diff options
Diffstat (limited to 'results/babyai_shared/pickup_p2/seed4101_sdil.json')
| -rw-r--r-- | results/babyai_shared/pickup_p2/seed4101_sdil.json | 295 |
1 files changed, 295 insertions, 0 deletions
diff --git a/results/babyai_shared/pickup_p2/seed4101_sdil.json b/results/babyai_shared/pickup_p2/seed4101_sdil.json new file mode 100644 index 0000000..6be8f4b --- /dev/null +++ b/results/babyai_shared/pickup_p2/seed4101_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": 1.4496507589440597 + }, + { + "epoch": 2, + "train_loss": 1.4869623249455501 + }, + { + "epoch": 3, + "train_loss": 1.4099679946899415 + }, + { + "epoch": 4, + "train_loss": 1.4097933264782554 + }, + { + "epoch": 5, + "train_loss": 1.5673320049988595 + }, + { + "epoch": 6, + "train_loss": 1.3529665086143896 + }, + { + "epoch": 7, + "train_loss": 1.6511782859501085 + }, + { + "epoch": 8, + "train_loss": 1.4052464414897718 + }, + { + "epoch": 9, + "train_loss": 1.5705512192374782 + }, + { + "epoch": 10, + "train_loss": 1.3675857054559808 + }, + { + "epoch": 11, + "train_loss": 1.384743598887795 + }, + { + "epoch": 12, + "train_loss": 1.3992685817417345 + }, + { + "epoch": 13, + "train_loss": 1.3667162430913824 + }, + { + "epoch": 14, + "train_loss": 1.4492891484812687 + }, + { + "epoch": 15, + "train_loss": 1.4339373264814679 + }, + { + "epoch": 16, + "train_loss": 1.4766567792390521 + }, + { + "epoch": 17, + "train_loss": 1.415901366534986 + }, + { + "epoch": 18, + "train_loss": 1.4241125407971833 + }, + { + "epoch": 19, + "train_loss": 1.4059961216073287 + }, + { + "epoch": 20, + "train_loss": 1.6259911338906539 + }, + { + "epoch": 21, + "train_loss": 1.5026495936042383 + }, + { + "epoch": 22, + "train_loss": 1.492478413832815 + }, + { + "epoch": 23, + "train_loss": 1.3337757732993678 + }, + { + "epoch": 24, + "train_loss": 1.4029922046159442 + }, + { + "epoch": 25, + "train_loss": 1.4238720023004632 + }, + { + "epoch": 26, + "train_loss": 1.6601720375763742 + }, + { + "epoch": 27, + "train_loss": 1.3956078170475208 + }, + { + "epoch": 28, + "train_loss": 1.3906170870128431 + }, + { + "epoch": 29, + "train_loss": 1.5026050532491584 + }, + { + "epoch": 30, + "train_loss": 1.356335889414737 + }, + { + "epoch": 31, + "train_loss": 1.4696443364494725 + }, + { + "epoch": 32, + "train_loss": 1.4274836944278917 + }, + { + "epoch": 33, + "train_loss": 1.3983684896167956 + }, + { + "epoch": 34, + "train_loss": 1.416925176821257 + }, + { + "epoch": 35, + "train_loss": 1.6563161368119088 + }, + { + "epoch": 36, + "train_loss": 1.4001352036626715 + }, + { + "epoch": 37, + "train_loss": 1.425227477926957 + }, + { + "epoch": 38, + "train_loss": 1.3887689311880815 + }, + { + "epoch": 39, + "train_loss": 1.364363015074479 + }, + { + "epoch": 40, + "train_loss": 1.377758667845475 + } + ], + "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": 1.3227376229581325 + }, + "predictor": [ + { + "action_observations": 0, + "context_rms": 0.20984283089637756, + "mean_per_cell_r2": 0.07823440432548523, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.6220114827156067, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.2008315920829773, + "mean_per_cell_r2": 0.3734843134880066, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.5552217364311218, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.19726455211639404, + "mean_per_cell_r2": 0.33756017684936523, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.5612375736236572, + "teaching_observations": 0 + }, + { + "action_observations": 0, + "context_rms": 0.19451536238193512, + "mean_per_cell_r2": 0.4691427946090698, + "neutral_observations": 4096, + "residual_context_rms_ratio": 0.5088029503822327, + "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": 4101, + "neutral_examples_per_epoch": 4096, + "parameter_count_including_reciprocal_context_predictor": 1438215, + "shuffle_seed": 4101, + "total_wall_seconds_including_rollouts": 346.5359559059143, + "training_wall_seconds": 322.71134519577026 + }, + "validation": { + "accuracy": 0.49890661699198185, + "loss": 1.3229210194647723 + } +} |