| { | |
| "_name_or_path": "./checkpoints/vicuna-v1-3-7b", | |
| "architectures": [ | |
| "HandsOnVLMForCausalLM" | |
| ], | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "freeze_mm_mlp_adapter": false, | |
| "fuse_input_mode": "origin", | |
| "hand_token_id": 32100, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "hoi_lambda": 1.0, | |
| "image_aspect_ratio": "square", | |
| "image_grid_pinpoints": null, | |
| "initializer_range": 0.02, | |
| "input_type": "video", | |
| "intermediate_size": 11008, | |
| "lambda_last_hand": 0.001, | |
| "lambda_obj": 0.1, | |
| "lambda_obj_kl": 0.001, | |
| "lambda_traj": 1, | |
| "lambda_traj_kl": 0.001, | |
| "max_position_embeddings": 2048, | |
| "mm_hidden_size": 1024, | |
| "mm_use_im_patch_token": false, | |
| "mm_use_im_start_end": false, | |
| "mm_vision_select_feature": "patch", | |
| "mm_vision_select_layer": -2, | |
| "mm_vision_tower": "openai/clip-vit-large-patch14", | |
| "model_type": "handsonvlm", | |
| "num_attention_heads": 32, | |
| "num_frames": 100, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 32, | |
| "num_time_tokens": 100, | |
| "pad_token_id": 0, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "tie_word_embeddings": false, | |
| "time_token_ids": [ | |
| 32000, | |
| 32001, | |
| 32002, | |
| 32003, | |
| 32004, | |
| 32005, | |
| 32006, | |
| 32007, | |
| 32008, | |
| 32009, | |
| 32010, | |
| 32011, | |
| 32012, | |
| 32013, | |
| 32014, | |
| 32015, | |
| 32016, | |
| 32017, | |
| 32018, | |
| 32019, | |
| 32020, | |
| 32021, | |
| 32022, | |
| 32023, | |
| 32024, | |
| 32025, | |
| 32026, | |
| 32027, | |
| 32028, | |
| 32029, | |
| 32030, | |
| 32031, | |
| 32032, | |
| 32033, | |
| 32034, | |
| 32035, | |
| 32036, | |
| 32037, | |
| 32038, | |
| 32039, | |
| 32040, | |
| 32041, | |
| 32042, | |
| 32043, | |
| 32044, | |
| 32045, | |
| 32046, | |
| 32047, | |
| 32048, | |
| 32049, | |
| 32050, | |
| 32051, | |
| 32052, | |
| 32053, | |
| 32054, | |
| 32055, | |
| 32056, | |
| 32057, | |
| 32058, | |
| 32059, | |
| 32060, | |
| 32061, | |
| 32062, | |
| 32063, | |
| 32064, | |
| 32065, | |
| 32066, | |
| 32067, | |
| 32068, | |
| 32069, | |
| 32070, | |
| 32071, | |
| 32072, | |
| 32073, | |
| 32074, | |
| 32075, | |
| 32076, | |
| 32077, | |
| 32078, | |
| 32079, | |
| 32080, | |
| 32081, | |
| 32082, | |
| 32083, | |
| 32084, | |
| 32085, | |
| 32086, | |
| 32087, | |
| 32088, | |
| 32089, | |
| 32090, | |
| 32091, | |
| 32092, | |
| 32093, | |
| 32094, | |
| 32095, | |
| 32096, | |
| 32097, | |
| 32098, | |
| 32099 | |
| ], | |
| "torch_dtype": "bfloat16", | |
| "traj_decoder": "CVAE", | |
| "transformers_version": "4.31.0", | |
| "tune_mm_mlp_adapter": false, | |
| "use_cache": false, | |
| "use_mm_proj": true, | |
| "video_compress_mode": "temporal_spatial_pool", | |
| "vocab_size": 32101 | |
| } | |