{ "checkpoint": { "algorithm_id": "ppo", "filename": "model.zip", "kind": "checkpoint", "model_class": "stable_baselines3.ppo.ppo.PPO", "sha256": "1c87c2b670eb3281cf65dc4b515edf8da3002cc93dd9ca3b065fb622cb22d233", "size_bytes": 21054226, "step": 4500000 }, "document_type": "rlab.model", "format_version": 1, "policy": { "algorithm_id": "ppo", "model_class": "stable_baselines3.ppo.ppo.PPO", "training_backend_config_hash": "e8cc4250603c745c57df370c3a352f81ed820072a5d3556c7ffbbf1b433ebd86", "training_backend_id": "sb3.ppo" }, "provenance": { "goal_slug": "Level1-4", "kind": "checkpoint", "metadata_version": 2, "queue_train_job_id": 5, "recipe_slug": "base", "repo_git_commit": "62e08d739b44f950cc7762b7319a75c2eba527ce", "run_description": "Seed trains the Level1-4 B257 soft-update no-bonus recipe, created from the historical Level1-4 screen winner to keep the score/progress reward contract while lowering PPO update pressure for more stable World 1-4 clears.", "run_name": "Level1-4_base_s_20260704T081304Z", "runtime_image_ref": "docker:ghcr.io/tsilva/gradlab/gradlab-train@sha256:d0bbbdacbfa96658f19d939ca0e5065eb989de9c030d6d7695fbad7055bf6851", "seed": 123, "training_metadata": { "env_config": { "action_set": "simple", "clip_rewards": false, "completion_reward": 0.0, "death_penalty": 25.0, "done_on_events": [ "life_loss", "level_change" ], "env_provider": "supermariobrosnes-turbo", "env_threads": 4, "env_wrappers": [ { "id": "SuperMarioBrosNesProgressInfoWrapper", "kwargs": {} }, { "id": "SuperMarioBrosNesRewardEnvWrapper", "kwargs": { "completion_reward": 0.0, "death_penalty": 25.0, "progress_reward_scale": 1.0, "reward_mode": "score", "reward_scale": 10.0, "score_progress_clipped": false, "terminal_reward": 50.0 } } ], "frame_skip": 4, "game": "SuperMarioBros-Nes-v0", "hud_crop_top": -1, "info_events": {}, "max_episode_steps": 4500, "max_pool_frames": false, "no_progress_min_delta": 0, "no_progress_timeout_steps": 0, "obs_crop": [ 32, 0, 0, 0 ], "obs_resize_algorithm": "area", "observation_size": 84, "progress_reward_cap": 30.0, "progress_reward_scale": 1.0, "reward_mode": "score", "reward_scale": 10.0, "score_progress_clipped": false, "state": "Level1-4", "state_distribution": [], "state_probs": [], "state_sampling_mode": "single", "states": [], "sticky_action_prob": 0.0, "task_conditioning": false, "task_conditioning_info_values": [], "task_conditioning_info_vars": [], "terminal_reward": 50.0, "time_penalty": 0.0, "use_retro_reward": false }, "environment": { "env_id": "supermariobrosnes-turbo:SuperMarioBros-Nes-v0", "preprocessing": { "frame_skip": 4, "frame_stack": 4, "max_pool_frames": false, "obs_copy": "safe_view", "obs_crop": [ 32, 0, 0, 0 ], "obs_crop_fill": 0, "obs_crop_mode": "remove", "obs_grayscale": true, "obs_resize": [ 84, 84 ], "obs_resize_algorithm": "area", "pipeline": "supermariobrosnes_turbo_native_vec_env", "policy_observation_layout": "channel_first", "sticky_action_prob": 0.0 }, "provider_args": { "frame_stack": 4, "info": "data", "info_filter": "all", "inttype": "stable", "noop_reset_max": 0, "obs_copy": "safe_view", "obs_grayscale": true, "obs_layout": "chw", "obs_type": "image", "players": 1, "record": false, "render_mode": "rgb_array", "reward_clip": false, "rom_path": null, "scenario": "scenario", "use_restricted_actions": "filtered" }, "schema_version": 2, "state": "Level1-4", "task": { "action": { "set": "simple" }, "events": { "level_change": { "operation": "change", "signal": "level" }, "life_loss": { "operation": "decrease", "signal": "lives" } }, "id": "mario", "reward": { "clip_rewards": false, "completion_reward": 0.0, "death_penalty": 25.0, "progress_reward_cap": 30.0, "progress_reward_scale": 1.0, "reward_mode": "score", "reward_scale": 10.0, "score_progress_clipped": false, "terminal_reward": 50.0, "time_penalty": 0.0, "use_native_reward": false }, "signals": { "level": [ "levelHi", "levelLo" ], "lives": "lives", "score": "score", "x": [ "xscrollHi", "xscrollLo" ] }, "termination": { "failure": [ "life_loss" ], "max_episode_steps": 4500, "success": [ "level_change" ] } } }, "environment_hash": "sha256:603e8f5a693bb5c7e56fbf316879e5858b1534e56cbbd11be79cfeec219b02eb", "preprocessing": { "frame_skip": 4, "frame_stack": 4, "max_pool_frames": false, "obs_copy": "safe_view", "obs_crop": [ 32, 0, 0, 0 ], "obs_crop_fill": 0, "obs_crop_mode": "remove", "obs_grayscale": true, "obs_resize": [ 84, 84 ], "obs_resize_algorithm": "area", "pipeline": "supermariobrosnes_turbo_native_vec_env", "policy_observation_layout": "channel_first", "sticky_action_prob": 0.0 }, "versions": { "stable_baselines3": "2.8.0", "stable_retro_turbo": "1.0.1.post4", "supermariobrosnes_turbo": "0.2.4" } }, "training_metadata_hash": "fa61efc868409bffa657116266d6031704cd891dbf5d12e3602cb5d745876c20", "wandb_project": "SuperMarioBros-Nes-v0", "wandb_run_id": "7gjw67kl", "wandb_run_path": "tsilva/SuperMarioBros-Nes-v0/7gjw67kl" }, "recipe": { "document_type": "rlab.recipe", "filename": "recipe.json", "format_version": 1, "sha256": "99c6c2881b3e4f5e1faa639c68004df40ea01322009c5c3e0264fb4a5b2dbaf1", "size_bytes": 10798 } }