| { | |
| "type": "wam", | |
| "n_obs_steps": 2, | |
| "input_features": { | |
| "observation.state": { | |
| "type": "STATE", | |
| "shape": [ | |
| 6 | |
| ] | |
| }, | |
| "observation.images.front": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 480, | |
| 640 | |
| ] | |
| } | |
| }, | |
| "output_features": { | |
| "action": { | |
| "type": "ACTION", | |
| "shape": [ | |
| 6 | |
| ] | |
| } | |
| }, | |
| "device": "cuda", | |
| "use_amp": false, | |
| "use_peft": false, | |
| "push_to_hub": true, | |
| "repo_id": "maximellerbach/test_fm", | |
| "private": null, | |
| "tags": null, | |
| "license": null, | |
| "pretrained_path": "outputs/train/pickandplace_fm/checkpoints/last/pretrained_model", | |
| "time_between_frames": 4, | |
| "lag_offset": 3, | |
| "image_size": [ | |
| 256, | |
| 256 | |
| ], | |
| "horizon": 30, | |
| "n_action_steps": 30, | |
| "autoregressive_steps": 2, | |
| "normalization_mapping": { | |
| "VISUAL": "IDENTITY", | |
| "STATE": "MEAN_STD", | |
| "ACTION": "MEAN_STD" | |
| }, | |
| "hidden_dim": 512, | |
| "action_head_num_layers": 4, | |
| "action_head_num_heads": 8, | |
| "action_head_dropout": 0.1, | |
| "n_flow_steps": 10, | |
| "predictor_num_layers": 6, | |
| "predictor_num_heads": 8, | |
| "predictor_dropout": 0.1, | |
| "predictor_max_tokens": 1024, | |
| "optimizer_lr": 5e-05, | |
| "optimizer_betas": [ | |
| 0.9, | |
| 0.999 | |
| ], | |
| "optimizer_eps": 1e-08, | |
| "optimizer_weight_decay": 0.01, | |
| "optimizer_grad_clip_norm": 1.0, | |
| "scheduler_decay_lr": 1e-05, | |
| "scheduler_warmup_steps": 500, | |
| "scheduler_decay_steps": 5000 | |
| } |