{ "iteration": 2, "timestamp": "2026-03-08T04:52:48.689881+00:00", "buffer_size": 2813, "config": { "iterations": 220, "episodes_per_iter": 20, "mcts_sims": 160, "c_puct": 1.5, "temp_threshold": 28, "add_noise": true, "seed": 42, "verbose_logs": true, "episode_log_every": 25, "epochs": 2, "batch_size": 224, "learning_rate": 0.0003, "weight_decay": 0.0001, "value_loss_coeff": 0.5, "buffer_size": 50000, "val_split": 0.1, "shuffle_train_val_split": true, "train_recent_fraction": 0.7, "train_recent_window_fraction": 0.4, "d_model": 128, "nhead": 8, "num_layers": 6, "dim_feedforward": 512, "dropout": 0.1, "log_dir": "/kaggle/working/logs", "checkpoint_dir": "/kaggle/working/checkpoints", "save_every": 1, "keep_last_n_local_checkpoints": 256, "keep_last_n_log_versions": 2, "keep_last_n_hf_checkpoints": 3, "onnx_path": "ataxx_model.onnx", "export_onnx": false, "hf_enabled": true, "hf_repo_id": "dieg0code/ataxx-zero", "hf_run_id": "policy_spatial_v3", "hf_bootstrap_run_id": "policy_spatial_v2", "hf_reset_iteration": true, "hf_token_env": "HF_TOKEN", "hf_local_dir": "hf_checkpoints", "max_pending_hf_uploads": 6, "hf_upload_future_timeout_s": 900.0, "show_progress_bar": false, "trainer_log_every_n_steps": 99999, "monitor_log_every": 1, "epoch_pulse_every": 1, "num_workers": 2, "persistent_workers": true, "strict_probs": false, "trainer_devices": 2, "trainer_strategy": "ddp_spawn", "trainer_precision": "16-mixed", "trainer_benchmark": true, "ddp_timeout_seconds": 120, "mcts_use_amp": true, "mcts_cache_size": 100000, "mcts_leaf_batch_size": 32, "fail_on_selfplay_parallel_error": false, "fail_on_hf_upload_error": false, "opponent_self_prob": 0.45, "opponent_heuristic_prob": 0.5, "opponent_random_prob": 0.05, "opponent_heuristic_level": "normal", "opponent_heuristic_easy_prob": 0.2, "opponent_heuristic_normal_prob": 0.5, "opponent_heuristic_hard_prob": 0.3, "model_side_swap_prob": 0.5, "eval_enabled": true, "eval_every": 6, "eval_games": 12, "eval_sims": 160, "eval_heuristic_level": "hard", "eval_heuristic_levels": "hard,apex,sentinel", "restore_best_on_regression": true, "eval_regression_delta": 0.03, "eval_regression_patience": 2, "selfplay_workers": 2, "selfplay_progress_every_s": 45.0, "selfplay_episode_timeout_s": 600.0, "compile_model": false, "quiet_mode": false, "warmup_games": 240, "warmup_epochs": 3, "warmup_heuristic_level": "sentinel", "warmup_heuristic_levels": "hard,apex,sentinel" }, "stats": { "replay_size": 2813, "best_eval_score": -1.0 } }