OpenGVLab/InternVL2-40B

1{2  "_commit_hash": null,3  "architectures": [4    "InternVLChatModel"5  ],6  "auto_map": {7    "AutoConfig": "configuration_internvl_chat.InternVLChatConfig",8    "AutoModel": "modeling_internvl_chat.InternVLChatModel",9    "AutoModelForCausalLM": "modeling_internvl_chat.InternVLChatModel"10  },11  "downsample_ratio": 0.5,12  "dynamic_image_size": true,13  "force_image_size": 448,14  "llm_config": {15    "_name_or_path": "NousResearch/Nous-Hermes-2-Yi-34B",16    "add_cross_attention": false,17    "architectures": [18      "LlamaForCausalLM"19    ],20    "_attn_implementation": "flash_attention_2",21    "attention_bias": false,22    "attention_dropout": 0.0,23    "bad_words_ids": null,24    "begin_suppress_tokens": null,25    "bos_token_id": 1,26    "chunk_size_feed_forward": 0,27    "cross_attention_hidden_size": null,28    "decoder_start_token_id": null,29    "diversity_penalty": 0.0,30    "do_sample": false,31    "early_stopping": false,32    "encoder_no_repeat_ngram_size": 0,33    "eos_token_id": 7,34    "exponential_decay_length_penalty": null,35    "finetuning_task": null,36    "forced_bos_token_id": null,37    "forced_eos_token_id": null,38    "hidden_act": "silu",39    "hidden_size": 7168,40    "id2label": {41      "0": "LABEL_0",42      "1": "LABEL_1"43    },44    "initializer_range": 0.02,45    "intermediate_size": 20480,46    "is_decoder": false,47    "is_encoder_decoder": false,48    "label2id": {49      "LABEL_0": 0,50      "LABEL_1": 151    },52    "length_penalty": 1.0,53    "max_length": 20,54    "max_position_embeddings": 8192,55    "min_length": 0,56    "model_type": "llama",57    "no_repeat_ngram_size": 0,58    "num_attention_heads": 56,59    "num_beam_groups": 1,60    "num_beams": 1,61    "num_hidden_layers": 60,62    "num_key_value_heads": 8,63    "num_return_sequences": 1,64    "output_attentions": false,65    "output_hidden_states": false,66    "output_scores": false,67    "pad_token_id": 0,68    "prefix": null,69    "pretraining_tp": 1,70    "problem_type": null,71    "pruned_heads": {},72    "remove_invalid_values": false,73    "repetition_penalty": 1.0,74    "return_dict": true,75    "return_dict_in_generate": false,76    "rms_norm_eps": 1e-05,77    "rope_scaling": {78      "factor": 3.0,79      "type": "dynamic"80    },81    "rope_theta": 5000000.0,82    "sep_token_id": null,83    "suppress_tokens": null,84    "task_specific_params": null,85    "temperature": 1.0,86    "tf_legacy_loss": false,87    "tie_encoder_decoder": false,88    "tie_word_embeddings": false,89    "tokenizer_class": null,90    "top_k": 50,91    "top_p": 1.0,92    "torch_dtype": "bfloat16",93    "torchscript": false,94    "transformers_version": "4.37.2",95    "typical_p": 1.0,96    "use_bfloat16": true,97    "use_cache": true,98    "vocab_size": 6400799  },100  "max_dynamic_patch": 12,101  "min_dynamic_patch": 1,102  "model_type": "internvl_chat",103  "ps_version": "v2",104  "select_layer": -1,105  "template": "Hermes-2",106  "torch_dtype": "bfloat16",107  "use_backbone_lora": 0,108  "use_llm_lora": 0,109  "use_thumbnail": true,110  "vision_config": {111    "architectures": [112      "InternVisionModel"113    ],114    "attention_dropout": 0.0,115    "drop_path_rate": 0.0,116    "dropout": 0.0,117    "hidden_act": "gelu",118    "hidden_size": 3200,119    "image_size": 448,120    "initializer_factor": 0.1,121    "initializer_range": 1e-10,122    "intermediate_size": 12800,123    "layer_norm_eps": 1e-06,124    "model_type": "intern_vit_6b",125    "norm_type": "rms_norm",126    "num_attention_heads": 25,127    "num_channels": 3,128    "num_hidden_layers": 45,129    "output_attentions": false,130    "output_hidden_states": false,131    "patch_size": 14,132    "qk_normalization": true,133    "qkv_bias": false,134    "return_dict": true,135    "torch_dtype": "bfloat16",136    "transformers_version": "4.37.2",137    "use_bfloat16": true,138    "use_flash_attn": true139  }140}141