TurboVLA-base for SO-100 from H-EmbodVis/TurboVLA/checkpoints/libero/turbovla_libero.pth
c15fa68 verified | { | |
| "type": "turbovla", | |
| "paper": "https://arxiv.org/abs/2607.27205", | |
| "init_from": "H-EmbodVis/TurboVLA/checkpoints/libero/turbovla_libero.pth", | |
| "cameras": [ | |
| "front", | |
| "wrist" | |
| ], | |
| "image_size": 256, | |
| "chunk_size": 12, | |
| "state_dim": 6, | |
| "action_dim": 6, | |
| "text_length": 32, | |
| "vision": { | |
| "dropout": 0.1 | |
| }, | |
| "interaction": { | |
| "hidden_dim": 256, | |
| "nheads": 8, | |
| "num_layers": 6, | |
| "dim_feedforward": 2048, | |
| "enhancer_inner_dim": 1024, | |
| "text_dropout": 0.0, | |
| "fusion_dropout": 0.0, | |
| "fusion_droppath": 0.1 | |
| }, | |
| "action": { | |
| "num_state_tokens": 2, | |
| "num_layers": 3, | |
| "mlp_hidden_dim": 512, | |
| "state_hidden_dim": 256, | |
| "dropout": 0.1 | |
| }, | |
| "dinov3": { | |
| "transformers_version": "5.16.1", | |
| "architectures": null, | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "dtype": null, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "patch_size": 16, | |
| "hidden_size": 768, | |
| "intermediate_size": 3072, | |
| "num_hidden_layers": 12, | |
| "num_attention_heads": 12, | |
| "hidden_act": "gelu", | |
| "attention_dropout": 0.0, | |
| "initializer_range": 0.02, | |
| "layer_norm_eps": 1e-05, | |
| "rope_theta": 100.0, | |
| "image_size": 224, | |
| "num_channels": 3, | |
| "query_bias": true, | |
| "key_bias": false, | |
| "value_bias": true, | |
| "proj_bias": true, | |
| "mlp_bias": true, | |
| "layerscale_value": 1.0, | |
| "drop_path_rate": 0.0, | |
| "use_gated_mlp": false, | |
| "num_register_tokens": 4, | |
| "pos_embed_shift": null, | |
| "pos_embed_jitter": null, | |
| "pos_embed_rescale": 2.0, | |
| "apply_layernorm": true, | |
| "reshape_hidden_states": true, | |
| "stage_names": [ | |
| "stem", | |
| "stage1", | |
| "stage2", | |
| "stage3", | |
| "stage4", | |
| "stage5", | |
| "stage6", | |
| "stage7", | |
| "stage8", | |
| "stage9", | |
| "stage10", | |
| "stage11", | |
| "stage12" | |
| ], | |
| "_name_or_path": "", | |
| "model_type": "dinov3_vit", | |
| "output_attentions": false, | |
| "out_features": [ | |
| "stage12" | |
| ], | |
| "out_indices": [ | |
| 12 | |
| ] | |
| }, | |
| "bert": { | |
| "transformers_version": "5.16.1", | |
| "architectures": [ | |
| "BertForMaskedLM" | |
| ], | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "dtype": null, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "vocab_size": 30522, | |
| "hidden_size": 768, | |
| "num_hidden_layers": 12, | |
| "num_attention_heads": 12, | |
| "intermediate_size": 3072, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "attention_probs_dropout_prob": 0.1, | |
| "max_position_embeddings": 512, | |
| "type_vocab_size": 2, | |
| "initializer_range": 0.02, | |
| "layer_norm_eps": 1e-12, | |
| "pad_token_id": 0, | |
| "use_cache": true, | |
| "classifier_dropout": null, | |
| "is_decoder": false, | |
| "add_cross_attention": false, | |
| "bos_token_id": null, | |
| "eos_token_id": null, | |
| "tie_word_embeddings": true, | |
| "_name_or_path": "", | |
| "gradient_checkpointing": false, | |
| "model_type": "bert", | |
| "position_embedding_type": "absolute", | |
| "output_attentions": false | |
| } | |
| } | |