diff options
Diffstat (limited to 'results/babyai_shared/pickup_p2/seed4103_sdil.json')
| -rw-r--r-- | results/babyai_shared/pickup_p2/seed4103_sdil.json | 295 |
1 files changed, 295 insertions, 0 deletions
diff --git a/results/babyai_shared/pickup_p2/seed4103_sdil.json b/results/babyai_shared/pickup_p2/seed4103_sdil.json new file mode 100644 index 0000000..acbef8b --- /dev/null +++ b/results/babyai_shared/pickup_p2/seed4103_sdil.json @@ -0,0 +1,295 @@ +{ + "condition": "sdil", + "config": { + "action_dim": 7, + "context_gain": 1.0, + "hidden_layers": 4, + "input_dim": 3964, + "learning_rate": 0.03, + "mission_dim": 23, + "momentum": 0.9, + "reciprocal_learning_rate": 0.03, + "weight_decay": 0.0001, + "width": 256 + }, + "data": { + "color_cardinality": 6, + "elapsed_seconds": 81.82925486564636, + "env_id": "BabyAI-PickupLoc-v0", + "minigrid_version": "3.1.0", + "object_cardinality": 11, + "protocol": "babyai_shared_feedback_b0", + "rollout_episodes": 500, + "rollout_seed_start": 200000, + "state_cardinality": 3, + "train_episodes": 20000, + "train_steps": 121467, + "validation_episodes": 2000, + "validation_steps": 12347, + "vocabulary": [ + "<unk>", + "a", + "ball", + "behind", + "blue", + "box", + "front", + "green", + "grey", + "in", + "key", + "left", + "of", + "on", + "pick", + "purple", + "red", + "right", + "the", + "up", + "yellow", + "you", + "your" + ] + }, + "epoch_history": [ + { + "epoch": 1, + "train_loss": 1.4704136446902627 + }, + { + "epoch": 2, + "train_loss": 1.4733195139232436 + }, + { + "epoch": 3, + "train_loss": 1.6606671679647345 + }, + { + "epoch": 4, + "train_loss": 1.6223741378282246 + }, + { + "epoch": 5, + "train_loss": 1.380738465158563 + }, + { + "epoch": 6, + "train_loss": 1.4576800007569162 + }, + { + "epoch": 7, + "train_loss": 1.3935401956658615 + }, + { + "epoch": 8, + "train_loss": 1.5138149407035426 + }, + { + "epoch": 9, + "train_loss": 1.5381833568372225 + }, + { + "epoch": 10, + "train_loss": 1.602822558503402 + }, + { + "epoch": 11, + "train_loss": 1.4934680660147417 + }, + { + "epoch": 12, + "train_loss": 1.462449248213517 + }, + { + "epoch": 13, + "train_loss": 1.5229607145409836 + }, + { + "epoch": 14, + "train_loss": 1.4338193808103863 + }, + { + "epoch": 15, + "train_loss": 1.7763569934744583 + }, + { + "epoch": 16, + "train_loss": 1.4266947914424695 + }, + { + "epoch": 17, + "train_loss": 1.4615587412683588 + }, + { + "epoch": 18, + "train_loss": 1.3746792825899625 + }, + { + "epoch": 19, + "train_loss": 1.4826162985751503 + }, + { + "epoch": 20, + "train_loss": 1.3551955336018613 + }, + { + "epoch": 21, + "train_loss": 1.6226069836867483 + }, + { + "epoch": 22, + "train_loss": 1.381537736340573 + }, + { + "epoch": 23, + "train_loss": 1.5173151382647063 + }, + { + "epoch": 24, + "train_loss": 1.54268351931321 + }, + { + "epoch": 25, + "train_loss": 1.5190619021967837 + }, + { + "epoch": 26, + "train_loss": 1.66076710274345 + }, + { + "epoch": 27, + "train_loss": 1.6404679448981034 + }, + { + "epoch": 28, + "train_loss": 1.3772655429338154 + }, + { + "epoch": 29, + "train_loss": 1.4842414411745573 + }, + { + "epoch": 30, + "train_loss": 1.382061290239033 + }, + { + "epoch": 31, + "train_loss": 1.3829637015493292 + }, + { + "epoch": 32, + "train_loss": 1.428046850907175 + }, + { + "epoch": 33, + "train_loss": 1.4464430650911833 + }, + { + "epoch": 34, + "train_loss": 1.4072801037838585 + }, + { + "epoch": 35, + "train_loss": 1.3329616769991424 + }, + { + "epoch": 36, + "train_loss": 1.436322965872915 + }, + { + "epoch": 37, + "train_loss": 1.3869416563134445 + }, + { + "epoch": 38, + "train_loss": 1.469730708975541 + }, + { + "epoch": 39, + "train_loss": 1.48009900143272 + }, + { + "epoch": 40, + "train_loss": 1.498611428110223 + } + ], + "epochs_completed": 40, + "finite": true, + "first_nonfinite_epoch": null, + "mission_lesion_rollout": { + "episodes": 500, + "mean_length": 64.0, + "mean_return": 0.0, + "success": 0.0 + }, + "mission_lesion_validation": { + "accuracy": 0.49890661699198185, + "loss": 1.3523273018484532 + }, + "predictor": [ + { + "action_observations": 0, + "context_rms": 0.20582586526870728, + "mean_per_cell_r2": 0.1324767768383026, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.6195260882377625, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.2105173021554947, + "mean_per_cell_r2": 0.40293416380882263, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.5007978081703186, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.2122984677553177, + "mean_per_cell_r2": 0.2902011275291443, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.5484216213226318, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.21755045652389526, + "mean_per_cell_r2": 0.4205133020877838, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.4854767322540283, + "teaching_observations": 0 + } + ], + "provenance": { + "cuda_device_name": "NVIDIA GeForce GTX 1080", + "cuda_peak_allocated_bytes": 435783168, + "cuda_visible_devices": "3", + "device": "cuda:0", + "git_commit": "7b7d7a08a5e78c3379397f01f0c9af97d5d84fbd", + "git_dirty_tracked": false, + "minigrid_version": "3.1.0", + "torch_version": "2.3.1+cu118" + }, + "rollout": { + "episodes": 500, + "mean_length": 64.0, + "mean_return": 0.0, + "success": 0.0 + }, + "stage": "babyai_pickup_p2", + "training": { + "batch_size": 256, + "history_steps": 4, + "model_seed": 4103, + "neutral_examples_per_epoch": 4096, + "parameter_count_including_reciprocal_context_predictor": 1438215, + "shuffle_seed": 4103, + "total_wall_seconds_including_rollouts": 171.52505588531494, + "training_wall_seconds": 152.92210721969604 + }, + "validation": { + "accuracy": 0.49890661699198185, + "loss": 1.351929626687293 + } +} |