| { |
| "type": "internvla_a1_5", |
| "n_obs_steps": 1, |
| "input_features": { |
| "observation.state": { |
| "type": "STATE", |
| "shape": [ |
| 32 |
| ] |
| } |
| }, |
| "output_features": { |
| "action": { |
| "type": "ACTION", |
| "shape": [ |
| 32 |
| ] |
| } |
| }, |
| "device": "cuda:0", |
| "use_amp": false, |
| "push_to_hub": false, |
| "repo_id": null, |
| "private": null, |
| "tags": null, |
| "license": null, |
| "pretrained_path": null, |
| "vlm_model_name_or_path": "Wan-AI/Wan2.2-TI2V-5B", |
| "action_expert_hidden_size": 1024, |
| "action_expert_intermediate_size": 3072, |
| "dtype": "bfloat16", |
| "chunk_size": 8, |
| "n_action_steps": 8, |
| "max_state_dim": 32, |
| "max_action_dim": 32, |
| "num_inference_steps": 10, |
| "time_sampling_beta_alpha": 1.5, |
| "time_sampling_beta_beta": 1.0, |
| "time_sampling_scale": 0.999, |
| "time_sampling_offset": 0.001, |
| "min_period": 0.004, |
| "max_period": 4.0, |
| "image_resolution": [ |
| 224, |
| 224 |
| ], |
| "empty_cameras": 0, |
| "normalization_mapping": { |
| "VISUAL": "IDENTITY", |
| "STATE": "IDENTITY", |
| "ACTION": "IDENTITY" |
| }, |
| "gradient_checkpointing": false, |
| "compile_model": false, |
| "compile_mode": "max-autotune", |
| "optimizer_lr": 5e-05, |
| "optimizer_betas": [ |
| 0.9, |
| 0.95 |
| ], |
| "optimizer_eps": 1e-08, |
| "optimizer_weight_decay": 0.01, |
| "optimizer_grad_clip_norm": 1.0, |
| "scheduler_warmup_steps": 2000, |
| "scheduler_decay_steps": 100000, |
| "scheduler_decay_lr": 5e-06, |
| "tokenizer_max_length": 48, |
| "freeze_vision_encoder": false, |
| "train_expert_only": false, |
| "enable_vqa_loss": true, |
| "lambda_vqa": 1.0, |
| "tokenize_state": true, |
| "action_token_min": 248077, |
| "action_token_max": 250124, |
| "knowledge_insulation": false, |
| "block_action_attend_fast_tokens": true, |
| "inference_action_type": "fm", |
| "use_sdpa": false, |
| "num_learnable_tokens": 50, |
| "wan_checkpoint_path": "Wan-AI/Wan2.2-TI2V-5B", |
| "wan_config_path": "Wan-AI/Wan2.2-TI2V-5B", |
| "vae_path": "Wan-AI/Wan2.2-TI2V-5B/Wan2.2_VAE.pth", |
| "video_precision": "bfloat16", |
| "freeze_wan_dit": true, |
| "num_video_frames": 4, |
| "video_height": 224, |
| "video_width": 224, |
| "video_loss_weight": 1.0, |
| "video_loss_only": false, |
| "action_loss_only": false, |
| "freeze_learnable_tokens": false, |
| "inference_backend": "standard" |
| } |
|
|
