| { |
| "type": "gaze_act", |
| "n_obs_steps": 1, |
| "input_features": { |
| "observation.images.right": { |
| "type": "VISUAL", |
| "shape": [ |
| 240, |
| 320, |
| 3 |
| ] |
| }, |
| "observation.images.left": { |
| "type": "VISUAL", |
| "shape": [ |
| 240, |
| 320, |
| 3 |
| ] |
| }, |
| "observation.images.user": { |
| "type": "VISUAL", |
| "shape": [ |
| 240, |
| 320, |
| 3 |
| ] |
| } |
| }, |
| "output_features": { |
| "action": { |
| "type": "ACTION", |
| "shape": [ |
| 26 |
| ] |
| } |
| }, |
| "device": "cuda", |
| "use_amp": false, |
| "use_peft": false, |
| "push_to_hub": true, |
| "repo_id": "fecasado/gact-cubes-27a", |
| "private": null, |
| "tags": null, |
| "license": null, |
| "pretrained_path": null, |
| "chunk_size": 30, |
| "n_action_steps": 1, |
| "normalization_mapping": { |
| "VISUAL": "MEAN_STD", |
| "STATE": "MEAN_STD", |
| "ACTION": "MEAN_STD" |
| }, |
| "vision_backbone": "resnet18", |
| "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1", |
| "replace_final_stride_with_dilation": false, |
| "pre_norm": false, |
| "dim_model": 512, |
| "n_heads": 8, |
| "dim_feedforward": 3200, |
| "feedforward_activation": "relu", |
| "n_encoder_layers": 4, |
| "n_decoder_layers": 1, |
| "use_vae": true, |
| "latent_dim": 32, |
| "n_vae_encoder_layers": 4, |
| "temporal_ensemble_coeff": 0.01, |
| "dropout": 0.1, |
| "kl_weight": 10.0, |
| "optimizer_lr": 1e-05, |
| "optimizer_weight_decay": 0.0001, |
| "optimizer_lr_backbone": 1e-05, |
| "vit_token_size": 16, |
| "vit_foveation_strides": [ |
| 1, |
| 2, |
| 6 |
| ], |
| "vit_foveation_grid_sizes": [ |
| 2, |
| 3, |
| 3 |
| ], |
| "vit_model_size": "base", |
| "vit_pretrained_repo_id": "iantc104/mae_vitb_foveated_vit", |
| "vit_freeze": true, |
| "gaze_feature_dim": 3, |
| "gaze_state_indices": [ |
| -3, |
| -2, |
| -1 |
| ], |
| "gaze_temporal_window": 1, |
| "gaze_cross_camera": false, |
| "gaze_cross_camera_method_train": "roma", |
| "gaze_cross_camera_method_inference": "roma", |
| "gaze_cross_camera_match_every": 5, |
| "gaze_cross_camera_timeout": 0.05, |
| "gaze_cross_camera_cache": null, |
| "gaze_token": false, |
| "gaze_heatmap": false, |
| "gaze_heatmap_sigma": 60.0, |
| "gaze_heatmap_decay": 1.0, |
| "gaze_film": false, |
| "gaze_film_multilayer": false, |
| "gaze_overlay": false, |
| "gaze_overlay_radius": 12, |
| "gaze_overlay_alpha": 0.5, |
| "gaze_camera": "user", |
| "gaze_attention_guidance": false, |
| "gaze_attention_weight": 1.0, |
| "gaze_contrastive": false, |
| "gaze_contrastive_weight": 0.1, |
| "gaze_contrastive_negatives": 16, |
| "gaze_contrastive_tau": 0.07, |
| "gaze_prediction": false, |
| "gaze_prediction_weight": 0.1 |
| } |
