+            exp_name = run.config["exp_name"]
+            for param in args.hyper_params_tuned:
+                if param in run.config:
+                    exp_name += "-" + param + "-" + str(run.config[param]) + "-"
+
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "algo", exp_name)
+            exp_names += [exp_name]
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "seed", run.config["seed"])
+
+            data += [metrics_dataframe]
+            if run.config["env_id"] not in envs:
+                envs[run.config["env_id"]] = [metrics_dataframe]
+                envs[run.config["env_id"] + "total_timesteps"] = run.config["total_timesteps"]
+            else:
+                envs[run.config["env_id"]] += [metrics_dataframe]
+
+            # run.summary are the output key/values like accuracy.  We call ._json_dict to omit large files
+            summary_list.append(run.summary._json_dict)
+
+            # run.config is the input metrics.  We remove special values that start with _.
+            config_list.append({k: v for k, v in run.config.items() if not k.startswith("_")})
+
+            # run.name is the name of the run.
+            name_list.append(run.name)
+
+
+summary_df = pd.DataFrame.from_records(summary_list)
+config_df = pd.DataFrame.from_records(config_list)
+name_df = pd.DataFrame({"name": name_list})
+all_df = pd.concat([name_df, config_df, summary_df], axis=1)
+# data = pd.concat(data, ignore_index=True)
+with open(f"{feature_name}/cache.pkl", "wb") as handle:
+    pickle.dump([summary_list, config_list, name_list, envs, exp_names, ids], handle, protocol=pickle.HIGHEST_PROTOCOL)
+print("data loaded")
+
+
+# https://stackoverflow.com/questions/42281844/what-is-the-mathematics-behind-the-smoothing-parameter-in-tensorboards-scalar#_=_
+def smooth(scalars, weight):  # Weight between 0 and 1
+    last = scalars[0]  # First value in the plot (first timestep)
+    smoothed = list()
+    for point in scalars:
+        smoothed_val = last * weight + (1 - weight) * point  # Calculate smoothed value
+        smoothed.append(smoothed_val)  # Save it
+        last = smoothed_val  # Anchor the last smoothed value
+
+    return smoothed
+
+
+# smoothing
+for env in envs:
+    if not env.endswith("total_timesteps"):
+        for idx, metrics_dataframe in enumerate(envs[env]):
+            envs[env][idx] = metrics_dataframe.dropna(subset=[args.feature_of_interest])
+#             envs[env][idx][args.feature_of_interest] = smooth(metrics_dataframe[args.feature_of_interest], 0.85)
+
+sns.set(style="darkgrid")
+
+
+def get_df_for_env(env_id):
+    env_total_timesteps = envs[env_id + "total_timesteps"]
+    env_increment = env_total_timesteps / 500
+    envs_same_x_axis = []
+    for sampled_run in envs[env_id]:
+        df = pd.DataFrame(columns=sampled_run.columns)
+        x_axis = [i * env_increment for i in range(500 - 2)]
+        current_row = 0
+        for timestep in x_axis:
+            while sampled_run.iloc[current_row]["global_step"] < timestep:
+                current_row += 1
+                if current_row > len(sampled_run) - 2:
+                    break
+            if current_row > len(sampled_run) - 2:
+                break
+            temp_row = sampled_run.iloc[current_row].copy()
+            temp_row["global_step"] = timestep
+            df = df.append(temp_row)
+
+        envs_same_x_axis += [df]
+    return pd.concat(envs_same_x_axis, ignore_index=True)
+
+
+def export_legend(ax, filename="legend.pdf"):
+    try:
+        # import matplotlib as mpl
+        # mpl.rcParams['text.usetex'] = True
+        # mpl.rcParams['text.latex.preamble'] = [r'\usepackage{amsmath}'] #for \text command
+        fig2 = plt.figure()
+        ax2 = fig2.add_subplot()
+        ax2.axis("off")
+        handles, labels = ax.get_legend_handles_labels()
+
+        legend = ax2.legend(
+            handles=handles, labels=labels, frameon=False, loc="lower center", ncol=4, fontsize=20, handlelength=1
+        )
+        for text in legend.get_texts():
+            if text.get_text() in exp_convert_dict:
+                text.set_text(exp_convert_dict[text.get_text()])
+            text.set_text(text.get_text().replace("_", "-"))
+        for line in legend.get_lines():
+            line.set_linewidth(4.0)
+        fig = legend.figure
+        fig.canvas.draw()
+
+        bbox = legend.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
+        fig.savefig(filename, dpi="figure", bbox_inches=bbox)
+        fig.clf()
+    except:
+        print(f"export legend failed: {filename}")
+
+
+if not os.path.exists(f"{feature_name}/data"):
+    os.makedirs(f"{feature_name}/data")
+if not os.path.exists(f"{feature_name}/plots"):
+    os.makedirs(f"{feature_name}/plots")
+if not os.path.exists(f"{feature_name}/legends"):
+    os.makedirs(f"{feature_name}/legends")
+
+
+interested_exp_names = sorted(list(set(exp_names)))  # ['ppo_continuous_action', 'ppo_atari_visual']
+current_palette = sns.color_palette(n_colors=len(interested_exp_names))
+current_palette_dict = dict(zip(interested_exp_names, current_palette))
+if args.interested_exp_names:
+    interested_exp_names = args.interested_exp_names
+print(interested_exp_names)
+# print(current_palette_dict)
+legend_df = pd.DataFrame()
+
+if args.font_size:
+    plt.rc("axes", titlesize=args.font_size)  # fontsize of the axes title
+    plt.rc("axes", labelsize=args.font_size)  # fontsize of the x and y labels
+    plt.rc("xtick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("ytick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("legend", fontsize=args.font_size)  # legend fontsize
+
+stats = {item: [] for item in ["env_id", "exp_name", args.feature_of_interest]}
+# uncommenet the following to generate all figures
+for env in set(all_df["env_id"]):
+    if not path.exists(f"{feature_name}/data/{env}.pkl"):
+        with open(f"{feature_name}/data/{env}.pkl", "wb") as handle:
+            data = get_df_for_env(env)
+            data["seed"] = data["seed"].astype(float)
+            data[args.feature_of_interest] = data[args.feature_of_interest].astype(float)
+            pickle.dump(data, handle, protocol=pickle.HIGHEST_PROTOCOL)
+    else:
+        with open(f"{feature_name}/data/{env}.pkl", "rb") as handle:
+            data = pickle.load(handle)
+            print(f"{env}'s data loaded")
+
+    def _smooth(df):
+        df[args.feature_of_interest] = smooth(list(df[args.feature_of_interest]), args.smooth_weight)
+        return df
+
+    plot_data = data.groupby(["seed", "algo"]).apply(_smooth).loc[data["algo"].isin(interested_exp_names)]
+    ax = sns.lineplot(
+        data=plot_data,
+        x="global_step",
+        y=args.feature_of_interest,
+        hue="algo",
+        ci="sd",
+    )  # , palette=current_palette_dict
+    ax.ticklabel_format(style="sci", scilimits=(0, 0), axis="x")
+    ax.set(xlabel=args.x_label, ylabel=args.y_label)
+    ax.legend().remove()
+    if args.y_lim_bottom:
+        plt.ylim(bottom=args.y_lim_bottom)
+    # plt.title(env)
+    plt.tight_layout()
+    plt.savefig(f"{feature_name}/plots/{env}.{args.output_format}")
+    plt.clf()
+
+    # export legend
+    legend_df = pd.DataFrame()
+    legend_df = legend_df.append(plot_data)
+    legend_df = legend_df.reset_index()
+    ax = sns.lineplot(
+        data=legend_df,
+        x="global_step",
+        y=args.feature_of_interest,
+        hue="algo",
+        ci="sd",
+    )  # , palette=current_palette_dict
+    ax.set(xlabel="Time Steps", ylabel="Average Episode Reward")
+    ax.legend().remove()
+    export_legend(ax, f"{feature_name}/legends/{env}.{args.output_format}")
+    plt.clf()
+
+    env_algos = data["algo"].unique()
+    for algo in env_algos:
+        algo_data = data.loc[data["algo"].isin([algo])]
+        last_n_episodes_global_step = sorted(algo_data["global_step"].unique())[-args.last_n_episodes]
+        last_n_episodes_features = (
+            algo_data[algo_data["global_step"] > last_n_episodes_global_step]
+            .groupby(["seed"])
+            .mean()[args.feature_of_interest]
+        )
+
+        for item in last_n_episodes_features:
+            stats[args.feature_of_interest] += [item]
+            if algo in exp_convert_dict:
+                stats["exp_name"] += [exp_convert_dict[algo]]
+            else:
+                stats["exp_name"] += [algo]
+            stats["env_id"] += [env]
+
+
+# analysis
+stats_df = pd.DataFrame(stats)
+g = stats_df.groupby(["env_id", "exp_name"]).agg(lambda x: f"{np.mean(x):.2f} ± {np.std(x):.2f}")
+print(g.reset_index().pivot("exp_name", "env_id", args.feature_of_interest).to_latex().replace("±", r"$\pm$"))
diff --git a/cleanrl/cleanrl_utils/plot_individual.py b/cleanrl/cleanrl_utils/plot_individual.py
new file mode 100644
index 0000000000000000000000000000000000000000..2808d386c408af36972d2c27300d603ef0d9f621
--- /dev/null
+++ b/cleanrl/cleanrl_utils/plot_individual.py
@@ -0,0 +1,329 @@
+import argparse
+import os
+import pickle
+from os import path
+
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import wandb
+
+parser = argparse.ArgumentParser(description="CleanRL Plots")
+# Common arguments
+parser.add_argument(
+    "--wandb-project",
+    type=str,
+    default="anonymous-rl-code/action-guidance",
+    help="the name of wandb project (e.g. cleanrl/cleanrl)",
+)
+parser.add_argument(
+    "--feature-of-interest",
+    type=str,
+    default="charts/episodic_return/ProduceCombatUnitRewardFunction",
+    help="which feature to be plotted on the y-axis",
+)
+parser.add_argument("--hyper-params-tuned", nargs="+", default=["shift", "adaptation"], help="the hyper parameters tuned")
+# parser.add_argument('--scan-history', type=lambda x:bool(strtobool(x)), default=False, nargs='?', const=True,
+#                     help='if toggled, cuda will not be enabled by default')
+parser.add_argument(
+    "--interested-exp-names",
+    nargs="+",
+    default=[
+        "ppo_ac_positive_reward-shift-800000--adaptation-1000000--positive_likelihood-1-",
+        "ppo_ac_positive_reward-shift-800000--adaptation-1000000--positive_likelihood-0-",
+    ],
+    help="the hyper parameters tuned",
+)
+parser.add_argument("--samples", type=int, default=500, help="the sampled point of the run")
+parser.add_argument("--smooth-weight", type=float, default=0.90, help="the weight parameter of the exponential moving average")
+parser.add_argument(
+    "--last-n-episodes",
+    type=int,
+    default=50,
+    help="for analysis only; the last n episodes from which the mean of the feature of interest is calculated",
+)
+parser.add_argument("--num-points-x-axis", type=int, default=500, help="the number of points in the x-axis")
+parser.add_argument("--font-size", type=int, default=13, help="the font size of the plots")
+parser.add_argument("--x-label", type=str, default="Time Steps", help="the label of x-axis")
+parser.add_argument("--y-label", type=str, default="Average Episode Reward", help="the label of y-axis")
+parser.add_argument("--y-lim-bottom", type=float, default=0.0, help="the bottom limit for the y-axis")
+parser.add_argument("--output-format", type=str, default="pdf", help="either `pdf`, `png`, or `svg`")
+parser.add_argument("--seed", type=int, default=6, help="seed of color palette shuffle")
+args = parser.parse_args()
+api = wandb.Api()
+np.random.seed(args.seed)
+
+# hacks
+env_dict = {
+    "MicrortsAttackShapedReward-v1": "MicrortsAttackHRL-v1",
+    "MicrortsProduceCombatUnitsShapedReward-v1": "MicrortsProduceCombatUnitHRL-v1",
+    "MicrortsRandomEnemyShapedReward3-v1": "MicrortsRandomEnemyHRL3-v1",
+}
+exp_convert_dict = {
+    "ppo_positive_reward-positive_likelihood-0-": "sparse reward - no PLO",
+    "ppo": "sparse reward",
+    "ppo_ac_positive_reward-shift-2000000--adaptation-2000000--positive_likelihood-0-": "action guidance - multi-agent  w/ PLO",
+    "ppo_ac_positive_reward-shift-2000000--adaptation-7000000--positive_likelihood-0-": "action guidance - long adaptation w/ PLO",
+    "ppo_ac_positive_reward-shift-800000--adaptation-1000000--positive_likelihood-0-": "action guidance - short adaptation w/ PLO",
+    "ppo_ac_positive_reward-shift-2000000--adaptation-7000000--positive_likelihood-1-": "action guidance - long adaptation",
+    "ppo_ac_positive_reward-shift-800000--adaptation-1000000--positive_likelihood-1-": "action guidance - short adaptation",
+    "pposhaped": "shaped reward",
+}
+
+# args.feature_of_interest = 'charts/episodic_return'
+feature_name = args.feature_of_interest.replace("/", "_")
+if not os.path.exists(feature_name):
+    os.makedirs(feature_name)
+
+if not path.exists(f"{feature_name}/all_df_cache.pkl"):
+    # Change oreilly-class/cifar to <entity/project-name>
+    runs = api.runs(args.wandb_project)
+    summary_list = []
+    config_list = []
+    name_list = []
+    envs = {}
+    data = []
+    exp_names = []
+
+    for idx, run in enumerate(runs):
+        if args.feature_of_interest in run.summary:
+            # if args.scan_history:
+            #     ls =
+            # else:
+            ls = run.history(keys=[args.feature_of_interest, "global_step"], pandas=False, samples=args.samples)
+            metrics_dataframe = pd.DataFrame(ls[0])
+            exp_name = run.config["exp_name"]
+            for param in args.hyper_params_tuned:
+                if param in run.config:
+                    exp_name += "-" + param + "-" + str(run.config[param]) + "-"
+
+            # hacks
+            if run.config["env_id"] in env_dict:
+                exp_name += "shaped"
+                run.config["env_id"] = env_dict[run.config["env_id"]]
+
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "algo", exp_name)
+            exp_names += [exp_name]
+            metrics_dataframe.insert(len(metrics_dataframe.columns), "seed", run.config["seed"])
+
+            data += [metrics_dataframe]
+            if run.config["env_id"] not in envs:
+                envs[run.config["env_id"]] = [metrics_dataframe]
+                envs[run.config["env_id"] + "total_timesteps"] = run.config["total_timesteps"]
+            else:
+                envs[run.config["env_id"]] += [metrics_dataframe]
+
+            # run.summary are the output key/values like accuracy.  We call ._json_dict to omit large files
+            summary_list.append(run.summary._json_dict)
+
+            # run.config is the input metrics.  We remove special values that start with _.
+            config_list.append({k: v for k, v in run.config.items() if not k.startswith("_")})
+
+            # run.name is the name of the run.
+            name_list.append(run.name)
+
+    summary_df = pd.DataFrame.from_records(summary_list)
+    config_df = pd.DataFrame.from_records(config_list)
+    name_df = pd.DataFrame({"name": name_list})
+    all_df = pd.concat([name_df, config_df, summary_df], axis=1)
+    data = pd.concat(data, ignore_index=True)
+
+    with open(f"{feature_name}/all_df_cache.pkl", "wb") as handle:
+        pickle.dump(all_df, handle, protocol=pickle.HIGHEST_PROTOCOL)
+    with open(f"{feature_name}/envs_cache.pkl", "wb") as handle:
+        pickle.dump(envs, handle, protocol=pickle.HIGHEST_PROTOCOL)
+    with open(f"{feature_name}/exp_names_cache.pkl", "wb") as handle:
+        pickle.dump(exp_names, handle, protocol=pickle.HIGHEST_PROTOCOL)
+else:
+    with open(f"{feature_name}/all_df_cache.pkl", "rb") as handle:
+        all_df = pickle.load(handle)
+    with open(f"{feature_name}/envs_cache.pkl", "rb") as handle:
+        envs = pickle.load(handle)
+    with open(f"{feature_name}/exp_names_cache.pkl", "rb") as handle:
+        exp_names = pickle.load(handle)
+print("data loaded")
+
+
+# https://stackoverflow.com/questions/42281844/what-is-the-mathematics-behind-the-smoothing-parameter-in-tensorboards-scalar#_=_
+def smooth(scalars, weight):  # Weight between 0 and 1
+    last = scalars[0]  # First value in the plot (first timestep)
+    smoothed = list()
+    for point in scalars:
+        smoothed_val = last * weight + (1 - weight) * point  # Calculate smoothed value
+        smoothed.append(smoothed_val)  # Save it
+        last = smoothed_val  # Anchor the last smoothed value
+
+    return smoothed
+
+
+# smoothing
+for env in envs:
+    if not env.endswith("total_timesteps"):
+        for idx, metrics_dataframe in enumerate(envs[env]):
+            envs[env][idx] = metrics_dataframe.dropna(subset=[args.feature_of_interest])
+#             envs[env][idx][args.feature_of_interest] = smooth(metrics_dataframe[args.feature_of_interest], 0.85)
+
+sns.set(style="darkgrid")
+
+
+def get_df_for_env(env_id):
+    env_total_timesteps = envs[env_id + "total_timesteps"]
+    env_increment = env_total_timesteps / 500
+    envs_same_x_axis = []
+    for sampled_run in envs[env_id]:
+        df = pd.DataFrame(columns=sampled_run.columns)
+        x_axis = [i * env_increment for i in range(500 - 2)]
+        current_row = 0
+        for timestep in x_axis:
+            while sampled_run.iloc[current_row]["global_step"] < timestep:
+                current_row += 1
+                if current_row > len(sampled_run) - 2:
+                    break
+            if current_row > len(sampled_run) - 2:
+                break
+            temp_row = sampled_run.iloc[current_row].copy()
+            temp_row["global_step"] = timestep
+            df = df.append(temp_row)
+
+        envs_same_x_axis += [df]
+    return pd.concat(envs_same_x_axis, ignore_index=True)
+
+
+def export_legend(ax, filename="legend.pdf"):
+    # import matplotlib as mpl
+    # mpl.rcParams['text.usetex'] = True
+    # mpl.rcParams['text.latex.preamble'] = [r'\usepackage{amsmath}'] #for \text command
+    fig2 = plt.figure()
+    ax2 = fig2.add_subplot()
+    ax2.axis("off")
+    handles, labels = ax.get_legend_handles_labels()
+
+    legend = ax2.legend(
+        handles=handles[1:], labels=labels[1:], frameon=False, loc="lower center", ncol=3, fontsize=20, handlelength=1
+    )
+    for text in legend.get_texts():
+        text.set_text(exp_convert_dict[text.get_text()])
+    for line in legend.get_lines():
+        line.set_linewidth(4.0)
+    fig = legend.figure
+    fig.canvas.draw()
+    bbox = legend.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
+    fig.savefig(filename, dpi="figure", bbox_inches=bbox)
+    fig.clf()
+
+
+if not os.path.exists(f"{feature_name}/data"):
+    os.makedirs(f"{feature_name}/data")
+if not os.path.exists(f"{feature_name}/plots"):
+    os.makedirs(f"{feature_name}/plots")
+if not os.path.exists(f"{feature_name}/legends"):
+    os.makedirs(f"{feature_name}/legends")
+
+
+interested_exp_names = sorted(list(set(exp_names)))  # ['ppo_continuous_action', 'ppo_atari_visual']
+current_palette = sns.color_palette(n_colors=len(interested_exp_names))
+np.random.shuffle(current_palette)
+current_palette_dict = dict(zip(interested_exp_names, current_palette))
+if args.interested_exp_names:
+    interested_exp_names = args.interested_exp_names
+print(current_palette_dict)
+legend_df = pd.DataFrame()
+
+if args.font_size:
+    plt.rc("axes", titlesize=args.font_size)  # fontsize of the axes title
+    plt.rc("axes", labelsize=args.font_size)  # fontsize of the x and y labels
+    plt.rc("xtick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("ytick", labelsize=args.font_size)  # fontsize of the tick labels
+    plt.rc("legend", fontsize=args.font_size)  # legend fontsize
+
+stats = {item: [] for item in ["env_id", "exp_name", args.feature_of_interest]}
+# uncommenet the following to generate all figures
+for env in set(all_df["env_id"]):
+    if not path.exists(f"{feature_name}/data/{env}.pkl"):
+        with open(f"{feature_name}/data/{env}.pkl", "wb") as handle:
+            data = get_df_for_env(env)
+            data["seed"] = data["seed"].astype(float)
+            data[args.feature_of_interest] = data[args.feature_of_interest].astype(float)
+            pickle.dump(data, handle, protocol=pickle.HIGHEST_PROTOCOL)
+    else:
+        with open(f"{feature_name}/data/{env}.pkl", "rb") as handle:
+            data = pickle.load(handle)
+            print(f"{env}'s data loaded")
+
+    def _smooth(df):
+        df[args.feature_of_interest] = smooth(list(df[args.feature_of_interest]), args.smooth_weight)
+        return df
+
+    legend_df = legend_df.append(data)
+    ax = sns.lineplot(
+        data=data.groupby(["seed", "algo"]).apply(_smooth).loc[data["algo"].isin(interested_exp_names)],
+        x="global_step",
+        y=args.feature_of_interest,
+        hue="algo",
+        units="seed",
+        estimator=None,
+        palette=current_palette_dict,
+        alpha=0.2,
+    )
+    sns.lineplot(
+        data=data.groupby(["seed", "algo"]).apply(_smooth).loc[data["algo"].isin(interested_exp_names)],
+        x="global_step",
+        y=args.feature_of_interest,
+        hue="algo",
+        ci=None,
+        palette=current_palette_dict,
+        linewidth=2.0,
+    )
+    ax.set(xlabel=args.x_label, ylabel=args.y_label)
+
+    handles, labels = ax.get_legend_handles_labels()
+    legend = ax.legend(
+        handles=handles[1 : len(labels) // 2],
+        labels=labels[1 : len(labels) // 2],
+        loc="upper center",
+        bbox_to_anchor=(0.5, -0.20),
+        fancybox=True,
+    )
+    for text in legend.get_texts():
+        text.set_text(exp_convert_dict[text.get_text()])
+    if args.y_lim_bottom:
+        plt.ylim(bottom=args.y_lim_bottom)
+    # plt.title(env)
+    plt.tight_layout()
+    plt.savefig(f"{feature_name}/plots/{env}.{args.output_format}")
+    plt.clf()
+
+    for algo in interested_exp_names:
+        algo_data = data.loc[data["algo"].isin([algo])]
+        last_n_episodes_global_step = sorted(algo_data["global_step"].unique())[-args.last_n_episodes]
+        last_n_episodes_features = (
+            algo_data[algo_data["global_step"] > last_n_episodes_global_step]
+            .groupby(["seed"])
+            .mean()[args.feature_of_interest]
+        )
+
+        for item in last_n_episodes_features:
+            stats[args.feature_of_interest] += [item]
+            stats["exp_name"] += [exp_convert_dict[algo]]
+            stats["env_id"] += [env]
+
+# export legend
+ax = sns.lineplot(
+    data=legend_df,
+    x="global_step",
+    y=args.feature_of_interest,
+    hue="algo",
+    ci="sd",
+    palette=current_palette_dict,
+)
+ax.set(xlabel="Time Steps", ylabel="Average Episode Reward")
+ax.legend().remove()
+export_legend(ax, f"{feature_name}/legend.{args.output_format}")
+plt.clf()
+
+
+# analysis
+stats_df = pd.DataFrame(stats)
+g = stats_df.groupby(["env_id", "exp_name"]).agg(lambda x: f"{np.mean(x):.2f} ± {np.std(x):.2f}")
+print(g.reset_index().pivot("exp_name", "env_id", args.feature_of_interest).to_latex().replace("±", r"$\pm$"))
diff --git a/cleanrl/cleanrl_utils/reproduce.py b/cleanrl/cleanrl_utils/reproduce.py
new file mode 100644
index 0000000000000000000000000000000000000000..c9dd65d0f4700267a59a779282da875e7fc1daba
--- /dev/null
+++ b/cleanrl/cleanrl_utils/reproduce.py
@@ -0,0 +1,54 @@
+import argparse
+from distutils.util import strtobool
+
+import requests
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="CleanRL Plots")
+    # Common arguments
+    parser.add_argument(
+        "--run",
+        type=str,
+        default="cleanrl/cleanrl.benchmark/runs/thq5rgnz",
+        help="the name of wandb project (e.g. cleanrl/cleanrl)",
+    )
+    parser.add_argument(
+        "--remove-entity",
+        type=lambda x: bool(strtobool(x)),
+        default=True,
+        nargs="?",
+        const=True,
+        help="if toggled, the wandb-entity will be removed",
+    )
+    args = parser.parse_args()
+    uri = args.run.replace("/runs", "")
+
+    requirements_txt_url = f"https://api.wandb.ai/files/{uri}/requirements.txt"
+    metadata_url = f"https://api.wandb.ai/files/{uri}/wandb-metadata.json"
+    metadata = requests.get(url=metadata_url).json()
+
+    if args.remove_entity:
+        a = []
+        wandb_entity_idx = None
+        for i in range(len(metadata["args"])):
+            if metadata["args"][i] == "--wandb-entity":
+                wandb_entity_idx = i
+                continue
+            if wandb_entity_idx and i == wandb_entity_idx + 1:
+                continue
+            a += [metadata["args"][i]]
+    else:
+        a = metadata["args"]
+
+    program = ["python"] + [metadata["program"]] + a
+
+    print(
+        f"""
+# run the following
+python3 -m venv venv
+source venv/bin/activate
+pip install -r {requirements_txt_url}
+curl -OL https://api.wandb.ai/files/{uri}/code/{metadata["codePath"]}
+{" ".join(program)}
+"""
+    )
diff --git a/cleanrl/cleanrl_utils/resume.py b/cleanrl/cleanrl_utils/resume.py
new file mode 100644
index 0000000000000000000000000000000000000000..a877ac933aa8e4188e2e815f394eb2365aca49e2
--- /dev/null
+++ b/cleanrl/cleanrl_utils/resume.py
@@ -0,0 +1,110 @@
+# pip install boto3
+import argparse
+import re
+import time
+from distutils.util import strtobool
+
+import boto3
+import requests
+import wandb
+
+client = boto3.client("batch")
+
+parser = argparse.ArgumentParser(description="CleanRL Experiment Submission")
+# Common arguments
+parser.add_argument(
+    "--wandb-project", type=str, default="vwxyzjn/gym-microrts", help="the name of wandb project (e.g. cleanrl/cleanrl)"
+)
+parser.add_argument("--run-state", type=str, default="crashed", help="the name of this experiment")
+
+parser.add_argument("--job-queue", type=str, default="cleanrl", help="the name of the job queue")
+parser.add_argument(
+    "--wandb-key", type=str, default="", help="the wandb key. If not provided, the script will try to read from `~/.netrc`"
+)
+parser.add_argument("--docker-repo", type=str, default="vwxyzjn/gym-microrts:latest", help="the name of the job queue")
+parser.add_argument("--job-definition", type=str, default="gym-microrts", help="the name of the job definition")
+parser.add_argument("--num-seed", type=int, default=2, help="number of random seeds for experiments")
+parser.add_argument("--num-vcpu", type=int, default=1, help="number of vcpu per experiment")
+parser.add_argument("--num-memory", type=int, default=2000, help="number of memory (MB) per experiment")
+parser.add_argument("--num-gpu", type=int, default=0, help="number of gpu per experiment")
+parser.add_argument("--num-hours", type=float, default=16.0, help="number of hours allocated experiment")
+parser.add_argument(
+    "--upload-files-baseurl", type=str, default="", help="the baseurl of your website if you decide to upload files"
+)
+parser.add_argument(
+    "--submit-aws",
+    type=lambda x: bool(strtobool(x)),
+    default=False,
+    nargs="?",
+    const=True,
+    help="if toggled, script will need to be uploaded",
+)
+args = parser.parse_args()
+
+api = wandb.Api()
+
+# Project is specified by <entity/project-name>
+runs = api.runs(args.wandb_project)
+run_ids = []
+final_run_cmds = []
+for run in runs:
+    if run.state == args.run_state:
+        run_ids += [run.path[-1]]
+        metadata = requests.get(url=run.file(name="wandb-metadata.json").url).json()
+        final_run_cmds += [["python", metadata["program"]] + metadata["args"]]
+        if args.upload_files_baseurl:
+            file_name = final_run_cmds[-1][1]
+            link = args.upload_files_baseurl + "/" + file_name
+            final_run_cmds[-1] = ["curl", "-O", link, ";"] + final_run_cmds[-1]
+
+if not args.wandb_key:
+    args.wandb_key = requests.utils.get_netrc_auth("https://api.wandb.ai")[-1]
+assert (
+    len(args.wandb_key) > 0
+), "set the environment variable `WANDB_KEY` to your WANDB API key, something like `export WANDB_KEY=fdsfdsfdsfads` "
+
+# use docker directly
+if not args.submit_aws:
+    cores = 40
+    current_core = 0
+    for run_id, final_run_cmd in zip(run_ids, final_run_cmds):
+        print(
+            f'docker run -d --cpuset-cpus="{current_core}" -e WANDB={wandb_key} -e WANDB_RESUME=must -e WANDB_RUN_ID={run_id} {args.docker_repo} '
+            + '/bin/bash -c "'
+            + " ".join(final_run_cmd)
+            + '"'
+        )
+        current_core = (current_core + 1) % cores
+
+# submit jobs
+if args.submit_aws:
+    for run_id, final_run_cmd in zip(run_ids, final_run_cmds):
+        job_name = re.findall("(python)(.+)(.py)", " ".join(final_run_cmd))[0][1].strip() + str(int(time.time()))
+        job_name = job_name.replace("/", "_").replace("_param ", "")
+        resources_requirements = []
+        if args.num_gpu:
+            resources_requirements = [
+                {"value": str(args.num_gpu), "type": "GPU"},
+            ]
+
+        response = client.submit_job(
+            jobName=job_name,
+            jobQueue=args.job_queue,
+            jobDefinition=args.job_definition,
+            containerOverrides={
+                "vcpus": args.num_vcpu,
+                "memory": args.num_memory,
+                "command": ["/bin/bash", "-c", " ".join(final_run_cmd)],
+                "environment": [
+                    {"name": "WANDB", "value": wandb_key},
+                    {"name": "WANDB_RESUME", "value": "must"},
+                    {"name": "WANDB_RUN_ID", "value": run_id},
+                ],
+                "resourceRequirements": resources_requirements,
+            },
+            retryStrategy={"attempts": 1},
+            timeout={"attemptDurationSeconds": int(args.num_hours * 60 * 60)},
+        )
+        if response["ResponseMetadata"]["HTTPStatusCode"] != 200:
+            print(response)
+            raise Exception("jobs submit failure")
diff --git a/cleanrl/cleanrl_utils/submit_exp.py b/cleanrl/cleanrl_utils/submit_exp.py
new file mode 100644
index 0000000000000000000000000000000000000000..301b8bdaa10efcb6bb6d0b8437dc4866a2d628e8
--- /dev/null
+++ b/cleanrl/cleanrl_utils/submit_exp.py
@@ -0,0 +1,142 @@
+import argparse
+import multiprocessing
+import subprocess
+import time
+from distutils.util import strtobool
+
+import boto3
+import requests
+import wandb
+
+# fmt: off
+parser = argparse.ArgumentParser(description='CleanRL Experiment Submission')
+# experiment generation
+parser.add_argument('--exp-script', type=str, default="debug.sh",
+    help='the file name of this experiment')
+parser.add_argument('--command', type=str, default="uv run python cleanrl/ppo.py",
+    help='the docker command')
+
+# CleanRL specific args
+parser.add_argument('--wandb-key', type=str, default="",
+    help='the wandb key. If not provided, the script will try to read from `netrc`')
+parser.add_argument('--num-seed', type=int, default=1,
+    help='number of random seeds for experiments')
+
+# experiment submission
+parser.add_argument('--job-queue', type=str, default="m6gd-medium",
+    help='the name of the job queue')
+parser.add_argument('--docker-tag', type=str, default="vwxyzjn/cleanrl:latest",
+    help='the name of the docker tag')
+parser.add_argument('--num-vcpu', type=int, default=1,
+    help='number of vcpu per experiment')
+parser.add_argument('--num-memory', type=int, default=2000,
+    help='number of memory (MB) per experiment')
+parser.add_argument('--num-gpu', type=int, default=0,
+    help='number of gpu per experiment')
+parser.add_argument('--num-hours', type=float, default=16.0,
+    help='number of hours allocated experiment')
+parser.add_argument('-b', '--build', type=lambda x:bool(strtobool(x)), default=False, nargs='?', const=True,
+    help='if toggled, the script will build a container')
+parser.add_argument('--archs', type=str, default="linux/amd64", # linux/arm64,linux/amd64
+    help='the archs to build the docker container for')
+parser.add_argument('-p', '--push', type=lambda x:bool(strtobool(x)), default=False, nargs='?', const=True,
+    help='if toggled, the script will push the built container')
+parser.add_argument('--provider', type=str, default="", choices=["aws"],
+    help='the cloud provider of choice (currently only `aws` is supported)')
+parser.add_argument('--aws-num-retries', type=int, default=1,
+    help='the number of job retries for `provider=="aws"`')
+args = parser.parse_args()
+# fmt: on
+
+if args.build:
+    output_type_str = "--output=type=registry" if args.push else "--output=type=docker"
+    subprocess.run(
+        f"docker buildx build {output_type_str} --platform {args.archs} -t {args.docker_tag} .",
+        shell=True,
+        check=True,
+    )
+
+if not args.wandb_key:
+    try:
+        args.wandb_key = requests.utils.get_netrc_auth("https://api.wandb.ai")[-1]
+    except:
+        pass
+assert len(args.wandb_key) > 0, "you have not logged into W&B; try do `wandb login`"
+
+# extract runs from bash scripts
+final_run_cmds = []
+for seed in range(1, 1 + args.num_seed):
+    final_run_cmds += [args.command + " --seed " + str(seed)]
+
+final_str = ""
+cores = multiprocessing.cpu_count()
+current_core = 0
+for final_run_cmd in final_run_cmds:
+    run_command = (
+        f'docker run -d --cpuset-cpus="{current_core}" -e WANDB_API_KEY={args.wandb_key} {args.docker_tag} '
+        + '/bin/bash -c "'
+        + final_run_cmd
+        + '"'
+        + "\n"
+    )
+    print(run_command)
+    final_str += run_command
+    current_core = (current_core + 1) % cores
+
+with open(f"{args.exp_script}.docker.sh", "w+") as f:
+    f.write(final_str)
+
+# submit jobs
+if args.provider == "aws":
+    client = boto3.client("batch")
+    for final_run_cmd in final_run_cmds:
+        job_name = args.docker_tag.replace(":", "").replace("/", "_").replace(" ", "").replace("-", "_") + str(
+            int(time.time())
+        )
+        resources_requirements = []
+        if args.num_gpu:
+            resources_requirements = [
+                {"value": str(args.num_gpu), "type": "GPU"},
+            ]
+        try:
+            job_def_name = args.docker_tag.replace(":", "_").replace("/", "_")
+            job_def = client.register_job_definition(
+                jobDefinitionName=job_def_name,
+                type="container",
+                containerProperties={
+                    "image": args.docker_tag,
+                    "vcpus": args.num_vcpu,
+                    "memory": args.num_memory,
+                    "command": [
+                        "/bin/bash",
+                    ],
+                },
+            )
+            response = client.submit_job(
+                jobName=job_name,
+                jobQueue=args.job_queue,
+                jobDefinition=job_def_name,
+                containerOverrides={
+                    "vcpus": args.num_vcpu,
+                    "memory": args.num_memory,
+                    "command": ["/bin/bash", "-c", final_run_cmd],
+                    "environment": [
+                        {"name": "WANDB_API_KEY", "value": args.wandb_key},
+                        {"name": "WANDB_RESUME", "value": "allow"},
+                        {"name": "WANDB_RUN_ID", "value": wandb.util.generate_id()},
+                    ],
+                    "resourceRequirements": resources_requirements,
+                },
+                retryStrategy={"attempts": args.aws_num_retries},
+                timeout={"attemptDurationSeconds": int(args.num_hours * 60 * 60)},
+            )
+            if response["ResponseMetadata"]["HTTPStatusCode"] != 200:
+                print(response)
+                raise Exception("jobs submit failure")
+        except Exception as e:
+            print(e)
+        finally:
+            response = client.deregister_job_definition(jobDefinition=job_def_name)
+            if response["ResponseMetadata"]["HTTPStatusCode"] != 200:
+                print(response)
+                raise Exception("jobs submit failure")
diff --git a/cleanrl/cleanrl_utils/tuner.py b/cleanrl/cleanrl_utils/tuner.py
new file mode 100644
index 0000000000000000000000000000000000000000..d72af58ab57812f0a8b0187055d49bef4c1695b3
--- /dev/null
+++ b/cleanrl/cleanrl_utils/tuner.py
@@ -0,0 +1,146 @@
+import os
+import runpy
+import sys
+import time
+from typing import Callable, Dict, List, Optional
+
+import numpy as np
+import optuna
+import wandb
+from rich import print
+from tensorboard.backend.event_processing import event_accumulator
+
+
+class HiddenPrints:
+    def __enter__(self):
+        self._original_stdout = sys.stdout
+        sys.stdout = open(os.devnull, "w")
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        sys.stdout.close()
+        sys.stdout = self._original_stdout
+
+
+class Tuner:
+    def __init__(
+        self,
+        script: str,
+        metric: str,
+        target_scores: Dict[str, Optional[List[float]]],
+        params_fn: Callable[[optuna.Trial], Dict],
+        direction: str = "maximize",
+        aggregation_type: str = "average",
+        metric_last_n_average_window: int = 50,
+        sampler: Optional[optuna.samplers.BaseSampler] = None,
+        pruner: Optional[optuna.pruners.BasePruner] = None,
+        storage: str = "sqlite:///cleanrl_hpopt.db",
+        study_name: str = "",
+        wandb_kwargs: Dict[str, any] = {},
+    ) -> None:
+        self.script = script
+        self.metric = metric
+        self.target_scores = target_scores
+        if len(self.target_scores) > 1:
+            if None in self.target_scores.values():
+                raise ValueError(
+                    "If there are multiple environments, the target scores must be specified for each environment."
+                )
+
+        self.params_fn = params_fn
+        self.direction = direction
+        self.aggregation_type = aggregation_type
+        if self.aggregation_type == "average":
+            self.aggregation_fn = np.average
+        elif self.aggregation_type == "median":
+            self.aggregation_fn = np.median
+        elif self.aggregation_type == "max":
+            self.aggregation_fn = np.max
+        elif self.aggregation_type == "min":
+            self.aggregation_fn = np.min
+        else:
+            raise ValueError(f"Unknown aggregation type {self.aggregation_type}")
+        self.metric_last_n_average_window = metric_last_n_average_window
+        self.pruner = pruner
+        self.sampler = sampler
+        self.storage = storage
+        self.study_name = study_name
+        if len(self.study_name) == 0:
+            self.study_name = f"tuner_{int(time.time())}"
+        self.wandb_kwargs = wandb_kwargs
+
+    def tune(self, num_trials: int, num_seeds: int) -> None:
+        def objective(trial: optuna.Trial):
+            params = self.params_fn(trial)
+            run = None
+            if len(self.wandb_kwargs.keys()) > 0:
+                run = wandb.init(
+                    **self.wandb_kwargs,
+                    config=params,
+                    name=f"{self.study_name}_{trial.number}",
+                    group=self.study_name,
+                    save_code=True,
+                    reinit=True,
+                )
+
+            algo_command = [f"--{key}={value}" for key, value in params.items()]
+            normalized_scoress = []
+            for seed in range(num_seeds):
+                normalized_scores = []
+                for env_id in self.target_scores.keys():
+                    sys.argv = algo_command + [f"--env-id={env_id}", f"--seed={seed}"]
+                    with HiddenPrints():
+                        experiment = runpy.run_path(path_name=self.script, run_name="__main__")
+
+                    # read metric from tensorboard
+                    ea = event_accumulator.EventAccumulator(f"runs/{experiment['run_name']}")
+                    ea.Reload()
+                    metric_values = [
+                        scalar_event.value for scalar_event in ea.Scalars(self.metric)[-self.metric_last_n_average_window :]
+                    ]
+                    print(
+                        f"The average episodic return on {env_id} is {np.average(metric_values)} averaged over the last {self.metric_last_n_average_window} episodes."
+                    )
+                    if self.target_scores[env_id] is not None:
+                        normalized_scores += [
+                            (np.average(metric_values) - self.target_scores[env_id][0])
+                            / (self.target_scores[env_id][1] - self.target_scores[env_id][0])
+                        ]
+                    else:
+                        normalized_scores += [np.average(metric_values)]
+                    if run:
+                        run.log({f"{env_id}_return": np.average(metric_values)})
+
+                normalized_scoress += [normalized_scores]
+                aggregated_normalized_score = self.aggregation_fn(normalized_scores)
+                print(f"The {self.aggregation_type} normalized score is {aggregated_normalized_score} with num_seeds={seed}")
+                trial.report(aggregated_normalized_score, step=seed)
+                if run:
+                    run.log({"aggregated_normalized_score": aggregated_normalized_score})
+                if trial.should_prune():
+                    if run:
+                        run.finish(quiet=True)
+                    raise optuna.TrialPruned()
+
+            if run:
+                run.finish(quiet=True)
+            return np.average(
+                self.aggregation_fn(normalized_scoress, axis=1)
+            )  # we alaways return the average of the aggregated normalized scores
+
+        study = optuna.create_study(
+            study_name=self.study_name,
+            direction=self.direction,
+            storage=self.storage,
+            pruner=self.pruner,
+            sampler=self.sampler,
+        )
+        print("==========================================================================================")
+        print("run another tuner with the following command:")
+        print(f"python -m cleanrl_utils.tuner --study-name {self.study_name}")
+        print("==========================================================================================")
+        study.optimize(
+            objective,
+            n_trials=num_trials,
+        )
+        print(f"The best trial obtains a normalized score of {study.best_trial.value}", study.best_trial.params)
+        return study.best_trial
diff --git a/cleanrl/cloud/.gitignore b/cleanrl/cloud/.gitignore
new file mode 100644
index 0000000000000000000000000000000000000000..d386bb58c075209b9a5270108b705cc925281bab
--- /dev/null
+++ b/cleanrl/cloud/.gitignore
@@ -0,0 +1,3 @@
+*.tfstate**
+*.lock.**
+*.terraform
\ No newline at end of file
diff --git a/cleanrl/cloud/main.tf b/cleanrl/cloud/main.tf
new file mode 100644
index 0000000000000000000000000000000000000000..69e11a312495056141936e5fc4e9d322d7499c1e
--- /dev/null
+++ b/cleanrl/cloud/main.tf
@@ -0,0 +1,29 @@
+terraform {
+  required_providers {
+    aws = {
+      source  = "hashicorp/aws"
+      version = "~> 3.27"
+    }
+  }
+
+  required_version = ">= 0.14.9"
+}
+
+provider "aws" {
+  profile = "default"
+  # region  = "us-west-2"
+}
+
+module "cleanrl" {
+  source              = "./modules/cleanrl"
+  spot_bid_percentage = "50"
+  instance_types = [
+    "g4dn.4xlarge", # 16 vCPU, 64GB, $1.204, GPU
+    "g4dn.xlarge",  # 4 vCPU, 16GB, $0.526, GPU
+    "r5ad.large",   # 2 vCPU, 16GB, $0.131
+    "c5a.large",    # 2 vCPU, 4GB, $0.077
+    # ARM-based
+    "c6g.medium",  # 1 vCPU, 2GB, $0.034
+    "m6gd.medium", # 1 vCPU, 4GB, $0.0452
+  ]
+}
\ No newline at end of file
diff --git a/cleanrl/cloud/modules/cleanrl/main.tf b/cleanrl/cloud/modules/cleanrl/main.tf
new file mode 100644
index 0000000000000000000000000000000000000000..bc49ec3daa7011c50664db6e7f54248c0bd01515
--- /dev/null
+++ b/cleanrl/cloud/modules/cleanrl/main.tf
@@ -0,0 +1,73 @@
+############
+# On-demand resources
+############
+
+resource "aws_batch_compute_environment" "on_demand" {
+  count                    = length(var.instance_types)
+  compute_environment_name = replace(var.instance_types[count.index], ".", "-")
+  compute_resources {
+    instance_role = aws_iam_instance_profile.ecs_instance_role.arn
+    instance_type = [
+      var.instance_types[count.index],
+    ]
+    max_vcpus = var.max_vcpus
+    min_vcpus = 0
+    security_group_ids = [
+      aws_security_group.sample.id,
+    ]
+    subnets             = data.aws_subnet_ids.all_default_subnets.ids
+    type                = "EC2"
+    allocation_strategy = var.on_demand_allocation_strategy
+  }
+  service_role = aws_iam_role.aws_batch_service_role.arn
+  type         = "MANAGED"
+  depends_on   = [aws_iam_role_policy_attachment.aws_batch_service_role]
+}
+
+resource "aws_batch_job_queue" "on_demand" {
+  count    = length(var.instance_types)
+  name     = replace(var.instance_types[count.index], ".", "-")
+  state    = "ENABLED"
+  priority = 100
+  compute_environments = [
+    aws_batch_compute_environment.on_demand[count.index].arn,
+  ]
+}
+
+############
+# Spot resources
+############
+
+resource "aws_batch_compute_environment" "spot" {
+  count                    = length(var.instance_types)
+  compute_environment_name = replace("${var.instance_types[count.index]}-spot", ".", "-")
+  compute_resources {
+    instance_role = aws_iam_instance_profile.ecs_instance_role.arn
+    instance_type = [
+      var.instance_types[count.index],
+    ]
+    max_vcpus = var.max_vcpus
+    min_vcpus = 0
+    security_group_ids = [
+      aws_security_group.sample.id,
+    ]
+    subnets             = data.aws_subnet_ids.all_default_subnets.ids
+    type                = "SPOT"
+    bid_percentage      = var.spot_bid_percentage
+    allocation_strategy = var.spot_allocation_strategy
+    spot_iam_fleet_role = aws_iam_role.AWS_EC2_spot_fleet_role.arn
+  }
+  service_role = aws_iam_role.aws_batch_service_role.arn
+  type         = "MANAGED"
+  depends_on   = [aws_iam_role_policy_attachment.aws_batch_service_role]
+}
+
+resource "aws_batch_job_queue" "spot" {
+  count    = length(var.instance_types)
+  name     = replace("${var.instance_types[count.index]}-spot", ".", "-")
+  state    = "ENABLED"
+  priority = 100
+  compute_environments = [
+    aws_batch_compute_environment.spot[count.index].arn,
+  ]
+}
diff --git a/cleanrl/cloud/modules/cleanrl/setups.tf b/cleanrl/cloud/modules/cleanrl/setups.tf
new file mode 100644
index 0000000000000000000000000000000000000000..8c95aeb2be8f1656539ed2ef358784bc91e77cc2
--- /dev/null
+++ b/cleanrl/cloud/modules/cleanrl/setups.tf
@@ -0,0 +1,96 @@
+resource "aws_iam_role" "ecs_instance_role" {
+  name               = "ecs_instance_role"
+  assume_role_policy = <<EOF
+{
+    "Version": "2012-10-17",
+    "Statement": [
+    {
+        "Action": "sts:AssumeRole",
+        "Effect": "Allow",
+        "Principal": {
+            "Service": "ec2.amazonaws.com"
+        }
+    }
+    ]
+}
+EOF
+}
+
+resource "aws_iam_role_policy_attachment" "ecs_instance_role" {
+  role       = aws_iam_role.ecs_instance_role.name
+  policy_arn = "arn:aws:iam::aws:policy/service-role/AmazonEC2ContainerServiceforEC2Role"
+}
+
+resource "aws_iam_instance_profile" "ecs_instance_role" {
+  name = "ecs_instance_role"
+  role = aws_iam_role.ecs_instance_role.name
+}
+
+resource "aws_iam_role" "aws_batch_service_role" {
+  name = "aws_batch_service_role"
+
+  assume_role_policy = <<EOF
+{
+    "Version": "2012-10-17",
+    "Statement": [
+    {
+        "Action": "sts:AssumeRole",
+        "Effect": "Allow",
+        "Principal": {
+        "Service": "batch.amazonaws.com"
+        }
+    }
+    ]
+}
+EOF
+}
+
+resource "aws_iam_role_policy_attachment" "aws_batch_service_role" {
+  role       = aws_iam_role.aws_batch_service_role.name
+  policy_arn = "arn:aws:iam::aws:policy/service-role/AWSBatchServiceRole"
+}
+
+resource "aws_iam_role" "AWS_EC2_spot_fleet_role" {
+  name = "aws_spot_fleet_service_role"
+
+  assume_role_policy = <<POLICY
+{
+    "Version": "2012-10-17",
+    "Statement": [
+    {
+        "Action": "sts:AssumeRole",
+        "Effect": "Allow",
+        "Principal": {
+          "Service": "spotfleet.amazonaws.com"
+        }
+    }
+    ]
+}
+POLICY
+}
+
+resource "aws_iam_role_policy_attachment" "AWS_EC2_spot_fleet_role" {
+  role       = aws_iam_role.AWS_EC2_spot_fleet_role.name
+  policy_arn = "arn:aws:iam::aws:policy/service-role/AmazonEC2SpotFleetTaggingRole"
+}
+
+resource "aws_security_group" "sample" {
+  name = "aws_batch_compute_environment_security_group"
+
+  egress {
+    from_port   = 0
+    to_port     = 0
+    protocol    = "-1"
+    cidr_blocks = ["0.0.0.0/0"]
+  }
+}
+
+resource "aws_default_vpc" "sample" {
+  tags = {
+    Name = "Default VPC"
+  }
+}
+
+data "aws_subnet_ids" "all_default_subnets" {
+  vpc_id = aws_default_vpc.sample.id
+}
\ No newline at end of file
diff --git a/cleanrl/cloud/modules/cleanrl/variables.tf b/cleanrl/cloud/modules/cleanrl/variables.tf
new file mode 100644
index 0000000000000000000000000000000000000000..e8b04b22fac43ff0df3fe0df14d85ce1f2d34df4
--- /dev/null
+++ b/cleanrl/cloud/modules/cleanrl/variables.tf
@@ -0,0 +1,36 @@
+variable "max_vcpus" {
+  description = "The maximum number of vcpus in each computing environment"
+  type        = number
+  default     = 2000
+}
+
+variable "on_demand_allocation_strategy" {
+  description = "The allocation strateg for the on-demand computing environment (e.g. `BEST_FIT`, `BEST_FIT_PROGRESSIVE`; see https://docs.aws.amazon.com/batch/latest/userguide/allocation-strategies.html)"
+  type        = string
+  default     = "BEST_FIT"
+}
+
+variable "spot_allocation_strategy" {
+  description = "The allocation strateg for the on-demand computing environment (e.g. `BEST_FIT`, `BEST_FIT_PROGRESSIVE`, or `SPOT_CAPACITY_OPTIMIZED`; see https://docs.aws.amazon.com/batch/latest/userguide/allocation-strategies.html)"
+  type        = string
+  default     = "BEST_FIT"
+}
+
+variable "spot_bid_percentage" {
+  description = "The spot bid percentage"
+  type        = string
+  default     = "50"
+}
+
+variable "instance_types" {
+  type = list(string)
+  default = [
+    "g4dn.4xlarge", # 16 vCPU, 64GB, $1.204, GPU
+    "g4dn.xlarge",  # 4 vCPU, 16GB, $0.526, GPU
+    "r5ad.large",   # 2 vCPU, 16GB, $0.131
+    "c6g.medium",   # 1 vCPU, 2GB, $0.034
+    # ARM-based
+    "a1.medium",   # 1 vCPU, 2GB, $0.0255
+    "m6gd.medium", # 1 vCPU, 4GB, $0.0452
+  ]
+}
diff --git a/cleanrl/docs/CNAME b/cleanrl/docs/CNAME
new file mode 100644
index 0000000000000000000000000000000000000000..f02e91f2f2bd6abd8a2f246d28d4efd5790bffa0
--- /dev/null
+++ b/cleanrl/docs/CNAME
@@ -0,0 +1 @@
+docs.cleanrl.dev
\ No newline at end of file
diff --git a/cleanrl/docs/advanced/hyperparameter-tuning.md b/cleanrl/docs/advanced/hyperparameter-tuning.md
new file mode 100644
index 0000000000000000000000000000000000000000..2752119238b64da9e3088f4de80b8b35f743da1f
--- /dev/null
+++ b/cleanrl/docs/advanced/hyperparameter-tuning.md
@@ -0,0 +1,265 @@
+# Hyperparameter Tuning
+
+CleanRL comes with a lightweight hyperparameter tuning utility `Tuner`, with a primary purpose of helping researchers find **a single set of  hyperparameters that works well with multiple tasks of similar type** (e.g., one set of parameters for MuJoCo, another set for Atari). We try to avoid tuning hyperparameters per task, which is usually adhoc and do not generalize.
+
+## Get started
+
+Consider the [:material-github: `tuner_example.py`](https://github.com/vwxyzjn/cleanrl/blob/master/tuner_example.py) in the CleanRL's root directory, which has the following content:
+
+
+```python title="tuner_example.py"
+import optuna
+
+from cleanrl_utils.tuner import Tuner
+
+tuner = Tuner(
+    script="cleanrl/ppo.py",
+    metric="charts/episodic_return",
+    metric_last_n_average_window=50,
+    direction="maximize",
+    aggregation_type="average",
+    target_scores={
+        "CartPole-v1": [0, 500],
+        "Acrobot-v1": [-500, 0],
+    },
+    params_fn=lambda trial: {
+        "learning-rate": trial.suggest_float("learning-rate", 0.0003, 0.003, log=True),
+        "num-minibatches": trial.suggest_categorical("num-minibatches", [1, 2, 4]),
+        "update-epochs": trial.suggest_categorical("update-epochs", [1, 2, 4, 8]),
+        "num-steps": trial.suggest_categorical("num-steps", [5, 16, 32, 64, 128]),
+        "vf-coef": trial.suggest_float("vf-coef", 0, 5),
+        "max-grad-norm": trial.suggest_float("max-grad-norm", 0, 5),
+        "total-timesteps": 100000,
+        "num-envs": 16,
+    },
+    pruner=optuna.pruners.MedianPruner(n_startup_trials=5),
+    sampler=optuna.samplers.TPESampler(),
+)
+tuner.tune(
+    num_trials=100,
+    num_seeds=3,
+)
+```
+
+You can run the tuner with the following command:
+
+```bash
+uv pip install ".[optuna]"
+python tuner_example.py
+```
+
+<script id="asciicast-517226" src="https://asciinema.org/a/517226.js" async></script>
+
+
+Often it is useful to find a single set of hyper parameters that work well with multiple environments, but this is challenging **because each environment may have different reward scales**, so we need to normalize the reward scales. 
+
+This is where `target_scores` comes in. We can use it to specify an upper and lower threshold of rewards for each environment (they don't have to be exact boundaries and can be just ballpark estimates). For example, if we want to find a set of hyper parameters that work well with `CartPole-v1` and `Acrobot-v1`, we can set the `target_scores` and `aggregation_type` as follows:
+
+```python
+tuner = Tuner(
+    ...,
+    target_scores={
+        "CartPole-v1": [0, 500],
+        "Acrobot-v1": [-500, 0],
+    }
+    aggregation_type="average",
+)
+```
+
+
+Here is what happened when running `python tuner_example.py`:
+
+1. The `tuner_example.py` launches `num_trials=100` *trials* to find the best single set of hyperparameters for `CartPole-v1` and `Acrobot-v1` in `script="cleanrl/ppo.py"`.
+1. Each *trial* samples a set of hyperparameters from the `params_fn` to run `num_seeds=3` *experiments* with different random seeds, mitigating the impact of randomness on the results. 
+    * In each *experiment*, `tuner_example.py` averages the last `metric_last_n_average_window=50` reported `metric="charts/episodic_return"` to a number $x_i$ and calculate a normalized score $z_i$ according to the `target_scores`. In this case, $z_{i_0} = (x_{i_0} - 0) / (500 - 0)$ and $z_{i_1} = (x_{i_1} - -500) / (0 - -500)$. Then the tuner will `aggregation_type="average"` the two scores to get the normalized score $z_i = (z_{i_0} + z_{i_1}) / 2$.
+1. Each *trial* then averages the normalized scores $z_i$ of the `num_seeds=3` *experiments* to a number $z$ and the tuner optimizes $z$ according `direction="maximize"`.
+
+
+Note that we are using 3 random seeds for each environment in `["CartPole-v1","Acrobot-v1"]`, totalling `2*3=6` experiments per trial.
+
+
+???+ info
+
+    When optimizing Atari games, you can put the human normalized scores in `target_scores` (Mnih et al., 2015, Extended Data Table 2)[^1], as done in the following example. The first number for each environment is the score obtained by random play and the second number is the score obtained by professional game testers. Note here we using `aggregation_type="median"`, this means we will optimize for the **median** of the human normalized scores **averaged** over the `num_seeds=3` experiments for each trial (pay close attention to the phrasing).
+
+    ```python
+    tuner = Tuner(
+        script="cleanrl/ppo_atari_envpool.py",
+        metric="charts/episodic_return",
+        metric_last_n_average_window=50,
+        direction="maximize",
+        aggregation_type="median",
+        target_scores={
+            "Alien-v5": [227.8, 6875],
+            "Amidar-v5": [5.8, 1676],
+            'Assault-v5': (222.4, 1496),
+            'Asterix-v5': (210.0, 8503),
+            'Asteroids-v5': (719.1, 13157),
+            ...
+        },
+        num_seeds=3,
+        ...
+    )
+    ```
+
+## Visualization
+
+Running `python tuner_example.py` will create a sqlite database containing all of the hyperparameter trials in `./cleanrl_hpopt.db`  We can use [optuna-dashboard](https://github.com/optuna/optuna-dashboard) to visualize the process.
+
+```bash
+uv run optuna-dashboard sqlite:///cleanrl_hpopt.db
+```
+
+![](./optuna-dashboard-1.png)
+
+In the panel above, the y-axis shows the normalized score of the two environments (`CartPole-v1` and `Acrobot-v1`) and the x-axis shows the number of trials.
+
+![](./optuna-dashboard-2.png)
+
+The panel above shows the parallel coordinates plot of the normalized score — for example, we can see having a lower learning rate and a higher number of minibatches is a good choice.
+
+
+???+ info
+
+    You can use a different database by passing `Tuner(..., storage="mysql://root@localhost/example")` to use a more persistent storage.
+
+## Work w/o knowing the reward scales
+
+What if we don't know the reward scales for the environments? We can optionally set the value `target_scores` to `None`, but this will only work with one environment at a time because it's difficult to optimize without knowing the scale of the rewards in each environment.
+
+
+```python title="tuner_example.py" hl_lines="11-13"
+import optuna
+
+from cleanrl_utils.tuner import Tuner
+
+tuner = Tuner(
+    script="cleanrl/ppo.py",
+    metric="charts/episodic_return",
+    metric_last_n_average_window=50,
+    direction="maximize",
+    aggregation_type="average",
+    target_scores={
+        "CartPole-v1": None,
+    },
+    params_fn=lambda trial: {
+        "learning-rate": trial.suggest_float("learning-rate", 0.0003, 0.003, log=True),
+        "num-minibatches": trial.suggest_categorical("num-minibatches", [1, 2, 4]),
+        "update-epochs": trial.suggest_categorical("update-epochs", [1, 2, 4, 8]),
+        "num-steps": trial.suggest_categorical("num-steps", [5, 16, 32, 64, 128]),
+        "vf-coef": trial.suggest_float("vf-coef", 0, 5),
+        "max-grad-norm": trial.suggest_float("max-grad-norm", 0, 5),
+        "total-timesteps": 100000,
+        "num-envs": 16,
+    },
+    pruner=optuna.pruners.MedianPruner(n_startup_trials=5),
+    sampler=optuna.samplers.TPESampler(),
+)
+tuner.tune(
+    num_trials=100,
+    num_seeds=3,
+)
+```
+
+In the example above, we will set the normalized score $z_i$ to be the average of the last `metric_last_n_average_window=50` reported `metric="charts/episodic_return"` for `CartPole-v1`.
+
+
+## Work w/ pruners and samplers
+
+You can use `Tuner` with any [pruner](https://optuna.readthedocs.io/en/stable/reference/pruners.html) from `optuna` to prune less promising experiments or [samplers](https://optuna.readthedocs.io/en/stable/reference/samplers.html) to sample new hyperparameters. If you don't specify them explicitly, the script will use the [default ones](https://optuna.readthedocs.io/en/stable/reference/generated/optuna.create_study.html).
+
+```python title="tuner_example.py" hl_lines="1 24 25"
+import optuna
+
+from cleanrl_utils.tuner import Tuner
+
+tuner = Tuner(
+    script="cleanrl/ppo.py",
+    metric="charts/episodic_return",
+    metric_last_n_average_window=50,
+    direction="maximize",
+    aggregation_type="average",
+    target_scores={
+        "CartPole-v1": None,
+    },
+    params_fn=lambda trial: {
+        "learning-rate": trial.suggest_float("learning-rate", 0.0003, 0.003, log=True),
+        "num-minibatches": trial.suggest_categorical("num-minibatches", [1, 2, 4]),
+        "update-epochs": trial.suggest_categorical("update-epochs", [1, 2, 4, 8]),
+        "num-steps": trial.suggest_categorical("num-steps", [5, 16, 32, 64, 128]),
+        "vf-coef": trial.suggest_float("vf-coef", 0, 5),
+        "max-grad-norm": trial.suggest_float("max-grad-norm", 0, 5),
+        "total-timesteps": 100000,
+        "num-envs": 16,
+    },
+    pruner=optuna.pruners.MedianPruner(n_startup_trials=5),
+    sampler=optuna.samplers.TPESampler(),
+)
+tuner.tune(
+    num_trials=100,
+    num_seeds=3,
+)
+```
+
+
+## Track experiments w/ Weights and Biases
+
+The `Tuner` can track all the experiments into [Weights and Biases](https://wandb.ai) to help you visualize the progress of the tuning.
+
+
+```python title="tuner_example.py" hl_lines="26"
+import optuna
+
+from cleanrl_utils.tuner import Tuner
+
+tuner = Tuner(
+    script="cleanrl/ppo.py",
+    metric="charts/episodic_return",
+    metric_last_n_average_window=50,
+    direction="maximize",
+    aggregation_type="average",
+    target_scores={
+        "CartPole-v1": None,
+    },
+    params_fn=lambda trial: {
+        "learning-rate": trial.suggest_float("learning-rate", 0.0003, 0.003, log=True),
+        "num-minibatches": trial.suggest_categorical("num-minibatches", [1, 2, 4]),
+        "update-epochs": trial.suggest_categorical("update-epochs", [1, 2, 4, 8]),
+        "num-steps": trial.suggest_categorical("num-steps", [5, 16, 32, 64, 128]),
+        "vf-coef": trial.suggest_float("vf-coef", 0, 5),
+        "max-grad-norm": trial.suggest_float("max-grad-norm", 0, 5),
+        "total-timesteps": 100000,
+        "num-envs": 16,
+    },
+    pruner=optuna.pruners.MedianPruner(n_startup_trials=5),
+    sampler=optuna.samplers.TPESampler(),
+    wandb_kwargs={"project": "cleanrl"},
+)
+tuner.tune(
+    num_trials=100,
+    num_seeds=3,
+)
+```
+
+
+<!-- 
+```
+$ wandb sweep -p cleanrl cleanrl_utils/tuner_config/ppo_continuous_action_envpool_jax.yaml
+
+wandb: Creating sweep from: cleanrl_utils/tuner_config/ppo_continuous_action_envpool_jax.yaml
+wandb: Created sweep with ID: pvux95ci
+wandb: View sweep at: https://wandb.ai/costa-huang/cleanRL/sweeps/pvux95ci
+wandb: Run sweep agent with: wandb agent costa-huang/cleanrl/pvux95ci
+```
+
+Then run `wandb agent costa-huang/cleanrl/pvux95ci`
+
+## Optuna
+
+```
+optuna create-study --study-name "distributed-example" --storage "sqlite:///cleanrl_hpopt.db"
+optuna-dashboard sqlite:///cleanrl_hpopt.db
+``` -->
+
+
+[^1]:Mnih, V., Kavukcuoglu, K., Silver, D. et al. Human-level control through deep reinforcement learning. Nature 518, 529–533 (2015). https://doi.org/10.1038/nature14236
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_atari.md b/cleanrl/docs/benchmark/ppo_atari.md
new file mode 100644
index 0000000000000000000000000000000000000000..f8a0094cd80065efb7dc8e05f1857a3ef7a75931
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_atari.md
@@ -0,0 +1,5 @@
+|                         | openrlbenchmark/cleanrl/ppo_atari ({'tag': ['pr-424']})   |
+|:------------------------|:----------------------------------------------------------|
+| PongNoFrameskip-v4      | 20.36 ± 0.20                                              |
+| BeamRiderNoFrameskip-v4 | 1915.93 ± 484.58                                          |
+| BreakoutNoFrameskip-v4  | 414.66 ± 28.09                                            |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_atari_envpool.md b/cleanrl/docs/benchmark/ppo_atari_envpool.md
new file mode 100644
index 0000000000000000000000000000000000000000..4f6f20afcbea238edf874ff290f4e64fc9192944
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_atari_envpool.md
@@ -0,0 +1,5 @@
+|              | openrlbenchmark/cleanrl/ppo_atari_envpool ({'tag': ['pr-424']})   | openrlbenchmark/cleanrl/ppo_atari ({'tag': ['pr-424']})   |
+|:-------------|:------------------------------------------------------------------|:----------------------------------------------------------|
+| Pong-v5      | 20.45 ± 0.09                                                      | 20.36 ± 0.20                                              |
+| BeamRider-v5 | 2501.85 ± 210.52                                                  | 1915.93 ± 484.58                                          |
+| Breakout-v5  | 211.24 ± 151.84                                                   | 414.66 ± 28.09                                            |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_atari_envpool_xla_jax_runtimes.md b/cleanrl/docs/benchmark/ppo_atari_envpool_xla_jax_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..09fe628f0cb6bf51121bee29f123a118e6c4dbeb
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_atari_envpool_xla_jax_runtimes.md
@@ -0,0 +1,59 @@
+|                     |   openrlbenchmark/envpool-atari/ppo_atari_envpool_xla_jax ({}) |   openrlbenchmark/baselines/baselines-ppo2-cnn ({}) |
+|:--------------------|---------------------------------------------------------------:|----------------------------------------------------:|
+| Alien-v5            |                                                        50.3275 |                                             117.397 |
+| Amidar-v5           |                                                        42.8176 |                                             114.093 |
+| Assault-v5          |                                                        35.9245 |                                             108.094 |
+| Asterix-v5          |                                                        37.7117 |                                             113.386 |
+| Asteroids-v5        |                                                        39.9731 |                                             114.409 |
+| Atlantis-v5         |                                                        40.1527 |                                             123.05  |
+| BankHeist-v5        |                                                        38.7443 |                                             137.308 |
+| BattleZone-v5       |                                                        45.0654 |                                             138.489 |
+| BeamRider-v5        |                                                        42.0778 |                                             119.437 |
+| Berzerk-v5          |                                                        38.7173 |                                             135.316 |
+| Bowling-v5          |                                                        35.0156 |                                             131.365 |
+| Boxing-v5           |                                                        48.8149 |                                             151.607 |
+| Breakout-v5         |                                                        42.3547 |                                             122.828 |
+| Centipede-v5        |                                                        43.6886 |                                             150.112 |
+| ChopperCommand-v5   |                                                        45.9308 |                                             131.192 |
+| CrazyClimber-v5     |                                                        36.0841 |                                             127.942 |
+| Defender-v5         |                                                        35.1029 |                                             132.29  |
+| DemonAttack-v5      |                                                        35.41   |                                             128.476 |
+| DoubleDunk-v5       |                                                        41.4521 |                                             108.028 |
+| Enduro-v5           |                                                        44.9909 |                                             142.046 |
+| FishingDerby-v5     |                                                        51.6075 |                                             151.286 |
+| Freeway-v5          |                                                        50.7103 |                                             154.163 |
+| Frostbite-v5        |                                                        47.5474 |                                             146.092 |
+| Gopher-v5           |                                                        36.2977 |                                             139.496 |
+| Gravitar-v5         |                                                        41.9322 |                                             138.746 |
+| Hero-v5             |                                                        50.5106 |                                             152.413 |
+| IceHockey-v5        |                                                        43.0228 |                                             144.455 |
+| Jamesbond-v5        |                                                        38.8264 |                                             137.321 |
+| Kangaroo-v5         |                                                        44.4304 |                                             142.436 |
+| Krull-v5            |                                                        47.7748 |                                             147.313 |
+| KungFuMaster-v5     |                                                        43.1534 |                                             141.903 |
+| MontezumaRevenge-v5 |                                                        44.8838 |                                             146.777 |
+| MsPacman-v5         |                                                        42.6463 |                                             138.382 |
+| NameThisGame-v5     |                                                        43.8473 |                                             136.264 |
+| Phoenix-v5          |                                                        36.7586 |                                             129.716 |
+| Pitfall-v5          |                                                        44.6369 |                                             137.36  |
+| Pong-v5             |                                                        36.7657 |                                             118.745 |
+| PrivateEye-v5       |                                                        43.3399 |                                             143.957 |
+| Qbert-v5            |                                                        40.1475 |                                             135.255 |
+| Riverraid-v5        |                                                        44.2555 |                                             142.627 |
+| RoadRunner-v5       |                                                        46.1059 |                                             145.451 |
+| Robotank-v5         |                                                        48.3364 |                                             149.681 |
+| Seaquest-v5         |                                                        38.3639 |                                             136.942 |
+| Skiing-v5           |                                                        38.6402 |                                             132.061 |
+| Solaris-v5          |                                                        50.2944 |                                             136.9   |
+| SpaceInvaders-v5    |                                                        39.4931 |                                             125.83  |
+| StarGunner-v5       |                                                        33.7096 |                                             119.18  |
+| Surround-v5         |                                                        33.923  |                                             132.017 |
+| Tennis-v5           |                                                        39.6194 |                                              97.019 |
+| TimePilot-v5        |                                                        37.0124 |                                             130.693 |
+| Tutankham-v5        |                                                        36.9677 |                                             139.694 |
+| UpNDown-v5          |                                                        52.9895 |                                             140.876 |
+| Venture-v5          |                                                        37.9828 |                                             144.236 |
+| VideoPinball-v5     |                                                        47.1716 |                                             179.866 |
+| WizardOfWor-v5      |                                                        37.5751 |                                             142.086 |
+| YarsRevenge-v5      |                                                        36.5889 |                                             127.358 |
+| Zaxxon-v5           |                                                        41.9785 |                                             133.922 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_atari_lstm_runtimes.md b/cleanrl/docs/benchmark/ppo_atari_lstm_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..079df7642da07149cde64873120cd711eefa03c2
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_atari_lstm_runtimes.md
@@ -0,0 +1,5 @@
+|                         |   openrlbenchmark/cleanrl/ppo_atari_lstm ({'tag': ['pr-424']}) |
+|:------------------------|---------------------------------------------------------------:|
+| PongNoFrameskip-v4      |                                                        317.607 |
+| BeamRiderNoFrameskip-v4 |                                                        314.864 |
+| BreakoutNoFrameskip-v4  |                                                        383.724 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_continuous_action_runtimes.md b/cleanrl/docs/benchmark/ppo_continuous_action_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..4ffc2e28d3e8e461c85b9ef9c93a0fbee27b2e39
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_continuous_action_runtimes.md
@@ -0,0 +1,11 @@
+|                                      |   openrlbenchmark/cleanrl/ppo_continuous_action ({'tag': ['pr-424']}) |
+|:-------------------------------------|----------------------------------------------------------------------:|
+| HalfCheetah-v4                       |                                                               25.3589 |
+| Walker2d-v4                          |                                                               24.3157 |
+| Hopper-v4                            |                                                               25.7066 |
+| InvertedPendulum-v4                  |                                                               23.7672 |
+| Humanoid-v4                          |                                                               49.5592 |
+| Pusher-v4                            |                                                               28.8162 |
+| dm_control/acrobot-swingup-v0        |                                                               26.5793 |
+| dm_control/acrobot-swingup_sparse-v0 |                                                               25.1265 |
+| dm_control/ball_in_cup-catch-v0      |                                                               26.1947 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_procgen_runtimes.md b/cleanrl/docs/benchmark/ppo_procgen_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..5b956125e896ebb11d6fe9346ed835ff1de919a3
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_procgen_runtimes.md
@@ -0,0 +1,5 @@
+|           |   openrlbenchmark/cleanrl/ppo_procgen ({'tag': ['pr-424']}) |
+|:----------|------------------------------------------------------------:|
+| starpilot |                                                     114.649 |
+| bossfight |                                                     128.679 |
+| bigfish   |                                                     107.788 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/ppo_runtimes.md b/cleanrl/docs/benchmark/ppo_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..0277e1f85a274d7e1912fd48cdfbc59d78513d28
--- /dev/null
+++ b/cleanrl/docs/benchmark/ppo_runtimes.md
@@ -0,0 +1,5 @@
+|                |   openrlbenchmark/cleanrl/ppo ({'tag': ['pr-424']}) |
+|:---------------|----------------------------------------------------:|
+| CartPole-v1    |                                            10.4737  |
+| Acrobot-v1     |                                            15.4606  |
+| MountainCar-v0 |                                             6.95995 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/sac_runtimes.md b/cleanrl/docs/benchmark/sac_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..b35f211217d9e1e8ff6a9abef3cb4a8a3c93be58
--- /dev/null
+++ b/cleanrl/docs/benchmark/sac_runtimes.md
@@ -0,0 +1,8 @@
+|                     |   openrlbenchmark/cleanrl/sac_continuous_action ({'tag': ['pr-424']}) |
+|:--------------------|----------------------------------------------------------------------:|
+| HalfCheetah-v4      |                                                               174.778 |
+| Walker2d-v4         |                                                               161.161 |
+| Hopper-v4           |                                                               173.242 |
+| InvertedPendulum-v4 |                                                               179.042 |
+| Humanoid-v4         |                                                               177.31  |
+| Pusher-v4           |                                                               172.123 |
\ No newline at end of file
diff --git a/cleanrl/docs/benchmark/td3_runtimes.md b/cleanrl/docs/benchmark/td3_runtimes.md
new file mode 100644
index 0000000000000000000000000000000000000000..76451881e8c3199370562f6476c0cb706f889b8c
--- /dev/null
+++ b/cleanrl/docs/benchmark/td3_runtimes.md
@@ -0,0 +1,8 @@
+|                     |   openrlbenchmark/cleanrl/td3_continuous_action ({'tag': ['pr-424']}) |   openrlbenchmark/cleanrl/td3_continuous_action_jax ({'tag': ['pr-424']}) |
+|:--------------------|----------------------------------------------------------------------:|--------------------------------------------------------------------------:|
+| HalfCheetah-v4      |                                                               87.353  |                                                                   39.5119 |
+| Walker2d-v4         |                                                               80.8592 |                                                                   34.0497 |
+| Hopper-v4           |                                                               90.9921 |                                                                   33.4079 |
+| InvertedPendulum-v4 |                                                               70.4218 |                                                                   30.2624 |
+| Humanoid-v4         |                                                               79.1624 |                                                                   70.2437 |
+| Pusher-v4           |                                                               95.2208 |                                                                   39.6051 |
\ No newline at end of file
diff --git a/cleanrl/docs/blog/.authors.yml b/cleanrl/docs/blog/.authors.yml
new file mode 100644
index 0000000000000000000000000000000000000000..f8e1d1bc7741504cbf5796533dbea8407d302805
--- /dev/null
+++ b/cleanrl/docs/blog/.authors.yml
@@ -0,0 +1,4 @@
+costa:
+  name: Costa Huang
+  description: Lead dev of CleanRL
+  avatar: https://avatars.githubusercontent.com/u/5555347
diff --git a/cleanrl/docs/blog/index.md b/cleanrl/docs/blog/index.md
new file mode 100644
index 0000000000000000000000000000000000000000..05761ac57f6bcedf28c34c3867849df97ef89a8b
--- /dev/null
+++ b/cleanrl/docs/blog/index.md
@@ -0,0 +1 @@
+# Blog
diff --git a/cleanrl/docs/blog/posts/cleanrl-v1.md b/cleanrl/docs/blog/posts/cleanrl-v1.md
new file mode 100644
index 0000000000000000000000000000000000000000..ce0cda7478f3fa55dae2229540909151eb0f7c6b
--- /dev/null
+++ b/cleanrl/docs/blog/posts/cleanrl-v1.md
@@ -0,0 +1,280 @@
+---
+date: 2022-10-05
+authors: [costa]
+description: >
+  🎉 We are thrilled to announce the v1.0.0 CleanRL Release. Along with our CleanRL paper's recent publication in Journal of Machine Learning Research (https://www.jmlr.org/papers/v23/21-1342.html), our v1.0.0 release includes reworked documentation, new algorithm variants, support for google's new ML framework JAX, hyperparameter tuning utilities, and more.
+categories:
+  - Blog
+---
+
+
+
+# CleanRL v1 Release
+
+
+🎉 We are thrilled to announce the v1.0.0 CleanRL Release. Along with our [CleanRL paper's recent publication in Journal of Machine Learning Research](https://www.jmlr.org/papers/v23/21-1342.html), our v1.0.0 release includes reworked documentation, new algorithm variants, support for google's new ML framework [JAX](https://github.com/google/jax), hyperparameter tuning utilities, and more. CleanRL has come a long way making high-quality deep reinforcement learning implementations easy to understand and reproducible. This release is a major milestone for the project and we are excited to share it with you. Over 90 PRs were merged to make this release possible. We would like to thank all the contributors who made this release possible.
+
+More detailed release notes are available at [v1.0.0b1](https://github.com/vwxyzjn/cleanrl/releases/tag/v1.0.0b1), [v1.0.0b2](https://github.com/vwxyzjn/cleanrl/releases/tag/v1.0.0b2), and [v1.0.0](https://github.com/vwxyzjn/cleanrl/releases/tag/v1.0.0).
+
+
+<!-- more -->
+
+## Reworked documentation
+
+One of the biggest change of the v1 release is the added documentation at [docs.cleanrl.dev](https://docs.cleanrl.dev). Having great documentation is important for building a reliable and reproducible project. We have reworked the documentation to make it easier to understand and use. For each implemented algorithm, we have documented as much as we can to promote transparency:
+
+* [Short description of the algorithm and references](/rl-algorithms/ppo/#overview)
+* [A list of implemented variant](/rl-algorithms/ppo/#implemented-variants)
+* [The usage information](/rl-algorithms/ppo/#usage)
+* [The explanation of the logged metrics](/rl-algorithms/ppo/#explanation-of-the-logged-metrics)
+* [The documentation of implementation details](/rl-algorithms/ppo/#implementation-details)
+* [Experimental results](/rl-algorithms/ppo/#experiment-results)
+
+Here is a list of the algorithm variants and their documentation:
+
+
+| Algorithm      | Variants Implemented |
+| ----------- | ----------- |
+| ✅ [Proximal Policy Gradient (PPO)](https://arxiv.org/pdf/1707.06347.pdf)  | :material-github: [`ppo.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppopy) |
+| | :material-github: [`ppo_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_ataripy)
+| | :material-github: [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_actionpy)
+| | :material-github: [`ppo_atari_lstm.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_lstmpy)
+| | :material-github: [`ppo_atari_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpoolpy)
+| | :material-github: [`ppo_atari_envpool_xla_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy)
+| | :material-github: [`ppo_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_procgenpy)
+| | :material-github: [`ppo_atari_multigpu.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_multigpupy)
+| | :material-github: [`ppo_pettingzoo_ma_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_pettingzoo_ma_ataripy)
+| | :material-github: [`ppo_continuous_action_isaacgym.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action_isaacgym/ppo_continuous_action_isaacgym.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_action_isaacgympy)
+| ✅ [Deep Q-Learning (DQN)](https://web.stanford.edu/class/psych209/Readings/MnihEtAlHassibis15NatureControlDeepRL.pdf) | :material-github: [`dqn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqnpy) |
+| | :material-github: [`dqn_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_ataripy) |
+| | :material-github: [`dqn_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_jaxpy) |
+| | :material-github: [`dqn_atari_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_atari_jaxpy) |
+| ✅ [Categorical DQN (C51)](https://arxiv.org/pdf/1707.06887.pdf) | :material-github: [`c51.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51.py), :material-file-document: [docs](/rl-algorithms/c51/#c51py) |
+| | :material-github: [`c51_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_atari.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_ataripy) |
+| ✅ [Soft Actor-Critic (SAC)](https://arxiv.org/pdf/1812.05905.pdf) | :material-github: [`sac_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/sac_continuous_action.py), :material-file-document: [docs](/rl-algorithms/sac/#sac_continuous_actionpy) |
+| ✅ [Deep Deterministic Policy Gradient (DDPG)](https://arxiv.org/pdf/1509.02971.pdf) | :material-github: [`ddpg_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action.py), :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_actionpy) |
+| | :material-github: [`ddpg_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action_jax.py),  :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_action_jaxpy)
+| ✅ [Twin Delayed Deep Deterministic Policy Gradient (TD3)](https://arxiv.org/pdf/1802.09477.pdf) | :material-github: [`td3_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_actionpy) |
+|  | :material-github: [`td3_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action_jax.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_action_jaxpy) |
+| ✅ [Phasic Policy Gradient (PPG)](https://arxiv.org/abs/2009.04416) | :material-github: [`ppg_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppg_procgen.py), :material-file-document: [docs](/rl-algorithms/ppg/#ppg_procgenpy) |
+| ✅ [Random Network Distillation (RND)](https://arxiv.org/abs/1810.12894) | :material-github: [`ppo_rnd_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_rnd_envpool.py), :material-file-document: [docs](/rl-algorithms/ppo-rnd/#ppo_rnd_envpoolpy) |
+
+
+We also improved the [contribution guide](https://github.com/vwxyzjn/cleanrl/blob/master/CONTRIBUTING.md) to make it easier for new contributors to get started. We are still working on improving the documentation. If you have any suggestions, please let us know in the [GitHub Issues](https://github.com/vwxyzjn/cleanrl/issues).
+
+
+## New algorithm variants, support for JAX
+
+We now support JAX-based learning algorithm variants, which are usually faster than the `torch` equivalent! Here are the docs of the new JAX-based DQN, TD3, and DDPG implementations:
+
+
+* [`dqn_atari_jax.py`](https://docs.cleanrl.dev/rl-algorithms/dqn/#dqn_atari_jaxpy) [@kinalmehta](https://github.com/kinalmehta) in [:material-github: vwxyzjn/cleanrl#222](https://github.com/vwxyzjn/cleanrl/pull/222)
+    * about 25% faster than `dqn_atari.py`.
+* [`td3_continuous_action_jax.py`](https://docs.cleanrl.dev/rl-algorithms/td3/#td3_continuous_action_jaxpy) by [@joaogui1](https://github.com/joaogui1) in [:material-github: vwxyzjn/cleanrl#225](https://github.com/vwxyzjn/cleanrl/pull/225)
+    * about 2.5-4x faster than `td3_continuous_action.py`.
+* [`ddpg_continuous_action_jax.py`](https://docs.cleanrl.dev/rl-algorithms/ddpg/#ddpg_continuous_action_jaxpy) by [@vwxyzjn](https://github.com/vwxyzjn) in [:material-github: vwxyzjn/cleanrl#187](https://github.com/vwxyzjn/cleanrl/pull/187)
+    * about 2.5-4x faster than `ddpg_continuous_action.py`.
+* [`ppo_atari_envpool_xla_jax.py`](https://docs.cleanrl.dev/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) by [@vwxyzjn](https://github.com/vwxyzjn) in [:material-github: vwxyzjn/cleanrl#227](https://github.com/vwxyzjn/cleanrl/pull/227)
+    * about 3x faster than openai/baselines' PPO.
+
+For example, below are the benchmark of DDPG + JAX (see docs [here](/rl-algorithms/ddpg/#ddpg_continuous_action_jaxpy) for further detail):
+
+
+<div class="grid-container">
+<img src="/rl-algorithms/ddpg-jax/HalfCheetah-v2.png">
+<img src="/rl-algorithms/ddpg-jax/HalfCheetah-v2-time.png">
+</div>
+
+Other new algorithm variants include multi-GPU PPO, PPO prototype that works with [Isaac Gym](https://github.com/NVIDIA-Omniverse/IsaacGymEnvs), multi-agent Atari PPO, and refactored PPG and PPO-RND implementations:
+
+* [`ppo_atari_multigpu.pu`](https://docs.cleanrl.dev/rl-algorithms/ppo/#ppo_atari_multigpupy) by [@vwxyzjn](https://github.com/vwxyzjn) in [:material-github: vwxyzjn/cleanrl#178]( https://github.com/vwxyzjn/cleanrl/pull/178)
+    * about 34% faster than `ppo_atari.py` which uses `SyncVectorEnv`.
+* [`ppo_continuous_action_isaacgym.py`](https://docs.cleanrl.dev/rl-algorithms/ppo/#ppo_continuous_action_isaacgympy) by [@vwxyzjn](https://github.com/vwxyzjn) in [:material-github: vwxyzjn/cleanrl#233](https://github.com/vwxyzjn/cleanrl/pull/233)
+    * achieves 4000+ score and 30M steps on IsaacGymEnvs' `Ant` in 4 mins. 
+* [`ppo_pettingzoo_ma_atari.py`](https://docs.cleanrl.dev/rl-algorithms/ppo/#ppo_pettingzoo_ma_ataripy) by [@vwxyzjn](https://github.com/vwxyzjn) in [:material-github: vwxyzjn/cleanrl#188](https://github.com/vwxyzjn/cleanrl/pull/188)
+    * achieves ~4000 *episodic length* (not episodic return) in Pong, creating competitive self play agents.
+* [`ppg_procgen.py`](https://docs.cleanrl.dev/rl-algorithms/ppg/#ppg_procgenpy) by [@Dipamc77](https://github.com/Dipamc77) in [:material-github: vwxyzjn/cleanrl#186](https://github.com/vwxyzjn/cleanrl/pull/186)
+    * matches openai/baselines' PPO performance in StarPilot (easy), BossFight (easy), and BigFish (easy).
+* [`ppo_rnd_envpoolpy.py`](https://docs.cleanrl.dev/rl-algorithms/ppo-rnd/#ppo_rnd_envpoolpy) by [@yooceii](https://github.com/yooceii) in [:material-github: vwxyzjn/cleanrl#151](https://github.com/vwxyzjn/cleanrl/pull/151)
+    * achieves ~7100 in `MontezumaRevengeNoFrameSkip-v4`.
+
+
+
+## Tooling improvements
+
+We love tools! The v1.0.0 release comes with a series of DevOps improvements, including pre-commit utilities, CI integration with GitHub to run end-to-end test cases. We also make available a new hyperparameter tuning tool and a new tool for running benchmark experiments.
+
+### DevOps
+
+We added a pre-commit utility to help contributors to format their code, check for spelling, and removing unused variables and imports before submitting a pull request (see [Contribution guide](/contribution/#pre-commit-utilities) for more detail).
+
+<img src="/static/pre-commit.png">
+
+
+To ensure our single-file implementations can run without error, we also added CI/CD pipeline which now runs end-to-end test cases for all the algorithm variants. The pipeline also tests builds across different operating systems, such as Linux, macOS, and Windows (see [here](https://github.com/vwxyzjn/cleanrl/actions/runs/3401991711/usage) as an example). GitHub actions are free for open source projects, and we are very happy to have this tool to help us maintain the project. 
+
+<img src="/static/blog/cleanrl-v1/github-action.png">
+
+### Hyperparameter tuning utilities
+
+We now have preliminary support for hyperparameter tuning via `optuna` (see [docs](https://docs.cleanrl.dev/advanced/hyperparameter-tuning/)), which is designed to help researchers to find **a single set** of hyperparameters that work well with a kind of games. The current API looks like below:
+
+```python
+import optuna
+from cleanrl_utils.tuner import Tuner
+tuner = Tuner(
+    script="cleanrl/ppo.py",
+    metric="charts/episodic_return",
+    metric_last_n_average_window=50,
+    direction="maximize",
+    aggregation_type="average",
+    target_scores={
+        "CartPole-v1": [0, 500],
+        "Acrobot-v1": [-500, 0],
+    },
+    params_fn=lambda trial: {
+        "learning-rate": trial.suggest_float("learning-rate", 0.0003, 0.003, log=True),
+        "num-minibatches": trial.suggest_categorical("num-minibatches", [1, 2, 4]),
+        "update-epochs": trial.suggest_categorical("update-epochs", [1, 2, 4, 8]),
+        "num-steps": trial.suggest_categorical("num-steps", [5, 16, 32, 64, 128]),
+        "vf-coef": trial.suggest_float("vf-coef", 0, 5),
+        "max-grad-norm": trial.suggest_float("max-grad-norm", 0, 5),
+        "total-timesteps": 100000,
+        "num-envs": 16,
+    },
+    pruner=optuna.pruners.MedianPruner(n_startup_trials=5),
+    sampler=optuna.samplers.TPESampler(),
+)
+tuner.tune(
+    num_trials=100,
+    num_seeds=3,
+)
+```
+
+### Benchmarking utilities
+
+We also added a new tool for running benchmark experiments. The tool is designed to help researchers to quickly run benchmark experiments across different algorithms environments with some random seeds. The tool lives in the `cleanrl_utils.benchmark` module, and the users can run commands such as:
+
+```bash
+OMP_NUM_THREADS=1 xvfb-run -a python -m cleanrl_utils.benchmark \
+    --env-ids CartPole-v1 Acrobot-v1 MountainCar-v0 \
+    --command "uv run python cleanrl/ppo.py --no_cuda --track --capture_video" \
+    --num-seeds 3 \
+    --workers 5
+```
+
+which will run the `ppo.py` script with `--no_cuda --track --capture_video` arguments across 3 random seeds for 3 environments. It uses `multiprocessing` to create a pool of 5 workers run the experiments in parallel.
+
+## What’s next?
+
+It is an exciting time and new improvements are coming to CleanRL. We plan to add more JAX-based implementations, huggingface integration, some RLops prototypes, and support Gymnasium. CleanRL is a community-based project and we always welcome new contributors. If there is an algorithm or new feature you would like to contribute, feel free to chat with us on our [discord channel](https://discord.gg/D6RCjA6sVT) or raise a GitHub issue. 
+
+### More JAX implementations
+
+More JAX-based implementation are coming. [Antonin Raffin](https://github.com/araffin), the core maintainer of [Stable-baselines3](https://github.com/DLR-RM/stable-baselines3), [SBX](https://github.com/araffin/sbx), and [rl-baselines3-zoo](https://github.com/DLR-RM/rl-baselines3-zoo), is contributing an optimized Soft Actor Critic implementation in JAX ([:material-github: vwxyzjn/cleanrl#300](https://github.com/vwxyzjn/cleanrl/pull/300)) and TD3+TQC, and DroQ ([:material-github: vwxyzjn/cleanrl#272](https://github.com/vwxyzjn/cleanrl/pull/272). These are incredibly exciting new algorithms. For example, DroQ is extremely sample efficient and can obtain ~5000 return in `HalfCheetah-v3` in just 100k steps ([tracked sbx experiment](https://wandb.ai/openrlbenchmark/sbx/runs/1tyzq3tu)).
+
+### Huggingface integration
+
+[Huggingface Hub 🤗](https://huggingface.co/models) is a great platform for sharing and collaborating models. We are working on a new integration with Huggingface Hub to make it easier for researchers to share their RL models and benchmark them against other models ([:material-github: vwxyzjn/cleanrl#292](https://github.com/vwxyzjn/cleanrl/pull/292)). Stay tuned! In the future, we will have a simple snippet for loading models like below:
+
+```python
+import random
+from typing import Callable
+
+import gym
+import numpy as np
+import torch
+
+
+def evaluate(
+    model_path: str,
+    make_env: Callable,
+    env_id: str,
+    eval_episodes: int,
+    run_name: str,
+    Model: torch.nn.Module,
+    device: torch.device,
+    epsilon: float = 0.05,
+    capture_video: bool = True,
+):
+    envs = gym.vector.SyncVectorEnv([make_env(env_id, 0, 0, capture_video, run_name)])
+    model = Model(envs).to(device)
+    model.load_state_dict(torch.load(model_path))
+    model.eval()
+
+    obs = envs.reset()
+    episodic_returns = []
+    while len(episodic_returns) < eval_episodes:
+        if random.random() < epsilon:
+            actions = np.array([envs.single_action_space.sample() for _ in range(envs.num_envs)])
+        else:
+            q_values = model(torch.Tensor(obs).to(device))
+            actions = torch.argmax(q_values, dim=1).cpu().numpy()
+        next_obs, _, _, infos = envs.step(actions)
+        for info in infos:
+            if "episode" in info.keys():
+                print(f"eval_episode={len(episodic_returns)}, episodic_return={info['episode']['r']}")
+                episodic_returns += [info["episode"]["r"]]
+        obs = next_obs
+
+    return episodic_returns
+
+
+if __name__ == "__main__":
+    from huggingface_hub import hf_hub_download
+
+    from cleanrl.dqn import QNetwork, make_env
+
+    model_path = hf_hub_download(repo_id="cleanrl/CartPole-v1-dqn-seed1", filename="q_network.pth")
+```
+
+
+<img src="/static/blog/cleanrl-v1/hf.png">
+
+### RLops
+
+How do we know the effect of a new feature / bug fix? DRL is brittle and has a series of reproducibility issues — even bug fixes sometimes could introduce performance regression (e.g., see [how a bug fix of contact force in MuJoCo results in worse performance for PPO](https://github.com/openai/gym/pull/2762#discussion_r853488897)). Therefore, it is essential to understand how the proposed changes impact the performance of the algorithms. 
+
+We are working a prototype tool that allows us to compare the performance of the library at different versions of the tracked experiment ([:material-github: vwxyzjn/cleanrl#307](https://github.com/vwxyzjn/cleanrl/pull/307)). With this tool, we can confidently merge new features / bug fixes without worrying about introducing catastrophic regression. The users can run commands such as:
+
+```bash
+python -m cleanrl_utils.rlops --exp-name ddpg_continuous_action \
+    --wandb-project-name cleanrl \
+    --wandb-entity openrlbenchmark \
+    --tags 'pr-299' 'rlops-pilot' \
+    --env-ids HalfCheetah-v2 Walker2d-v2 Hopper-v2 InvertedPendulum-v2 Humanoid-v2 Pusher-v2 \
+    --output-filename compare.png \
+    --scan-history \
+    --metric-last-n-average-window 100 \
+    --report
+```
+
+which generates the following image
+
+<img max-width="500px" src="/static/blog/cleanrl-v1/rlops.png">
+
+
+### Support for Gymnasium
+
+[Farama-Foundation/Gymnasium](https://github.com/Farama-Foundation/Gymnasium) is the next generation of [`openai/gym`](https://github.com/openai/gym) that will continue to be maintained and introduce new features. Please see their [announcement](https://farama.org/Announcing-The-Farama-Foundation) for further detail. We are migrating to `gymnasium` and the progress can be tracked in [:material-github: vwxyzjn/cleanrl#277](https://github.com/vwxyzjn/cleanrl/pull/277).
+
+<blockquote class="twitter-tweet"><p lang="en" dir="ltr">Today we&#39;re launching the Farama Foundation, a new nonprofit dedicated to open source reinforcement learning, and we&#39;re beginning by maintaining and standardizing all the major open source reinforcement learning environments. Read more here: <a href="https://t.co/kQqFMQdVqn">https://t.co/kQqFMQdVqn</a></p>&mdash; Farama Foundation (@FaramaFound) <a href="https://twitter.com/FaramaFound/status/1584936111461502977?ref_src=twsrc%5Etfw">October 25, 2022</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
+
+Also, the Farama foundation is working a project called [Shimmy](https://github.com/Farama-Foundation/Shimmy) which offers conversion wrapper for [`deepmind/dm_env`](https://github.com/deepmind/dm_env) environments, such as [`dm_control`](https://github.com/deepmind/dm_control) and [`deepmind/lab`](https://github.com/deepmind/lab). This is an exciting project that will allow us to support `deepmind/dm_env` in the future.
+
+
+## Contributions
+
+CleanRL has benefited from the contributions of many awesome folks. I would like to cordially thank the core dev members [@dosssman](https://github.com/dosssman) [@yooceii](https://github.com/yooceii) [@Dipamc](https://github.com/Dipamc) [@kinalmehta](https://github.com/kinalmehta) [@bragajj](https://github.com/bragajj) for their efforts in helping maintain the CleanRL repository.  I would also like to give a shout-out to our new contributors [@cool](https://github.com/cool)-RR, [@Howuhh](https://github.com/Howuhh), [@jseppanen](https://github.com/jseppanen), [@joaogui1](https://github.com/joaogui1), [@ALPH2H](https://github.com/ALPH2H), [@ElliotMunro200](https://github.com/ElliotMunro200), [@WillDudley](https://github.com/WillDudley), and [@sdpkjc](https://github.com/sdpkjc).
+
+We always welcome new contributors to the project. If you are interested in contributing to CleanRL (e.g., new features, bug fixes, new algorithms), please check out our reworked [contributing guide](https://docs.cleanrl.dev/contribution/).
+
+
+## New CleanRL Supported Publications
+
+* Md Masudur Rahman and Yexiang Xue. "Bootstrap Advantage Estimation for Policy Optimization in Reinforcement Learning." In Proceedings of the IEEE International Conference on Machine Learning and Applications (ICMLA), 2022. [https://arxiv.org/pdf/2210.07312.pdf](https://arxiv.org/pdf/2210.07312.pdf)
+* Weng, Jiayi, Min Lin, Shengyi Huang, Bo Liu, Denys Makoviichuk, Viktor Makoviychuk, Zichen Liu et al. "Envpool: A highly parallel reinforcement learning environment execution engine." In Thirty-sixth Conference on Neural Information Processing Systems Datasets and Benchmarks Track. [https://openreview.net/forum?id=BubxnHpuMbG](https://openreview.net/forum?id=BubxnHpuMbG)
+* Huang, Shengyi, Rousslan Fernand Julien Dossa, Antonin Raffin, Anssi Kanervisto, and Weixun Wang. "The 37 Implementation Details of Proximal Policy Optimization." International Conference on Learning Representations 2022 Blog Post Track, [https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/)
+* Huang, Shengyi, and Santiago Ontañón. "A closer look at invalid action masking in policy gradient algorithms." The International FLAIRS Conference Proceedings, 35. [https://journals.flvc.org/FLAIRS/article/view/130584](https://journals.flvc.org/FLAIRS/article/view/130584)
+* Schmidt, Dominik, and Thomas Schmied. "Fast and Data-Efficient Training of Rainbow: an Experimental Study on Atari." Deep Reinforcement Learning Workshop at the 35th Conference on Neural Information Processing Systems, [https://arxiv.org/abs/2111.10247](https://arxiv.org/abs/2111.10247)
+
diff --git a/cleanrl/docs/cleanrl-supported-papers-projects.md b/cleanrl/docs/cleanrl-supported-papers-projects.md
new file mode 100644
index 0000000000000000000000000000000000000000..6247b4ddb23fd36973cf10274155ef3ba8f421ad
--- /dev/null
+++ b/cleanrl/docs/cleanrl-supported-papers-projects.md
@@ -0,0 +1,32 @@
+# CleanRL-supported Papers / Projects
+
+CleanRL has become an increasingly popular deep reinforcement learning library, especially among practitioners who prefer more customizable code. Since its debut in July 2019, CleanRL has supported many open source projects and publications. Below are some CleanRL-supported projects and publications.
+
+**Feel free to edit this list if your project or paper has used CleanRL.**
+
+## Publications
+
+* Md Masudur Rahman and Yexiang Xue. "Bootstrap Advantage Estimation for Policy Optimization in Reinforcement Learning." In Proceedings of the IEEE International Conference on Machine Learning and Applications (ICMLA), 2022. [https://arxiv.org/pdf/2210.07312.pdf](https://arxiv.org/pdf/2210.07312.pdf)
+
+* Centa, Matheus, and Philippe Preux. "Soft Action Priors: Towards Robust Policy Transfer." arXiv preprint arXiv:2209.09882 (2022). [https://arxiv.org/pdf/2209.09882.pdf](https://arxiv.org/pdf/2209.09882.pdf)
+
+* Weng, Jiayi, Min Lin, Shengyi Huang, Bo Liu, Denys Makoviichuk, Viktor Makoviychuk, Zichen Liu et al. "Envpool: A highly parallel reinforcement learning environment execution engine." In Thirty-sixth Conference on Neural Information Processing Systems Datasets and Benchmarks Track. [https://openreview.net/forum?id=BubxnHpuMbG](https://openreview.net/forum?id=BubxnHpuMbG)
+
+* Huang, Shengyi, Rousslan Fernand Julien Dossa, Antonin Raffin, Anssi Kanervisto, and Weixun Wang. "The 37 Implementation Details of Proximal Policy Optimization." International Conference on Learning Representations 2022 Blog Post Track, [https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/)
+
+* Huang, Shengyi, and Santiago Ontañón. "A closer look at invalid action masking in policy gradient algorithms." The International FLAIRS Conference Proceedings, 35. [https://journals.flvc.org/FLAIRS/article/view/130584](https://journals.flvc.org/FLAIRS/article/view/130584)
+
+* Schmidt, Dominik, and Thomas Schmied. "Fast and Data-Efficient Training of Rainbow: an Experimental Study on Atari." Deep Reinforcement Learning Workshop at the 35th Conference on Neural Information Processing Systems, [https://arxiv.org/abs/2111.10247](https://arxiv.org/abs/2111.10247)
+
+
+* Dossa, Rousslan Fernand Julien, Shengyi Huang, Santiago Ontañón, and Takashi Matsubara. "An Empirical Investigation of Early Stopping Optimizations in Proximal Policy Optimization." IEEE Access 9 (2021): 117981-117992. [https://ieeexplore.ieee.org/abstract/document/9520424](https://ieeexplore.ieee.org/abstract/document/9520424)
+
+* Huang, Shengyi, Santiago Ontañón, Chris Bamford, and Lukasz Grela. "Gym-µRTS: Toward Affordable Full Game Real-time Strategy Games Research with Deep Reinforcement Learning." In 2021 IEEE Conference on Games (CoG), pp. 1-8. IEEE, 2021. [https://ieeexplore.ieee.org/abstract/document/9619076](https://ieeexplore.ieee.org/abstract/document/9619076)
+
+* Huang, Shengyi, and Santiago Ontañón. "Measuring Generalization of Deep Reinforcement Learning Applied to Real-time Strategy Games", AAAI 2021 Reinforcement Learning in Games Workshop, http://aaai-rlg.mlanctot.info/papers/AAAI21-RLG_paper_33.pdf
+
+* Bamford, Chris, Huang, Shengyi, and Lucas, Simon, "Griddly: A platform for AI research in games", *AAAI 2021 Reinforcement Learning in Games Workshop*, [https://arxiv.org/abs/2011.](https://arxiv.org/abs/2011.)06363
+
+* Huang, Shengyi, and Santiago Ontañón. "Action guidance: Getting the best of sparse rewards and shaped rewards for real-time strategy games." AIIDE Workshop on Artificial Intelligence for Strategy Games, [https://arxiv.org/abs/2010.03956](https://arxiv.org/abs/2010.03956)
+
+* Huang, Shengyi, and Santiago Ontañón. "Comparing Observation and Action Representations for Deep Reinforcement Learning in $\mu $ RTS." AIIDE Workshop on Artificial Intelligence for Strategy Gamee, October 2019 [https://arxiv.org/abs/1910.12134](https://arxiv.org/abs/1910.12134)
diff --git a/cleanrl/docs/cloud/installation.md b/cleanrl/docs/cloud/installation.md
new file mode 100644
index 0000000000000000000000000000000000000000..b57a20e20cfb58d145d809c1f60f3c3ddff60d28
--- /dev/null
+++ b/cleanrl/docs/cloud/installation.md
@@ -0,0 +1,38 @@
+# Installation
+
+The rough idea behind the cloud integration is to package our code into a docker container and use AWS Batch to
+run thousands of experiments concurrently. 
+
+## Prerequisites
+
+* Terraform (see installation tutorial [here](https://learn.hashicorp.com/tutorials/terraform/install-cli))
+
+We use Terraform to define our infrastructure with AWS Batch, which you can spin up as follows
+
+```bash
+# assuming you are at the root of the CleanRL project
+uv pip install ".[cloud]"
+cd cloud
+python -m awscli configure
+terraform init
+export AWS_DEFAULT_REGION=$(aws configure get region --profile default)
+terraform apply
+```
+
+<script id="asciicast-445048" src="https://asciinema.org/a/445048.js" async></script>
+
+!!! note
+    Don't worry about the cost of spinning up these AWS Batch compute environments and job queues. They are completely free and you are only charged when you submit experiments.
+
+
+Then your AWS Batch console should look like
+
+![aws_batch1.png](aws_batch1.png)
+
+
+### Clean Up
+Uninstalling/Deleting the infrastructure is pretty straightforward:
+```
+export AWS_DEFAULT_REGION=$(aws configure get region --profile default)
+terraform destroy
+```
diff --git a/cleanrl/docs/cloud/submit-experiments.md b/cleanrl/docs/cloud/submit-experiments.md
new file mode 100644
index 0000000000000000000000000000000000000000..59f8450a9d11cb36d06cb90b259fced558644099
--- /dev/null
+++ b/cleanrl/docs/cloud/submit-experiments.md
@@ -0,0 +1,123 @@
+# Submit Experiments
+
+### Inspection
+
+Dry run to inspect the generated docker command
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video" \
+    --num-seed 1
+```
+
+The generated docker command should look like
+```
+docker run -d --cpuset-cpus="0" -e WANDB_API_KEY=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx vwxyzjn/cleanrl:latest /bin/bash -c "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video --seed 1"
+```
+
+### Run on AWS
+
+Submit a job using AWS's compute-optimized spot instances 
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video" \
+    --job-queue c5a-large-spot \
+    --num-seed 1 \
+    --num-vcpu 1 \
+    --num-memory 2000 \
+    --num-hours 48.0 \
+    --provider aws
+```
+
+Submit a job using AWS's accelerated-computing spot instances 
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo_atari.py --env-id BreakoutNoFrameskip-v4 --track --capture_video" \
+    --job-queue g4dn-xlarge-spot \
+    --num-seed 1 \
+    --num-vcpu 1 \
+    --num-gpu 1 \
+    --num-memory 4000 \
+    --num-hours 48.0 \
+    --provider aws
+```
+
+Submit a job using AWS's compute-optimized on-demand instances 
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video" \
+    --job-queue c5a-large \
+    --num-seed 1 \
+    --num-vcpu 1 \
+    --num-memory 2000 \
+    --num-hours 48.0 \
+    --provider aws
+```
+
+Submit a job using AWS's accelerated-computing on-demand instances 
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo_atari.py --env-id BreakoutNoFrameskip-v4 --track --capture_video" \
+    --job-queue g4dn-xlarge \
+    --num-seed 1 \
+    --num-vcpu 1 \
+    --num-gpu 1 \
+    --num-memory 4000 \
+    --num-hours 48.0 \
+    --provider aws
+```
+
+<script id="asciicast-445050" src="https://asciinema.org/a/445050.js" async></script>
+
+Then you should see:
+
+![aws_batch1.png](aws_batch1.png)
+![aws_batch2.png](aws_batch2.png)
+
+![wandb.png](wandb.png)
+
+## Customize the Docker Container
+
+Set up docker's `buildx` and login in to your preferred registry.
+
+```
+docker buildx create --use
+docker login
+```
+
+Then you could build a container using the `--build` flag based on the `Dockerfile` in the current directory. Also, `--push` will auto-push to the docker registry.
+
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video" \
+    --build --push
+```
+
+To build a multi-arch image using `--archs linux/arm64,linux/amd64`:
+
+```
+uv run python -m cleanrl_utils.submit_exp \
+    --docker-tag vwxyzjn/cleanrl:latest \
+    --command "uv run python cleanrl/ppo.py --env-id CartPole-v1 --total-timesteps 100000 --track --capture_video" \
+    --archs linux/arm64,linux/amd64
+    --build --push
+```
+
+!!! note
+    Building an multi-arch image is quite slow but will allow you to use ARM instances such as `m6gd.medium` that is 20-70% cheaper than X86 instances.
+    However, note there is no cloud providers that give ARM instances with Nvidia's GPU (to my knowledge), so this effort might not be worth it.
+
+    If you still wants to pursue multi-arch, you can speed things up by using a native ARM server and connect it to your `buildx` instance:
+
+    ```
+    docker -H ssh://costa@gpu info
+    docker buildx create --name remote --use
+    docker buildx create --name remote --append ssh://costa@gpu
+    docker buildx inspect --bootstrap
+    python -m cleanrl_utils.submit_exp -b --archs linux/arm64,linux/amd64
+    ```
diff --git a/cleanrl/docs/contribution.md b/cleanrl/docs/contribution.md
new file mode 100644
index 0000000000000000000000000000000000000000..71521eb5db8d184c0851b6358fd214ef791e8949
--- /dev/null
+++ b/cleanrl/docs/contribution.md
@@ -0,0 +1,218 @@
+👍🎉 Thank you for taking the time to contribute! 🎉👍
+
+Feel free to open an issue or a Pull Request if you have any questions or suggestions. You can also [join our Discord](https://discord.gg/D6RCjA6sVT) and ask questions there. If you plan to work on an issue, let us know in the issue thread to avoid duplicate work.
+
+Good luck and have fun!
+
+## Development Environment Setup
+
+To setup the development environment, please clone the repository and follow the [installation docs](/get-started/installation/) and [usage docs](/get-started/basic-usage). They should help you set a working uv environment, so you have the same set-up as other contributors. Additionally, you may want to run the following command to install dev-dependencies for documentation:
+
+```bash
+uv pip install ".[docs]"
+```
+
+
+To build the documentation, you can run the following command:
+
+```bash
+uv run mkdocs serve
+```
+
+For testing, we generally recommend making a PR and let GitHub run the tests in [`.github/workflows/tests.yaml`](https://github.com/vwxyzjn/cleanrl/blob/master/.github/workflows/tests.yaml) automatically. These tests are cross-platform and run on Linux, macOS, and Windows. However, if you want to run the tests locally, you can run the following command:
+
+```bash
+uv run install --all-extras
+uv run pytest tests/.
+```
+
+
+
+Also, we use [pre-commit](https://pre-commit.com/) to helps us automate a sequence of short tasks (called pre-commit "hooks") such as code formatting. In particular, we always use the following hooks when submitting code to the main repository.
+
+* [**pyupgrade**](https://github.com/asottile/pyupgrade): pyupgrade upgrades syntax for newer versions of the language. 
+* [**isort**](https://github.com/PyCQA/isort): isort sorts imported dependencies according to their type (e.g, standard library vs third-party library) and name.
+* [**black**](https://black.readthedocs.io/en/stable/): black enforces an uniform code style across the codebase.
+* [**codespell**](https://github.com/codespell-project/codespell): codespell helps avoid common incorrect spelling.
+
+You can run the following command to run the following hooks:
+
+```bash
+uv run pre-commit run --all-files
+```
+
+which in most cases should automatically fix things as shown below: 
+
+![](static/pre-commit.png)
+
+
+
+## Contribution Process
+
+
+!!! warning
+
+    **Before opening a pull request**, please open an issue first to discuss with us since this is likely a sizable effort. Once we agree on the plan, feel free to make a PR to include the new algorithm.
+
+To ensure the validity of new features or bug fixes and prevent regressions, we use the "RLops" process. Deep Reinforcement Learning (DRL) is brittle and suffers from various reproducibility issues. Even bug fixes can sometimes lead to performance regressions (e.g., see [how a bug fix of contact force in MuJoCo results in worse performance for PPO](https://github.com/openai/gym/pull/2762#discussion_r853488897)).
+
+
+Therefore, it is essential to understand how the proposed changes impact the performance of the algorithms. Broadly, we categorize contributions into two types: 1) non-performance-impacting changes and 2) performance-impacting changes.
+
+* **non-performance-impacting changes**: this type of change does *not* impact the performance of the algorithm, such as documentation fixes ([:material-github: #282](https://github.com/vwxyzjn/cleanrl/pull/282)), renaming variables ([:material-github: #257](https://github.com/vwxyzjn/cleanrl/pull/257)), and removing unused code ([:material-github: #287](https://github.com/vwxyzjn/cleanrl/pull/287)). For this type of change, we can easily merge them without worrying too much about the consequences.
+* **performance-impacting changes**: this type of change influences the algorithm's performance. Examples include making a slight modification to the `gamma` parameter in PPO ([:material-github: #209](https://github.com/vwxyzjn/cleanrl/pull/209)), properly handling action bounds in DDPG ([:material-github: #211](https://github.com/vwxyzjn/cleanrl/pull/211)), and fixing bugs ([:material-github: #281](https://github.com/vwxyzjn/cleanrl/pull/281))
+
+
+**Non-performance-impacting changes** are relatively straightforward — we just need to make sure the tests pass and the code is formatted correctly. However, **performance-impacting changes** are more complicated. We need to make sure that the proposed changes do not lead to performance regressions. To do so, we use the "RLops" process detailed below, which is a set of procedures to help us run experiments and understand the impacts of the proposed changes.
+
+
+## RLops for Performance-impacting Changes:
+
+
+Importantly, **regardless of the slight difference in performance-impacting changes, we need to re-run the benchmark to ensure there is no regression**. To do so, we are going to leverage the `openrlbenchmark.rlops` CLI utility. This utility is a command-line interface that helps us run experiments and compare the performance of different versions of the codebase.
+
+### (Step 1) Run the benchmark
+
+Given a new feature, we create a PR and then run the benchmark experiments through [`benchmark.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl_utils/benchmark.py), such as the following:
+
+```bash
+uv pip install ".[docs, mujoco]"
+xvfb-run -a python -m cleanrl_utils.benchmark \
+    --env-ids HalfCheetah-v4 Walker2d-v4 Hopper-v4 \
+    --command "uv run python cleanrl/ddpg_continuous_action.py --track --capture_video" \
+    --num-seeds 3 \
+    --workers 1
+```
+
+under the hood, this script by default invokes an `--autotag` feature that tries to tag the the experiments with version control information, such as the git tag (e.g., `v1.0.0b2-8-g6081d30`) and the github PR number (e.g., `pr-299`). This is useful for us to compare the performance of the same algorithm across different versions.
+
+![](./rlops/tags.png)
+
+### (Step 2) Regression check
+
+Let's say our latest experiments is tagged with `pr-299`. We can then run the following command to compare its performance with our pilot experiments `rlops-pilot`. Note that the pilot experiments include all experiments before we started using RLops (i.e., `rlops-pilot` is the baseline).
+
+
+```bash
+python -m openrlbenchmark.rlops \
+    --filters '?we=openrlbenchmark&wpn=cleanrl&ceik=env_id&cen=exp_name&metric=charts/episodic_return' \
+        'ddpg_continuous_action?tag=pr-299' \
+        'ddpg_continuous_action?tag=rlops-pilot' \
+    --output-filename compare \
+    --scan-history \
+    --report
+```
+
+Here, we created a filter `'?we=openrlbenchmark&wpn=cleanrl&ceik=env_id&cen=exp_name&metric=charts/episodic_return'`, which is a query string that specifies the following:
+
+* `we`: the W&B entity name
+* `wpn`: the W&B project name
+* `ceik`: the custom key for the environment id
+* `cen`: the custom key for the experiment name
+* `metric`: the metric we are interested in
+
+So we are fetching metrics from [https://wandb.ai/openrlbenchmark/cleanrl](https://wandb.ai/openrlbenchmark/cleanrl). The environment id is stored in the `env` key, and the experiment name is stored in the `exp_name` key. The metric we are interested in is `charts/episodic_return`.
+
+
+<script async id="asciicast-570481" src="https://asciinema.org/a/570481.js"></script>
+
+
+It generates a table comparing the runtime (in minutes):
+
+|                |   openrlbenchmark/cleanrl/ddpg_continuous_action ({'tag': ['pr-299']}) |   openrlbenchmark/cleanrl/ddpg_continuous_action ({'tag': ['rlops-pilot']}) |
+|:---------------|-----------------------------------------------------------------------:|----------------------------------------------------------------------------:|
+| Hopper-v2      |                                                                49.596  |                                                                     64.7033 |
+| Walker2d-v2    |                                                                50.1181 |                                                                     64.8198 |
+| HalfCheetah-v2 |                                                                50.97   |                                                                     65.8884 |
+
+It also generates a table comparing the episodic return in the last 100 episodes:
+
+|                | openrlbenchmark/cleanrl/ddpg_continuous_action ({'tag': ['pr-299']})   | openrlbenchmark/cleanrl/ddpg_continuous_action ({'tag': ['rlops-pilot']})   |
+|:---------------|:-----------------------------------------------------------------------|:----------------------------------------------------------------------------|
+| Hopper-v2      | 1007.44 ± 148.29                                                       | 1126.37 ± 278.02                                                            |
+| Walker2d-v2    | 1661.14 ± 250.01                                                       | 1447.09 ± 260.24                                                            |
+| HalfCheetah-v2 | 10210.57 ± 196.22                                                      | 9205.65 ± 1093.88                                                           |
+
+
+!!! info
+
+    For more documentation on `openrlbenchmark.rlops`, please refer to its [documentation](https://github.com/openrlbenchmark/openrlbenchmark). 
+
+!!! tip
+
+    To make the script run faster, we can choose not to use `--scan-history` which allows wandb to sample 500 data points from the training data. This is the default behavior and is much faster.
+
+
+It also generates the following image and a wandb report.
+
+![](./rlops/rlops.png)
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/cleanrl-cache/reports/Regression-Report-ddpg_continuous_action-pr-299-rlops-pilot---VmlldzoyODkyNTcy" style="width:100%; height:500px" title="Regression Report: ddpg_continuous_action (['pr-299', 'rlops-pilot'])"></iframe>
+
+
+
+### (Step 3) Update the documentation
+
+Once we confirm there is no regression in the performance, we can update the documentation to display the new benchmark results. Run the previous command without comparing previous tags:
+
+```bash
+python -m cleanrl_utils.rlops --exp-name ddpg_continuous_action \
+    --wandb-project-name cleanrl \
+    --wandb-entity openrlbenchmark \
+    --tags 'pr-299' \
+    --env-ids HalfCheetah-v2 Walker2d-v2 Hopper-v2 \
+    --output-filename compare.png \
+    --scan-history \
+    --metric-last-n-average-window 100
+```
+
+which gives us a table like below and a `compare.png` as the learning curve.
+
+```
+               CleanRL's ddpg_continuous_action (pr-299)
+HalfCheetah-v2                         10210.57 ± 196.22
+Walker2d-v2                             1661.14 ± 250.01
+Hopper-v2                               1007.44 ± 148.29
+```
+
+We will use them to update the [experimental result section](https://github.com/vwxyzjn/cleanrl/blob/master/docs/rl-algorithms/ddpg.md#experiment-results) in the docs and replace the learning curves with the new ones. 
+
+
+![](./rlops/docs-update.png)
+
+### (Step 4) Update test cases
+
+In CleanRL, we have end-to-end test cases that at least ensure the code can run without crashing. When applicable, you should update the test cases in `tests` and CI setting at [`.github/workflows/tests.yaml`](https://github.com/vwxyzjn/cleanrl/blob/master/.github/workflows/tests.yaml).
+
+
+
+### (Step 5) Merge the PR
+
+Finally, we can merge the PR.
+
+
+## Checklist
+
+Here is a checklist of the contribution process. See [:material-github: #331](https://github.com/vwxyzjn/cleanrl/pull/331/files) as an example for the list of deliverables.
+
+- [ ] I've read the [CONTRIBUTION](https://docs.cleanrl.dev/contribution/) guide (**required**).
+- [ ] I have ensured `pre-commit run --all-files` passes (**required**).
+- [ ] I have updated the tests accordingly (if applicable).
+- [ ] I have updated the documentation and previewed the changes via `mkdocs serve`.
+    - [ ] I have explained note-worthy implementation details.
+    - [ ] I have explained the logged metrics.
+    - [ ] I have added links to the original paper and related papers.
+
+
+If you need to run benchmark experiments for a performance-impacting changes:
+
+- [ ] I have contacted @vwxyzjn to obtain access to the [openrlbenchmark W&B team](https://wandb.ai/openrlbenchmark).
+- [ ] I have used the [benchmark utility](/get-started/benchmark-utility/) to submit the tracked experiments to the [openrlbenchmark/cleanrl](https://wandb.ai/openrlbenchmark/cleanrl) W&B project, optionally with `--capture_video`.
+- [ ] I have performed RLops with `python -m openrlbenchmark.rlops`.
+    - For new feature or bug fix:
+        - [ ] I have used the RLops utility to understand the performance impact of the changes and confirmed there is no regression.
+    - For new algorithm:
+        - [ ] I have created a table comparing my results against those from reputable sources (i.e., the original paper or other reference implementation).
+    - [ ] I have added the learning curves generated by the `python -m openrlbenchmark.rlops` utility to the documentation.
+    - [ ] I have added links to the tracked experiments in W&B, generated by `python -m openrlbenchmark.rlops ....your_args... --report`,  to the documentation.
+
diff --git a/cleanrl/docs/css/custom.css b/cleanrl/docs/css/custom.css
new file mode 100644
index 0000000000000000000000000000000000000000..7d3503e492b8669c419a83c60a6150dfaa66d159
--- /dev/null
+++ b/cleanrl/docs/css/custom.css
@@ -0,0 +1,120 @@
+.termynal-comment {
+  color: #4a968f;
+  font-style: italic;
+  display: block;
+}
+
+.termy [data-termynal] {
+  white-space: pre-wrap;
+}
+
+a.external-link::after {
+  /* \00A0 is a non-breaking space
+        to make the mark be on the same line as the link
+    */
+  content: "\00A0[↪]";
+}
+
+a.internal-link::after {
+  /* \00A0 is a non-breaking space
+        to make the mark be on the same line as the link
+    */
+  content: "\00A0↪";
+}
+
+.shadow {
+  box-shadow: 5px 5px 10px #999;
+}
+
+/* Give space to lower icons so Gitter chat doesn't get on top of them */
+.md-footer-meta {
+  padding-bottom: 2em;
+}
+
+.user-list {
+  display: flex;
+  flex-wrap: wrap;
+  margin-bottom: 2rem;
+}
+
+.user-list-center {
+  justify-content: space-evenly;
+}
+
+.user {
+  margin: 1em;
+  min-width: 7em;
+}
+
+.user .avatar-wrapper {
+  width: 80px;
+  height: 80px;
+  margin: 10px auto;
+  overflow: hidden;
+  border-radius: 50%;
+  position: relative;
+}
+
+.user .avatar-wrapper img {
+  position: absolute;
+  top: 50%;
+  left: 50%;
+  transform: translate(-50%, -50%);
+}
+
+.user .title {
+  text-align: center;
+}
+
+.user .count {
+  font-size: 80%;
+  text-align: center;
+}
+
+a.announce-link:link,
+a.announce-link:visited {
+  color: #fff;
+}
+
+a.announce-link:hover {
+  color: var(--md-accent-fg-color);
+}
+
+.announce-wrapper {
+  display: flex;
+  justify-content: space-between;
+  flex-wrap: wrap;
+  align-items: center;
+}
+
+.announce-wrapper div.item {
+  display: none;
+}
+
+.announce-wrapper .sponsor-badge {
+  display: block;
+  position: absolute;
+  top: -5px;
+  right: 0;
+  font-size: 0.5rem;
+  color: #999;
+  background-color: #666;
+  border-radius: 10px;
+  padding: 0 10px;
+  z-index: 10;
+}
+
+.announce-wrapper .sponsor-image {
+  display: block;
+  border-radius: 20px;
+}
+
+.announce-wrapper>div {
+  min-height: 40px;
+  display: flex;
+  align-items: center;
+}
+
+.twitter {
+  color: #00acee;
+}
diff --git a/cleanrl/docs/get-started/basic-usage.md b/cleanrl/docs/get-started/basic-usage.md
new file mode 100644
index 0000000000000000000000000000000000000000..9722854eeba43572baf35bb17ba5f7a1c27fcdb5
--- /dev/null
+++ b/cleanrl/docs/get-started/basic-usage.md
@@ -0,0 +1,171 @@
+# Basic Usage
+
+## Two Ways to Run
+After the dependencies have been installed, there are **two ways** to run 
+the CleanRL script under the uv virtual environments.
+
+
+1. Using `uv run`:
+
+    ```bash
+    uv run python cleanrl/ppo.py \
+        --seed 1 \
+        --env-id CartPole-v0 \
+        --total-timesteps 50000
+    ```
+    <script id="asciicast-443649" src="https://asciinema.org/a/443649.js" async></script>
+
+
+2. Using `uv venv`:
+
+    1. We first activate the virtual environment by using
+    `uv venv`
+    2. Then, run any desired CleanRL script
+   
+    Attention: Each step must be executed separately!
+
+
+    ```bash
+    uv venv
+    ```
+    ```bash
+    python cleanrl/ppo.py \
+        --seed 1 \
+        --env-id CartPole-v0 \
+        --total-timesteps 50000
+    ```
+    <script id="asciicast-JL1FR00I2JNklAhMd2dwEAQuz" src="https://asciinema.org/a/JL1FR00I2JNklAhMd2dwEAQuz.js" async></script>
+
+!!! note
+
+    We recommend `uv venv` workflow for development. When the shell is activated, you should
+    be seeing a prefix like `(cleanrl-iXg02GqF-py3.9)` in your shell's prompt, which is the name
+    of the poetry's virtual environment.
+    **We will assume to run other commands (e.g. `tensorboard`) in the documentation within the poetry's shell.**
+
+
+!!! warning
+
+    If you are using NVIDIA ampere GPUs (e.g., 3060 TI), you might meet the following error
+
+    ```bash
+    NVIDIA GeForce RTX 3060 Ti with CUDA capability sm_86 is not compatible with the current PyTorch installation.
+    The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70.
+    If you want to use the NVIDIA GeForce RTX 3060 Ti GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
+
+    warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))
+    Traceback (most recent call last):
+    File "ppo_atari_envpool.py", line 240, in <module>
+        action, logprob, _, value = agent.get_action_and_value(next_obs)
+    File "ppo_atari_envpool.py", line 156, in get_action_and_value
+        hidden = self.network(x / 255.0)
+    RuntimeError: CUDA error: no kernel image is available for execution on the device
+    CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
+    For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
+    ```
+
+    This is because the `torch` wheel on PyPi is built with cuda 10.2. You would need to manually install the cuda 11.3 wheel like this:
+    ```bash
+    uv pip install torch==1.12.1 --upgrade --extra-index-url https://download.pytorch.org/whl/cu113
+    ```
+    Then, you can run the script again.
+
+## Visualize Training Metrics
+
+By default, the CleanRL scripts record all the training metrics via Tensorboard
+into the `runs` folder. So, after running the training script above, feel free to run
+
+```bash
+tensorboard --logdir runs
+```
+
+![Tensorboard](tensorboard.png)
+
+
+## Visualize the Agent's Gameplay Videos
+
+CleanRL helps record the agent's gameplay videos with a `--capture_video` flag,
+which will save the videos in the `videos/{$run_name}` folder.
+
+```bash linenums="1" hl_lines="5"
+python cleanrl/ppo.py \
+    --seed 1 \
+    --env-id CartPole-v0 \
+    --total-timesteps 50000 \
+    --capture_video
+```
+
+![videos](videos.png)
+![videos2](videos2.png)
+
+## Get Documentation
+
+You can directly obtained the documentation by using the `--help` flag.
+
+```bash
+python cleanrl/ppo.py --help
+
+usage: ppo.py [-h] [--exp-name EXP_NAME] [--env-id ENV_ID]
+              [--learning-rate LEARNING_RATE] [--seed SEED]
+              [--total-timesteps TOTAL_TIMESTEPS]
+              [--torch-deterministic [TORCH_DETERMINISTIC]] [--cuda [CUDA]]
+              [--track [TRACK]] [--wandb-project-name WANDB_PROJECT_NAME]
+              [--wandb-entity WANDB_ENTITY] [--capture_video [CAPTURE_VIDEO]]
+              [--num-envs NUM_ENVS] [--num-steps NUM_STEPS]
+              [--anneal-lr [ANNEAL_LR]] [--gae [GAE]] [--gamma GAMMA]
+              [--gae-lambda GAE_LAMBDA] [--num-minibatches NUM_MINIBATCHES]
+              [--update-epochs UPDATE_EPOCHS] [--norm-adv [NORM_ADV]]
+              [--clip-coef CLIP_COEF] [--clip-vloss [CLIP_VLOSS]]
+              [--ent-coef ENT_COEF] [--vf-coef VF_COEF]
+              [--max-grad-norm MAX_GRAD_NORM] [--target-kl TARGET_KL]
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --exp-name EXP_NAME   the name of this experiment
+  --env-id ENV_ID       the id of the environment
+  --learning-rate LEARNING_RATE
+                        the learning rate of the optimizer
+  --seed SEED           seed of the experiment
+  --total-timesteps TOTAL_TIMESTEPS
+                        total timesteps of the experiments
+  --torch-deterministic [TORCH_DETERMINISTIC]
+                        if toggled, `torch.backends.cudnn.deterministic=False`
+  --cuda [CUDA]         if toggled, cuda will be enabled by default
+  --track [TRACK]       if toggled, this experiment will be tracked with Weights
+                        and Biases
+  --wandb-project-name WANDB_PROJECT_NAME
+                        the wandb's project name
+  --wandb-entity WANDB_ENTITY
+                        the entity (team) of wandb's project
+  --capture_video [CAPTURE_VIDEO]
+                        weather to capture videos of the agent performances (check
+                        out `videos` folder)
+  --num-envs NUM_ENVS   the number of parallel game environments
+  --num-steps NUM_STEPS
+                        the number of steps to run in each environment per policy
+                        rollout
+  --anneal-lr [ANNEAL_LR]
+                        Toggle learning rate annealing for policy and value
+                        networks
+  --gae [GAE]           Use GAE for advantage computation
+  --gamma GAMMA         the discount factor gamma
+  --gae-lambda GAE_LAMBDA
+                        the lambda for the general advantage estimation
+  --num-minibatches NUM_MINIBATCHES
+                        the number of mini-batches
+  --update-epochs UPDATE_EPOCHS
+                        the K epochs to update the policy
+  --norm-adv [NORM_ADV]
+                        Toggles advantages normalization
+  --clip-coef CLIP_COEF
+                        the surrogate clipping coefficient
+  --clip-vloss [CLIP_VLOSS]
+                        Toggles whether or not to use a clipped loss for the value
+                        function, as per the paper.
+  --ent-coef ENT_COEF   coefficient of the entropy
+  --vf-coef VF_COEF     coefficient of the value function
+  --max-grad-norm MAX_GRAD_NORM
+                        the maximum norm for the gradient clipping
+  --target-kl TARGET_KL
+                        the target KL divergence threshold
+```
diff --git a/cleanrl/docs/get-started/benchmark-utility.md b/cleanrl/docs/get-started/benchmark-utility.md
new file mode 100644
index 0000000000000000000000000000000000000000..49db86d7fed52beec7f5246e54eb097331a5d367
--- /dev/null
+++ b/cleanrl/docs/get-started/benchmark-utility.md
@@ -0,0 +1,159 @@
+# Benchmark Utility
+
+CleanRL comes with a utility module `cleanrl_utils.benchmark` to help schedule and run benchmark experiments on your local machine.
+
+## Usage
+
+Try running `python -m cleanrl_utils.benchmark --help` to get the help text.
+
+```bash
+$ python -m cleanrl_utils.benchmark --help
+usage: benchmark.py [-h] --env-ids [STR
+                    [STR ...]] --command STR [--num-seeds INT]
+                    [--start-seed INT] [--workers INT]
+                    [--auto-tag | --no-auto-tag]
+                    [--slurm-template-path {None}|STR]
+                    [--slurm-gpus-per-task {None}|INT]
+                    [--slurm-total-cpus {None}|INT]
+                    [--slurm-ntasks {None}|INT] [--slurm-nodes {None}|INT]
+
+╭─ arguments ──────────────────────────────────────────────────────────────╮
+│ -h, --help                                                               │
+│     show this help message and exit                                      │
+│ --env-ids [STR [STR ...]]                                                │
+│     the ids of the environment to compare (required)                     │
+│ --command STR                                                            │
+│     the command to run (required)                                        │
+│ --num-seeds INT                                                          │
+│     the number of random seeds (default: 3)                              │
+│ --start-seed INT                                                         │
+│     the number of the starting seed (default: 1)                         │
+│ --workers INT                                                            │
+│     the number of workers to run benchmark experimenets (default: 0)     │
+│ --auto-tag, --no-auto-tag                                                │
+│     if toggled, the runs will be tagged with git tags, commit, and pull  │
+│     request number if possible (default: True)                           │
+│ --slurm-template-path {None}|STR                                         │
+│     the path to the slurm template file (see docs for more details)      │
+│     (default: None)                                                      │
+│ --slurm-gpus-per-task {None}|INT                                         │
+│     the number of gpus per task to use for slurm jobs (default: None)    │
+│ --slurm-total-cpus {None}|INT                                            │
+│     the number of gpus per task to use for slurm jobs (default: None)    │
+│ --slurm-ntasks {None}|INT                                                │
+│     the number of tasks to use for slurm jobs (default: None)            │
+│ --slurm-nodes {None}|INT                                                 │
+│     the number of nodes to use for slurm jobs (default: None)            │
+╰──────────────────────────────────────────────────────────────────────────╯
+```
+
+## Examples
+
+The following example demonstrates how to run classic control benchmark experiments.
+
+```bash
+OMP_NUM_THREADS=1 xvfb-run -a python -m cleanrl_utils.benchmark \
+    --env-ids CartPole-v1 Acrobot-v1 MountainCar-v0 \
+    --command "uv run python cleanrl/ppo.py --no_cuda --track --capture_video" \
+    --num-seeds 3 \
+    --workers 5
+```
+
+What just happened here? In principle the helps run the following commands in 5 subprocesses:
+
+```bash
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id CartPole-v1 --seed 1
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id Acrobot-v1 --seed 1
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id MountainCar-v0 --seed 1
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id CartPole-v1 --seed 2
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id Acrobot-v1 --seed 2
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id MountainCar-v0 --seed 2
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id CartPole-v1 --seed 3
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id Acrobot-v1 --seed 3
+uv run python cleanrl/ppo.py --no_cuda --track --capture_video --env-id MountainCar-v0 --seed 3
+```
+
+More specifically:
+
+1. `--env-ids CartPole-v1 Acrobot-v1 MountainCar-v0` specifies that running experiments against these three environments
+1. `--command "uv run python cleanrl/ppo.py --no_cuda --track --capture_video"` suggests running `ppo.py` with these settings:
+    * turn off GPU usage via `--no_cuda`: because `ppo.py` has such as small neural network it often runs faster on CPU only
+    * track the experiments via `--track`
+    * render the agent gameplay videos via `--capture_video`; these videos algo get saved to the tracked experiments
+        * ` xvfb-run -a` virtualizes a display for video recording, enabling these commands on a headless linux system
+1. `--num-seeds 3` suggests running the the command with 3 random seeds for each `env-id`
+1. `--workers 5` suggests at maximum using 5 subprocesses to run the experiments
+    * `OMP_NUM_THREADS=1` suggests `torch` to use only 1 thread for each subprocesses; this way we don't have processes fighting each other.
+1. `--autotag` tries to tag the the experiments with version control information, such as the git tag (e.g., `v1.0.0b2-8-g6081d30`) and the github PR number (e.g., `pr-299`). This is useful for us to compare the performance of the same algorithm across different versions.
+
+
+Note that when you run with high-throughput environments such as `envpool` or `procgen`, it's recommended to set `--workers 1` to maximuize SPS (steps per second), such as
+
+```bash
+xvfb-run -a python -m cleanrl_utils.benchmark \
+    --env-ids Pong-v5 BeamRider-v5 Breakout-v5 \
+    --command "uv run python cleanrl/ppo_atari_envpool.py --track --capture_video" \
+    --num-seeds 3 \
+    --workers 1
+```
+
+For more example usage, see [https://github.com/vwxyzjn/cleanrl/blob/master/benchmark](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark)
+
+
+## Slurm integration
+
+If you have access to a slurm cluster, you can use `cleanrl_utils.benchmark` to schedule jobs on the cluster. The following example demonstrates how to run classic control benchmark experiments on a slurm cluster.
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:3:12"
+```
+
+```
+poetry install
+OMP_NUM_THREADS=1 xvfb-run -a uv run python -m cleanrl_utils.benchmark \
+    --env-ids CartPole-v1 Acrobot-v1 MountainCar-v0 \
+    --command "uv run python cleanrl/ppo.py --no_cuda --track --capture_video" \
+    --num-seeds 3 \
+    --workers 9 \
+    --slurm-gpus-per-task 1 \
+    --slurm-ntasks 1 \
+    --slurm-total-cpus 10 \
+    --slurm-template-path benchmark/cleanrl_1gpu.slurm_template
+```
+
+Here, we have
+* `--slurm-gpus-per-task 1` suggests that each slurm job should use 1 GPU
+* `--slurm-ntasks 1` suggests that each slurm job should use 1 CPU
+* `--slurm-total-cpus 10` suggests that each slurm job should use 10 CPUs in total
+* `--slurm-template-path benchmark/cleanrl_1gpu.slurm_template` suggests that we should use the template file `benchmark/cleanrl_1gpu.slurm_template` to generate the slurm job scripts. The template file looks like this:
+
+``` title="benchmark/cleanrl_1gpu.slurm_template" linenums="1"
+--8<-- "benchmark/cleanrl_1gpu.slurm_template"
+```
+
+The utility will generate a slurm script based on the template file and submit the job to the cluster. The generated slurm script looks like this:
+
+```
+#!/bin/bash
+#SBATCH --job-name=low-priority
+#SBATCH --partition=production-cluster
+#SBATCH --gpus-per-task=1
+#SBATCH --cpus-per-gpu=10
+#SBATCH --ntasks=1
+#SBATCH --output=slurm/logs/%x_%j.out
+#SBATCH --array=0-8%9
+#SBATCH --mem-per-cpu=12G
+#SBATCH --exclude=ip-26-0-147-[245,247],ip-26-0-156-239
+##SBATCH --nodelist=ip-26-0-156-13
+
+
+env_ids=(CartPole-v1 Acrobot-v1 MountainCar-v0)
+seeds=(1 2 3)
+env_id=${env_ids[$SLURM_ARRAY_TASK_ID / 3]}
+seed=${seeds[$SLURM_ARRAY_TASK_ID % 3]}
+
+echo "Running task $SLURM_ARRAY_TASK_ID with env_id: $env_id and seed: $seed"
+
+srun uv run python cleanrl/ppo.py --no_cuda --track --env-id $env_id --seed $seed # 
+```
+
diff --git a/cleanrl/docs/get-started/colab-badge.svg b/cleanrl/docs/get-started/colab-badge.svg
new file mode 100644
index 0000000000000000000000000000000000000000..e5830d5332975c03acc2a9715bd880097083e91d
--- /dev/null
+++ b/cleanrl/docs/get-started/colab-badge.svg
@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="117" height="20"><linearGradient id="b" x2="0" y2="100%"><stop offset="0" stop-color="#bbb" stop-opacity=".1"/><stop offset="1" stop-opacity=".1"/></linearGradient><clipPath id="a"><rect width="117" height="20" rx="3" fill="#fff"/></clipPath><g clip-path="url(#a)"><path fill="#555" d="M0 0h30v20H0z"/><path fill="#007ec6" d="M30 0h87v20H30z"/><path fill="url(#b)" d="M0 0h117v20H0z"/></g><g fill="#fff" text-anchor="middle" font-family="DejaVu Sans,Verdana,Geneva,sans-serif" font-size="110"><svg x="4px" y="0px" width="22px" height="20px" viewBox="-2 0 28 24" style="background-color: #fff;border-radius: 1px;"><path style="fill:#e8710a;" d="M1.977,16.77c-2.667-2.277-2.605-7.079,0-9.357C2.919,8.057,3.522,9.075,4.49,9.691c-1.152,1.6-1.146,3.201-0.004,4.803C3.522,15.111,2.918,16.126,1.977,16.77z"/><path style="fill:#f9ab00;" d="M12.257,17.114c-1.767-1.633-2.485-3.658-2.118-6.02c0.451-2.91,2.139-4.893,4.946-5.678c2.565-0.718,4.964-0.217,6.878,1.819c-0.884,0.743-1.707,1.547-2.434,2.446C18.488,8.827,17.319,8.435,16,8.856c-2.404,0.767-3.046,3.241-1.494,5.644c-0.241,0.275-0.493,0.541-0.721,0.826C13.295,15.939,12.511,16.3,12.257,17.114z"/><path style="fill:#e8710a;" d="M19.529,9.682c0.727-0.899,1.55-1.703,2.434-2.446c2.703,2.783,2.701,7.031-0.005,9.764c-2.648,2.674-6.936,2.725-9.701,0.115c0.254-0.814,1.038-1.175,1.528-1.788c0.228-0.285,0.48-0.552,0.721-0.826c1.053,0.916,2.254,1.268,3.6,0.83C20.502,14.551,21.151,11.927,19.529,9.682z"/><path style="fill:#f9ab00;" d="M4.49,9.691C3.522,9.075,2.919,8.057,1.977,7.413c2.209-2.398,5.721-2.942,8.476-1.355c0.555,0.32,0.719,0.606,0.285,1.128c-0.157,0.188-0.258,0.422-0.391,0.631c-0.299,0.47-0.509,1.067-0.929,1.371C8.933,9.539,8.523,8.847,8.021,8.746C6.673,8.475,5.509,8.787,4.49,9.691z"/><path style="fill:#f9ab00;" d="M1.977,16.77c0.941-0.644,1.545-1.659,2.509-2.277c1.373,1.152,2.85,1.433,4.45,0.499c0.332-0.194,0.503-0.088,0.673,0.19c0.386,0.635,0.753,1.285,1.181,1.89c0.34,0.48,0.222,0.715-0.253,1.006C7.84,19.73,4.205,19.188,1.977,16.77z"/></svg><text x="245" y="140" transform="scale(.1)" textLength="30"> </text><text x="725" y="150" fill="#010101" fill-opacity=".3" transform="scale(.1)" textLength="770">Open in Colab</text><text x="725" y="140" transform="scale(.1)" textLength="770">Open in Colab</text></g> </svg>
diff --git a/cleanrl/docs/get-started/examples.md b/cleanrl/docs/get-started/examples.md
new file mode 100644
index 0000000000000000000000000000000000000000..c2c62fffff8059c0f07badb664122c120d07a2f2
--- /dev/null
+++ b/cleanrl/docs/get-started/examples.md
@@ -0,0 +1,51 @@
+# Examples
+
+## Atari
+```
+uv venv
+
+uv pip install ".[atari]"
+python cleanrl/dqn_atari.py --env-id BreakoutNoFrameskip-v4
+python cleanrl/c51_atari.py --env-id BreakoutNoFrameskip-v4
+python cleanrl/ppo_atari.py --env-id BreakoutNoFrameskip-v4
+python cleanrl/sac_atari.py --env-id BreakoutNoFrameskip-v4
+
+# NEW: 3-4x side-effects free speed up with envpool's atari (only available to linux)
+uv pip install ".[envpool]"
+python cleanrl/ppo_atari_envpool.py --env-id BreakoutNoFrameskip-v4
+# Learn Pong-v5 in ~5-10 mins
+# Side effects such as lower sample efficiency might occur
+uv run python ppo_atari_envpool.py --clip-coef=0.2 --num-envs=16 --num-minibatches=8 --num-steps=128 --update-epochs=3
+```
+### Demo
+
+<script id="asciicast-443625" src="https://asciinema.org/a/443625.js" async></script>
+
+You can also run training scripts in other games, such as:
+
+## Classic Control
+```
+uv venv
+
+python cleanrl/dqn.py --env-id CartPole-v1
+python cleanrl/ppo.py --env-id CartPole-v1
+python cleanrl/c51.py --env-id CartPole-v1
+```
+
+## Procgen 
+```
+uv venv
+
+uv pip install ".[procgen]"
+python cleanrl/ppo_procgen.py --env-id starpilot
+python cleanrl/ppg_procgen.py --env-id starpilot
+```
+
+
+## PPO + LSTM
+```
+uv venv
+
+uv pip install ".[atari]"
+python cleanrl/ppo_atari_lstm.py --env-id BreakoutNoFrameskip-v4
+```
diff --git a/cleanrl/docs/get-started/experiment-tracking.md b/cleanrl/docs/get-started/experiment-tracking.md
new file mode 100644
index 0000000000000000000000000000000000000000..94b5b85926a3ebd4f89cf1685c99e0554997bd9e
--- /dev/null
+++ b/cleanrl/docs/get-started/experiment-tracking.md
@@ -0,0 +1,24 @@
+# Experiment tracking
+
+To use experiment tracking with wandb, run with the `--track` flag, which will also
+upload the videos recorded by the `--capture_video` flag.
+```bash
+uv venv
+wandb login # only required for the first time
+python cleanrl/ppo.py --track --capture_video
+```
+
+
+<script id="asciicast-443626" src="https://asciinema.org/a/443626.js" async></script>
+
+The console will output the url for the tracked experiment like the following
+
+```bash
+wandb:  View project at https://wandb.ai/costa-huang/cleanRL                            
+wandb:  View run at https://wandb.ai/costa-huang/cleanRL/runs/10dwbgeh
+```
+
+When you open the URL, it's going to look like the following page:
+
+<iframe src="https://wandb.ai/costa-huang/cleanRL/runs/10dwbgeh" style="width:100%; height:1000px" title="CleanRL CartPole-v1 Example"></iframe>
+
diff --git a/cleanrl/docs/get-started/installation.md b/cleanrl/docs/get-started/installation.md
new file mode 100644
index 0000000000000000000000000000000000000000..9afc83594b03dffa59787102b94a8387cd21b4c2
--- /dev/null
+++ b/cleanrl/docs/get-started/installation.md
@@ -0,0 +1,67 @@
+# Installation
+
+## Prerequisites
+
+* Python >=3.7.1,<3.11
+* [uv 0.7.19+](https://docs.astral.sh/uv/)
+
+Simply run the following command for a quick start
+
+```bash
+git clone https://github.com/vwxyzjn/cleanrl.git && cd cleanrl
+uv pip install .
+```
+
+<script id="asciicast-443647" src="https://asciinema.org/a/443647.js" async></script>
+
+
+!!! note "Working with different CUDA versions for `torch`"
+
+    By default, the `torch` wheel is built with CUDA 10.2. If you are using newer NVIDIA GPUs (e.g., 3060 TI), you may need to specifically install CUDA 11.3 wheels by overriding the `torch` dependency with `pip`:
+
+    ```bash
+    uv pip install "torch==1.12.1" --upgrade --extra-index-url https://download.pytorch.org/whl/cu113
+    ```
+
+
+## Install via `pip`
+
+While we recommend using `uv` to manage environments and dependencies, the traditional `requirements.txt` are available:
+
+```bash
+# core dependencies
+pip install -r requirements/requirements.txt
+
+# optional dependencies
+pip install -r requirements/requirements-atari.txt
+pip install -r requirements/requirements-mujoco.txt
+pip install -r requirements/requirements-mujoco_py.txt
+pip install -r requirements/requirements-procgen.txt
+pip install -r requirements/requirements-envpool.txt
+pip install -r requirements/requirements-pettingzoo.txt
+pip install -r requirements/requirements-jax.txt
+pip install -r requirements/requirements-docs.txt
+pip install -r requirements/requirements-cloud.txt
+```
+
+
+## Optional Dependencies
+
+CleanRL makes it easy to install optional dependencies for common RL environments
+and various development utilities. These optional dependencies are defined at the
+[`pyproject.toml`](https://github.com/vwxyzjn/cleanrl/blob/6afb51624a6fd51775b8351dd25099bd778cb1b1/pyproject.toml#L22-L37) as optional dependencies
+
+You can install them using the following command
+
+```bash
+uv pip install ".[atari]"
+uv pip install ".[mujoco]"
+uv pip install ".[dm_control]"
+uv pip install ".[procgen]"
+uv pip install ".[envpool]"
+uv pip install ".[pettingzoo]"
+uv pip install ".[jax]"
+uv pip install ".[optuna]"
+uv pip install ".[docs]"
+uv pip install ".[cloud]"
+```
diff --git a/cleanrl/docs/get-started/zoo.md b/cleanrl/docs/get-started/zoo.md
new file mode 100644
index 0000000000000000000000000000000000000000..29c590051be63135e6d33a405ad3851ea6610e61
--- /dev/null
+++ b/cleanrl/docs/get-started/zoo.md
@@ -0,0 +1,57 @@
+# 🤗 Model Zoo
+
+[<img src="https://img.shields.io/badge/%F0%9F%A4%97%20Models-Huggingface-F8D521">](https://huggingface.co/cleanrl)
+[![Open In Colab](https://github.com/vwxyzjn/cleanrl/raw/master/docs/get-started/colab-badge.svg)](https://colab.research.google.com/github/vwxyzjn/cleanrl/blob/master/docs/get-started/CleanRL_Huggingface_Integration_Demo.ipynb)
+
+CleanRL now has 🧪 experimental support for saving and loading models from 🤗 HuggingFace's [Model Hub](https://huggingface.co/models). We are rolling out this feature in phases, and currently only support saving and loading models from the following algorithm variants:
+
+
+| Algorithm      | Variants Implemented |
+| ----------- | ----------- |
+| ✅ [Deep Q-Learning (DQN)](https://web.stanford.edu/class/psych209/Readings/MnihEtAlHassibis15NatureControlDeepRL.pdf) | :material-github: [`dqn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqnpy) |
+| | :material-github: [`dqn_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_ataripy) |
+| | :material-github: [`dqn_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_jaxpy) |
+| | :material-github: [`dqn_atari_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_atari_jaxpy) |
+| ✅ [Categorical DQN (C51)](https://arxiv.org/pdf/1707.06887.pdf) | :material-github: [`c51.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51.py), :material-file-document: [docs](/rl-algorithms/c51/#c51py) |
+| | :material-github: [`c51_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_atari.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_ataripy) |
+| | :material-github: [`c51_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_jax.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_jaxpy) |
+| | :material-github: [`c51_atari_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_atari_jax.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_atari_jaxpy) |
+| ✅ [Deep Deterministic Policy Gradient (DDPG)](https://arxiv.org/pdf/1509.02971.pdf) | :material-github: [`ddpg_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action.py), :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_actionpy) |
+| | :material-github: [`ddpg_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action_jax.py),  :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_action_jaxpy)
+| ✅ [Twin Delayed Deep Deterministic Policy Gradient (TD3)](https://arxiv.org/pdf/1802.09477.pdf) | :material-github: [`td3_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_actionpy) |
+|  | :material-github: [`td3_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action_jax.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_action_jaxpy) |
+
+
+## Load models from the Model Hub
+
+We have a simple utility `enjoy.py` to load models from the hub and run them in an environment. We currently support the following commands:
+
+```bash
+uv pip install ".[dqn]"
+uv run python -m cleanrl_utils.enjoy --exp-name dqn --env-id CartPole-v1
+uv pip install ".[dqn, jax]"
+uv run python -m cleanrl_utils.enjoy --exp-name dqn_jax --env-id CartPole-v1
+
+uv pip install ".[atari]"
+uv run python -m cleanrl_utils.enjoy --exp-name dqn_atari --env-id BreakoutNoFrameskip-v4
+uv pip install ".[atari, jax]"
+uv run python -m cleanrl_utils.enjoy --exp-name dqn_atari_jax --env-id BreakoutNoFrameskip-v4
+```
+
+To see a list of supported models, please visit 🤗 [https://huggingface.co/cleanrl](https://huggingface.co/cleanrl).
+
+
+???+ info "What happens under the hood?"
+    
+    The `cleanrl_utils.enjoy` is a simple wrapper to load the models from the hub and run them in an environment. A minimal version of the script can be found at [cleanrl_utils/evals/dqn_eval.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl_utils/evals/dqn_eval.py), which may give you a more fine-grained control and access to the model.
+    
+    <script src="https://emgithub.com/embed-v2.js?target=https%3A%2F%2Fgithub.com%2Fvwxyzjn%2Fcleanrl%2Fblob%2Fmaster%2Fcleanrl_utils%2Fevals%2Fdqn_eval.py&style=github&type=code&showBorder=on&showLineNumbers=on&showFileMeta=on&showFullPath=on&showCopy=on"></script>
+
+## Save model to Model Hub
+
+In the supported algorithm variants, you can run the script with the `--save-model` flag, which saves a model to the `runs` folder, and the `--upload-model` flag, which upload the model to huggingface under your default entity (username). Optionally, you may override the default entity with `--hf-entity` flag.
+
+```bash
+uv run python cleanrl/dqn_jax.py --env-id CartPole-v1 --save-model --upload-model # --hf-entity cleanrl
+uv run python cleanrl/dqn_atari_jax.py --env-id SeaquestNoFrameskip-v4  --save-model --upload-model # --hf-entity cleanrl
+```
diff --git a/cleanrl/docs/js/custom.js b/cleanrl/docs/js/custom.js
new file mode 100644
index 0000000000000000000000000000000000000000..8e3be4c130704a12174ac56f7a325706bc4d7a80
--- /dev/null
+++ b/cleanrl/docs/js/custom.js
@@ -0,0 +1,180 @@
+const div = document.querySelector('.github-topic-projects')
+
+async function getDataBatch(page) {
+    const response = await fetch(`https://api.github.com/search/repositories?q=topic:fastapi&per_page=100&page=${page}`, { headers: { Accept: 'application/vnd.github.mercy-preview+json' } })
+    const data = await response.json()
+    return data
+}
+
+async function getData() {
+    let page = 1
+    let data = []
+    let dataBatch = await getDataBatch(page)
+    data = data.concat(dataBatch.items)
+    const totalCount = dataBatch.total_count
+    while (data.length < totalCount) {
+        page += 1
+        dataBatch = await getDataBatch(page)
+        data = data.concat(dataBatch.items)
+    }
+    return data
+}
+
+function setupTermynal() {
+    document.querySelectorAll(".use-termynal").forEach(node => {
+        node.style.display = "block";
+        new Termynal(node, {
+            lineDelay: 500
+        });
+    });
+    const progressLiteralStart = "---> 100%";
+    const promptLiteralStart = "$ ";
+    const customPromptLiteralStart = "# ";
+    const termynalActivateClass = "termy";
+    let termynals = [];
+
+    function createTermynals() {
+        document
+            .querySelectorAll(`.${termynalActivateClass} .highlight`)
+            .forEach(node => {
+                const text = node.textContent;
+                const lines = text.split("\n");
+                const useLines = [];
+                let buffer = [];
+                function saveBuffer() {
+                    if (buffer.length) {
+                        let isBlankSpace = true;
+                        buffer.forEach(line => {
+                            if (line) {
+                                isBlankSpace = false;
+                            }
+                        });
+                        dataValue = {};
+                        if (isBlankSpace) {
+                            dataValue["delay"] = 0;
+                        }
+                        if (buffer[buffer.length - 1] === "") {
+                            // A last single <br> won't have effect
+                            // so put an additional one
+                            buffer.push("");
+                        }
+                        const bufferValue = buffer.join("<br>");
+                        dataValue["value"] = bufferValue;
+                        useLines.push(dataValue);
+                        buffer = [];
+                    }
+                }
+                for (let line of lines) {
+                    if (line === progressLiteralStart) {
+                        saveBuffer();
+                        useLines.push({
+                            type: "progress"
+                        });
+                    } else if (line.startsWith(promptLiteralStart)) {
+                        saveBuffer();
+                        const value = line.replace(promptLiteralStart, "").trimEnd();
+                        useLines.push({
+                            type: "input",
+                            value: value
+                        });
+                    } else if (line.startsWith("// ")) {
+                        saveBuffer();
+                        const value = "💬 " + line.replace("// ", "").trimEnd();
+                        useLines.push({
+                            value: value,
+                            class: "termynal-comment",
+                            delay: 0
+                        });
+                    } else if (line.startsWith(customPromptLiteralStart)) {
+                        saveBuffer();
+                        const promptStart = line.indexOf(promptLiteralStart);
+                        if (promptStart === -1) {
+                            console.error("Custom prompt found but no end delimiter", line)
+                        }
+                        const prompt = line.slice(0, promptStart).replace(customPromptLiteralStart, "")
+                        let value = line.slice(promptStart + promptLiteralStart.length);
+                        useLines.push({
+                            type: "input",
+                            value: value,
+                            prompt: prompt
+                        });
+                    } else {
+                        buffer.push(line);
+                    }
+                }
+                saveBuffer();
+                const div = document.createElement("div");
+                node.replaceWith(div);
+                const termynal = new Termynal(div, {
+                    lineData: useLines,
+                    noInit: true,
+                    lineDelay: 500
+                });
+                termynals.push(termynal);
+            });
+    }
+
+    function loadVisibleTermynals() {
+        termynals = termynals.filter(termynal => {
+            if (termynal.container.getBoundingClientRect().top - innerHeight <= 0) {
+                termynal.init();
+                return false;
+            }
+            return true;
+        });
+    }
+    window.addEventListener("scroll", loadVisibleTermynals);
+    createTermynals();
+    loadVisibleTermynals();
+}
+
+function shuffle(array) {
+    var currentIndex = array.length, temporaryValue, randomIndex;
+    while (0 !== currentIndex) {
+        randomIndex = Math.floor(Math.random() * currentIndex);
+        currentIndex -= 1;
+        temporaryValue = array[currentIndex];
+        array[currentIndex] = array[randomIndex];
+        array[randomIndex] = temporaryValue;
+    }
+    return array;
+}
+
+async function showRandomAnnouncement(groupId, timeInterval) {
+    const announceFastAPI = document.getElementById(groupId);
+    if (announceFastAPI) {
+        let children = [].slice.call(announceFastAPI.children);
+        children = shuffle(children)
+        let index = 0
+        const announceRandom = () => {
+            children.forEach((el, i) => {el.style.display = "none"});
+            children[index].style.display = "block"
+            index = (index + 1) % children.length
+        }
+        announceRandom()
+        setInterval(announceRandom, timeInterval
+        )
+    }
+}
+
+async function main() {
+    if (div) {
+        data = await getData()
+        div.innerHTML = '<ul></ul>'
+        const ul = document.querySelector('.github-topic-projects ul')
+        data.forEach(v => {
+            if (v.full_name === 'tiangolo/fastapi') {
+                return
+            }
+            const li = document.createElement('li')
+            li.innerHTML = `<a href="${v.html_url}" target="_blank">★ ${v.stargazers_count} - ${v.full_name}</a> by <a href="${v.owner.html_url}" target="_blank">@${v.owner.login}</a>`
+            ul.append(li)
+        })
+    }
+
+    setupTermynal();
+    showRandomAnnouncement('announce-left', 5000)
+    showRandomAnnouncement('announce-right', 10000)
+}
+
+main()
diff --git a/cleanrl/docs/js/termynal.js b/cleanrl/docs/js/termynal.js
new file mode 100644
index 0000000000000000000000000000000000000000..8b0e9339e8772022d0e89babbf922835a643eead
--- /dev/null
+++ b/cleanrl/docs/js/termynal.js
@@ -0,0 +1,264 @@
+/**
+ * termynal.js
+ * A lightweight, modern and extensible animated terminal window, using
+ * async/await.
+ *
+ * @author Ines Montani <ines@ines.io>
+ * @version 0.0.1
+ * @license MIT
+ */
+
+'use strict';
+
+/** Generate a terminal widget. */
+class Termynal {
+    /**
+     * Construct the widget's settings.
+     * @param {(string|Node)=} container - Query selector or container element.
+     * @param {Object=} options - Custom settings.
+     * @param {string} options.prefix - Prefix to use for data attributes.
+     * @param {number} options.startDelay - Delay before animation, in ms.
+     * @param {number} options.typeDelay - Delay between each typed character, in ms.
+     * @param {number} options.lineDelay - Delay between each line, in ms.
+     * @param {number} options.progressLength - Number of characters displayed as progress bar.
+     * @param {string} options.progressChar – Character to use for progress bar, defaults to █.
+	 * @param {number} options.progressPercent - Max percent of progress.
+     * @param {string} options.cursor – Character to use for cursor, defaults to ▋.
+     * @param {Object[]} lineData - Dynamically loaded line data objects.
+     * @param {boolean} options.noInit - Don't initialise the animation.
+     */
+    constructor(container = '#termynal', options = {}) {
+        this.container = (typeof container === 'string') ? document.querySelector(container) : container;
+        this.pfx = `data-${options.prefix || 'ty'}`;
+        this.originalStartDelay = this.startDelay = options.startDelay
+            || parseFloat(this.container.getAttribute(`${this.pfx}-startDelay`)) || 600;
+        this.originalTypeDelay = this.typeDelay = options.typeDelay
+            || parseFloat(this.container.getAttribute(`${this.pfx}-typeDelay`)) || 90;
+        this.originalLineDelay = this.lineDelay = options.lineDelay
+            || parseFloat(this.container.getAttribute(`${this.pfx}-lineDelay`)) || 1500;
+        this.progressLength = options.progressLength
+            || parseFloat(this.container.getAttribute(`${this.pfx}-progressLength`)) || 40;
+        this.progressChar = options.progressChar
+            || this.container.getAttribute(`${this.pfx}-progressChar`) || '█';
+		this.progressPercent = options.progressPercent
+            || parseFloat(this.container.getAttribute(`${this.pfx}-progressPercent`)) || 100;
+        this.cursor = options.cursor
+            || this.container.getAttribute(`${this.pfx}-cursor`) || '▋';
+        this.lineData = this.lineDataToElements(options.lineData || []);
+        this.loadLines()
+        if (!options.noInit) this.init()
+    }
+
+    loadLines() {
+        // Load all the lines and create the container so that the size is fixed
+        // Otherwise it would be changing and the user viewport would be constantly
+        // moving as she/he scrolls
+        const finish = this.generateFinish()
+        finish.style.visibility = 'hidden'
+        this.container.appendChild(finish)
+        // Appends dynamically loaded lines to existing line elements.
+        this.lines = [...this.container.querySelectorAll(`[${this.pfx}]`)].concat(this.lineData);
+        for (let line of this.lines) {
+            line.style.visibility = 'hidden'
+            this.container.appendChild(line)
+        }
+        const restart = this.generateRestart()
+        restart.style.visibility = 'hidden'
+        this.container.appendChild(restart)
+        this.container.setAttribute('data-termynal', '');
+    }
+
+    /**
+     * Initialise the widget, get lines, clear container and start animation.
+     */
+    init() {
+        /** 
+         * Calculates width and height of Termynal container.
+         * If container is empty and lines are dynamically loaded, defaults to browser `auto` or CSS.
+         */ 
+        const containerStyle = getComputedStyle(this.container);
+        this.container.style.width = containerStyle.width !== '0px' ? 
+            containerStyle.width : undefined;
+        this.container.style.minHeight = containerStyle.height !== '0px' ? 
+            containerStyle.height : undefined;
+
+        this.container.setAttribute('data-termynal', '');
+        this.container.innerHTML = '';
+        for (let line of this.lines) {
+            line.style.visibility = 'visible'
+        }
+        this.start();
+    }
+
+    /**
+     * Start the animation and rener the lines depending on their data attributes.
+     */
+    async start() {
+        this.addFinish()
+        await this._wait(this.startDelay);
+
+        for (let line of this.lines) {
+            const type = line.getAttribute(this.pfx);
+            const delay = line.getAttribute(`${this.pfx}-delay`) || this.lineDelay;
+
+            if (type == 'input') {
+                line.setAttribute(`${this.pfx}-cursor`, this.cursor);
+                await this.type(line);
+                await this._wait(delay);
+            }
+
+            else if (type == 'progress') {
+                await this.progress(line);
+                await this._wait(delay);
+            }
+
+            else {
+                this.container.appendChild(line);
+                await this._wait(delay);
+            }
+
+            line.removeAttribute(`${this.pfx}-cursor`);
+        }
+        this.addRestart()
+        this.finishElement.style.visibility = 'hidden'
+        this.lineDelay = this.originalLineDelay
+        this.typeDelay = this.originalTypeDelay
+        this.startDelay = this.originalStartDelay
+    }
+
+    generateRestart() {
+        const restart = document.createElement('a')
+        restart.onclick = (e) => {
+            e.preventDefault()
+            this.container.innerHTML = ''
+            this.init()
+        }
+        restart.href = '#'
+        restart.setAttribute('data-terminal-control', '')
+        restart.innerHTML = "restart ↻"
+        return restart
+    }
+    
+    generateFinish() {
+        const finish = document.createElement('a')
+        finish.onclick = (e) => {
+            e.preventDefault()
+            this.lineDelay = 0
+            this.typeDelay = 0
+            this.startDelay = 0
+        }
+        finish.href = '#'
+        finish.setAttribute('data-terminal-control', '')
+        finish.innerHTML = "fast →"
+        this.finishElement = finish
+        return finish
+    }
+
+    addRestart() {
+        const restart = this.generateRestart()
+        this.container.appendChild(restart)
+    }
+
+    addFinish() {
+        const finish = this.generateFinish()
+        this.container.appendChild(finish)
+    }
+
+    /**
+     * Animate a typed line.
+     * @param {Node} line - The line element to render.
+     */
+    async type(line) {
+        const chars = [...line.textContent];
+        line.textContent = '';
+        this.container.appendChild(line);
+
+        for (let char of chars) {
+            const delay = line.getAttribute(`${this.pfx}-typeDelay`) || this.typeDelay;
+            await this._wait(delay);
+            line.textContent += char;
+        }
+    }
+
+    /**
+     * Animate a progress bar.
+     * @param {Node} line - The line element to render.
+     */
+    async progress(line) {
+        const progressLength = line.getAttribute(`${this.pfx}-progressLength`)
+            || this.progressLength;
+        const progressChar = line.getAttribute(`${this.pfx}-progressChar`)
+            || this.progressChar;
+        const chars = progressChar.repeat(progressLength);
+		const progressPercent = line.getAttribute(`${this.pfx}-progressPercent`)
+			|| this.progressPercent;
+        line.textContent = '';
+        this.container.appendChild(line);
+
+        for (let i = 1; i < chars.length + 1; i++) {
+            await this._wait(this.typeDelay);
+            const percent = Math.round(i / chars.length * 100);
+            line.textContent = `${chars.slice(0, i)} ${percent}%`;
+			if (percent>progressPercent) {
+				break;
+			}
+        }
+    }
+
+    /**
+     * Helper function for animation delays, called with `await`.
+     * @param {number} time - Timeout, in ms.
+     */
+    _wait(time) {
+        return new Promise(resolve => setTimeout(resolve, time));
+    }
+
+    /**
+     * Converts line data objects into line elements.
+     * 
+     * @param {Object[]} lineData - Dynamically loaded lines.
+     * @param {Object} line - Line data object.
+     * @returns {Element[]} - Array of line elements.
+     */
+    lineDataToElements(lineData) {
+        return lineData.map(line => {
+            let div = document.createElement('div');
+            div.innerHTML = `<span ${this._attributes(line)}>${line.value || ''}</span>`;
+
+            return div.firstElementChild;
+        });
+    }
+
+    /**
+     * Helper function for generating attributes string.
+     * 
+     * @param {Object} line - Line data object.
+     * @returns {string} - String of attributes.
+     */
+    _attributes(line) {
+        let attrs = '';
+        for (let prop in line) {
+            // Custom add class
+            if (prop === 'class') {
+                attrs += ` class=${line[prop]} `
+                continue
+            }
+            if (prop === 'type') {
+                attrs += `${this.pfx}="${line[prop]}" `
+            } else if (prop !== 'value') {
+                attrs += `${this.pfx}-${prop}="${line[prop]}" `
+            }
+        }
+
+        return attrs;
+    }
+}
+
+/**
+* HTML API: If current script has container(s) specified, initialise Termynal.
+*/
+if (document.currentScript.hasAttribute('data-termynal-container')) {
+    const containers = document.currentScript.getAttribute('data-termynal-container');
+    containers.split('|')
+        .forEach(container => new Termynal(container))
+}
diff --git a/cleanrl/docs/rl-algorithms/overview.md b/cleanrl/docs/rl-algorithms/overview.md
new file mode 100644
index 0000000000000000000000000000000000000000..6c2d94ea85acc85a0fcd2089e4fd1d70e2dabccb
--- /dev/null
+++ b/cleanrl/docs/rl-algorithms/overview.md
@@ -0,0 +1,33 @@
+# Overview
+
+| Algorithm      | Variants Implemented |
+| ----------- | ----------- |
+| ✅ [Proximal Policy Gradient (PPO)](https://arxiv.org/pdf/1707.06347.pdf)  | :material-github: [`ppo.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppopy) |
+| | :material-github: [`ppo_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_ataripy)
+| | :material-github: [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_actionpy)
+| | :material-github: [`ppo_atari_lstm.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_lstmpy)
+| | :material-github: [`ppo_atari_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpoolpy)
+| | :material-github: [`ppo_atari_envpool_xla_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy)
+| | :material-github: [`ppo_atari_envpool_xla_jax_scan.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jax_scanpy)
+| | :material-github: [`ppo_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_procgenpy)
+| | :material-github: [`ppo_atari_multigpu.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_multigpupy)
+| | :material-github: [`ppo_pettingzoo_ma_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_pettingzoo_ma_ataripy)
+| | :material-github: [`ppo_continuous_action_isaacgym.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action_isaacgym/ppo_continuous_action_isaacgym.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_action_isaacgympy)
+| ✅ [Deep Q-Learning (DQN)](https://web.stanford.edu/class/psych209/Readings/MnihEtAlHassibis15NatureControlDeepRL.pdf) | :material-github: [`dqn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqnpy) |
+| | :material-github: [`dqn_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_ataripy) |
+| | :material-github: [`dqn_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_jaxpy) |
+| | :material-github: [`dqn_atari_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/dqn_atari_jax.py), :material-file-document: [docs](/rl-algorithms/dqn/#dqn_atari_jaxpy) |
+| ✅ [Categorical DQN (C51)](https://arxiv.org/pdf/1707.06887.pdf) | :material-github: [`c51.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51.py), :material-file-document: [docs](/rl-algorithms/c51/#c51py) |
+| | :material-github: [`c51_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_atari.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_ataripy) |
+| | :material-github: [`c51_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_jax.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_jaxpy) |
+| | :material-github: [`c51_atari_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/c51_atari_jax.py), :material-file-document: [docs](/rl-algorithms/c51/#c51_atari_jaxpy) |
+| ✅ [Soft Actor-Critic (SAC)](https://arxiv.org/pdf/1812.05905.pdf) | :material-github: [`sac_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/sac_continuous_action.py), :material-file-document: [docs](/rl-algorithms/sac/#sac_continuous_actionpy) |
+| | :material-github: [`sac_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/sac_atari.py), :material-file-document: [docs](https://docs.cleanrl.dev/rl-algorithms/sac/#sac_atarinpy) |
+| ✅ [Deep Deterministic Policy Gradient (DDPG)](https://arxiv.org/pdf/1509.02971.pdf) | :material-github: [`ddpg_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action.py), :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_actionpy) |
+| | :material-github: [`ddpg_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ddpg_continuous_action_jax.py),  :material-file-document: [docs](/rl-algorithms/ddpg/#ddpg_continuous_action_jaxpy)
+| ✅ [Twin Delayed Deep Deterministic Policy Gradient (TD3)](https://arxiv.org/pdf/1802.09477.pdf) | :material-github: [`td3_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_actionpy) |
+|  | :material-github: [`td3_continuous_action_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/td3_continuous_action_jax.py), :material-file-document: [docs](/rl-algorithms/td3/#td3_continuous_action_jaxpy) |
+| ✅ [Phasic Policy Gradient (PPG)](https://arxiv.org/abs/2009.04416) | :material-github: [`ppg_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppg_procgen.py), :material-file-document: [docs](/rl-algorithms/ppg/#ppg_procgenpy) |
+| ✅ [Random Network Distillation (RND)](https://arxiv.org/abs/1810.12894) | :material-github: [`ppo_rnd_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_rnd_envpool.py), :material-file-document: [docs](/rl-algorithms/ppo-rnd/#ppo_rnd_envpoolpy) |
+| ✅ [Qdagger](https://arxiv.org/abs/2206.01626) | :material-github: [`qdagger_dqn_atari_impalacnn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_impalacnn.py), :material-file-document: [docs](/rl-algorithms/qdagger/#qdagger_dqn_atari_impalacnnpy) |
+|  | :material-github: [`qdagger_dqn_atari_jax_impalacnn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_jax_impalacnn.py), :material-file-document: [docs](/rl-algorithms/qdagger/#qdagger_dqn_atari_jax_impalacnnpy) |
\ No newline at end of file
diff --git a/cleanrl/docs/rl-algorithms/ppo.md b/cleanrl/docs/rl-algorithms/ppo.md
new file mode 100644
index 0000000000000000000000000000000000000000..37e65a352fd56a1f9db7b5cdd42f7564d26894c9
--- /dev/null
+++ b/cleanrl/docs/rl-algorithms/ppo.md
@@ -0,0 +1,1180 @@
+# Proximal Policy Gradient (PPO)
+
+
+## Overview
+
+PPO is one of the most popular DRL algorithms. It runs reasonably fast by leveraging vector (parallel) environments and naturally works well with different action spaces, therefore supporting a variety of games. It also has good sample efficiency compared to algorithms such as DQN.
+
+
+Original paper: 
+
+* [Proximal Policy Optimization Algorithms](https://arxiv.org/abs/1707.06347)
+
+Reference resources:
+
+* [Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO](https://arxiv.org/abs/2005.12729)
+* [What Matters In On-Policy Reinforcement Learning? A Large-Scale Empirical Study](https://arxiv.org/abs/2006.05990)
+* ⭐ [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/)
+
+All our PPO implementations below are augmented with the same code-level optimizations presented in `openai/baselines`'s [PPO](https://github.com/openai/baselines/tree/master/baselines/ppo2). To achieve this, see how we matched the implementation details in our blog post [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/).
+
+## Implemented Variants
+
+
+| Variants Implemented      | Description |
+| ----------- | ----------- |
+| :material-github: [`ppo.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppopy) | For classic control tasks like `CartPole-v1`. |
+| :material-github: [`ppo_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_ataripy) |  For Atari games. It uses convolutional layers and common atari-based pre-processing techniques. |
+| :material-github: [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_continuous_actionpy) | For continuous action space. Also implemented Mujoco-specific code-level optimizations. |
+| :material-github: [`ppo_atari_lstm.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_lstmpy) | For Atari games using LSTM without stacked frames. |
+| :material-github: [`ppo_atari_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpoolpy) | Uses the blazing fast Envpool Atari vectorized environment. |
+| :material-github: [`ppo_atari_envpool_xla_jax.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) | Uses the blazing fast Envpool Atari vectorized environment with EnvPool's XLA interface and JAX. |
+| :material-github: [`ppo_atari_envpool_xla_jax_scan.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jax_scanpy) | Uses native `jax.scan` as opposed to python loops for faster compilation time. |
+| :material-github: [`ppo_procgen.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py), :material-file-document: [docs](/rl-algorithms/ppo/#ppo_procgenpy) | For the procgen environments. |
+| :material-github: [`ppo_atari_multigpu.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_atari_multigpupy)| For Atari environments leveraging multi-GPUs. |
+| :material-github: [`ppo_pettingzoo_ma_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py),  :material-file-document: [docs](/rl-algorithms/ppo/#ppo_pettingzoo_ma_ataripy)| For Pettingzoo's multi-agent Atari environments. |
+
+Below are our single-file implementations of PPO:
+
+## `ppo.py`
+
+The [ppo.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py) has the following features:
+
+* Works with the `Box` observation space of low-level features
+* Works with the `Discrete` action space
+* Works with envs like `CartPole-v1`
+
+### Usage
+
+=== "uv"
+
+    ```bash
+    uv pip install .
+    uv run python cleanrl/ppo.py --help
+    uv run python cleanrl/ppo.py --env-id CartPole-v1
+    ```
+
+=== "pip"
+
+    ```bash
+    python cleanrl/ppo.py --help
+    python cleanrl/ppo.py --env-id CartPole-v1
+    ```
+
+### Explanation of the logged metrics
+
+Running `python cleanrl/ppo.py` will automatically record various metrics such as actor or value losses in Tensorboard. Below is the documentation for these metrics:
+
+* `charts/episodic_return`: episodic return of the game
+* `charts/episodic_length`: episodic length of the game
+* `charts/SPS`: number of steps per second
+* `charts/learning_rate`: the current learning rate
+* `losses/value_loss`: the mean value loss across all data points
+* `losses/policy_loss`: the mean policy loss across all data points
+* `losses/entropy`: the mean entropy value across all data points
+* `losses/old_approx_kl`: the approximate Kullback–Leibler divergence, measured by `(-logratio).mean()`, which corresponds to the k1 estimator in John Schulman’s blog post on [approximating KL](http://joschu.net/blog/kl-approx.html)
+* `losses/approx_kl`: better alternative to `olad_approx_kl` measured by `(logratio.exp() - 1) - logratio`, which corresponds to the k3 estimator in [approximating KL](http://joschu.net/blog/kl-approx.html)
+* `losses/clipfrac`: the fraction of the training data that triggered the clipped objective
+* `losses/explained_variance`: the explained variance for the value function
+
+
+### Implementation details
+
+[ppo.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo.py) is based on the "13 core implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
+
+1. Vectorized architecture (:material-github: [common/cmd_util.py#L22](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/cmd_util.py#L22))
+1. Orthogonal Initialization of Weights and Constant Initialization of biases (:material-github: [a2c/utils.py#L58)](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L58))
+1. The Adam Optimizer's Epsilon Parameter (:material-github: [ppo2/model.py#L100](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L100))
+1. Adam Learning Rate Annealing (:material-github: [ppo2/ppo2.py#L133-L135](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/ppo2.py#L133-L135))
+1. Generalized Advantage Estimation (:material-github: [ppo2/runner.py#L56-L65](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/runner.py#L56-L65))
+1. Mini-batch Updates (:material-github: [ppo2/ppo2.py#L157-L166](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/ppo2.py#L157-L166))
+1. Normalization of Advantages (:material-github: [ppo2/model.py#L139](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L139))
+1. Clipped surrogate objective (:material-github: [ppo2/model.py#L81-L86](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L81-L86))
+1. Value Function Loss Clipping (:material-github: [ppo2/model.py#L68-L75](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L68-L75))
+1. Overall Loss and Entropy Bonus (:material-github: [ppo2/model.py#L91](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L91))
+1. Global Gradient Clipping (:material-github: [ppo2/model.py#L102-L108](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L102-L108))
+1. Debug variables (:material-github: [ppo2/model.py#L115-L116](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/ppo2/model.py#L115-L116))
+1. Separate MLP networks for policy and value functions (:material-github: [common/policies.py#L156-L160](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/policies.py#L156-L160), [baselines/common/models.py#L75-L103](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L75-L103))
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:3:8"
+```
+
+Below are the average episodic returns for `ppo.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
+
+| Environment      | `ppo.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
+| ----------- | ----------- | ----------- |
+| CartPole-v1      | 490.04 ± 6.12     |497.54 ± 4.02  |
+| Acrobot-v1       | -86.36 ± 1.32     |  -81.82 ± 5.58 |
+| MountainCar-v0   | -200.00 ± 0.00         | -200.00 ± 0.00 |
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh::9"
+```
+
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo-time.png">
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Classic-Control-CleanRL-s-PPO--VmlldzoxODU5MDY1" style="width:100%; height:500px" title="Classic-Control-CleanRL-s-PPO"></iframe>
+
+### Video tutorial
+
+If you'd like to learn `ppo.py` in-depth, consider checking out the following video tutorial:
+
+
+<div style="text-align: center;"><iframe width="560" height="315" src="https://www.youtube.com/embed/MEt6rrxH8W4" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></div>
+
+
+## `ppo_atari.py`
+
+The [ppo_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py) has the following features:
+
+* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari]"
+    uv run python cleanrl/ppo_atari.py --help
+    uv run python cleanrl/ppo_atari.py --env-id BreakoutNoFrameskip-v4
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/ppo_atari.py --help
+    python cleanrl/ppo_atari.py --env-id BreakoutNoFrameskip-v4
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari.py) is based on the "9 Atari implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
+
+1. The Use of `NoopResetEnv` (:material-github: [common/atari_wrappers.py#L12](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L12)) 
+1. The Use of `MaxAndSkipEnv` (:material-github: [common/atari_wrappers.py#L97](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L97)) 
+1. The Use of `EpisodicLifeEnv` (:material-github: [common/atari_wrappers.py#L61](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L61)) 
+1. The Use of `FireResetEnv` (:material-github: [common/atari_wrappers.py#L41](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L41)) 
+1. The Use of `WarpFrame` (Image transformation) [common/atari_wrappers.py#L134](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L134) 
+1. The Use of `ClipRewardEnv` (:material-github: [common/atari_wrappers.py#L125](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L125)) 
+1. The Use of `FrameStack` (:material-github: [common/atari_wrappers.py#L188](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/atari_wrappers.py#L188)) 
+1. Shared Nature-CNN network for the policy and value functions (:material-github: [common/policies.py#L157](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/policies.py#L157), [common/models.py#L15-L26](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L15-L26))
+1. Scaling the Images to Range [0, 1] (:material-github: [common/models.py#L19](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/models.py#L19))
+
+### Experiment results
+
+
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:14:19"
+```
+
+
+Below are the average episodic returns for `ppo_atari.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
+
+| Environment      | `ppo_atari.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
+| ----------- | ----------- | ----------- |
+| BreakoutNoFrameskip-v4      | 414.66 ± 28.09     | 406.57 ± 31.554  |
+| PongNoFrameskip-v4   | 20.36 ± 0.20    |  20.512 ± 0.50 |
+| BeamRiderNoFrameskip-v4   | 1915.93 ± 484.58         | 2642.97 ± 670.37 |
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:11:19"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari-time.png">
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-s-PPO--VmlldzoxNjk3NjYy" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO"></iframe>
+
+### Video tutorial
+
+If you'd like to learn `ppo_atari.py` in-depth, consider checking out the following video tutorial:
+
+
+<div style="text-align: center;"><iframe width="560" height="315" src="https://www.youtube.com/embed/05RMTj-2K_Y" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></div>
+
+
+## `ppo_continuous_action.py`
+
+The [ppo_continuous_action.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) has the following features:
+
+* For continuous action space. Also implemented Mujoco-specific code-level optimizations
+* Works with the `Box` observation space of low-level features
+* Works with the `Box` (continuous) action space
+* adding experimental support for [Gymnasium](https://gymnasium.farama.org/)
+* 🧪 support `dm_control` environments via [Shimmy](https://github.com/Farama-Foundation/Shimmy)
+
+
+???+ warning
+
+    We are now recommending users to use [`rpo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/rpo_continuous_action.py) instead of [`ppo_continuous_action.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) because `rpo_continuous_action.py` empirically performs better than `ppo_continuous_action.py` in 93% of the environments we tested. Please see [experiment results](/rl-algorithms/rpo/#experiment-results) for detailed analysis.
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    # mujoco v4 environments
+    uv pip install ".[mujoco]"
+    python cleanrl/ppo_continuous_action.py --help
+    python cleanrl/ppo_continuous_action.py --env-id Hopper-v4
+    # dm_control environments
+    uv pip install ".[mujoco, dm_control]"
+    python cleanrl/ppo_continuous_action.py --env-id dm_control/cartpole-balance-v0
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-mujoco.txt
+    python cleanrl/ppo_continuous_action.py --help
+    python cleanrl/ppo_continuous_action.py --env-id Hopper-v4
+    pip install -r requirements/requirements-dm_control.txt
+    python cleanrl/ppo_continuous_action.py --env-id dm_control/cartpole-balance-v0
+    ```
+
+???+ warning "dm_control installation issue"
+
+    If you run into error like `AttributeError: 'GLFWContext' object has no attribute '_context'` in Linux, it's because the rendering dependencies are not installed properly. To fix it, try running
+
+    ```
+    sudo apt-get update && sudo apt-get -y install libgl1-mesa-glx libosmesa6 libglfw3 
+    ```
+
+    See [https://github.com/deepmind/dm_control#rendering](https://github.com/deepmind/dm_control#rendering) for more detail.
+
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_continuous_action.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_continuous_action.py) is based on the "9 details for continuous action domains (e.g. Mujoco)" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
+
+
+1. Continuous actions via normal distributions (:material-github: [common/distributions.py#L103-L104](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L103-L104))
+2. State-independent log standard deviation (:material-github: [common/distributions.py#L104](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L104))
+3. Independent action components (:material-github: [common/distributions.py#L238-L246](https://github.com/openai/baselines/blob/9b68103b737ac46bc201dfb3121cfa5df2127e53/baselines/common/distributions.py#L238-L246))
+4. Separate MLP networks for policy and value functions (:material-github: [common/policies.py#L160](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/policies.py#L160), [baselines/common/models.py#L75-L103](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L75-L103)
+5. Handling of action clipping to valid range and storage (:material-github: [common/cmd_util.py#L99-L100](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/cmd_util.py#L99-L100)) 
+6. Normalization of Observation (:material-github: [common/vec_env/vec_normalize.py#L4](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L4))
+7. Observation Clipping (:material-github: [common/vec_env/vec_normalize.py#L39](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L39))
+8. Reward Scaling (:material-github: [common/vec_env/vec_normalize.py#L28](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L28))
+9. Reward Clipping (:material-github: [common/vec_env/vec_normalize.py#L32](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/vec_env/vec_normalize.py#L32))
+
+
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+MuJoCo v4
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:25:30"
+```
+
+{!benchmark/ppo_continuous_action.md!}
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:11:19"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_continuous_action.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_continuous_action-time.png">
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/costa-huang/cleanRL/reports/MuJoCo-v4-CleanRL-s-PPO--VmlldzozMTIxOTI5" style="width:100%; height:500px" title="MuJoCo-CleanRL-s-PPO"></iframe>
+
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:36:41"
+```
+
+Below are the average episodic returns for `ppo_continuous_action.py` in `dm_control` environments.
+
+|                                       | ppo_continuous_action ({'tag': ['v1.0.0-13-gcbd83f6']})   |
+|:--------------------------------------|:----------------------------------------------------------|
+| dm_control/acrobot-swingup-v0         | 27.84 ± 9.25                                              |
+| dm_control/acrobot-swingup_sparse-v0  | 1.60 ± 1.17                                               |
+| dm_control/ball_in_cup-catch-v0       | 900.78 ± 5.26                                             |
+| dm_control/cartpole-balance-v0        | 855.47 ± 22.06                                            |
+| dm_control/cartpole-balance_sparse-v0 | 999.93 ± 0.10                                             |
+| dm_control/cartpole-swingup-v0        | 640.86 ± 11.44                                            |
+| dm_control/cartpole-swingup_sparse-v0 | 51.34 ± 58.35                                             |
+| dm_control/cartpole-two_poles-v0      | 203.86 ± 11.84                                            |
+| dm_control/cartpole-three_poles-v0    | 164.59 ± 3.23                                             |
+| dm_control/cheetah-run-v0             | 432.56 ± 82.54                                            |
+| dm_control/dog-stand-v0               | 307.79 ± 46.26                                            |
+| dm_control/dog-walk-v0                | 120.05 ± 8.80                                             |
+| dm_control/dog-trot-v0                | 76.56 ± 6.44                                              |
+| dm_control/dog-run-v0                 | 60.25 ± 1.33                                              |
+| dm_control/dog-fetch-v0               | 34.26 ± 2.24                                              |
+| dm_control/finger-spin-v0             | 590.49 ± 171.09                                           |
+| dm_control/finger-turn_easy-v0        | 180.42 ± 44.91                                            |
+| dm_control/finger-turn_hard-v0        | 61.40 ± 9.59                                              |
+| dm_control/fish-upright-v0            | 516.21 ± 59.52                                            |
+| dm_control/fish-swim-v0               | 87.91 ± 6.83                                              |
+| dm_control/hopper-stand-v0            | 2.72 ± 1.72                                               |
+| dm_control/hopper-hop-v0              | 0.52 ± 0.48                                               |
+| dm_control/humanoid-stand-v0          | 6.59 ± 0.18                                               |
+| dm_control/humanoid-walk-v0           | 1.73 ± 0.03                                               |
+| dm_control/humanoid-run-v0            | 1.11 ± 0.04                                               |
+| dm_control/humanoid-run_pure_state-v0 | 0.98 ± 0.03                                               |
+| dm_control/humanoid_CMU-stand-v0      | 4.79 ± 0.18                                               |
+| dm_control/humanoid_CMU-run-v0        | 0.88 ± 0.05                                               |
+| dm_control/manipulator-bring_ball-v0  | 0.50 ± 0.29                                               |
+| dm_control/manipulator-bring_peg-v0   | 1.80 ± 1.58                                               |
+| dm_control/manipulator-insert_ball-v0 | 35.50 ± 13.04                                             |
+| dm_control/manipulator-insert_peg-v0  | 60.40 ± 21.76                                             |
+| dm_control/pendulum-swingup-v0        | 242.81 ± 245.95                                           |
+| dm_control/point_mass-easy-v0         | 273.95 ± 362.28                                           |
+| dm_control/point_mass-hard-v0         | 143.25 ± 38.12                                            |
+| dm_control/quadruped-walk-v0          | 239.03 ± 66.17                                            |
+| dm_control/quadruped-run-v0           | 180.44 ± 32.91                                            |
+| dm_control/quadruped-escape-v0        | 28.92 ± 11.21                                             |
+| dm_control/quadruped-fetch-v0         | 193.97 ± 22.20                                            |
+| dm_control/reacher-easy-v0            | 626.28 ± 15.51                                            |
+| dm_control/reacher-hard-v0            | 443.80 ± 9.64                                             |
+| dm_control/stacker-stack_2-v0         | 75.68 ± 4.83                                              |
+| dm_control/stacker-stack_4-v0         | 68.02 ± 4.02                                              |
+| dm_control/swimmer-swimmer6-v0        | 158.19 ± 10.22                                            |
+| dm_control/swimmer-swimmer15-v0       | 131.94 ± 0.88                                             |
+| dm_control/walker-stand-v0            | 564.46 ± 235.22                                           |
+| dm_control/walker-walk-v0             | 392.51 ± 56.25                                            |
+| dm_control/walker-run-v0              | 125.92 ± 10.01                                            |
+
+Note that the dm_control/lqr-lqr_2_1-v0 dm_control/lqr-lqr_6_2-v0 environments are never terminated or truncated. See https://wandb.ai/openrlbenchmark/cleanrl/runs/3tm00923 and https://wandb.ai/openrlbenchmark/cleanrl/runs/1z9us07j as an example.
+
+Learning curves:
+
+![](../ppo/ppo_continuous_action_gymnasium_dm_control.png)
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/dm_control-CleanRL-s-PPO-part-1---VmlldzozMTI2MjE2" style="width:100%; height:500px" title="dm_control-CleanRL-s-PPO-part-1"></iframe>
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/dm_control-CleanRL-s-PPO-part-2---VmlldzozMTI2MjI1" style="width:100%; height:500px" title="dm_control-CleanRL-s-PPO-part-2"></iframe>
+
+
+???+ info
+
+    In the gymnasium environments, we use the v4 mujoco environments, which roughly results in the same performance as the v2 mujoco environments.
+
+    ![](../ppo/ppo_continuous_action_v2_vs_v4.png)
+
+
+### Video tutorial
+
+If you'd like to learn `ppo_continuous_action.py` in-depth, consider checking out the following video tutorial:
+
+
+<div style="text-align: center;"><iframe width="560" height="315" src="https://www.youtube.com/embed/BvZvx7ENZBw" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></div>
+
+
+## `ppo_atari_lstm.py`
+
+The [ppo_atari_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py) has the following features:
+
+* For Atari games using LSTM without stacked frames. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+### Usage
+
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari]"
+    uv run python cleanrl/ppo_atari_lstm.py --help
+    uv run python cleanrl/ppo_atari_lstm.py --env-id BreakoutNoFrameskip-v4
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/ppo_atari_lstm.py --help
+    python cleanrl/ppo_atari_lstm.py --env-id BreakoutNoFrameskip-v4
+    ```
+
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_atari_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_lstm.py) is based on the "5 LSTM implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
+
+1. Layer initialization for LSTM layers (:material-github: [a2c/utils.py#L84-L86](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L84-L86))
+2. Initialize the LSTM states to be zeros (:material-github: [common/models.py#L179](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L179))
+3. Reset LSTM states at the end of the episode (:material-github: [common/models.py#L141](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L141))
+4. Prepare sequential rollouts in mini-batches (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
+5. Reconstruct LSTM states during training (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
+
+To help test out the memory, we remove the 4 stacked frames from the observation (i.e., using `env = gym.wrappers.FrameStack(env, 1)` instead of `env = gym.wrappers.FrameStack(env, 4)` like in `ppo_atari.py` )
+
+
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:47:52"
+```
+
+Below are the average episodic returns for `ppo_atari_lstm.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
+
+
+| Environment      | `ppo_atari_lstm.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
+| ----------- | ----------- | ----------- |
+| BreakoutNoFrameskip-v4      | 128.92 ± 31.10    | 138.98 ± 50.76  |
+| PongNoFrameskip-v4   | 19.78 ± 1.58    | 19.79 ± 0.67 |
+| BeamRiderNoFrameskip-v4   | 1536.20 ± 612.21         | 1591.68 ± 372.95|
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:11:19"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_lstm.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_lstm-time.png">
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-s-PPO-LSTM--VmlldzoxODcxMzE4" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO-LSTM"></iframe>
+
+
+
+## `ppo_atari_envpool.py`
+
+The [ppo_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py) has the following features:
+
+* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
+* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+???+ warning
+
+    Note that `ppo_atari_envpool.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
+
+???+ bug
+
+    EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PPO implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
+
+    * Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
+    * Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
+
+    This causes the $s_{last}$ to be off by one. 
+    See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
+
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[envpool]"
+    uv run python cleanrl/ppo_atari_envpool.py --help
+    uv run python cleanrl/ppo_atari_envpool.py --env-id Breakout-v5
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-envpool.txt
+    python cleanrl/ppo_atari_envpool.py --help
+    python cleanrl/ppo_atari_envpool.py --env-id Breakout-v5
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool.py) uses a customized `RecordEpisodeStatistics` to work with envpool but has the same other implementation details as `ppo_atari.py` (see [related docs](/rl-algorithms/ppo/#implementation-details_1)).
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:58:63"
+```
+
+{!benchmark/ppo_atari_envpool.md!}
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:51:62"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool-time.png">
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-s-PPO-Envpool--VmlldzoxODcxMzI3" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO-Envpool"></iframe>
+
+
+## `ppo_atari_envpool_xla_jax.py`
+
+The [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) has the following features:
+
+* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
+    * Uses EnvPool's experimental [XLA interface](https://envpool.readthedocs.io/en/latest/content/xla_interface.html).
+* Uses [Jax](https://github.com/google/jax), [Flax](https://github.com/google/flax), and [Optax](https://github.com/deepmind/optax) instead of `torch`.  
+* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+???+ warning
+
+    Note that `ppo_atari_envpool_xla_jax.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
+
+
+???+ bug
+
+    EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PPO implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
+
+    * Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
+    * Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
+
+    This causes the $s_{last}$ to be off by one. 
+    See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
+
+
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[envpool, jax]"
+    uv pip install --upgrade "jax[cuda11_cudnn82]==0.4.8" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    uv run python cleanrl/ppo_atari_envpool_xla_jax.py --help
+    uv run python cleanrl/ppo_atari_envpool_xla_jax.py --env-id Breakout-v5
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-envpool.txt
+    pip install -r requirements/requirements-jax.txt
+    pip install --upgrade "jax[cuda]==0.3.17" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    python cleanrl/ppo_atari_envpool_xla_jax.py --help
+    python cleanrl/ppo_atari_envpool_xla_jax.py --env-id Breakout-v5
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`. In [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) we omit logging `losses/old_approx_kl` and `losses/clipfrac` for brevity.
+
+Additionally, we record the following metric:
+
+* `charts/avg_episodic_return`: the average value of the *latest* episodic returns of `args.num_envs=8` envs
+* `charts/avg_episodic_length`: the average value of the *latest* episodic lengths of `args.num_envs=8` envs
+
+???+ info
+
+    Note that we use `charts/avg_episodic_return` and `charts/avg_episodic_length` in place of `charts/episodic_return` and `charts/episodic_length` because under the EnvPool's XLA interface, we can only record fixed-shape metrics where as there could be a variable number of raw episodic returns / lengths. To resolve this challenge, we create variables (e.g., `returned_episode_returns`, `returned_episode_lengths`) to keep track of the *latest* episodic returns / lengths of each environment and average them for reporting purposes.
+
+### Implementation details
+
+[ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) uses the same other implementation details as `ppo_atari.py` (see [related docs](/rl-algorithms/ppo/#implementation-details_1)), with two differences
+
+1. [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) does not use the value function clipping by default, because there is no sufficient evidence that value function clipping actually improves performance.
+1. [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) uses a customized `EpisodeStatistics` to record episode statistics instead of the `RecordEpisodeStatistics` used in other variants. `RecordEpisodeStatistics` is a *stateful* python wrapper which is incompatible with EnvPool's *stateless* XLA interface. To address this issue, we used a `EpisodeStatistics` dataclass and simply implement the logic of `RecordEpisodeStatistics`. However, `EpisodeStatistics` comes with a major limitation: its storage has a fixed shape and can only record the *latest* episodic return of the sub-environments. Furthermore, the default episodic return values in `EpisodeStatistics` are set to zeros, which does not necessarily correspond to the episodic return obtained by a random policy. For example, we would report `charts/avg_episodic_return=0` for `Pong-v5`, even if they should have been `charts/avg_episodic_return=-21`. That said, this issue goes away as soon as the sub-environments finished their first episodes, therefore not impacting the reported results.
+
+
+???+ info
+    
+    We benchmarked the PPO implementation w/ and w/o value function clipping, finding no significant difference in performance, which is consistent with the findings in Andrychowicz et al.[^2]. See the related report [part 1](https://wandb.ai/costa-huang/cleanRL/reports/CleanRL-PPO-JAX-EnvPool-s-XLA-w-and-w-o-value-loss-clipping-vs-openai-baselins-PPO-part-1---VmlldzoyNzQ3MzQ1) and [part 2](https://wandb.ai/costa-huang/cleanRL/reports/CleanRL-PPO-JAX-EnvPool-s-XLA-w-and-w-o-value-loss-clipping-vs-openai-baselins-PPO-part-2---VmlldzoyNzQ3MzUw).
+
+    ![](../ppo/ppo_atari_envpool_xla_jax/hns_ppo_vs_baselines2.svg)
+
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:69:74"
+```
+
+
+{!benchmark/ppo_atari_envpool_xla_jax.md!}
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:64:85"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool_xla_jax_sample_walltime_efficiency.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool_xla_jax.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool_xla_jax-time.png">
+
+
+
+???+ info
+
+    Note the original openai/baselines uses `atari-py==0.2.6` which hangs on `gym.make("DefenderNoFrameskip-v4")` and does not support SurroundNoFrameskip-v4 (see issue [:material-github: openai/atari-py#73](https://github.com/openai/atari-py/issues/73)). To get results on these environments, we use `gym==0.23.1 ale-py==0.7.4 "AutoROM[accept-rom-license]==0.4.2` and [manually register `SurroundNoFrameskip-v4` in our fork](https://github.com/vwxyzjn/baselines/blob/e2cb1c938a62fa8d7fe98187246cde08dfd57bd1/baselines/common/register_all_atari_envs.py#L2). 
+
+
+Median Human Normalized Score (HNS) compared to SEEDRL's R2D2 (data available [here](https://github.com/google-research/seed_rl/blob/66e8890261f09d0355e8bf5f1c5e41968ca9f02b/docs/seed_r2d2_atari_graphs.csv)). 
+
+![](../ppo/ppo_atari_envpool_xla_jax/hns_ppo_vs_r2d2.svg)
+
+???+ info
+
+    Note the SEEDRL's R2D2's median HNS data does not include learning curves for `Defender` and `Surround` (see [google-research/seed_rl#78](https://github.com/google-research/seed_rl/issues/78)). Also note the SEEDRL's R2D2 uses slightly different Atari preprocessing than our `ppo_atari_envpool_xla_jax.py`, so we may be comparing apples and oranges; however, the results are still informative at the scale of 57 Atari games — we would be at least comparing similar apples.
+
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-PPO-JAX-EnvPool-s-XLA-vs-openai-baselins-PPO-part-1---VmlldzoyNjE2ODMz" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO-Envpool-vs-openai-baselines-Part-1"></iframe>
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-PPO-JAX-EnvPool-s-XLA-vs-openai-baselins-PPO-part-2---VmlldzoyNjE2ODM1" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO-Envpool-vs-openai-baselines-Part-2"></iframe>
+
+
+
+## `ppo_atari_envpool_xla_jax_scan.py`
+
+The [ppo_atari_envpool_xla_jax_scan.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py) has the following features:
+
+* Replaces python loops in `compute_gae`, `update_ppo`, and `rollout` functions of [ppo_atari_envpool_xla_jax.py](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) with native `jax.scan`
+* Warnings and caveats from [ppo_atari_envpool_xla_jax.py](/rl-algorithms/ppo/#ppo_atari_envpool_xla_jaxpy) also apply here
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[envpool, jax]"
+    uv pip install --upgrade "jax[cuda11_cudnn82]==0.4.8" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    uv run python cleanrl/ppo_atari_envpool_xla_jax_scan.py --help
+    uv run python cleanrl/ppo_atari_envpool_xla_jax_scan.py --env-id Breakout-v5
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-envpool.txt
+    pip install -r requirements/requirements-jax.txt
+    pip install --upgrade "jax[cuda]==0.3.17" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    python cleanrl/ppo_atari_envpool_xla_jax_scan.py --help
+    python cleanrl/ppo_atari_envpool_xla_jax_scan.py --env-id Breakout-v5
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`. The metrics are the same as those in [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py).
+
+### Implementation details
+
+[ppo_atari_envpool_xla_jax_scan.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax_scan.py) is a clone of [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) that replaces the python loops with native `jax.scan`.
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:80:85"
+```
+
+
+{!benchmark/ppo_atari_envpool_xla_jax_scan.md!}
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:87:96"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool_xla_jax_scan.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_envpool_xla_jax_scan-time.png">
+
+Learning curves:
+
+???+ info
+
+    The training time of this variant and that of [ppo_atari_envpool_xla_jax.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_envpool_xla_jax.py) are very similar but the compilation time is reduced significantly (see [vwxyzjn/cleanrl#328](https://github.com/vwxyzjn/cleanrl/pull/328#issuecomment-1340474894)). Note that the hardware also affects the speed in the learning curve below. Runs from [`costa-huang`](https://github.com/vwxyzjn/) (red) are slower from those of [`51616`](https://github.com/51616/) (blue and orange) because of hardware differences.
+
+    ![](../ppo/ppo_atari_envpool_xla_jax_scan/compare.png)
+    ![](../ppo/ppo_atari_envpool_xla_jax_scan/compare-time.png)
+
+
+Tracked experiments:
+
+<iframe src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Regression-Report-ppo_atari_envpool_xla_jax_scan--VmlldzozMTk2MzM2" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO-Envpool-Jax-scan"></iframe>
+
+
+## `ppo_procgen.py`
+
+The [ppo_procgen.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py) has the following features:
+
+* For the procgen environments
+* Uses IMPALA-style neural network
+* Works with the `Discrete` action space
+
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[procgen]"
+    uv run python cleanrl/ppo_procgen.py --help
+    uv run python cleanrl/ppo_procgen.py --env-id starpilot
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-procgen.txt
+    python cleanrl/ppo_procgen.py --help
+    python cleanrl/ppo_procgen.py --env-id starpilot
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_procgen.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_procgen.py) is based on the details in "Appendix" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/), which are as follows:
+
+1. IMPALA-style Neural Network (:material-github: [common/models.py#L28](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L28))
+1. Use the same `gamma` parameter in the `NormalizeReward` wrapper. Note that the original implementation from [openai/train-procgen](https://github.com/openai/train-procgen) uses the default `gamma=0.99` in [the `VecNormalize` wrapper](https://github.com/openai/train-procgen/blob/1a2ae2194a61f76a733a39339530401c024c3ad8/train_procgen/train.py#L43) but `gamma=0.999` as PPO's parameter. The mismatch between the `gamma`s is technically incorrect. See [#209](https://github.com/vwxyzjn/cleanrl/pull/209)
+
+### Experiment results
+
+
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:91:100"
+```
+
+We try to match the default setting in [openai/train-procgen](https://github.com/openai/train-procgen) except that we use the `easy` distribution mode and `total_timesteps=25e6` to save compute. Notice [openai/train-procgen](https://github.com/openai/train-procgen) has the following settings:
+
+1. Learning rate annealing is turned off by default
+1. Reward scaling and reward clipping is used
+
+
+Below are the average episodic returns for `ppo_procgen.py`. To ensure the quality of the implementation, we compared the results against `openai/baselies`' PPO.
+
+| Environment      | `ppo_procgen.py` | `openai/baselies`' PPO (Huang et al., 2022)[^1]
+| ----------- | ----------- | ----------- |
+| StarPilot (easy)      | 30.99 ± 1.96      | 33.97 ± 7.86  |
+| BossFight (easy)   | 8.85 ± 0.33    |  9.35 ± 2.04 |
+| BigFish  (easy)  | 16.46 ± 2.71         | 20.06 ± 5.34 |
+
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:98:106"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_procgen.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_procgen-time.png">
+
+
+???+ info
+
+    Note that we have run the procgen experiments using the `easy` distribution for reducing the computational cost.
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Procgen-CleanRL-s-PPO--VmlldzoxODcxMzUy" style="width:100%; height:500px" title="Procgen-CleanRL-s-PPO"></iframe>
+
+
+
+## `ppo_atari_multigpu.py`
+
+The [ppo_atari_multigpu.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py) leverages data parallelism to speed up training time *at no cost of sample efficiency*. 
+
+`ppo_atari_multigpu.py` has the following features:
+
+* Allows the users to use do training leveraging data parallelism
+* For playing Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+???+ warning
+
+    Note that `ppo_atari_multigpu.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. It will error out with `NOTE: Redirects are currently not supported in Windows or MacOs.` See [pytorch/pytorch#20380](https://github.com/pytorch/pytorch/issues/20380)
+
+### Usage
+
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari]"
+    uv run python cleanrl/ppo_atari_multigpu.py --help
+
+    # `--nproc_per_node=2` specifies how many subprocesses we spawn for training with data parallelism
+    # note it is possible to run this with a *single GPU*: each process will simply share the same GPU
+    uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4
+
+    # by default we use the `gloo` backend, but you can use the `nccl` backend for better multi-GPU performance
+    uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --backend nccl
+
+    # it is possible to spawn more processes than the amount of GPUs you have via `--device-ids`
+    # e.g., the command below spawns two processes using GPU 0 and two processes using GPU 1
+    uv run torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --device-ids 0 0 1 1
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/ppo_atari_multigpu.py --help
+
+    # `--nproc_per_node=2` specifies how many subprocesses we spawn for training with data parallelism
+    # note it is possible to run this with a *single GPU*: each process will simply share the same GPU
+    torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4
+
+    # by default we use the `gloo` backend, but you can use the `nccl` backend for better multi-GPU performance
+    torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --backend nccl
+
+    # it is possible to spawn more processes than the amount of GPUs you have via `--device-ids`
+    # e.g., the command below spawns two processes using GPU 0 and two processes using GPU 1
+    torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --device-ids 0 0 1 1
+    ```
+
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_atari_multigpu.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_atari_multigpu.py) is based on `ppo_atari.py` (see its [related docs](/rl-algorithms/ppo/#implementation-details_1)).
+
+We use [Pytorch's distributed API](https://pytorch.org/tutorials/intermediate/dist_tuto.html) to implement the data parallelism paradigm. The basic idea is that the user can spawn $N$ processes each running a copy of `ppo_atari.py`,  holding a copy of the model, stepping the environments, and averaging their gradients together for the backward pass. Here are a few note-worthy implementation details.
+
+1. **Local versus global parameters**: All of the parameters in `ppo_atari.py` are global (such as batch size), but in `ppo_atari_multigpu.py` we have local parameters as well. Say we run `torchrun --standalone --nnodes=1 --nproc_per_node=2 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --local-num-envs=4`; here are how all multi-gpu related parameters are adjusted:
+    * **number of environments**: `num_envs = local_num_envs * world_size = 4 * 2 = 8`
+    * **batch size**: `local_batch_size = local_num_envs * num_steps = 4 * 128 = 512`, `batch_size = num_envs * num_steps) = 8 * 128 = 1024`
+    * **minibatch size**:  `local_minibatch_size = int(args.local_batch_size // args.num_minibatches) = 512 // 4 = 128`, `minibatch_size = int(args.batch_size // args.num_minibatches) = 1024 // 4 = 256`
+    * **number of updates**: `num_iterations = args.total_timesteps // args.batch_size = 10000000 // 1024 = 9765`
+1. **Adjust seed per process**: we need be very careful with seeding: we could have used the exact same seed for each subprocess. To ensure this does not happen, we do the following
+
+    ```python hl_lines="2 5 16"
+    # CRUCIAL: note that we needed to pass a different seed for each data parallelism worker
+    args.seed += local_rank
+    random.seed(args.seed)
+    np.random.seed(args.seed)
+    torch.manual_seed(args.seed - local_rank)
+    torch.backends.cudnn.deterministic = args.torch_deterministic
+
+    # ...
+
+    envs = gym.vector.SyncVectorEnv(
+        [make_env(args.env_id, args.seed + i, i, args.capture_video, run_name) for i in range(args.num_envs)]
+    )
+    assert isinstance(envs.single_action_space, gym.spaces.Discrete), "only discrete action space is supported"
+
+    agent = Agent(envs).to(device)
+    torch.manual_seed(args.seed)
+    optimizer = optim.Adam(agent.parameters(), lr=args.learning_rate, eps=1e-5)
+    ```
+
+    Notice that we adjust the seed with `args.seed += local_rank` (line 2), where `local_rank` is the index of the subprocesses. This ensures we seed packages and envs with uncorrealted seeds. However, we do need to use the same `torch` seed for all process to initialize same weights for the `agent` (line 5), after which we can use a different seed for `torch` (line 16).
+1. **Efficient gradient averaging**: PyTorch recommends to average the gradient across the whole world via the following (see [docs](https://pytorch.org/tutorials/intermediate/dist_tuto.html#distributed-training))
+
+    ```python
+    for param in agent.parameters():
+        dist.all_reduce(param.grad.data, op=dist.ReduceOp.SUM)
+        param.grad.data /= world_size
+    ```
+
+    However, [@cswinter](https://github.com/cswinter) introduces a more efficient gradient averaging scheme with proper batching (see :material-github: [entity-neural-network/incubator#220](https://github.com/entity-neural-network/incubator/pull/220)), which looks like:
+
+    ```python
+    all_grads_list = []
+    for param in agent.parameters():
+        if param.grad is not None:
+            all_grads_list.append(param.grad.view(-1))
+    all_grads = torch.cat(all_grads_list)
+    dist.all_reduce(all_grads, op=dist.ReduceOp.SUM)
+    offset = 0
+    for param in agent.parameters():
+        if param.grad is not None:
+            param.grad.data.copy_(
+                all_grads[offset : offset + param.numel()].view_as(param.grad.data) / world_size
+            )
+            offset += param.numel()
+    ```
+
+    In our previous empirical testing (see :material-github: [vwxyzjn/cleanrl#162](https://github.com/vwxyzjn/cleanrl/pull/162#issuecomment-1107909696)), we have found [@cswinter](https://github.com/cswinter)'s implementation to be faster, hence we adopt it in our implementation.
+
+
+
+
+### Experiment results
+
+
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+
+``` title="benchmark/ppo.sh" linenums="1"
+--8<-- "benchmark/ppo.sh:102:107"
+```
+
+Below are the average episodic returns for `ppo_atari_multigpu.py`. To ensure no loss of sample efficiency, we compared the results against `ppo_atari.py`.
+
+
+{!benchmark/ppo_atari_multigpu.md!}
+
+
+Learning curves:
+
+``` title="benchmark/ppo_plot.sh" linenums="1"
+--8<-- "benchmark/ppo_plot.sh:108:117"
+```
+
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_multigpu.png">
+<img loading="lazy" src="https://huggingface.co/datasets/cleanrl/benchmark/resolve/main/benchmark/pr-424/ppo_atari_multigpu-time.png">
+
+
+
+
+Under the same hardware, we see that `ppo_atari_multigpu.py` is about **30% faster** than `ppo_atari.py` with no loss of sample efficiency. 
+
+
+???+ info
+
+    The experiments above is to show correctness -- we show that by aligning the same hyperparameters of `ppo_atari.py` and `ppo_atari_multigpu.py`, we can achieve the same sample efficiency. However, we can train even faster by simply running a much larger batch size. For example, we can run `torchrun --standalone --nnodes=1 --nproc_per_node=8 cleanrl/ppo_atari_multigpu.py --env-id BreakoutNoFrameskip-v4 --local-num-envs=8`, which will run 8 x 8 = 64 environments in parallel and achieve a batch size of 64 x 128 = 8192. This will likely result in a sample efficiency but should increase the wall time efficiency.
+
+
+???+ info
+
+    Although `ppo_atari_multigpu.py` is 30% faster than `ppo_atari.py`, `ppo_atari_multigpu.py` is still slower than `ppo_atari_envpool.py`, as shown below.  This comparison really highlights the different kinds of optimization possible.
+
+
+    <div class="grid-container">
+        <img src="../ppo/Breakout-a.png">
+        <img src="../ppo/Breakout-time-a.png">
+    </div>
+
+    The purpose of `ppo_atari_multigpu.py` is not (yet) to achieve the fastest PPO + Atari example. Rather, its purpose is to *rigorously validate data parallelism does provide performance benefits*. We could do something like `ppo_atari_multigpu_envpool.py` to possibly obtain the fastest PPO + Atari possible, but that is for another day. Note we may need `numba` to pin the threads `envpool` is using in each subprocess to avoid threads fighting each other and lowering the throughput.
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-s-PPO-MultiGPU--VmlldzoxOTM2NDUx" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO"></iframe>
+
+
+
+
+
+## `ppo_pettingzoo_ma_atari.py`
+[ppo_pettingzoo_ma_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py) trains an agent to learn playing Atari games via selfplay. The selfplay environment is implemented as a vectorized environment from [PettingZoo.ml](https://www.pettingzoo.ml/atari). The basic idea is to create vectorized environment $E$ with `num_envs = N`, where $N$ is the number of players in the game. Say $N = 2$, then the 0-th sub environment of $E$ will return the observation for player 0 and 1-th sub environment will return the observation of player 1. Then the two environments takes a batch of 2 actions and execute them for player 0 and player 1, respectively. See "Vectorized architecture" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/) for more detail.
+
+`ppo_pettingzoo_ma_atari.py` has the following features:
+
+* For playing the pettingzoo's multi-agent Atari game.
+* Works with the pixel-based observation space
+* Works with the `Box` action space
+
+???+ warning
+
+    Note that `ppo_pettingzoo_ma_atari.py` does not work in Windows :fontawesome-brands-windows:. See [https://pypi.org/project/multi-agent-ale-py/#files](https://pypi.org/project/multi-agent-ale-py/#files)
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[pettingzoo, atari]"
+    uv run AutoROM --accept-license
+    uv run  cleanrl/ppo_pettingzoo_ma_atari.py --help
+    uv run  cleanrl/ppo_pettingzoo_ma_atari.py --env-id pong_v3
+    uv run  cleanrl/ppo_pettingzoo_ma_atari.py --env-id surround_v2
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-pettingzoo.txt
+    pip install -r requirements/requirements-atari.txt
+    AutoROM --accept-license
+    python cleanrl/ppo_pettingzoo_ma_atari.py --help
+    python cleanrl/ppo_pettingzoo_ma_atari.py --env-id pong_v3
+    python cleanrl/ppo_pettingzoo_ma_atari.py --env-id surround_v2
+    ```
+
+See [https://www.pettingzoo.ml/atari](https://www.pettingzoo.ml/atari) for a full-list of supported environments such as `basketball_pong_v3`. Notice pettingzoo sometimes introduces breaking changes, so make sure to install the pinned dependencies via `poetry`.
+
+### Explanation of the logged metrics
+
+Additionally, it logs the following metrics
+
+* `charts/episodic_return-player0`: episodic return of the game for player 0
+* `charts/episodic_return-player1`: episodic return of the game for player 1
+* `charts/episodic_length-player0`: episodic length of the game for player 0
+* `charts/episodic_length-player1`: episodic length of the game for player 1
+
+See other logged metrics in the [related docs](/rl-algorithms/ppo/#explanation-of-the-logged-metrics) for `ppo.py`.
+
+### Implementation details
+
+[ppo_pettingzoo_ma_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/ppo_pettingzoo_ma_atari.py) is based on `ppo_atari.py` (see its [related docs](/rl-algorithms/ppo/#implementation-details_1)).
+
+`ppo_pettingzoo_ma_atari.py` additionally has the following implementation details:
+
+1. **`supersuit` wrappers**:  uses preprocessing wrappers from `supersuit` instead of from `stable_baselines3`, which looks like the following. In particular note that the `supersuit` does not offer a wrapper similar to `NoopResetEnv`, and that it uses the `agent_indicator_v0` to add two channels indicating the which player the agent controls.
+
+    ```diff
+    -env = gym.make(env_id)
+    -env = NoopResetEnv(env, noop_max=30)
+    -env = MaxAndSkipEnv(env, skip=4)
+    -env = EpisodicLifeEnv(env)
+    -if "FIRE" in env.unwrapped.get_action_meanings():
+    -    env = FireResetEnv(env)
+    -env = ClipRewardEnv(env)
+    -env = gym.wrappers.ResizeObservation(env, (84, 84))
+    -env = gym.wrappers.GrayScaleObservation(env)
+    -env = gym.wrappers.FrameStack(env, 4)
+    +env = importlib.import_module(f"pettingzoo.atari.{args.env_id}").parallel_env()
+    +env = ss.max_observation_v0(env, 2)
+    +env = ss.frame_skip_v0(env, 4)
+    +env = ss.clip_reward_v0(env, lower_bound=-1, upper_bound=1)
+    +env = ss.color_reduction_v0(env, mode="B")
+    +env = ss.resize_v1(env, x_size=84, y_size=84)
+    +env = ss.frame_stack_v1(env, 4)
+    +env = ss.agent_indicator_v0(env, type_only=False)
+    +env = ss.pettingzoo_env_to_vec_env_v1(env)
+    +envs = ss.concat_vec_envs_v1(env, args.num_envs // 2, num_cpus=0, base_class="gym")
+    ```
+1. **A more detailed note on the `agent_indicator_v0` wrapper**: let's dig deeper into how `agent_indicator_v0` works. We do `print(envs.reset(), envs.reset().shape)`
+    ```python
+    [  0.,   0.,   0., 236.,   1,   0.]],
+
+    [[  0.,   0.,   0., 236.,   0.,   1.],
+    [  0.,   0.,   0., 236.,   0.,   1.],
+    [  0.,   0.,   0., 236.,   0.,   1.],
+    ...,
+    [  0.,   0.,   0., 236.,   0.,   1.],
+    [  0.,   0.,   0., 236.,   0.,   1.],
+    [  0.,   0.,   0., 236.,   0.,   1.]]]]) torch.Size([16, 84, 84, 6])
+    ```
+    
+    So the `agent_indicator_v0` adds the last two columns, where `[  0.,   0.,   0., 236.,   1,   0.]]` means this observation is for player 0, and `[  0.,   0.,   0., 236.,   0.,   1.]` is for player 1. Notice the observation still has the range of $[0, 255]$ but the agent indicator channel has the range of $[0,1]$, so we need to be careful when dividing the observation by 255. In particular, we would only divide the first four channels by 255 and leave the agent indicator channels untouched as follows:
+
+    ```py
+    def get_action_and_value(self, x, action=None):
+        x = x.clone()
+        x[:, :, :, [0, 1, 2, 3]] /= 255.0
+        hidden = self.network(x.permute((0, 3, 1, 2)))
+    ```
+
+
+### Experiment results
+
+
+
+To run benchmark experiments, see :material-github: [benchmark/ppo.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/ppo.sh). Specifically, execute the following command:
+
+<script src="https://emgithub.com/embed.js?target=https%3A%2F%2Fgithub.com%2Fvwxyzjn%2Fcleanrl%2Fblob%2Fmaster%2Fbenchmark%2Fppo.sh%23L53-L59&style=github&showBorder=on&showLineNumbers=on&showFileMeta=on&showCopy=on"></script>
+
+???+ info
+
+    Note that evaluation is usually tricker in in selfplay environments. The usual episodic return is not a good indicator of the agent's performance in zero-sum games because the episodic return converges to zero. To evaluate the agent's ability, an intuitive approach is to take a look at the videos of the agents playing the game (included below), visually inspect the agent's behavior. The best scheme, however, is rating systems like [Trueskill](https://www.microsoft.com/en-us/research/project/trueskill-ranking-system/) or [ELO scores](https://en.wikipedia.org/wiki/Elo_rating_system). However, they are more difficult to implement and are outside the scode of `ppo_pettingzoo_ma_atari.py`. 
+    
+    
+    For simplicity, we measure the **episodic length** instead, which in a sense measures how many "back and forth" the agent can create. In other words, the longer the agent can play the game, the better the agent can play. Empirically, we have found episodic length to be a good indicator of the agent's skill, especially in `pong_v3` and `surround_v2`. However, it is not the case for `tennis_v3` and we'd need to visually inspect the agents' game play videos.
+
+
+Below are the average **episodic length** for `ppo_pettingzoo_ma_atari.py`. To ensure no loss of sample efficiency, we compared the results against `ppo_atari.py`.
+
+| Environment      | `ppo_pettingzoo_ma_atari.py`  | 
+| ----------- | ----------- | 
+| pong_v3 | 4153.60 ± 190.80      | 
+| surround_v2 | 3055.33 ± 223.68  | 
+| tennis_v3 | 14538.02 ± 7005.54   | 
+
+
+Learning curves:
+
+<div class="grid-container">
+<img src="../ppo/pong_v3.png">
+
+<img src="../ppo/surround_v2.png">
+
+<img src="../ppo/tennis_v3.png">
+</div>
+
+
+
+Tracked experiments and game play videos:
+
+<iframe loading="lazy" src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Pettingzoo-s-Multi-agent-Atari-CleanRL-s-PPO--VmlldzoyMDkxNTE5" style="width:100%; height:500px" title="Atari-CleanRL-s-PPO"></iframe>
+
+
+
+{!rl-algorithms/ppo-isaacgymenvs.md!}
+
+[^1]: Huang, Shengyi; Dossa, Rousslan Fernand Julien; Raffin, Antonin; Kanervisto, Anssi; Wang, Weixun (2022). The 37 Implementation Details of Proximal Policy Optimization. ICLR 2022 Blog Track https://iclr-blog-track.github.io/2022/03/25/ppo-implementation-details/
+
+[^2]: Andrychowicz, Marcin, Anton Raichuk, Piotr Stańczyk, Manu Orsini, Sertan Girgin, Raphael Marinier, Léonard Hussenot et al. "What matters in on-policy reinforcement learning? a large-scale empirical study." International Conference on Learning Representations 2021, https://openreview.net/forum?id=nIAxjsniDzg
diff --git a/cleanrl/docs/rl-algorithms/pqn.md b/cleanrl/docs/rl-algorithms/pqn.md
new file mode 100644
index 0000000000000000000000000000000000000000..ed532b58765058e88c4074c605845c33dbb46cd9
--- /dev/null
+++ b/cleanrl/docs/rl-algorithms/pqn.md
@@ -0,0 +1,298 @@
+# Parallel Q Network (PQN)
+
+
+## Overview
+
+PQN is a parallelized version of the Deep Q-learning algorithm. It is designed to be more efficient than DQN by using multiple agents to interact with the environment in parallel. PQN can be thought of as DQN (1) without replay buffer and target networks, and (2) with layer normalizations and parallel environments.
+
+Original paper: 
+
+* [Simplifying Deep Temporal Difference Learning](https://arxiv.org/html/2407.04811v2)
+
+Reference resources:
+
+* :material-github: [purejaxql](https://github.com/mttga/purejaxql)
+
+## Implemented Variants
+
+
+| Variants Implemented      | Description |
+| ----------- | ----------- |
+| :material-github: [`pqn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn.py), :material-file-document: [docs](/rl-algorithms/pqn/#pqnpy) | For classic control tasks like `CartPole-v1`. |
+| :material-github: [`pqn_atari_envpool.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool.py), :material-file-document: [docs](/rl-algorithms/pqn/#pqn_atari_envpoolpy) |  For Atari games. Uses the blazing fast Envpool Atari vectorized environment. It uses convolutional layers and common atari-based pre-processing techniques. |
+| :material-github: [`pqn_atari_envpool_lstm.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool_lstm.py), :material-file-document: [docs](/rl-algorithms/pqn/#pqn_atari_envpool_lstmpy) | For Atari games. Uses the blazing fast Envpool Atari vectorized environment. Using LSTM without stacked frames. |
+
+Below are our single-file implementations of PQN:
+
+## `pqn.py`
+
+The [pqn.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn.py) has the following features:
+
+* Works with the `Box` observation space of low-level features
+* Works with the `Discrete` action space
+* Works with envs like `CartPole-v1`
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    poetry install
+    uv run python cleanrl/pqn.py --help
+    uv run python cleanrl/pqn.py --env-id CartPole-v1
+    ```
+
+=== "pip"
+
+    ```bash
+    python cleanrl/pqn.py --help
+    python cleanrl/pqn.py --env-id CartPole-v1
+    ```
+
+### Explanation of the logged metrics
+
+Running `python cleanrl/pqn.py` will automatically record various metrics such as actor or value losses in Tensorboard. Below is the documentation for these metrics:
+
+* `charts/episodic_return`: episodic return of the game
+* `charts/episodic_length`: episodic length of the game
+* `charts/SPS`: number of steps per second
+* `charts/learning_rate`: the current learning rate
+* `losses/td_loss`: the mean squared error (MSE) between the Q values at timestep $t$ and the Bellman update target estimated using the $Q(\lambda)$ returns.
+* `losses/q_values`: it is the average Q values of the sampled data in the replay buffer; useful when gauging if under or over estimation happens.
+
+### Implementation details
+
+1. Vectorized architecture (:material-github: [common/cmd_util.py#L22](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/cmd_util.py#L22))
+2. Orthogonal Initialization of Weights and Constant Initialization of biases (:material-github: [a2c/utils.py#L58)](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L58))
+3. Normalized Q Network (:material-github: [purejaxql/pqn_atari.py#L200](https://github.com/mttga/purejaxql/blob/2205ae5308134d2cedccd749074bff2871832dc8/purejaxql/pqn_atari.py#L200))
+4. Uses the RAdam Optimizer with the default epsilon parameter(:material-github: [purejaxql/pqn_atari.py#L362](https://github.com/mttga/purejaxql/blob/2205ae5308134d2cedccd749074bff2871832dc8/purejaxql/pqn_atari.py#L362))
+5. Adam Learning Rate Annealing (:material-github: [pqn2/pqn2.py#L133-L135](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/pqn2/pqn2.py#L133-L135))
+6. Q Lambda Returns (:material-github: [purejaxql/pqn_atari.py#L446](https://github.com/mttga/purejaxql/blob/2205ae5308134d2cedccd749074bff2871832dc8/purejaxql/pqn_atari.py#L446))
+7. Mini-batch Updates (:material-github: [pqn2/pqn2.py#L157-L166](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/pqn2/pqn2.py#L157-L166))
+8. Global Gradient Clipping (:material-github: [purejaxql/pqn_atari.py#L360](https://github.com/mttga/purejaxql/blob/2205ae5308134d2cedccd749074bff2871832dc8/purejaxql/pqn_atari.py#L360))
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/pqn.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/pqn.sh). Specifically, execute the following command:
+
+``` title="benchmark/pqn.sh" linenums="1"
+--8<-- "benchmark/pqn.sh:0:6"
+```
+
+Episode Rewards:
+
+|                | CleanRL PQN    |
+|:---------------|:---------------|
+| CartPole-v1    | 495.13 ± 6.89  |
+| Acrobot-v1     | -95.63 ± 5.73  |
+| MountainCar-v0 | -200.00 ± 0.00 |
+
+Runtime:
+
+|                |   CleanRL PQN |
+|:---------------|--------------:|
+| CartPole-v1    |      0.833548 |
+| Acrobot-v1     |      1.35797  |
+| MountainCar-v0 |      1.02083  |
+
+Learning curves:
+
+``` title="benchmark/pqn_plot.sh" linenums="1"
+--8<-- "benchmark/pqn_plot.sh:1:9"
+```
+
+<img src="../pqn/pqn_state.png">
+
+Tracked experiments: 
+
+<iframe src="https://api.wandb.ai/links/rogercreus/9s50xw0j" style="width:100%; height:500px" title="CleanRL-s-PPO-TrXL"></iframe>
+
+## `pqn_atari_envpool.py`
+
+The [pqn_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool.py) has the following features:
+
+* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
+* For Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+???+ warning
+
+    Note that `pqn_atari_envpool.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
+
+???+ bug
+
+    EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PQN implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
+
+    * Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
+    * Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
+
+    This causes the $s_{last}$ to be off by one. 
+    See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
+
+
+### Usage
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[envpool]"
+    uv run python cleanrl/pqn_atari_envpool.py --help
+    uv run python cleanrl/pqn_atari_envpool.py --env-id Breakout-v5
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-envpool.txt
+    python cleanrl/pqn_atari_envpool.py --help
+    python cleanrl/pqn_atari_envpool.py --env-id Breakout-v5
+    ```
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/pqn/#explanation-of-the-logged-metrics) for `pqn.py`.
+
+### Implementation details
+
+[pqn_atari_envpool.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool.py) uses a customized `RecordEpisodeStatistics` to work with envpool but has the same other implementation details as `ppo_atari.py`.
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/pqn.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/pqn.sh). Specifically, execute the following command:
+
+``` title="benchmark/pqn.sh" linenums="1"
+--8<-- "benchmark/pqn.sh:12:17"
+```
+
+Episode Rewards:
+
+|                  | CleanRL PQN       |
+|:-----------------|:------------------|
+| Breakout-v5      | 384.85 ± 12.39    |
+| SpaceInvaders-v5 | 1325.20 ± 78.49   |
+| BeamRider-v5     | 5753.03 ± 2394.70 |
+| Pong-v5          | 20.49 ± 0.11      |
+| MsPacman-v5      | 2298.83 ± 128.24  |
+
+Runtime:
+
+|                  |   CleanRL PQN |
+|:-----------------|--------------:|
+| Breakout-v5      |       42.8203 |
+| SpaceInvaders-v5 |       41.2196 |
+| BeamRider-v5     |       43.0951 |
+| Pong-v5          |       40.7316 |
+| MsPacman-v5      |       43.7812 |
+
+
+Learning curves:
+
+``` title="benchmark/pqn_plot.sh" linenums="1"
+--8<-- "benchmark/pqn_plot.sh:11:29"
+```
+
+<img src="../pqn/pqn.png">
+
+Tracked experiments: 
+
+<iframe src="https://wandb.ai/rogercreus/cleanRL/reports/PQN-PR-472--Vmlldzo5ODg1NTkx" style="width:100%; height:500px" title="CleanRL-s-PPO-TrXL"></iframe>
+
+## `pqn_atari_envpool_lstm.py`
+
+The [pqn_atari_envpool_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool_lstm.py) has the following features:
+
+* Uses the blazing fast [Envpool](https://github.com/sail-sg/envpool) vectorized environment.
+* For Atari games using LSTM without stacked frames. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+???+ warning
+
+    Note that `pqn_atari_envpool.py` does not work in Windows :fontawesome-brands-windows: and MacOs :fontawesome-brands-apple:. See envpool's built wheels here: [https://pypi.org/project/envpool/#files](https://pypi.org/project/envpool/#files)
+
+???+ bug
+
+    EnvPool's vectorized environment **does not behave the same** as gym's vectorized environment, which causes a compatibility bug in our PQN implementation. When an action $a$ results in an episode termination or truncation, the environment generates $s_{last}$ as the terminated or truncated state; we then use $s_{new}$ to denote the initial state of the new episodes. Here is how the bahviors differ:
+
+    * Under the vectorized environment of `envpool<=0.6.4`, the `obs` in `obs, reward, done, info = env.step(action)` is the truncated state $s_{last}$
+    * Under the vectorized environment of `gym==0.23.1`, the `obs` in `obs, reward, done, info = env.step(action)` is the initial state $s_{new}$.
+
+    This causes the $s_{last}$ to be off by one. 
+    See [:material-github: sail-sg/envpool#194](https://github.com/sail-sg/envpool/issues/194) for more detail. However, it does not seem to impact performance, so we take a note here and await for the upstream fix.
+
+### Usage
+
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari]"
+    uv run python cleanrl/pqn_atari_envpool_lstm.py --help
+    uv run python cleanrl/pqn_atari_envpool_lstm.py --env-id Breakout-v5
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/pqn_atari_envpool_lstm.py --help
+    python cleanrl/pqn_atari_envpool_lstm.py --env-id Breakout-v5
+    ```
+
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/pqn/#explanation-of-the-logged-metrics) for `pqn.py`.
+
+### Implementation details
+
+[pqn_atari_envpool_lstm.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/pqn_atari_envpool_lstm.py) is based on the "5 LSTM implementation details" in [The 37 Implementation Details of Proximal Policy Optimization](https://iclr-blog-track.github.io/2022/03/25/pqn-implementation-details/), which are as follows:
+
+1. Layer initialization for LSTM layers (:material-github: [a2c/utils.py#L84-L86](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L84-L86))
+2. Initialize the LSTM states to be zeros (:material-github: [common/models.py#L179](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L179))
+3. Reset LSTM states at the end of the episode (:material-github: [common/models.py#L141](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/common/models.py#L141))
+4. Prepare sequential rollouts in mini-batches (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
+5. Reconstruct LSTM states during training (:material-github: [a2c/utils.py#L81](https://github.com/openai/baselines/blob/ea25b9e8b234e6ee1bca43083f8f3cf974143998/baselines/a2c/utils.py#L81))
+
+To help test out the memory, we remove the 4 stacked frames from the observation (i.e., using `env = gym.wrappers.FrameStack(env, 1)` instead of `env = gym.wrappers.FrameStack(env, 4)` like in `ppo_atari.py` )
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/pqn.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/pqn.sh). Specifically, execute the following command:
+
+``` title="benchmark/pqn.sh" linenums="1"
+--8<-- "benchmark/pqn.sh:23:28"
+```
+
+
+Episode Rewards:
+
+|                  | CleanRL PQN       |
+|:-----------------|:------------------|
+| Breakout-v5      | 400.35 ± 9.08     |
+| SpaceInvaders-v5 | 813.47 ± 58.14    |
+| BeamRider-v5     | 11161.43 ± 579.88 |
+| Pong-v5          | 20.43 ± 0.11      |
+| MsPacman-v5      | 1649.63 ± 135.80  |
+
+Runtime:
+
+|                  |   CleanRL PQN |
+|:-----------------|--------------:|
+| Breakout-v5      |       178.144 |
+| SpaceInvaders-v5 |       209.603 |
+| BeamRider-v5     |       174.153 |
+| Pong-v5          |       160.462 |
+| MsPacman-v5      |       162.222 |
+Learning curves:
+
+``` title="benchmark/pqn_plot.sh" linenums="1"
+--8<-- "benchmark/pqn_plot.sh:32:50"
+```
+
+<img src="../pqn/pqn_lstm.png">
+
+Tracked experiments: 
+
+<iframe src="https://wandb.ai/rogercreus/cleanRL/reports/PQN-PR-472--Vmlldzo5ODg1NTkx" style="width:100%; height:500px" title="CleanRL-s-PPO-TrXL"></iframe>
\ No newline at end of file
diff --git a/cleanrl/docs/rl-algorithms/qdagger.md b/cleanrl/docs/rl-algorithms/qdagger.md
new file mode 100644
index 0000000000000000000000000000000000000000..925219d1384856dfb4e3186a340c100a85e36cd7
--- /dev/null
+++ b/cleanrl/docs/rl-algorithms/qdagger.md
@@ -0,0 +1,201 @@
+# QDagger
+
+## Overview
+
+QDagger is an extension of the DQN algorithm that uses previously computed results, like teacher policy and teacher replay buffer, to help train student policy. This method eliminates the need for learning from scratch, improving sample efficiency and reducing computational effort in training new policy.
+
+Original paper:
+
+* [Reincarnating Reinforcement Learning: Reusing Prior Computation to Accelerate Progress](https://arxiv.org/abs/2206.01626)
+
+Reference resources:
+
+* :material-github: [google-research/reincarnating_rl](https://github.com/google-research/reincarnating_rl)
+* [Original Paper's Website](https://agarwl.github.io/reincarnating_rl/)
+
+## Implemented Variants
+
+| Variants Implemented      | Description |
+| ----------- | ----------- |
+| :material-github: [`qdagger_dqn_atari_impalacnn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_impalacnn.py), :material-file-document: [docs](/rl-algorithms/qdagger/#qdagger_dqn_atari_impalacnnpy) | For playing Atari games. It uses Impala-CNN from RainbowDQN and common atari-based pre-processing techniques. |
+| :material-github: [`qdagger_dqn_atari_jax_impalacnn.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_jax_impalacnn.py), :material-file-document: [docs](/rl-algorithms/qdagger/#qdagger_dqn_atari_jax_impalacnnpy) | For playing Atari games. It uses Impala-CNN from RainbowDQN and common atari-based pre-processing techniques. |
+
+
+Below are our single-file implementations of QDagger:
+
+
+## `qdagger_dqn_atari_impalacnn.py`
+
+The [qdagger_dqn_atari_impalacnn.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_impalacnn.py) has the following features:
+
+* For playing Atari games. It uses Impala-CNN from RainbowDQN and common atari-based pre-processing techniques.
+* Its teacher policy uses CleanRL's `dqn_atari` policy from the [huggingface/cleanrl](https://huggingface.co/cleanrl) repository.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+### Usage
+
+```bash
+uv pip install ".[atari]"
+python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id BreakoutNoFrameskip-v4
+python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id PongNoFrameskip-v4
+```
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari]"
+    uv run python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id BreakoutNoFrameskip-v4
+    uv run python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id PongNoFrameskip-v4
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id BreakoutNoFrameskip-v4
+    python cleanrl/qdagger_dqn_atari_impalacnn.py --env-id PongNoFrameskip-v4
+    ```
+
+
+### Explanation of the logged metrics
+
+Running `python cleanrl/qdagger_dqn_atari_impalacnn.py` will automatically record various metrics such as value or distillation losses in Tensorboard. Below is the documentation for these metrics:
+
+* `charts/episodic_return`: episodic return of the game
+* `charts/SPS`: number of steps per second
+* `losses/td_loss`: the mean squared error (MSE) between the Q values at timestep $t$ and the Bellman update target estimated using the reward $r_t$ and the Q values at timestep $t+1$, thus minimizing the *one-step* temporal difference. Formally, it can be expressed by the equation below.
+$$
+    J(\theta^{Q}) = \mathbb{E}_{(s,a,r,s') \sim \mathcal{D}} \big[ (Q(s, a) - y)^2 \big],
+$$
+with the Bellman update target is $y = r + \gamma \, Q^{'}(s', a')$ and the replay buffer is $\mathcal{D}$.
+* `losses/q_values`: implemented as `qf1(data.observations, data.actions).view(-1)`, it is the average Q values of the sampled data in the replay buffer; useful when gauging if under or over estimation happens.
+* `losses/distill_loss`: the distillation loss, which is the KL divergence between the teacher policy $\pi_T$ and the student policy $\pi$. Formally, it can be expressed by the equation below.
+$$
+    L_{\text{distill}} = \lambda_t \mathbb{E}_{(s,a,r,s') \sim \mathcal{D}} \left[ \sum_a \pi_T(a|s)\log\pi(a|s)\right]
+$$
+* `Charts/distill_coeff`: the coefficient $\lambda_t$ for the distillation loss, which is a function of the ratio between the teacher policy $\pi_T$ and the student policy $\pi$. Formally, it can be expressed by the equation below.
+$$
+\lambda_t = 1_{t<t_0}\max(1 - G^\pi/G^{\pi_T}, 0)
+$$
+* `losses/loss`: the total loss, which is the sum of the TD loss and the distillation loss.
+$$
+    L_{\text{qdagger}} = J(\theta^{Q}) + L_{\text{distill}}
+$$
+* `charts/teacher/avg_episodic_return`: average episodic return of teacher policy evaluation
+* `charts/offline/avg_episodic_return`: average episodic return of policy evaluation in offline training phase
+
+
+### Implementation details
+
+`qdagger_dqn_atari_impalacnn.py` is based on (Agarwal et al., 2022)[^1] but presents a few implementation differences:
+
+* (Agarwal et al., 2022)[^1] uses the teacher replay buffer data saved during the training of the teacher policy, but our teacher policy, which uses CleanRL's `dqn_atari` from the [huggingface/cleanrl](https://huggingface.co/cleanrl), does not contain replay buffer data, so we populate the teacher replay buffer with the teacher policy before training. For more details, see section A.5 "Additional ablations for QDagger" in the original paper.
+* (Agarwal et al., 2022)[^1] uses DQN (Adam) @ 400M frames, but we use CleanRL's `dqn_atari`, which is DQN (Adam) @ 10M steps(40M frames).
+* We have used an old set of Atari preprocessing techniques that doesn't use sticky action, but the original paper does.
+
+### Experiment results
+
+Below are the average episodic returns for `qdagger_dqn_atari_impalacnn.py`. 
+
+
+| Environment | `qdagger_dqn_atari_impalacnn.py` 10M steps(40M frames) | (Agarwal et al., 2022)[^1] 10M frames |
+| ----------- | ----------- | ----------- |
+| BreakoutNoFrameskip-v4 | 295.55 ± 12.30 | 275.15 ± 20.65 |
+| PongNoFrameskip-v4  | 19.72 ± 0.20 | - |
+| BeamRiderNoFrameskip-v4 | 9284.99 ± 242.28 | 6514.25 ± 411.10 |
+
+
+Learning curves:
+
+<div class="grid-container">
+<img src="../qdagger/BreakoutNoFrameskip-v4.png">
+
+<img src="../qdagger/PongNoFrameskip-v4.png">
+
+<img src="../qdagger/BeamRiderNoFrameskip-v4.png">
+</div>
+
+Learning curve comparison with `dqn_atari`:
+
+<img src="../qdagger/compare.png">
+
+Tracked experiments and game play videos:
+
+<iframe src="https://wandb.ai/openrlbenchmark/openrlbenchmark/reports/Atari-CleanRL-s-Qdagger--Vmlldzo0NTg1ODY5" style="width:100%; height:500px" title="CleanRL DQN + Atari Tracked Experiments"></iframe>
+
+
+## `qdagger_dqn_atari_jax_impalacnn.py`
+
+
+The [qdagger_dqn_atari_jax_impalacnn.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_jax_impalacnn.py) has the following features:
+
+* Uses [Jax](https://github.com/google/jax), [Flax](https://github.com/google/flax), and [Optax](https://github.com/deepmind/optax) instead of `torch`.  [qdagger_dqn_atari_jax_impalacnn.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_jax_impalacnn.py) is roughly 25%-50% faster than  [qdagger_dqn_atari_impalacnn.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/qdagger_dqn_atari_impalacnn.py)
+* For playing Atari games. It uses Impala-CNN from RainbowDQN and common atari-based pre-processing techniques.
+* Its teacher policy uses CleanRL's `dqn_atari_jax` policy from the [huggingface/cleanrl](https://huggingface.co/cleanrl) repository.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+### Usage
+
+
+=== "poetry"
+
+    ```bash
+    uv pip install ".[atari, jax]"
+    uv pip install --upgrade "jax[cuda11_cudnn82]==0.4.8" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    uv run python cleanrl/qdagger_dqn_atari_jax_impalacnn.py --env-id BreakoutNoFrameskip-v4
+    uv run python cleanrl/qdagger_dqn_atari_jax_impalacnn.py --env-id PongNoFrameskip-v4
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    pip install -r requirements/requirements-jax.txt
+    pip install --upgrade "jax[cuda]==0.3.17" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+    python cleanrl/qdagger_dqn_atari_jax_impalacnn.py --env-id BreakoutNoFrameskip-v4
+    python cleanrl/qdagger_dqn_atari_jax_impalacnn.py --env-id PongNoFrameskip-v4
+    ```
+
+
+???+ warning
+
+    Note that JAX does not work in Windows :fontawesome-brands-windows:. The official [docs](https://github.com/google/jax#installation) recommends using Windows Subsystem for Linux (WSL) to install JAX.
+
+### Explanation of the logged metrics
+
+See [related docs](/rl-algorithms/qdagger/#explanation-of-the-logged-metrics) for `qdagger_dqn_atari_impalacnn.py`.
+
+### Implementation details
+
+See [related docs](/rl-algorithms/qdagger/#implementation-details) for `qdagger_dqn_atari_impalacnn.py`.
+
+### Experiment results
+
+Below are the average episodic returns for `qdagger_dqn_atari_jax_impalacnn.py`.
+
+
+| Environment | `qdagger_dqn_atari_jax_impalacnn.py` 10M steps(40M frames) | (Agarwal et al., 2022)[^1] 10M frames |
+| ----------- | ----------- | ----------- |
+| BreakoutNoFrameskip-v4 | 335.08 ± 19.12 | 275.15 ± 20.65 |
+| PongNoFrameskip-v4  | 18.75 ± 0.19 | - |
+| BeamRiderNoFrameskip-v4 | 8024.75 ± 579.02 | 6514.25 ± 411.10 |
+
+
+Learning curves:
+
+<div class="grid-container">
+<img src="../qdagger/jax/BreakoutNoFrameskip-v4.png">
+
+<img src="../qdagger/jax/PongNoFrameskip-v4.png">
+
+<img src="../qdagger/jax/BeamRiderNoFrameskip-v4.png">
+</div>
+
+Learning curve comparison with `dqn_atari_jax`:
+
+<img src="../qdagger/jax/compare.png">
+
+
+[^1]:Agarwal, Rishabh, Max Schwarzer, Pablo Samuel Castro, Aaron Courville, and Marc G. Bellemare. “Reincarnating Reinforcement Learning: Reusing Prior Computation to Accelerate Progress.” arXiv, October 4, 2022. http://arxiv.org/abs/2206.01626.
diff --git a/cleanrl/docs/rl-algorithms/rainbow.md b/cleanrl/docs/rl-algorithms/rainbow.md
new file mode 100644
index 0000000000000000000000000000000000000000..0a186b20276c36a251e0c13240cc01b42712df4c
--- /dev/null
+++ b/cleanrl/docs/rl-algorithms/rainbow.md
@@ -0,0 +1,137 @@
+# Rainbow
+
+## Overview
+
+The Rainbow algorithm is an extension of DQN that combines multiple improvements:
+
+* Prioritized Experience Replay
+* Dueling Network Architecture
+* Noisy Networks
+* Distributional Q-Learning
+* N-step Learning
+* Double Q-Learning
+
+Original papers: 
+
+* [Rainbow: Combining Improvements in Deep Reinforcement Learning](https://arxiv.org/abs/1710.02298)
+
+Reference resources:
+
+* :material-github: [Dopamine](https://github.com/google/dopamine)
+
+* :material-github: [Kaixhin](https://github.com/Kaixhin/Rainbow)
+
+## Implemented Variants
+
+| Variants Implemented      | Description |
+| ----------- | ----------- |
+| :material-github: [`rainbow_atari.py`](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/rainbow_atari.py), :material-file-document: [docs](/rl-algorithms/rainbow/#rainbow_ataripy) |  For playing Atari games. It uses convolutional layers and common atari-based pre-processing techniques. |
+
+
+## `rainbow_atari.py`
+
+The [rainbow_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/rainbow_atari.py) has the following features:
+
+* For playing Atari games. It uses convolutional layers and common atari-based pre-processing techniques.
+* Works with the Atari's pixel `Box` observation space of shape `(210, 160, 3)`
+* Works with the `Discrete` action space
+
+### Usage
+
+```bash
+poetry install -E atari
+python cleanrl/rainbow_atari.py --env-id BreakoutNoFrameskip-v4
+python cleanrl/rainbow_atari.py --env-id PongNoFrameskip-v4
+```
+
+=== "poetry"
+
+    ```bash
+    poetry install -E atari
+    poetry run python cleanrl/rainbow_atari.py --env-id BreakoutNoFrameskip-v4
+    poetry run python cleanrl/rainbow_atari.py --env-id PongNoFrameskip-v4
+    ```
+
+=== "pip"
+
+    ```bash
+    pip install -r requirements/requirements-atari.txt
+    python cleanrl/rainbow_atari.py --env-id BreakoutNoFrameskip-v4
+    python cleanrl/rainbow_atari.py --env-id PongNoFrameskip-v4
+    ```
+
+
+### Explanation of the logged metrics
+
+Running `python cleanrl/rainbow_atari.py` will automatically record various metrics such as actor or value losses in Tensorboard. Below is the documentation for these metrics:
+
+* `charts/episodic_return`: episodic return of the game
+* `charts/SPS`: number of steps per second
+* `losses/td_loss`: the n-step distributional TD loss
+* `losses/q_values`: the mean Q values of the sampled data in the replay buffer
+* `charts/beta`: the beta value of the prioritized experience replay
+
+### Implementation details
+
+[rainbow_atari.py](https://github.com/vwxyzjn/cleanrl/blob/master/cleanrl/rainbow_atari.py) is based on (Hessel et al., 2018)[^1], and uses the same hyperparameters as (Hessel et al., 2018)[^1]. See Table 1 in (Hessel et al., 2018)[^1] for the hyperparameters. However, there are a few implementation differences:
+
+1. `rainbow_atari.py` uses the more popular Adam Optimizer with the `--learning-rate=0.0000625` as follows:
+    ```python
+    optim.Adam(q_network.parameters(), lr=0.0000625)
+    ```
+    whereas (Hessel et al., 2018)[^1] uses the RMSProp optimizer with `--learning-rate=0.0000625`, gradient momentum `0.95`, squared gradient momentum `0.95`, and min squared gradient `0.01` as follows:
+    ```python
+    optim.RMSprop(
+        q_network.parameters(),
+        lr=2.5e-4,
+        momentum=0.95,
+        # ... PyTorch's RMSprop does not directly support
+        # squared gradient momentum and min squared gradient
+        # so we are not sure what to put here.
+    )
+    ``` 
+
+### Experiment results
+
+To run benchmark experiments, see :material-github: [benchmark/rainbow.sh](https://github.com/vwxyzjn/cleanrl/blob/master/benchmark/rainbow.sh). Specifically, execute the following command:
+
+``` title="benchmark/rainbow.sh" linenums="1"
+--8<-- "benchmark/rainbow.sh:0:6"
+```
+
+Below are the average episodic returns for `rainbow_atari.py`. 
+
+|                             | Rainbow   | C51   | DQN   |
+|:----------------------------|:----------------------------------------|:------------------------------------|:------------------------------------|
+| AlienNoFrameskip-v4         | 2907.03 ± 355.53                        | 1831.00 ± 98.23                     | 1275.77 ± 65.41                     |
+| AssaultNoFrameskip-v4       | 7661.11 ± 226.51                        | 3322.54 ± 94.46                     | 3845.70 ± 443.31                    |
+| GopherNoFrameskip-v4        | 8111.07 ± 300.60                        | 8715.60 ± 492.23                    | 10415.53 ± 3438.12                  |
+| YarsRevengeNoFrameskip-v4   | 63536.39 ± 5432.22                      | 11010.99 ± 904.27                   | 15290.12 ± 8010.56                  |
+| SpaceInvadersNoFrameskip-v4 | 1835.52 ± 205.10                        | 2009.05 ± 226.96                    | 1441.68 ± 23.92                     |
+| MsPacmanNoFrameskip-v4      | 3113.30 ± 393.00                        | 2445.13 ± 30.16                     | 2109.43 ± 49.85                     |
+
+
+Learning curves:
+
+Rainbow shows better performance than C51 and DQN.
+<div class="grid-container">
+<img src="../rainbow/rainbow_env_curves.png">
+</div>
+
+Rainbow is also more sample efficient than C51 and DQN.
+
+<div class="grid-container">
+<img src="../rainbow/rainbow_sample_eff.png">
+</div>
+
+Rainbow obtains better aggregated performance than C51 and DQN.
+<div class="grid-container">
+<img src="../rainbow/rainbow_c51_dqn_bars.png">
+</div>
+
+<!-- Tracked experiments:
+
+<iframe src="https://wandb.ai/rogercreus/rainbow?nw=nwuserrogercreus" style="width:100%; height:500px" title="CleanRL Rainbow + Atari Tracked Experiments"></iframe> -->
+
+
+[^1]: Hessel, M., Modayil, J., Hasselt, H.V., Schaul, T., Ostrovski, G., Dabney, W., Horgan, D., Piot, B., Azar, M.G., & Silver, D. (2018). Rainbow: Combining Improvements in Deep Reinforcement Learning. AAAI.
\ No newline at end of file