Upload Connect4 checkpoint envstep_90000.pth.tar

README.md

@@ -1,16 +1,22 @@
 ---
-library_name: efficientzero
+library_name: pytorch
 tags:
-- muzero
 - efficientzero
-- connect4
+- muzero
+- board-game
+- combinatorial-reasoning
 - crpt
 ---
 
 # Connect4-CK
 
-Intermediate EfficientZero Connect4 checkpoint from `simplified5_balanced_1day_wandb_online_20260518`.
+Latest exported EfficientZero checkpoint for `connect4` from the simplified-five CRPT training runs.
 
-Uploaded checkpoint: `envstep_5000.pth.tar`.
+- Checkpoint: `checkpoints/envstep_90000.pth.tar`
+- Source path: `/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_90000.pth.tar`
+- Source attempt directory: `/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459`
+- W&B project: `crpt-simplified5-constant-lr-continuation`
+- W&B run: `constant_lr_20260519__connect4__a01`
+- Uploaded at: `2026-05-19T15:04:12Z`
 
-This is a recovery/export test from an ephemeral pod. The resolved training config is included at `configs/connect4.yaml`.
+Companion metadata is stored under `metadata/`, including the resolved LightZero config when available.

checkpoints/envstep_90000.pth.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bda258a15a1dbab2a7d8112baeb6cfdfd59ea1a34e2345eaef9b5db8d4373211
+size 139564171

metadata/checkpoint_index.yaml

@@ -0,0 +1,2234 @@
+checkpoints:
+  ckpt_best.pth.tar:
+    checkpoint_name: ckpt_best.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/ckpt_best.pth.tar
+    saved_at: '2026-05-19T10:13:05.573846+00:00'
+    train_iter: 910
+    envstep: 10031
+    trigger: null
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 910
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 10031
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 83
+        step: 910
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 83
+        step: 10031
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 8.3
+        step: 910
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 8.3
+        step: 10031
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 8.95365625
+        step: 910
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 8.95365625
+        step: 10031
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 9.269956058453774
+        step: 910
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 9.269956058453774
+        step: 10031
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 1.1168621757173223
+        step: 910
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 1.1168621757173223
+        step: 10031
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -0.8
+        step: 910
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -0.8
+        step: 10031
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.6000000000000001
+        step: 910
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.6000000000000001
+        step: 10031
+        walltime: null
+      evaluator_iter/reward_max:
+        value: 1.0
+        step: 910
+        walltime: null
+      evaluator_step/reward_max:
+        value: 1.0
+        step: 10031
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 910
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 10031
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -0.8
+        step: 910
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -0.8
+        step: 10031
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 368
+        step: 910
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 368
+        step: 910
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 10031
+        step: 910
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 6.2646942138671875
+        step: 910
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2058.0
+        step: 910
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 900
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 10031
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 900
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 10031
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 0.010909480466084047
+        step: 900
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 0.010909480466084047
+        step: 10031
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 0.010909480466084047
+        step: 900
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 0.010909480466084047
+        step: 10031
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 8.115459442138672
+        step: 900
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 8.115459442138672
+        step: 10031
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.429311665621671
+        step: 900
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.429311665621671
+        step: 10031
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.4289432727929319
+        step: 900
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.4289432727929319
+        step: 10031
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.21515980769287457
+        step: 900
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.21515980769287457
+        step: 10031
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 4.965807264501398
+        step: 900
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 4.965807264501398
+        step: 10031
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9561161474748091
+        step: 900
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9561161474748091
+        step: 10031
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.5874042294242166
+        step: 900
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.5874042294242166
+        step: 10031
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.060487691313028336
+        step: 900
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.060487691313028336
+        step: 10031
+        walltime: null
+      learner_iter/target_value_avg:
+        value: -0.001775568192401393
+        step: 900
+        walltime: null
+      learner_step/target_value_avg:
+        value: -0.001775568192401393
+        step: 10031
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.04929109527306123
+        step: 900
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.04929109527306123
+        step: 10031
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: -0.001578690177253024
+        step: 900
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: -0.001578690177253024
+        step: 10031
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.02511530691249804
+        step: 900
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.02511530691249804
+        step: 10031
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: -0.0007372399453412403
+        step: 900
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: -0.0007372399453412403
+        step: 10031
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.3114241280338981
+        step: 900
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.3114241280338981
+        step: 10031
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 900
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 10031
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1010
+        step: 900
+        walltime: null
+      collector_step/envstep_count:
+        value: 1010
+        step: 10031
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 25.25
+        step: 900
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 25.25
+        step: 10031
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 9.44826405658907
+        step: 900
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 9.44826405658907
+        step: 10031
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.374188675508478
+        step: 900
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.374188675508478
+        step: 10031
+        walltime: null
+      collector_iter/collect_time:
+        value: 106.89794378636591
+        step: 900
+        walltime: null
+      collector_step/collect_time:
+        value: 106.89794378636591
+        step: 10031
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.4000000059604645
+        step: 900
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.4000000059604645
+        step: 10031
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.9165151715278625
+        step: 900
+        walltime: null
+      collector_step/reward_std:
+        value: 0.9165151715278625
+        step: 10031
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 900
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 10031
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 900
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 10031
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 10031
+        step: 900
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 10031
+        step: 10031
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 368
+        step: 900
+        walltime: null
+      collector_step/total_episode_count:
+        value: 368
+        step: 10031
+        walltime: null
+      collector_iter/total_duration:
+        value: 5777.1309334583675
+        step: 900
+        walltime: null
+      collector_step/total_duration:
+        value: 5777.1309334583675
+        step: 10031
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 0.970061481600583
+        step: 900
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 0.970061481600583
+        step: 10031
+        walltime: null
+  envstep_70000.pth.tar:
+    checkpoint_name: envstep_70000.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_70000.pth.tar
+    saved_at: '2026-05-19T13:50:05.614710+00:00'
+    train_iter: 6180
+    envstep: 70000
+    trigger: envstep
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 5770
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 65076
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 56
+        step: 5770
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 56
+        step: 65076
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 5.6
+        step: 5770
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 5.6
+        step: 65076
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 7.194775390625
+        step: 5770
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 7.194775390625
+        step: 65076
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 7.783425744321305
+        step: 5770
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 7.783425744321305
+        step: 65076
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 1.3898974543430902
+        step: 5770
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 1.3898974543430902
+        step: 65076
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -1.0
+        step: 5770
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -1.0
+        step: 65076
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.0
+        step: 5770
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.0
+        step: 65076
+        walltime: null
+      evaluator_iter/reward_max:
+        value: -1.0
+        step: 5770
+        walltime: null
+      evaluator_step/reward_max:
+        value: -1.0
+        step: 65076
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 5770
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 65076
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -1.0
+        step: 5770
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -1.0
+        step: 65076
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 2476
+        step: 6180
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 1764
+        step: 6180
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 49968
+        step: 6180
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 30.9970703125
+        step: 6180
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2081.33984375
+        step: 6180
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 6100
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 69076
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 6100
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 69076
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 0.8979335264726118
+        step: 6100
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 0.8979335264726118
+        step: 69076
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 0.8979335264726118
+        step: 6100
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 0.8979335264726118
+        step: 69076
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 8.947579123757102
+        step: 6100
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 8.947579123757102
+        step: 69076
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.5767489346590908
+        step: 6100
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.5767489346590908
+        step: 69076
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.4807274847319634
+        step: 6100
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.4807274847319634
+        step: 69076
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.21087754585526206
+        step: 6100
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.21087754585526206
+        step: 69076
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 5.24790130962025
+        step: 6100
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 5.24790130962025
+        step: 69076
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9572498668323862
+        step: 6100
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9572498668323862
+        step: 69076
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.8108745000579141
+        step: 6100
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.8108745000579141
+        step: 69076
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.061316289346326484
+        step: 6100
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.061316289346326484
+        step: 69076
+        walltime: null
+      learner_iter/target_value_avg:
+        value: -0.000118371194482527
+        step: 6100
+        walltime: null
+      learner_step/target_value_avg:
+        value: -0.000118371194482527
+        step: 69076
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.05231790278445591
+        step: 6100
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.05231790278445591
+        step: 69076
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: 0.0001337626811370931
+        step: 6100
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: 0.0001337626811370931
+        step: 69076
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.025459352880716324
+        step: 6100
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.025459352880716324
+        step: 69076
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: -4.914934238926931e-05
+        step: 6100
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: -4.914934238926931e-05
+        step: 69076
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.3467647677118128
+        step: 6100
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.3467647677118128
+        step: 69076
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 6100
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 69076
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1181
+        step: 6100
+        walltime: null
+      collector_step/envstep_count:
+        value: 1181
+        step: 69076
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 29.525
+        step: 6100
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 29.525
+        step: 69076
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 9.84222987476683
+        step: 6100
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 9.84222987476683
+        step: 69076
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.3333524089675471
+        step: 6100
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.3333524089675471
+        step: 69076
+        walltime: null
+      collector_iter/collect_time:
+        value: 119.99313316465076
+        step: 6100
+        walltime: null
+      collector_step/collect_time:
+        value: 119.99313316465076
+        step: 69076
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.4749999940395355
+        step: 6100
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.4749999940395355
+        step: 69076
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.8656644821166992
+        step: 6100
+        walltime: null
+      collector_step/reward_std:
+        value: 0.8656644821166992
+        step: 69076
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 6100
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 69076
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 6100
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 69076
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 69076
+        step: 6100
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 69076
+        step: 69076
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 2448
+        step: 6100
+        walltime: null
+      collector_step/total_episode_count:
+        value: 2448
+        step: 69076
+        walltime: null
+      collector_iter/total_duration:
+        value: 38851.81019043435
+        step: 6100
+        walltime: null
+      collector_step/total_duration:
+        value: 38851.81019043435
+        step: 69076
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 0.9780714071106607
+        step: 6100
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 0.9780714071106607
+        step: 69076
+        walltime: null
+  envstep_75000.pth.tar:
+    checkpoint_name: envstep_75000.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_75000.pth.tar
+    saved_at: '2026-05-19T14:07:06.562230+00:00'
+    train_iter: 6610
+    envstep: 75000
+    trigger: envstep
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 6190
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 70062
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 69
+        step: 6190
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 69
+        step: 70062
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 6.9
+        step: 6190
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 6.9
+        step: 70062
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 12.7152763671875
+        step: 6190
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 12.7152763671875
+        step: 70062
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 5.426543474749669
+        step: 6190
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 5.426543474749669
+        step: 70062
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 0.7864555760506766
+        step: 6190
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 0.7864555760506766
+        step: 70062
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -1.0
+        step: 6190
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -1.0
+        step: 70062
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.0
+        step: 6190
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.0
+        step: 70062
+        walltime: null
+      evaluator_iter/reward_max:
+        value: -1.0
+        step: 6190
+        walltime: null
+      evaluator_step/reward_max:
+        value: -1.0
+        step: 70062
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 6190
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 70062
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -1.0
+        step: 6190
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -1.0
+        step: 70062
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 2648
+        step: 6610
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 1763
+        step: 6610
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 49982
+        step: 6610
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 31.00725555419922
+        step: 6610
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2078.375
+        step: 6610
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 6600
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 74963
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 6600
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 74963
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 0.9603883515704762
+        step: 6600
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 0.9603883515704762
+        step: 74963
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 0.9603883515704762
+        step: 6600
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 0.9603883515704762
+        step: 74963
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 9.02350971915505
+        step: 6600
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 9.02350971915505
+        step: 74963
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.580617500073982
+        step: 6600
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.580617500073982
+        step: 74963
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.484574101187966
+        step: 6600
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.484574101187966
+        step: 74963
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.2109665017236363
+        step: 6600
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.2109665017236363
+        step: 74963
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 5.287646987221458
+        step: 6600
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 5.287646987221458
+        step: 74963
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9595999804410067
+        step: 6600
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9595999804410067
+        step: 74963
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.8126288869164207
+        step: 6600
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.8126288869164207
+        step: 74963
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.06214488839561289
+        step: 6600
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.06214488839561289
+        step: 74963
+        walltime: null
+      learner_iter/target_value_avg:
+        value: -0.0002367424312978983
+        step: 6600
+        walltime: null
+      learner_step/target_value_avg:
+        value: -0.0002367424312978983
+        step: 74963
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.04919351637363434
+        step: 6600
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.04919351637363434
+        step: 74963
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: -0.00015135957668958739
+        step: 6600
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: -0.00015135957668958739
+        step: 74963
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.02580339749428359
+        step: 6600
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.02580339749428359
+        step: 74963
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: -9.829866537598518e-05
+        step: 6600
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: -9.829866537598518e-05
+        step: 74963
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.3910723734985698
+        step: 6600
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.3910723734985698
+        step: 74963
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 6600
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 74963
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1119
+        step: 6600
+        walltime: null
+      collector_step/envstep_count:
+        value: 1119
+        step: 74963
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 27.975
+        step: 6600
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 27.975
+        step: 74963
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 10.896117848612953
+        step: 6600
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 10.896117848612953
+        step: 74963
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.38949482926230394
+        step: 6600
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.38949482926230394
+        step: 74963
+        walltime: null
+      collector_iter/collect_time:
+        value: 102.6971271371157
+        step: 6600
+        walltime: null
+      collector_step/collect_time:
+        value: 102.6971271371157
+        step: 74963
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.25
+        step: 6600
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.25
+        step: 74963
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.9682458639144897
+        step: 6600
+        walltime: null
+      collector_step/reward_std:
+        value: 0.9682458639144897
+        step: 74963
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 6600
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 74963
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 6600
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 74963
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 74963
+        step: 6600
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 74963
+        step: 74963
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 2648
+        step: 6600
+        walltime: null
+      collector_step/total_episode_count:
+        value: 2648
+        step: 74963
+        walltime: null
+      collector_iter/total_duration:
+        value: 41918.13410943206
+        step: 6600
+        walltime: null
+      collector_step/total_duration:
+        value: 41918.13410943206
+        step: 74963
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 1.107473344376133
+        step: 6600
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 1.107473344376133
+        step: 74963
+        walltime: null
+  envstep_80000.pth.tar:
+    checkpoint_name: envstep_80000.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_80000.pth.tar
+    saved_at: '2026-05-19T14:25:15.596441+00:00'
+    train_iter: 7070
+    envstep: 80000
+    trigger: envstep
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 6620
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 75082
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 50
+        step: 6620
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 50
+        step: 75082
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 5.0
+        step: 6620
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 5.0
+        step: 75082
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 6.18236572265625
+        step: 6620
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 6.18236572265625
+        step: 75082
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 8.087518960058793
+        step: 6620
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 8.087518960058793
+        step: 75082
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 1.6175037920117583
+        step: 6620
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 1.6175037920117583
+        step: 75082
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -1.0
+        step: 6620
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -1.0
+        step: 75082
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.0
+        step: 6620
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.0
+        step: 75082
+        walltime: null
+      evaluator_iter/reward_max:
+        value: -1.0
+        step: 6620
+        walltime: null
+      evaluator_step/reward_max:
+        value: -1.0
+        step: 75082
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 6620
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 75082
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -1.0
+        step: 6620
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -1.0
+        step: 75082
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 2832
+        step: 7070
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 1769
+        step: 7070
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 49988
+        step: 7070
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 31.030921936035156
+        step: 7070
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2077.1328125
+        step: 7070
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 7000
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 79326
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 7000
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 79326
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 0.9828287308866327
+        step: 7000
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 0.9828287308866327
+        step: 79326
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 0.9828287308866327
+        step: 7000
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 0.9828287308866327
+        step: 79326
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 9.123100020668723
+        step: 7000
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 9.123100020668723
+        step: 79326
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.5887374444441362
+        step: 7000
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.5887374444441362
+        step: 79326
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.4960533488880505
+        step: 7000
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.4960533488880505
+        step: 79326
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.18883771787990222
+        step: 7000
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.18883771787990222
+        step: 79326
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 5.286637089469216
+        step: 7000
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 5.286637089469216
+        step: 79326
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9650768540122292
+        step: 7000
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9650768540122292
+        step: 79326
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.824982453476299
+        step: 7000
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.824982453476299
+        step: 79326
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.057054925709962845
+        step: 7000
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.057054925709962845
+        step: 79326
+        walltime: null
+      learner_iter/target_value_avg:
+        value: -0.0005918560676615347
+        step: 7000
+        walltime: null
+      learner_step/target_value_avg:
+        value: -0.0005918560676615347
+        step: 79326
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.04756909812038595
+        step: 7000
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.04756909812038595
+        step: 79326
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: -0.00309789812176446
+        step: 7000
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: -0.00309789812176446
+        step: 79326
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.02368997579271143
+        step: 7000
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.02368997579271143
+        step: 79326
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: -0.00024574662728065795
+        step: 7000
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: -0.00024574662728065795
+        step: 79326
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.3728354573249817
+        step: 7000
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.3728354573249817
+        step: 79326
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 7000
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 79326
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1078
+        step: 7000
+        walltime: null
+      collector_step/envstep_count:
+        value: 1078
+        step: 79326
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 26.95
+        step: 7000
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 26.95
+        step: 79326
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 10.314620505059771
+        step: 7000
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 10.314620505059771
+        step: 79326
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.38273174415806205
+        step: 7000
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.38273174415806205
+        step: 79326
+        walltime: null
+      collector_iter/collect_time:
+        value: 104.5118431134906
+        step: 7000
+        walltime: null
+      collector_step/collect_time:
+        value: 104.5118431134906
+        step: 79326
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.10000000149011612
+        step: 7000
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.10000000149011612
+        step: 79326
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.994987428188324
+        step: 7000
+        walltime: null
+      collector_step/reward_std:
+        value: 0.994987428188324
+        step: 79326
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 7000
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 79326
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 7000
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 79326
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 79326
+        step: 7000
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 79326
+        step: 79326
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 2808
+        step: 7000
+        walltime: null
+      collector_step/total_episode_count:
+        value: 2808
+        step: 79326
+        walltime: null
+      collector_iter/total_duration:
+        value: 44287.69188188307
+        step: 7000
+        walltime: null
+      collector_step/total_duration:
+        value: 44287.69188188307
+        step: 79326
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 1.0988913002798388
+        step: 7000
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 1.0988913002798388
+        step: 79326
+        walltime: null
+  envstep_85000.pth.tar:
+    checkpoint_name: envstep_85000.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_85000.pth.tar
+    saved_at: '2026-05-19T14:44:43.675350+00:00'
+    train_iter: 7580
+    envstep: 85000
+    trigger: envstep
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 7080
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 80045
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 73
+        step: 7080
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 73
+        step: 80045
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 7.3
+        step: 7080
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 7.3
+        step: 80045
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 10.966330078125
+        step: 7080
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 10.966330078125
+        step: 80045
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 6.656739262811008
+        step: 7080
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 6.656739262811008
+        step: 80045
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 0.9118820907960284
+        step: 7080
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 0.9118820907960284
+        step: 80045
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -1.0
+        step: 7080
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -1.0
+        step: 80045
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.0
+        step: 7080
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.0
+        step: 80045
+        walltime: null
+      evaluator_iter/reward_max:
+        value: -1.0
+        step: 7080
+        walltime: null
+      evaluator_step/reward_max:
+        value: -1.0
+        step: 80045
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 7080
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 80045
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -1.0
+        step: 7080
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -1.0
+        step: 80045
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 3036
+        step: 7580
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 1795
+        step: 7580
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 49996
+        step: 7580
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 31.115081787109375
+        step: 7580
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2077.03515625
+        step: 7580
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 7500
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 84336
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 7500
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 84336
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 1.1162791360508313
+        step: 7500
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 1.1162791360508313
+        step: 84336
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 1.1162791360508313
+        step: 7500
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 1.1162791360508313
+        step: 84336
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 9.150714440779252
+        step: 7500
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 9.150714440779252
+        step: 84336
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.615344033096776
+        step: 7500
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.615344033096776
+        step: 84336
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.5126509810938977
+        step: 7500
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.5126509810938977
+        step: 84336
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.22831940650939941
+        step: 7500
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.22831940650939941
+        step: 84336
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 5.274801861156117
+        step: 7500
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 5.274801861156117
+        step: 84336
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9581454797224566
+        step: 7500
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9581454797224566
+        step: 84336
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.818570922721516
+        step: 7500
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.818570922721516
+        step: 84336
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.05906723676757379
+        step: 7500
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.05906723676757379
+        step: 84336
+        walltime: null
+      learner_iter/target_value_avg:
+        value: 0.0010653409514237535
+        step: 7500
+        walltime: null
+      learner_step/target_value_avg:
+        value: 0.0010653409514237535
+        step: 84336
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.049273560162295
+        step: 7500
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.049273560162295
+        step: 84336
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: -0.001351071257059547
+        step: 7500
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: -0.001351071257059547
+        step: 84336
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.02452551421116699
+        step: 7500
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.02452551421116699
+        step: 84336
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: 0.0004423439650881019
+        step: 7500
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: 0.0004423439650881019
+        step: 84336
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.45741088823838666
+        step: 7500
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.45741088823838666
+        step: 84336
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 7500
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 84336
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1013
+        step: 7500
+        walltime: null
+      collector_step/envstep_count:
+        value: 1013
+        step: 84336
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 25.325
+        step: 7500
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 25.325
+        step: 84336
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 9.130103441466435
+        step: 7500
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 9.130103441466435
+        step: 84336
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.3605174113116065
+        step: 7500
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.3605174113116065
+        step: 84336
+        walltime: null
+      collector_iter/collect_time:
+        value: 110.95164545444588
+        step: 7500
+        walltime: null
+      collector_step/collect_time:
+        value: 110.95164545444588
+        step: 84336
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.375
+        step: 7500
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.375
+        step: 84336
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.9134413003921509
+        step: 7500
+        walltime: null
+      collector_step/reward_std:
+        value: 0.9134413003921509
+        step: 84336
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 7500
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 84336
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 7500
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 84336
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 84336
+        step: 7500
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 84336
+        step: 84336
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 3008
+        step: 7500
+        walltime: null
+      collector_step/total_episode_count:
+        value: 3008
+        step: 84336
+        walltime: null
+      collector_iter/total_duration:
+        value: 47114.3674302446
+        step: 7500
+        walltime: null
+      collector_step/total_duration:
+        value: 47114.3674302446
+        step: 84336
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 1.27042619265444
+        step: 7500
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 1.27042619265444
+        step: 84336
+        walltime: null
+  envstep_90000.pth.tar:
+    checkpoint_name: envstep_90000.pth.tar
+    checkpoint_path: /workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_90000.pth.tar
+    saved_at: '2026-05-19T15:02:38.872182+00:00'
+    train_iter: 8050
+    envstep: 90000
+    trigger: envstep
+    latest_scalars:
+      evaluator_iter/episode_count:
+        value: 10
+        step: 7590
+        walltime: null
+      evaluator_step/episode_count:
+        value: 10
+        step: 85111
+        walltime: null
+      evaluator_iter/envstep_count:
+        value: 55
+        step: 7590
+        walltime: null
+      evaluator_step/envstep_count:
+        value: 55
+        step: 85111
+        walltime: null
+      evaluator_iter/avg_envstep_per_episode:
+        value: 5.5
+        step: 7590
+        walltime: null
+      evaluator_step/avg_envstep_per_episode:
+        value: 5.5
+        step: 85111
+        walltime: null
+      evaluator_iter/evaluate_time:
+        value: 5.41702392578125
+        step: 7590
+        walltime: null
+      evaluator_step/evaluate_time:
+        value: 5.41702392578125
+        step: 85111
+        walltime: null
+      evaluator_iter/avg_envstep_per_sec:
+        value: 10.15317649572091
+        step: 7590
+        walltime: null
+      evaluator_step/avg_envstep_per_sec:
+        value: 10.15317649572091
+        step: 85111
+        walltime: null
+      evaluator_iter/avg_time_per_episode:
+        value: 1.8460320901310745
+        step: 7590
+        walltime: null
+      evaluator_step/avg_time_per_episode:
+        value: 1.8460320901310745
+        step: 85111
+        walltime: null
+      evaluator_iter/reward_mean:
+        value: -1.0
+        step: 7590
+        walltime: null
+      evaluator_step/reward_mean:
+        value: -1.0
+        step: 85111
+        walltime: null
+      evaluator_iter/reward_std:
+        value: 0.0
+        step: 7590
+        walltime: null
+      evaluator_step/reward_std:
+        value: 0.0
+        step: 85111
+        walltime: null
+      evaluator_iter/reward_max:
+        value: -1.0
+        step: 7590
+        walltime: null
+      evaluator_step/reward_max:
+        value: -1.0
+        step: 85111
+        walltime: null
+      evaluator_iter/reward_min:
+        value: -1.0
+        step: 7590
+        walltime: null
+      evaluator_step/reward_min:
+        value: -1.0
+        step: 85111
+        walltime: null
+      evaluator_iter/eval_episode_return_mean:
+        value: -1.0
+        step: 7590
+        walltime: null
+      evaluator_step/eval_episode_return_mean:
+        value: -1.0
+        step: 85111
+        walltime: null
+      Buffer/Task_0/num_collected_episodes:
+        value: 3224
+        step: 8050
+        walltime: null
+      Buffer/Task_0/num_game_segments:
+        value: 1800
+        step: 8050
+        walltime: null
+      Buffer/Task_0/num_transitions:
+        value: 49996
+        step: 8050
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/game_segment_buffer:
+        value: 31.134323120117188
+        step: 8050
+        walltime: null
+      Buffer/Task_0/memory_usage_mb/process:
+        value: 2076.12109375
+        step: 8050
+        walltime: null
+      learner_iter/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 8000
+        walltime: null
+      learner_step/collect_mcts_temperature_avg:
+        value: 0.25
+        step: 89525
+        walltime: null
+      learner_iter/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 8000
+        walltime: null
+      learner_step/cur_lr_avg:
+        value: 0.0029999999999999996
+        step: 89525
+        walltime: null
+      learner_iter/weighted_total_loss_avg:
+        value: 1.1734138293699785
+        step: 8000
+        walltime: null
+      learner_step/weighted_total_loss_avg:
+        value: 1.1734138293699785
+        step: 89525
+        walltime: null
+      learner_iter/total_loss_avg:
+        value: 1.1734138293699785
+        step: 8000
+        walltime: null
+      learner_step/total_loss_avg:
+        value: 1.1734138293699785
+        step: 89525
+        walltime: null
+      learner_iter/policy_loss_avg:
+        value: 9.202515255321156
+        step: 8000
+        walltime: null
+      learner_step/policy_loss_avg:
+        value: 9.202515255321156
+        step: 89525
+        walltime: null
+      learner_iter/policy_entropy_avg:
+        value: 1.6213162884567722
+        step: 8000
+        walltime: null
+      learner_step/policy_entropy_avg:
+        value: 1.6213162884567722
+        step: 89525
+        walltime: null
+      learner_iter/target_policy_entropy_avg:
+        value: 1.5226185683048135
+        step: 8000
+        walltime: null
+      learner_step/target_policy_entropy_avg:
+        value: 1.5226185683048135
+        step: 89525
+        walltime: null
+      learner_iter/value_prefix_loss_avg:
+        value: 0.2290004247968847
+        step: 8000
+        walltime: null
+      learner_step/value_prefix_loss_avg:
+        value: 0.2290004247968847
+        step: 89525
+        walltime: null
+      learner_iter/value_loss_avg:
+        value: 5.280927354639227
+        step: 8000
+        walltime: null
+      learner_step/value_loss_avg:
+        value: 5.280927354639227
+        step: 89525
+        walltime: null
+      learner_iter/consistency_loss_avg:
+        value: -0.9578333854675293
+        step: 8000
+        walltime: null
+      learner_step/consistency_loss_avg:
+        value: -0.9578333854675293
+        step: 89525
+        walltime: null
+      learner_iter/value_priority_avg:
+        value: 0.8299497799439863
+        step: 8000
+        walltime: null
+      learner_step/value_priority_avg:
+        value: 0.8299497799439863
+        step: 89525
+        walltime: null
+      learner_iter/target_value_prefix_avg:
+        value: 0.06202651695771651
+        step: 8000
+        walltime: null
+      learner_step/target_value_prefix_avg:
+        value: 0.06202651695771651
+        step: 89525
+        walltime: null
+      learner_iter/target_value_avg:
+        value: 0.00011837120506573807
+        step: 8000
+        walltime: null
+      learner_step/target_value_avg:
+        value: 0.00011837120506573807
+        step: 89525
+        walltime: null
+      learner_iter/predicted_value_prefixs_avg:
+        value: 0.05285426567901264
+        step: 8000
+        walltime: null
+      learner_step/predicted_value_prefixs_avg:
+        value: 0.05285426567901264
+        step: 89525
+        walltime: null
+      learner_iter/predicted_values_avg:
+        value: -0.0015143624815126796
+        step: 8000
+        walltime: null
+      learner_step/predicted_values_avg:
+        value: -0.0015143624815126796
+        step: 89525
+        walltime: null
+      learner_iter/transformed_target_value_prefix_avg:
+        value: 0.025754247707399456
+        step: 8000
+        walltime: null
+      learner_step/transformed_target_value_prefix_avg:
+        value: 0.025754247707399456
+        step: 89525
+        walltime: null
+      learner_iter/transformed_target_value_avg:
+        value: 4.914934238926931e-05
+        step: 8000
+        walltime: null
+      learner_step/transformed_target_value_avg:
+        value: 4.914934238926931e-05
+        step: 89525
+        walltime: null
+      learner_iter/total_grad_norm_before_clip_avg:
+        value: 0.4895417636091059
+        step: 8000
+        walltime: null
+      learner_step/total_grad_norm_before_clip_avg:
+        value: 0.4895417636091059
+        step: 89525
+        walltime: null
+      collector_iter/episode_count:
+        value: 40
+        step: 8000
+        walltime: null
+      collector_step/episode_count:
+        value: 40
+        step: 89525
+        walltime: null
+      collector_iter/envstep_count:
+        value: 1117
+        step: 8000
+        walltime: null
+      collector_step/envstep_count:
+        value: 1117
+        step: 89525
+        walltime: null
+      collector_iter/avg_envstep_per_episode:
+        value: 27.925
+        step: 8000
+        walltime: null
+      collector_step/avg_envstep_per_episode:
+        value: 27.925
+        step: 89525
+        walltime: null
+      collector_iter/avg_envstep_per_sec:
+        value: 10.071015591584729
+        step: 8000
+        walltime: null
+      collector_step/avg_envstep_per_sec:
+        value: 10.071015591584729
+        step: 89525
+        walltime: null
+      collector_iter/avg_episode_per_sec:
+        value: 0.3606451420442159
+        step: 8000
+        walltime: null
+      collector_step/avg_episode_per_sec:
+        value: 0.3606451420442159
+        step: 89525
+        walltime: null
+      collector_iter/collect_time:
+        value: 110.91234938940592
+        step: 8000
+        walltime: null
+      collector_step/collect_time:
+        value: 110.91234938940592
+        step: 89525
+        walltime: null
+      collector_iter/reward_mean:
+        value: 0.07500000298023224
+        step: 8000
+        walltime: null
+      collector_step/reward_mean:
+        value: 0.07500000298023224
+        step: 89525
+        walltime: null
+      collector_iter/reward_std:
+        value: 0.9845684170722961
+        step: 8000
+        walltime: null
+      collector_step/reward_std:
+        value: 0.9845684170722961
+        step: 89525
+        walltime: null
+      collector_iter/reward_max:
+        value: 1.0
+        step: 8000
+        walltime: null
+      collector_step/reward_max:
+        value: 1.0
+        step: 89525
+        walltime: null
+      collector_iter/reward_min:
+        value: -1.0
+        step: 8000
+        walltime: null
+      collector_step/reward_min:
+        value: -1.0
+        step: 89525
+        walltime: null
+      collector_iter/total_envstep_count:
+        value: 89525
+        step: 8000
+        walltime: null
+      collector_step/total_envstep_count:
+        value: 89525
+        step: 89525
+        walltime: null
+      collector_iter/total_episode_count:
+        value: 3208
+        step: 8000
+        walltime: null
+      collector_step/total_episode_count:
+        value: 3208
+        step: 89525
+        walltime: null
+      collector_iter/total_duration:
+        value: 49916.304257843265
+        step: 8000
+        walltime: null
+      collector_step/total_duration:
+        value: 49916.304257843265
+        step: 89525
+        walltime: null
+      collector_iter/visit_entropy_mean:
+        value: 1.024226404569014
+        step: 8000
+        walltime: null
+      collector_step/visit_entropy_mean:
+        value: 1.024226404569014
+        step: 89525
+        walltime: null
+metadata_version: 1

metadata/total_config.py

@@ -0,0 +1,236 @@
+exp_config = {
+    'env': {
+        'manager': {
+            'episode_num': float("inf"),
+            'max_retry': 1,
+            'step_timeout': None,
+            'auto_reset': True,
+            'reset_timeout': None,
+            'retry_type': 'reset',
+            'retry_waiting_time': 0.1,
+            'shared_memory': False,
+            'copy_on_get': True,
+            'context': 'fork',
+            'wait_num': float("inf"),
+            'step_wait_timeout': None,
+            'connect_timeout': 60,
+            'reset_inplace': False,
+            'cfg_type': 'SyncSubprocessEnvManagerDict',
+            'type': 'subprocess'
+        },
+        'stop_value': 2,
+        'n_evaluator_episode': 10,
+        'env_id': 'Connect4',
+        'battle_mode': 'self_play_mode',
+        'battle_mode_in_simulation_env': 'self_play_mode',
+        'render_mode': None,
+        'replay_path': None,
+        'bot_action_type': 'rule',
+        'agent_vs_human': False,
+        'prob_random_agent': 0,
+        'prob_expert_agent': 0,
+        'prob_random_action_in_bot': 0.0,
+        '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
+    },
+    'policy': {
+        'model': {
+            'model_type': 'conv',
+            'continuous_action_space': False,
+            'observation_shape': (3, 6, 7),
+            'self_supervised_learning_loss': True,
+            'categorical_distribution': True,
+            'image_channel': 3,
+            'frame_stack_num': 1,
+            'num_res_blocks': 1,
+            'num_channels': 64,
+            'reward_support_range': (-300.0, 301.0, 1.0),
+            'value_support_range': (-300.0, 301.0, 1.0),
+            'bias': True,
+            'discrete_action_encoding_type': 'one_hot',
+            'res_connection_in_dynamics': True,
+            'norm_type': 'BN',
+            'analysis_sim_norm': False,
+            'analysis_dormant_ratio': False,
+            'harmony_balance': False,
+            'lstm_hidden_size': 512,
+            'action_space_size': 7
+        },
+        'learn': {
+            'learner': {
+                'train_iterations': 1000000000,
+                'dataloader': {
+                    'num_workers': 0
+                },
+                'log_policy': True,
+                'hook': {
+                    'load_ckpt_before_run': '',
+                    'log_show_after_iter': 100,
+                    'save_ckpt_after_iter': 10000,
+                    'save_ckpt_after_run': True
+                },
+                'cfg_type': 'BaseLearnerDict'
+            },
+            'resume_training': False
+        },
+        'collect': {
+            'collector': {
+                'deepcopy_obs': False,
+                'transform_obs': False,
+                'collect_print_freq': 100,
+                'cfg_type': 'SampleSerialCollectorDict',
+                'type': 'sample'
+            }
+        },
+        'eval': {
+            'evaluator': {
+                'eval_freq': 1000,
+                'render': {
+                    'render_freq': -1,
+                    'mode': 'train_iter'
+                },
+                'figure_path': None,
+                'cfg_type': 'InteractionSerialEvaluatorDict',
+                'stop_value': 2,
+                'n_episode': 10
+            }
+        },
+        'other': {
+            'replay_buffer': {
+                'type': 'advanced',
+                'replay_buffer_size': 4096,
+                'max_use': float("inf"),
+                'max_staleness': float("inf"),
+                'alpha': 0.6,
+                'beta': 0.4,
+                'anneal_step': 100000,
+                'enable_track_used_data': False,
+                'deepcopy': False,
+                'thruput_controller': {
+                    'push_sample_rate_limit': {
+                        'max': float("inf"),
+                        'min': 0
+                    },
+                    'window_seconds': 30,
+                    'sample_min_limit_ratio': 1
+                },
+                'monitor': {
+                    'sampled_data_attr': {
+                        'average_range': 5,
+                        'print_freq': 200
+                    },
+                    'periodic_thruput': {
+                        'seconds': 60
+                    }
+                },
+                'cfg_type': 'AdvancedReplayBufferDict'
+            },
+            'commander': {
+                'cfg_type': 'BaseSerialCommanderDict'
+            }
+        },
+        'on_policy': False,
+        'cuda': True,
+        'multi_gpu': False,
+        'bp_update_sync': True,
+        'traj_len_inf': False,
+        'use_wandb': True,
+        'use_rnd_model': False,
+        'sampled_algo': False,
+        'gumbel_algo': False,
+        'mcts_ctree': True,
+        'collector_env_num': 4,
+        'evaluator_env_num': 10,
+        'env_type': 'board_games',
+        'action_type': 'varied_action_space',
+        'battle_mode': 'self_play_mode',
+        'monitor_extra_statistics': True,
+        'game_segment_length': 42,
+        'eval_offline': False,
+        'calculate_dormant_ratio': False,
+        'analysis_sim_norm': False,
+        'analysis_dormant_ratio': False,
+        'transform2string': False,
+        'gray_scale': False,
+        'use_augmentation': False,
+        'augmentation': ['shift', 'intensity'],
+        'ignore_done': False,
+        'update_per_collect': 10,
+        'replay_ratio': 0.25,
+        'batch_size': 128,
+        'optim_type': 'Adam',
+        'learning_rate': 0.003,
+        'target_update_freq': 100,
+        'target_update_freq_for_intrinsic_reward': 1000,
+        'weight_decay': 0.0001,
+        'momentum': 0.9,
+        'grad_clip_value': 0.5,
+        'n_episode': 4,
+        'num_segments': 8,
+        'num_simulations': 20,
+        'discount_factor': 1,
+        'td_steps': 42,
+        'num_unroll_steps': 5,
+        'reward_loss_weight': 1,
+        'value_loss_weight': 0.25,
+        'policy_loss_weight': 1,
+        'policy_entropy_weight': 0,
+        'ssl_loss_weight': 2,
+        'piecewise_decay_lr_scheduler': False,
+        'threshold_training_steps_for_final_lr': 50000,
+        'manual_temperature_decay': False,
+        'threshold_training_steps_for_final_temperature': 100000,
+        'fixed_temperature_value': 0.25,
+        'use_ture_chance_label_in_chance_encoder': False,
+        'reanalyze_noise': True,
+        'reuse_search': False,
+        'collect_with_pure_policy': False,
+        'use_priority': False,
+        'priority_prob_alpha': 0.6,
+        'priority_prob_beta': 0.4,
+        'root_dirichlet_alpha': 0.3,
+        'root_noise_weight': 0.25,
+        'random_collect_episode_num': 0,
+        'eps': {
+            'eps_greedy_exploration_in_collect': False,
+            'type': 'linear',
+            'start': 1.0,
+            'end': 0.05,
+            'decay': 100000
+        },
+        'cfg_type': 'EfficientZeroPolicyDict',
+        'lstm_horizon_len': 5,
+        'type': 'efficientzero',
+        'import_names': ['lzero.policy.efficientzero'],
+        'model_path':
+        '/workspace/combinatorial_reasoning_post_training/models/simplified5_immediate_300k_continuation_20260518/round-01/connect4/attempt-01_260518_231735/ckpt/envstep_165000.pth.tar',
+        'reanalyze_ratio': 0.0,
+        'eval_freq': 300001,
+        'replay_buffer_size': 50000,
+        'best_ckpt_strategy': 'raw',
+        'best_ckpt_ema_alpha': 0.3,
+        'best_ckpt_min_episodes': 20,
+        'battle_mode_in_simulation_env': 'self_play_mode',
+        'eval_opponent_type': 'env_bot',
+        'previous_best_checkpoint': {
+            'path': None,
+            'selector': 'best',
+            'update_policy': 'on_new_best',
+            'num_simulations': None,
+            'n_evaluator_episode': None,
+            'evaluator_env_num': None,
+            'promotion_threshold': 0.0,
+            'fallback_to_env_bot': False
+        },
+        'device': 'cuda'
+    },
+    'exp_name':
+    '/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459',
+    'seed': 0
+}

metadata/upload_manifest.json

@@ -0,0 +1,15 @@
+{
+  "checkpoint_mtime_utc": "2026-05-19T15:02:37Z",
+  "checkpoint_size_bytes": 139564171,
+  "display_name": "Connect4",
+  "game_id": "connect4",
+  "path_in_repo": "checkpoints/envstep_90000.pth.tar",
+  "repo_id": "LorMolf/Connect4-CK",
+  "source_attempt_dir": "/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459",
+  "source_checkpoint": "/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/envstep_90000.pth.tar",
+  "source_checkpoint_index": "/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/ckpt/checkpoint_index.yaml",
+  "source_total_config": "/workspace/combinatorial_reasoning_post_training/models/simplified5_constant_lr_fixed_20260519/round-01/connect4/attempt-01_260519_093459/total_config.py",
+  "uploaded_at_utc": "2026-05-19T15:04:12Z",
+  "wandb_project": "crpt-simplified5-constant-lr-continuation",
+  "wandb_run": "constant_lr_20260519__connect4__a01"
+}