code
stringlengths
17
6.64M
def main(args, config): utils.init_distributed_mode(args) device = torch.device(args.device) seed = (args.seed + utils.get_rank()) torch.manual_seed(seed) np.random.seed(seed) random.seed(seed) cudnn.benchmark = True print('Creating captioning dataset') (val_dataset, test_dataset) ...
def train(model, data_loader, optimizer, epoch, device, config): model.train() metric_logger = utils.MetricLogger(delimiter=' ') metric_logger.add_meter('lr', utils.SmoothedValue(window_size=50, fmt='{value:.6f}')) metric_logger.add_meter('loss_ita', utils.SmoothedValue(window_size=50, fmt='{value:.4...
def main(args, config): utils.init_distributed_mode(args) device = torch.device(args.gpu) seed = (args.seed + utils.get_rank()) torch.manual_seed(seed) np.random.seed(seed) random.seed(seed) cudnn.benchmark = True cudnn.deterministic = True print('Creating dataset') datasets = ...
def train(model, data_loader, optimizer, epoch, device, config): model.train() metric_logger = utils.MetricLogger(delimiter=' ') metric_logger.add_meter('lr', utils.SmoothedValue(window_size=50, fmt='{value:.6f}')) metric_logger.add_meter('loss_ita', utils.SmoothedValue(window_size=50, fmt='{value:.4...
def main(args, config): utils.init_distributed_mode(args) device = torch.device(args.gpu) seed = (args.seed + utils.get_rank()) torch.manual_seed(seed) np.random.seed(seed) random.seed(seed) cudnn.benchmark = True cudnn.deterministic = True print('Creating dataset') datasets = ...
def train(model, data_loader, optimizer, epoch, device, config): model.train() metric_logger = utils.MetricLogger(delimiter=' ') metric_logger.add_meter('lr', utils.SmoothedValue(window_size=50, fmt='{value:.6f}')) metric_logger.add_meter('loss', utils.SmoothedValue(window_size=50, fmt='{value:.4f}')...
@torch.no_grad() def evaluate(model, data_loader, device, config): model.eval() metric_logger = utils.MetricLogger(delimiter=' ') header = 'Evaluation:' print_freq = 200 for (image0, image1, text, targets) in metric_logger.log_every(data_loader, print_freq, header): images = torch.cat([im...
def main(args, config): utils.init_distributed_mode(args) device = torch.device(args.device) seed = (args.seed + utils.get_rank()) torch.manual_seed(seed) np.random.seed(seed) random.seed(seed) cudnn.benchmark = True print('Creating dataset') datasets = create_dataset('nlvr', confi...
def cosine_lr_schedule(optimizer, epoch, max_epoch, init_lr, min_lr): 'Decay the learning rate' lr = ((((init_lr - min_lr) * 0.5) * (1.0 + math.cos(((math.pi * epoch) / max_epoch)))) + min_lr) for param_group in optimizer.param_groups: param_group['lr'] = lr
def warmup_lr_schedule(optimizer, step, max_step, init_lr, max_lr): 'Warmup the learning rate' lr = min(max_lr, (init_lr + (((max_lr - init_lr) * step) / max_step))) for param_group in optimizer.param_groups: param_group['lr'] = lr
def step_lr_schedule(optimizer, epoch, init_lr, min_lr, decay_rate): 'Decay the learning rate' lr = max(min_lr, (init_lr * (decay_rate ** epoch))) for param_group in optimizer.param_groups: param_group['lr'] = lr
class SmoothedValue(object): 'Track a series of values and provide access to smoothed values over a\n window or the global series average.\n ' def __init__(self, window_size=20, fmt=None): if (fmt is None): fmt = '{median:.4f} ({global_avg:.4f})' self.deque = deque(maxlen=wi...
class MetricLogger(object): def __init__(self, delimiter='\t'): self.meters = defaultdict(SmoothedValue) self.delimiter = delimiter def update(self, **kwargs): for (k, v) in kwargs.items(): if isinstance(v, torch.Tensor): v = v.item() assert is...
class AttrDict(dict): def __init__(self, *args, **kwargs): super(AttrDict, self).__init__(*args, **kwargs) self.__dict__ = self
def compute_acc(logits, label, reduction='mean'): ret = (torch.argmax(logits, dim=1) == label).float() if (reduction == 'none'): return ret.detach() elif (reduction == 'mean'): return ret.mean().item()
def compute_n_params(model, return_str=True): tot = 0 for p in model.parameters(): w = 1 for x in p.shape: w *= x tot += w if return_str: if (tot >= 1000000.0): return '{:.1f}M'.format((tot / 1000000.0)) else: return '{:.1f}K'.for...
def setup_for_distributed(is_master): '\n This function disables printing when not in master process\n ' import builtins as __builtin__ builtin_print = __builtin__.print def print(*args, **kwargs): force = kwargs.pop('force', False) if (is_master or force): builtin_p...
def is_dist_avail_and_initialized(): if (not dist.is_available()): return False if (not dist.is_initialized()): return False return True
def get_world_size(): if (not is_dist_avail_and_initialized()): return 1 return dist.get_world_size()
def get_rank(): if (not is_dist_avail_and_initialized()): return 0 return dist.get_rank()
def is_main_process(): return (get_rank() == 0)
def save_on_master(*args, **kwargs): if is_main_process(): torch.save(*args, **kwargs)
def init_distributed_mode(args): '\n initialize the normal job\n ' if (('RANK' in os.environ) and ('WORLD_SIZE' in os.environ)): args.rank = int(os.environ['RANK']) args.world_size = int(os.environ['WORLD_SIZE']) args.gpu = int(os.environ.get('LOCAL_RANK', 0)) print('args...
def init_distributed_ddpjob(args=None): '\n initialize the ddp job\n ' if (dist.is_available() and dist.is_initialized()): return (dist.get_world_size(), dist.get_rank()) try: os.environ['MASTER_PORT'] = '40101' torch.distributed.init_process_group(backend='nccl') except ...
def generate_object_bbox(objects): object_bboxs = [] for (index, object) in enumerate(objects): if (index < 10): object_bboxs.append([int(coord) for coord in object['rect']]) return object_bboxs
def _loss_names(d): ret = {'itm': 0, 'mlm': 0, 'mpp': 0, 'vqa': 0, 'nlvr2': 0, 'irtr': 0} ret.update(d) return ret
@ex.config def config(): exp_name = 'vilt' seed = 0 datasets = ['coco', 'vg', 'sbu', 'gcc'] loss_names = _loss_names({'itm': 1, 'mlm': 1}) batch_size = 4096 train_transform_keys = ['pixelbert'] val_transform_keys = ['pixelbert'] image_size = 384 max_image_len = (- 1) patch_size...
@ex.named_config def env_dandelin(): data_root = '/data2/dsets/dataset' log_dir = '/data2/vilt/result' num_gpus = 8 num_nodes = 1
@ex.named_config def task_mlm_itm(): exp_name = 'mlm_itm' datasets = ['cc3m'] loss_names = _loss_names({'itm': 1, 'mlm': 1}) batch_size = 4096 max_epoch = 10 max_image_len = 200
@ex.named_config def task_mlm_itm_randaug(): exp_name = 'mlm_itm_randaug' datasets = ['coco', 'vg', 'sbu', 'gcc'] train_transform_keys = ['pixelbert_randaug'] loss_names = _loss_names({'itm': 1, 'mlm': 1}) batch_size = 4096 max_epoch = 10 max_image_len = 200
@ex.named_config def task_mlm_itm_mpp(): exp_name = 'mlm_itm_mpp' datasets = ['coco', 'vg', 'sbu', 'gcc'] loss_names = _loss_names({'itm': 1, 'mlm': 1, 'mpp': 1}) batch_size = 4096 max_epoch = 10 max_image_len = 200
@ex.named_config def task_finetune_nlvr2(): exp_name = 'finetune_nlvr2' datasets = ['nlvr2'] loss_names = _loss_names({'nlvr2': 1}) batch_size = 128 max_epoch = 10 max_steps = None warmup_steps = 0.1 draw_false_image = 0 learning_rate = 0.0001
@ex.named_config def task_finetune_nlvr2_randaug(): exp_name = 'finetune_nlvr2_randaug' datasets = ['nlvr2'] train_transform_keys = ['pixelbert_randaug'] loss_names = _loss_names({'nlvr2': 1}) batch_size = 128 max_epoch = 10 max_steps = None warmup_steps = 0.1 draw_false_image = 0 ...
@ex.named_config def task_finetune_vqa(): exp_name = 'finetune_vqa' datasets = ['vqa'] loss_names = _loss_names({'vqa': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 draw_false_image = 0 learning_rate = 0.0001 val_check_interval = 0.1 lr_mult = 10
@ex.named_config def task_finetune_vqa_randaug(): exp_name = 'finetune_vqa_randaug' datasets = ['vqa'] train_transform_keys = ['pixelbert_randaug'] loss_names = _loss_names({'vqa': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 draw_false_image = 0 lear...
@ex.named_config def task_finetune_irtr_coco(): exp_name = 'finetune_irtr_coco' datasets = ['coco'] loss_names = _loss_names({'itm': 0.5, 'irtr': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 get_recall_metric = True draw_false_text = 15 learning_rate ...
@ex.named_config def task_finetune_irtr_coco_randaug(): exp_name = 'finetune_irtr_coco_randaug' datasets = ['coco'] train_transform_keys = ['pixelbert_randaug'] loss_names = _loss_names({'itm': 0.5, 'irtr': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 get...
@ex.named_config def task_finetune_irtr_f30k(): exp_name = 'finetune_irtr_f30k' datasets = ['f30k'] loss_names = _loss_names({'itm': 0.5, 'irtr': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 get_recall_metric = True draw_false_text = 15 learning_rate ...
@ex.named_config def task_finetune_irtr_f30k_randaug(): exp_name = 'finetune_irtr_f30k_randaug' datasets = ['f30k'] train_transform_keys = ['pixelbert_randaug'] loss_names = _loss_names({'itm': 0.5, 'irtr': 1}) batch_size = 256 max_epoch = 10 max_steps = None warmup_steps = 0.1 get...
@ex.named_config def step25k(): max_epoch = 100 max_steps = 25000
@ex.named_config def step50k(): max_epoch = 100 max_steps = 50000
@ex.named_config def step100k(): max_epoch = 100 max_steps = 100000
@ex.named_config def step200k(): max_epoch = 200 max_steps = 200000
@ex.named_config def vit32_base(): vit = 'vit_base_patch32_384' patch_size = 32 hidden_size = 768 num_heads = 12 num_layers = 12
class NoopResetEnv(gym.Wrapper): 'Sample initial states by taking random number of no-ops on reset.\n\n No-op is assumed to be action 0.\n\n :param gym.Env env: the environment to wrap.\n :param int noop_max: the maximum value of no-ops to run.\n ' def __init__(self, env, noop_max=30): super().__...
class MaxAndSkipEnv(gym.Wrapper): 'Return only every `skip`-th frame (frameskipping) using most recent raw\n observations (for max pooling across time steps)\n\n :param gym.Env env: the environment to wrap.\n :param int skip: number of `skip`-th frame.\n ' def __init__(self, env, skip=4): super()...
class EpisodicLifeEnv(gym.Wrapper): 'Make end-of-life == end-of-episode, but only reset on true game over.\n\n It helps the value estimation.\n\n :param gym.Env env: the environment to wrap.\n ' def __init__(self, env): super().__init__(env) self.lives = 0 self.was_real_done = True...
class FireResetEnv(gym.Wrapper): 'Take action on reset for environments that are fixed until firing.\n\n Related discussion: https://github.com/openai/baselines/issues/240\n\n :param gym.Env env: the environment to wrap.\n ' def __init__(self, env): super().__init__(env) assert (env.unwrap...
class WarpFrame(gym.ObservationWrapper): 'Warp frames to 84x84 as done in the Nature paper and later work.\n\n :param gym.Env env: the environment to wrap.\n ' def __init__(self, env): super().__init__(env) self.size = 84 self.observation_space = gym.spaces.Box(low=np.min(env.observ...
class ScaledFloatFrame(gym.ObservationWrapper): 'Normalize observations to 0~1.\n\n :param gym.Env env: the environment to wrap.\n ' def __init__(self, env): super().__init__(env) low = np.min(env.observation_space.low) high = np.max(env.observation_space.high) self.bias = l...
class ClipRewardEnv(gym.RewardWrapper): 'clips the reward to {+1, 0, -1} by its sign.\n\n :param gym.Env env: the environment to wrap.\n ' def __init__(self, env): super().__init__(env) self.reward_range = ((- 1), 1) def reward(self, reward): 'Bin reward to {+1, 0, -1} by its s...
class FrameStack(gym.Wrapper): 'Stack n_frames last frames.\n\n :param gym.Env env: the environment to wrap.\n :param int n_frames: the number of frames to stack.\n ' def __init__(self, env, n_frames): super().__init__(env) self.n_frames = n_frames self.frames = deque([], maxlen=n_...
def wrap_deepmind(env, episode_life=True, clip_rewards=True, frame_stack=4, scale=False, warp_frame=True): 'Configure environment for DeepMind-style Atari.\n\n The observation is channel-first: (c, h, w) instead of (h, w, c).\n\n :param str env: the raw atari environment.\n :param bool episode_life: wrap the e...
def reset_data() -> None: global data data = {'Num. Workers': [], 'FPS': [], 'Env': [], 'System': [], 'Method': []}
def parse_table(env: str, system: str, suffix: str) -> None: private_copy = {'Num. Workers': [], 'FPS': [], 'Env': [], 'System': [], 'Method': []} sep = f'<!-- {env} - {system} -->' raw = open('README.md').read().split(sep)[1].strip().splitlines() worker_num = list(map(int, raw[0].split('|')[2:(- 1)])...
def benchmark(suffix: str) -> None: global data reset_data() for env in ['Atari', 'Mujoco']: for system in ['Laptop', 'Workstation', 'TPU-VM', 'DGX-A100']: parse_table(env, system, suffix) data = pd.DataFrame(data) print(data.groupby(['Env', 'Method', 'System']).max()) def...
def run_dmc(env, action, frame_skip, total_step): ts = env.reset() t = time.time() for i in tqdm.trange(total_step): if (ts.discount == 0): ts = env.reset() else: ts = env.step(action[i]) fps = ((frame_skip * total_step) / (time.time() - t)) print(f'FPS(dmc)...
def run_envpool(env, action, frame_skip, total_step): ts = env.reset() t = time.time() for i in tqdm.trange(total_step): if (ts.discount[0] == 0): ts = env.reset() else: ts = env.step(action[i:(i + 1)]) fps = ((frame_skip * total_step) / (time.time() - t)) p...
def run(env, num_envs, total_step, async_): if (env == 'atari'): task_id = 'PongNoFrameskip-v4' frame_skip = 4 if (num_envs == 1): env = wrap_deepmind(gym.make(task_id), episode_life=False, clip_rewards=False, frame_stack=4) else: env = gym.vector.make(task_...
def get_version() -> str: with open(os.path.join('..', 'envpool', '__init__.py'), 'r') as f: init = f.read().split() return init[(init.index('__version__') + 2)][1:(- 1)]
def setup(app): app.add_js_file('js/copybutton.js') app.add_css_file('css/style.css')
class _SpecTest(absltest.TestCase): @no_type_check def test_spec(self) -> None: action_nums = {'pong': 6, 'breakout': 4} for task in ['pong', 'breakout']: action_num = action_nums[task] spec = make_spec((task.capitalize() + '-v5')) logging.info(spec) ...
class _DMSyncTest(absltest.TestCase): @no_type_check def test_spec(self) -> None: action_nums = {'pong': 6, 'breakout': 4} for task in ['pong', 'breakout']: action_num = action_nums[task] env = make_dm((task.capitalize() + '-v5')) self.assertIsInstance(env,...
class _GymSyncTest(absltest.TestCase): @no_type_check def test_spec(self) -> None: action_nums = {'pong': 6, 'breakout': 4} for task in ['pong', 'breakout']: action_num = action_nums[task] env = make_gym((task.capitalize() + '-v5')) self.assertIsInstance(en...
class DQN(nn.Module): 'Reference: Human-level control through deep reinforcement learning.' def __init__(self, c: int, h: int, w: int, action_shape: Sequence[int], device: Union[(str, int, torch.device)]='cpu', features_only: bool=False) -> None: 'Constructor of DQN.' super().__init__() ...
class C51(DQN): 'Reference: A distributional perspective on reinforcement learning.' def __init__(self, c: int, h: int, w: int, action_shape: Sequence[int], num_atoms: int=51, device: Union[(str, int, torch.device)]='cpu') -> None: 'Constructor of C51.' self.action_num = np.prod(action_shape)...
class QRDQN(DQN): 'Reference: Distributional Reinforcement Learning with Quantile Regression.' def __init__(self, c: int, h: int, w: int, action_shape: Sequence[int], num_quantiles: int=200, device: Union[(str, int, torch.device)]='cpu') -> None: 'Constructor of QRDQN.' self.action_num = np...
class _AtariPretrainTest(absltest.TestCase): def eval_qrdqn(self, task: str, resume_path: str, num_envs: int=10, seed: int=0, target_reward: float=0.0) -> None: env = make_gym((task.capitalize() + '-v5'), num_envs=num_envs, seed=seed) state_shape = env.observation_space.shape action_shape...
class _Box2dEnvPoolCorrectnessTest(absltest.TestCase): def run_space_check(self, env0: gym.Env, env1: Any) -> None: 'Check observation_space and action space.' (obs0, obs1) = (env0.observation_space, env1.observation_space) np.testing.assert_allclose(obs0.shape, obs1.shape) (act0,...
class _Box2dEnvPoolDeterministicTest(absltest.TestCase): def run_deterministic_check(self, task_id: str, num_envs: int=4, **kwargs: Any) -> None: env0 = make_gym(task_id, num_envs=num_envs, seed=0, **kwargs) env1 = make_gym(task_id, num_envs=num_envs, seed=0, **kwargs) env2 = make_gym(tas...
class _ClassicControlEnvPoolTest(absltest.TestCase): @no_type_check def run_space_check(self, env0: gym.Env, env1: Any) -> None: 'Check if envpool.observation_space == gym.make().observation_space.' (obs0, obs1) = (env0.observation_space, env1.observation_space) np.testing.assert_allc...
class _DummyEnvPoolTest(absltest.TestCase): def test_config(self) -> None: ref_config_keys = ['num_envs', 'batch_size', 'num_threads', 'max_num_players', 'thread_affinity_offset', 'base_path', 'seed', 'gym_reset_return_info', 'state_num', 'action_num', 'max_episode_steps'] default_conf = _DummyEn...
class _MakeTest(absltest.TestCase): def test_version(self) -> None: print(envpool.__version__) def test_list_all_envs(self) -> None: pprint.pprint(envpool.list_all_envs()) def test_make_atari(self) -> None: self.assertRaises(TypeError, envpool.make, 'Pong-v5') spec = env...
class _MiniGridEnvPoolAlignTest(absltest.TestCase): def check_spec(self, spec0: gym.spaces.Space, spec1: gym.spaces.Space) -> None: self.assertEqual(spec0.dtype, spec1.dtype) if isinstance(spec0, gym.spaces.Discrete): self.assertEqual(spec0.n, spec1.n) elif isinstance(spec0, g...
class _MiniGridEnvPoolDeterministicTest(absltest.TestCase): def run_deterministic_check(self, task_id: str, num_envs: int=4, total: int=5000, seed: int=1, **kwargs: Any) -> None: env0 = make_gym(task_id, num_envs=num_envs, seed=0, **kwargs) env1 = make_gym(task_id, num_envs=num_envs, seed=0, **kw...
class _MujocoDmcAlignTest(absltest.TestCase): def run_space_check(self, env0: dm_env.Environment, env1: Any) -> None: 'Check observation_spec() and action_spec().' (obs0, obs1) = (env0.observation_spec(), env1.observation_spec()) for k in obs0: self.assertTrue(hasattr(obs1, k)...
class _MujocoDmcDeterministicTest(absltest.TestCase): def check(self, domain: str, task: str, obs_keys: List[str], blacklist: Optional[List[str]]=None, num_envs: int=4) -> None: domain_name = ''.join([(g[:1].upper() + g[1:]) for g in domain.split('_')]) task_name = ''.join([(g[:1].upper() + g[1:]...
class _MujocoDmcSuiteExtAlignTest(absltest.TestCase): def run_space_check(self, env0: dm_env.Environment, env1: Any) -> None: 'Check observation_spec() and action_spec().' (obs0, obs1) = (env0.observation_spec(), env1.observation_spec()) for k in obs0: self.assertTrue(hasattr(...
class _MujocoDmcSuiteExtDeterministicTest(absltest.TestCase): def check(self, domain: str, task: str, obs_keys: List[str], blacklist: Optional[List[str]]=None, num_envs: int=4) -> None: domain_name = ''.join([(g[:1].upper() + g[1:]) for g in domain.split('_')]) task_name = ''.join([(g[:1].upper()...
class _MujocoGymAlignTest(absltest.TestCase): @no_type_check def run_space_check(self, env0: gym.Env, env1: Any) -> None: 'Check observation_space and action space.' 'Check observation_space and action space.' (obs0, obs1) = (env0.observation_space, env1.observation_space) np....
class _MujocoGymDeterministicTest(absltest.TestCase): def check(self, task_id: str, num_envs: int=4) -> None: env0 = make_gym(task_id, num_envs=num_envs, seed=0) env1 = make_gym(task_id, num_envs=num_envs, seed=0) env2 = make_gym(task_id, num_envs=num_envs, seed=1) act_space = env...
class _ProcgenEnvPoolTest(absltest.TestCase): def deterministic_check(self, task_id: str, num_envs: int=4, total: int=200) -> None: logging.info(f'deterministic check for {task_id}') env0 = make_gym(task_id, num_envs=num_envs, seed=0) env1 = make_gym(task_id, num_envs=num_envs, seed=0) ...
def py_env(envspec: Type[EnvSpec], envpool: Type[EnvPool]) -> Tuple[(Type[EnvSpec], Type[EnvPool], Type[EnvPool], Type[EnvPool])]: 'Initialize EnvPool for users.' spec_name = envspec.__name__[1:] pool_name = envpool.__name__[1:] return (EnvSpecMeta(spec_name, (envspec,), {}), DMEnvPoolMeta(pool_name.r...
def to_nested_dict(flatten_dict: Dict[(str, Any)], generator: Type=dict) -> Dict[(str, Any)]: 'Convert a flat dict to a hierarchical dict.\n\n The input dict\'s hierarchy is denoted by ``.``.\n\n Example:\n ::\n\n >>> to_nested_dict({"a.b": 2333, "a.c": 666})\n {"a": {"b": 2333, "c": 666}}\n\n Arg...
def to_namedtuple(name: str, hdict: Dict) -> Tuple: 'Convert a hierarchical dict to namedtuple.' return namedtuple(name, hdict.keys())(*[(to_namedtuple(k, v) if isinstance(v, Dict) else v) for (k, v) in hdict.items()])
def dm_spec_transform(name: str, spec: ArraySpec, spec_type: str) -> dm_env.specs.Array: 'Transform ArraySpec to dm_env compatible specs.' if ((np.prod(np.abs(spec.shape)) == 1) and np.isclose(spec.minimum, 0) and (spec.maximum < ACTION_THRESHOLD)): return dm_env.specs.DiscreteArray(name=name, dtype=s...
def gym_spec_transform(name: str, spec: ArraySpec, spec_type: str) -> gym.Space: 'Transform ArraySpec to gym.Env compatible spaces.' if ((np.prod(np.abs(spec.shape)) == 1) and np.isclose(spec.minimum, 0) and (spec.maximum < ACTION_THRESHOLD)): discrete_range = int(((spec.maximum - spec.minimum) + 1)) ...
def gymnasium_spec_transform(name: str, spec: ArraySpec, spec_type: str) -> gymnasium.Space: 'Transform ArraySpec to gymnasium.Env compatible spaces.' if ((np.prod(np.abs(spec.shape)) == 1) and np.isclose(spec.minimum, 0) and (spec.maximum < ACTION_THRESHOLD)): discrete_range = int(((spec.maximum - sp...
def dm_structure(root_name: str, keys: List[str]) -> Tuple[(List[Tuple[(int, ...)]], List[int], PyTreeSpec)]: 'Convert flat keys into tree structure for namedtuple construction.' new_keys = [] for key in keys: if (key in ['obs', 'info']): key = f'obs:{key}' key = key.replace('i...
def gym_structure(keys: List[str]) -> Tuple[(List[Tuple[(str, ...)]], List[int], PyTreeSpec)]: 'Convert flat keys into tree structure for dict construction.' keys = [k.replace(':', '.') for k in keys] dict_tree = to_nested_dict(dict(zip(keys, list(range(len(keys)))))) (paths, indices, treespec) = optr...
class DMEnvPoolMixin(ABC): 'Special treatment for dm_env API.' def observation_spec(self: Any) -> Tuple: 'Observation spec from EnvSpec.' if (not hasattr(self, '_dm_observation_spec')): self._dm_observation_spec = self.spec.observation_spec() return self._dm_observation_sp...
class DMEnvPoolMeta(ABCMeta): 'Additional wrapper for EnvPool dm_env API.' def __new__(cls: Any, name: str, parents: Tuple, attrs: Dict) -> Any: 'Check internal config and initialize data format convertion.' base = parents[0] try: from .lax import XlaMixin pare...
class EnvSpecMixin(ABC): 'Mixin class for EnvSpec, exposed to EnvSpecMeta.' gen_config: Type @property def config(self: EnvSpec) -> NamedTuple: 'Configuration used to create the current EnvSpec.' return self.gen_config(*self._config_values) @property def reward_threshold(self...
class EnvSpecMeta(ABCMeta): 'Additional checker and wrapper for EnvSpec.' def __new__(cls: Any, name: str, parents: Tuple, attrs: Dict) -> Any: 'Check keys and initialize namedtuple config.' base = parents[0] parents = (base, EnvSpecMixin) config_keys = base._config_keys ...
class EnvPoolMixin(ABC): 'Mixin class for EnvPool, exposed to EnvPoolMeta.' _spec: EnvSpec def _check_action(self: EnvPool, actions: List[np.ndarray]) -> None: if hasattr(self, '_check_action_finished'): return self._check_action_finished = True for (a, (k, v)) in zip(...
class GymEnvPoolMixin(ABC): 'Special treatment for gym API.' @property def observation_space(self: Any) -> Union[(gym.Space, Dict[(str, Any)])]: 'Observation space from EnvSpec.' if (not hasattr(self, '_gym_observation_space')): self._gym_observation_space = self.spec.observat...
class GymEnvPoolMeta(ABCMeta, gym.Env.__class__): 'Additional wrapper for EnvPool gym.Env API.' def __new__(cls: Any, name: str, parents: Tuple, attrs: Dict) -> Any: 'Check internal config and initialize data format convertion.' base = parents[0] try: from .lax import XlaM...
class GymnasiumEnvPoolMixin(ABC): 'Special treatment for gymnasim API.' @property def observation_space(self: Any) -> Union[(gymnasium.Space, Dict[(str, Any)])]: 'Observation space from EnvSpec.' if (not hasattr(self, '_gym_observation_space')): self._gym_observation_space = s...
class GymnasiumEnvPoolMeta(ABCMeta, gymnasium.Env.__class__): 'Additional wrapper for EnvPool gymnasium.Env API.' def __new__(cls: Any, name: str, parents: Tuple, attrs: Dict) -> Any: 'Check internal config and initialize data format convertion.' base = parents[0] try: fro...
class XlaMixin(ABC): 'Mixin to provide XLA for envpool class.' def xla(self: Any) -> Tuple[(Any, Callable, Callable, Callable)]: 'Return the XLA version of send/recv/step functions.' (_handle, _recv, _send) = make_xla(self) def recv(handle: jnp.ndarray) -> Union[(TimeStep, Tuple)]: ...