Efficient-Large-Model/VILA1.5-3b

1{2  "_name_or_path": "./vlm",3  "architectures": [4    "LlavaLlamaModel"5  ],6  "drop_path_rate": 0.0,7  "hidden_size": 2560,8  "image_aspect_ratio": "resize",9  "interpolate_mode": "linear",10  "llm_cfg": {11    "_name_or_path": "./llm",12    "add_cross_attention": false,13    "architectures": [14      "LlamaForCausalLM"15    ],16    "attention_bias": false,17    "attention_dropout": 0.0,18    "bad_words_ids": null,19    "begin_suppress_tokens": null,20    "bos_token_id": 1,21    "chunk_size_feed_forward": 0,22    "cross_attention_hidden_size": null,23    "decoder_start_token_id": null,24    "diversity_penalty": 0.0,25    "do_sample": false,26    "early_stopping": false,27    "encoder_no_repeat_ngram_size": 0,28    "eos_token_id": 2,29    "exponential_decay_length_penalty": null,30    "finetuning_task": null,31    "forced_bos_token_id": null,32    "forced_eos_token_id": null,33    "hidden_act": "silu",34    "hidden_size": 2560,35    "id2label": {36      "0": "LABEL_0",37      "1": "LABEL_1"38    },39    "initializer_range": 0.02,40    "intermediate_size": 6912,41    "is_decoder": false,42    "is_encoder_decoder": false,43    "label2id": {44      "LABEL_0": 0,45      "LABEL_1": 146    },47    "length_penalty": 1.0,48    "max_length": 20,49    "max_position_embeddings": 4096,50    "min_length": 0,51    "model_max_length": 4096,52    "model_type": "llama",53    "no_repeat_ngram_size": 0,54    "num_attention_heads": 20,55    "num_beam_groups": 1,56    "num_beams": 1,57    "num_hidden_layers": 32,58    "num_key_value_heads": 20,59    "num_return_sequences": 1,60    "output_attentions": false,61    "output_hidden_states": false,62    "output_scores": false,63    "pad_token_id": 0,64    "prefix": null,65    "pretraining_tp": 1,66    "problem_type": null,67    "pruned_heads": {},68    "remove_invalid_values": false,69    "repetition_penalty": 1.0,70    "return_dict": true,71    "return_dict_in_generate": false,72    "rms_norm_eps": 1e-05,73    "rope_scaling": null,74    "rope_theta": 10000.0,75    "sep_token_id": null,76    "suppress_tokens": null,77    "task_specific_params": null,78    "temperature": 1.0,79    "tf_legacy_loss": false,80    "tie_encoder_decoder": false,81    "tie_word_embeddings": false,82    "tokenizer_class": null,83    "tokenizer_model_max_length": 4096,84    "tokenizer_padding_side": "right",85    "top_k": 50,86    "top_p": 1.0,87    "torch_dtype": "bfloat16",88    "torchscript": false,89    "typical_p": 1.0,90    "use_bfloat16": false,91    "use_cache": true,92    "vocab_size": 3200093  },94  "mm_hidden_size": 1152,95  "mm_projector_cfg": {96    "_name_or_path": "./mm_projector",97    "add_cross_attention": false,98    "architectures": [99      "MultimodalProjector"100    ],101    "bad_words_ids": null,102    "begin_suppress_tokens": null,103    "bos_token_id": null,104    "chunk_size_feed_forward": 0,105    "cross_attention_hidden_size": null,106    "decoder_start_token_id": null,107    "diversity_penalty": 0.0,108    "do_sample": false,109    "early_stopping": false,110    "encoder_no_repeat_ngram_size": 0,111    "eos_token_id": null,112    "exponential_decay_length_penalty": null,113    "finetuning_task": null,114    "forced_bos_token_id": null,115    "forced_eos_token_id": null,116    "id2label": {117      "0": "LABEL_0",118      "1": "LABEL_1"119    },120    "is_decoder": false,121    "is_encoder_decoder": false,122    "label2id": {123      "LABEL_0": 0,124      "LABEL_1": 1125    },126    "length_penalty": 1.0,127    "max_length": 20,128    "min_length": 0,129    "mm_projector_type": "mlp_downsample",130    "model_type": "v2l_projector",131    "no_repeat_ngram_size": 0,132    "num_beam_groups": 1,133    "num_beams": 1,134    "num_return_sequences": 1,135    "output_attentions": false,136    "output_hidden_states": false,137    "output_scores": false,138    "pad_token_id": null,139    "prefix": null,140    "problem_type": null,141    "pruned_heads": {},142    "remove_invalid_values": false,143    "repetition_penalty": 1.0,144    "return_dict": true,145    "return_dict_in_generate": false,146    "sep_token_id": null,147    "suppress_tokens": null,148    "task_specific_params": null,149    "temperature": 1.0,150    "tf_legacy_loss": false,151    "tie_encoder_decoder": false,152    "tie_word_embeddings": true,153    "tokenizer_class": null,154    "top_k": 50,155    "top_p": 1.0,156    "torch_dtype": "bfloat16",157    "torchscript": false,158    "typical_p": 1.0,159    "use_bfloat16": false160  },161  "mm_projector_lr": null,162  "mm_use_im_patch_token": false,163  "mm_use_im_start_end": false,164  "mm_vision_select_feature": "cls_patch",165  "mm_vision_select_layer": -2,166  "model_dtype": "torch.bfloat16",167  "model_type": "llava_llama",168  "num_video_frames": 8,169  "resume_path": "./vlm",170  "s2": false,171  "s2_max_split_size": 336,172  "s2_scales": "336,672,1008",173  "transformers_version": "4.36.2",174  "tune_language_model": true,175  "tune_mm_projector": true,176  "tune_vision_tower": true,177  "vision_resolution": -1,178  "vision_tower_cfg": {179    "_name_or_path": "./vision_tower",180    "add_cross_attention": false,181    "architectures": [182      "SiglipVisionModel"183    ],184    "attention_dropout": 0.0,185    "bad_words_ids": null,186    "begin_suppress_tokens": null,187    "bos_token_id": null,188    "chunk_size_feed_forward": 0,189    "cross_attention_hidden_size": null,190    "decoder_start_token_id": null,191    "diversity_penalty": 0.0,192    "do_sample": false,193    "early_stopping": false,194    "encoder_no_repeat_ngram_size": 0,195    "eos_token_id": null,196    "exponential_decay_length_penalty": null,197    "finetuning_task": null,198    "forced_bos_token_id": null,199    "forced_eos_token_id": null,200    "hidden_act": "gelu_pytorch_tanh",201    "hidden_size": 1152,202    "id2label": {203      "0": "LABEL_0",204      "1": "LABEL_1"205    },206    "image_size": 384,207    "intermediate_size": 4304,208    "is_decoder": false,209    "is_encoder_decoder": false,210    "label2id": {211      "LABEL_0": 0,212      "LABEL_1": 1213    },214    "layer_norm_eps": 1e-06,215    "length_penalty": 1.0,216    "max_length": 20,217    "min_length": 0,218    "model_type": "siglip_vision_model",219    "no_repeat_ngram_size": 0,220    "num_attention_heads": 16,221    "num_beam_groups": 1,222    "num_beams": 1,223    "num_channels": 3,224    "num_hidden_layers": 27,225    "num_return_sequences": 1,226    "output_attentions": false,227    "output_hidden_states": false,228    "output_scores": false,229    "pad_token_id": null,230    "patch_size": 14,231    "prefix": null,232    "problem_type": null,233    "pruned_heads": {},234    "remove_invalid_values": false,235    "repetition_penalty": 1.0,236    "return_dict": true,237    "return_dict_in_generate": false,238    "sep_token_id": null,239    "suppress_tokens": null,240    "task_specific_params": null,241    "temperature": 1.0,242    "tf_legacy_loss": false,243    "tie_encoder_decoder": false,244    "tie_word_embeddings": true,245    "tokenizer_class": null,246    "top_k": 50,247    "top_p": 1.0,248    "torch_dtype": "bfloat16",249    "torchscript": false,250    "typical_p": 1.0,251    "use_bfloat16": false252  }253}254