| { | |
| "_name_or_path": "./ckpt/Llava-7B-V1-1/", | |
| "architectures": [ | |
| "PixelLMForCausalLM" | |
| ], | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "freeze_mm_mlp_adapter": true, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "image_aspect_ratio": "square", | |
| "image_feature_scale_num": 2, | |
| "image_grid_pinpoints": null, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 11008, | |
| "max_position_embeddings": 2048, | |
| "max_sequence_length": 2048, | |
| "mm_hidden_size": 1024, | |
| "mm_hidden_size_config_from_config": 1280, | |
| "mm_projector_hidden_dim": 2, | |
| "mm_projector_out_dim": 1, | |
| "mm_use_im_patch_token": false, | |
| "mm_use_im_start_end": true, | |
| "mm_vision_select_feature": "patch", | |
| "mm_vision_select_layer": -2, | |
| "mm_vision_tower": "openai/clip-vit-large-patch14-336", | |
| "mm_vision_tower_config": "/mnt/bn/panxuran/LISA/configs/vit_huge_p16.json", | |
| "model_type": "llava", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 32, | |
| "out_dim": 256, | |
| "pad_token_id": 0, | |
| "pad_train_clip_images": true, | |
| "pretrain_mm_mlp_adapter": null, | |
| "pretraining_tp": 1, | |
| "resize_vision_tower": true, | |
| "resize_vision_tower_size": 448, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "separate_mm_projector": true, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "train_mask_decoder": true, | |
| "transformers_version": "4.31.0", | |
| "tune_mm_mlp_adapter": false, | |
| "use_cache": false, | |
| "use_mm_proj": true, | |
| "vision_tower": "openai/clip-vit-large-patch14-336", | |
| "vision_tower_for_mask": true, | |
| "vocab_size": 32009 | |
| } | |