{ "type": "dic_vla", "n_obs_steps": 1, "input_features": { "observation.images.image": { "type": "VISUAL", "shape": [ 3, 256, 256 ] }, "observation.images.image2": { "type": "VISUAL", "shape": [ 3, 256, 256 ] }, "observation.state": { "type": "STATE", "shape": [ 8 ] } }, "output_features": { "action": { "type": "ACTION", "shape": [ 7 ] } }, "device": "cuda", "use_amp": false, "push_to_hub": true, "repo_id": "radiuson/Dic-libero-test", "private": null, "tags": null, "license": null, "pretrained_path": null, "chunk_size": 64, "n_action_steps": 64, "normalization_mapping": { "VISUAL": "IDENTITY", "STATE": "MEAN_STD", "ACTION": "MEAN_STD" }, "max_state_dim": 14, "max_action_dim": 14, "resize_imgs_with_padding": [ 512, 512 ], "empty_cameras": 0, "vlm_model_name": "HuggingFaceTB/SmolVLM-256M-Instruct", "load_vlm_weights": false, "num_vlm_layers": 16, "freeze_vision_encoder": true, "train_state_proj": true, "dic_config": "dic-tiny", "dic_base_channels": null, "dic_num_groups": null, "dic_encoder_blocks": null, "dic_bottleneck_blocks": null, "dic_decoder_blocks": null, "dic_dropout": 0.2, "dic_sparse_skip_stride": 1, "num_inference_steps": 4, "min_period": 1.0, "max_period": 10000.0, "condition_pooling": "mean", "condition_dim": null, "optimizer_lr": 0.0001, "optimizer_betas": [ 0.9, 0.95 ], "optimizer_eps": 1e-08, "optimizer_weight_decay": 1e-06, "optimizer_grad_clip_norm": 1.0, "scheduler_warmup_steps": 1000, "scheduler_decay_steps": 30000, "scheduler_decay_lr": 1e-06, "tokenizer_max_length": 48, "add_image_special_tokens": false, "prefix_length": -1, "use_cache": true, "pad_language_to": "longest", "adapt_to_pi_aloha": false, "use_delta_joint_actions_aloha": false }