Update connect4 checkpoint metadata

metadata/total_config.py

@@ -19,9 +19,9 @@ exp_config = {
             'type': 'subprocess'
         },
         'stop_value': 2,
-        'n_evaluator_episode': 10,
+        'n_evaluator_episode': 20,
         'env_id': 'Connect4',
-        'battle_mode': 'self_play_mode',
+        'battle_mode': 'play_with_bot_mode',
         'battle_mode_in_simulation_env': 'self_play_mode',
         'render_mode': None,
         'replay_path': None,
@@ -30,14 +30,16 @@ exp_config = {
         'prob_random_agent': 0,
         'prob_expert_agent': 0,
         'prob_random_action_in_bot': 0.0,
+        'collector_bot_mode_seat_swap': True,
+        'bot_mode_live_seat': 1,
         'screen_scaling': 9,
         'channel_last': False,
         'scale': False,
         'cfg_type': 'Connect4EnvDict',
         'type': 'connect4',
         'import_names': ['zoo.board_games.connect4.envs.connect4_env'],
-        'collector_env_num': 4,
-        'evaluator_env_num': 10
+        'collector_env_num': 32,
+        'evaluator_env_num': 20
     },
     'policy': {
         'model': {
@@ -98,7 +100,7 @@ exp_config = {
                 'figure_path': None,
                 'cfg_type': 'InteractionSerialEvaluatorDict',
                 'stop_value': 2,
-                'n_episode': 10
+                'n_episode': 20
             }
         },
         'other': {
@@ -145,13 +147,13 @@ exp_config = {
         'sampled_algo': False,
         'gumbel_algo': False,
         'mcts_ctree': True,
-        'collector_env_num': 4,
-        'evaluator_env_num': 10,
+        'collector_env_num': 32,
+        'evaluator_env_num': 20,
         'env_type': 'board_games',
         'action_type': 'varied_action_space',
-        'battle_mode': 'self_play_mode',
+        'battle_mode': 'play_with_bot_mode',
         'monitor_extra_statistics': True,
-        'game_segment_length': 42,
+        'game_segment_length': 21,
         'eval_offline': False,
         'calculate_dormant_ratio': False,
         'analysis_sim_norm': False,
@@ -161,7 +163,7 @@ exp_config = {
         'use_augmentation': False,
         'augmentation': ['shift', 'intensity'],
         'ignore_done': False,
-        'update_per_collect': 50,
+        'update_per_collect': 25,
         'replay_ratio': 0.25,
         'batch_size': 256,
         'optim_type': 'Adam',
@@ -171,11 +173,11 @@ exp_config = {
         'weight_decay': 0.0001,
         'momentum': 0.9,
         'grad_clip_value': 0.5,
-        'n_episode': 4,
+        'n_episode': 32,
         'num_segments': 8,
         'num_simulations': 50,
         'discount_factor': 1,
-        'td_steps': 42,
+        'td_steps': 21,
         'num_unroll_steps': 5,
         'reward_loss_weight': 1,
         'value_loss_weight': 0.25,
@@ -209,7 +211,7 @@ exp_config = {
         'type': 'efficientzero',
         'import_names': ['lzero.policy.efficientzero'],
         'model_path':
-        '/workspace/combinatorial_reasoning_post_training/models/simplified5_two_machine_continuation_20260520/shard-0/connect4/attempt-01_260520_133222/ckpt/envstep_5000.pth.tar',
+        '/home/molfetta/combinatorial_reasoning_post_training/models/main5_hf_bot_mode_200k_collector_tuned_20260522/connect4/attempt-01_260522_142412/ckpt/ckpt_last.pth.tar',
         'reanalyze_ratio': 0.0,
         'eval_freq': 1000001,
         'replay_buffer_size': 100000,
@@ -218,6 +220,10 @@ exp_config = {
         'best_ckpt_min_episodes': 20,
         'battle_mode_in_simulation_env': 'self_play_mode',
         'eval_opponent_type': 'env_bot',
+        'fixed_bot_evaluator': {
+            'type': 'arena',
+            'seat_swap': True
+        },
         'previous_best_checkpoint': {
             'path': None,
             'selector': 'best',
@@ -231,6 +237,6 @@ exp_config = {
         'device': 'cuda'
     },
     'exp_name':
-    '/workspace/combinatorial_reasoning_post_training/models/simplified5_two_machine_continuation_20260520/shard-0/connect4/attempt-03_260520_145624',
+    '/home/molfetta/combinatorial_reasoning_post_training/models/connect4_hf_bot_mode_1m_20260522/connect4/attempt-01_260522_173857',
     'seed': 0
 }