OpenGVLab/Mini-InternVL2-4B-DA-DriveLM
5112
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": "microsoft/Phi-3-mini-128k-instruct",16 "add_cross_attention": false,17 "architectures": [18 "Phi3ForCausalLM"19 ],20 "attention_dropout": 0.0,21 "auto_map": {22 "AutoConfig": "configuration_phi3.Phi3Config",23 "AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"24 },25 "bad_words_ids": null,26 "begin_suppress_tokens": null,27 "bos_token_id": 1,28 "chunk_size_feed_forward": 0,29 "cross_attention_hidden_size": null,30 "decoder_start_token_id": null,31 "diversity_penalty": 0.0,32 "do_sample": false,33 "early_stopping": false,34 "embd_pdrop": 0.0,35 "encoder_no_repeat_ngram_size": 0,36 "eos_token_id": 32000,37 "exponential_decay_length_penalty": null,38 "finetuning_task": null,39 "forced_bos_token_id": null,40 "forced_eos_token_id": null,41 "hidden_act": "silu",42 "hidden_size": 3072,43 "id2label": {44 "0": "LABEL_0",45 "1": "LABEL_1"46 },47 "initializer_range": 0.02,48 "intermediate_size": 8192,49 "is_decoder": false,50 "is_encoder_decoder": false,51 "label2id": {52 "LABEL_0": 0,53 "LABEL_1": 154 },55 "length_penalty": 1.0,56 "max_length": 20,57 "max_position_embeddings": 131072,58 "min_length": 0,59 "model_type": "phi3",60 "no_repeat_ngram_size": 0,61 "num_attention_heads": 32,62 "num_beam_groups": 1,63 "num_beams": 1,64 "num_hidden_layers": 32,65 "num_key_value_heads": 32,66 "num_return_sequences": 1,67 "original_max_position_embeddings": 4096,68 "output_attentions": false,69 "output_hidden_states": false,70 "output_scores": false,71 "pad_token_id": 32000,72 "prefix": null,73 "problem_type": null,74 "pruned_heads": {},75 "remove_invalid_values": false,76 "repetition_penalty": 1.0,77 "resid_pdrop": 0.0,78 "return_dict": true,79 "return_dict_in_generate": false,80 "rms_norm_eps": 1e-05,81 "rope_scaling": {82 "long_factor": [83 1.0299999713897705,84 1.0499999523162842,85 1.0499999523162842,86 1.0799999237060547,87 1.2299998998641968,88 1.2299998998641968,89 1.2999999523162842,90 1.4499999284744263,91 1.5999999046325684,92 1.6499998569488525,93 1.8999998569488525,94 2.859999895095825,95 3.68999981880188,96 5.419999599456787,97 5.489999771118164,98 5.489999771118164,99 9.09000015258789,100 11.579999923706055,101 15.65999984741211,102 15.769999504089355,103 15.789999961853027,104 18.360000610351562,105 21.989999771118164,106 23.079999923706055,107 30.009998321533203,108 32.35000228881836,109 32.590003967285156,110 35.56000518798828,111 39.95000457763672,112 53.840003967285156,113 56.20000457763672,114 57.95000457763672,115 59.29000473022461,116 59.77000427246094,117 59.920005798339844,118 61.190006256103516,119 61.96000671386719,120 62.50000762939453,121 63.3700065612793,122 63.48000717163086,123 63.48000717163086,124 63.66000747680664,125 63.850006103515625,126 64.08000946044922,127 64.760009765625,128 64.80001068115234,129 64.81001281738281,130 64.81001281738281131 ],132 "short_factor": [133 1.05,134 1.05,135 1.05,136 1.1,137 1.1,138 1.1500000000000001,139 1.2000000000000002,140 1.2500000000000002,141 1.3000000000000003,142 1.3500000000000003,143 1.5000000000000004,144 2.000000000000001,145 2.000000000000001,146 2.000000000000001,147 2.000000000000001,148 2.000000000000001,149 2.000000000000001,150 2.000000000000001,151 2.000000000000001,152 2.000000000000001,153 2.000000000000001,154 2.000000000000001,155 2.000000000000001,156 2.000000000000001,157 2.000000000000001,158 2.000000000000001,159 2.000000000000001,160 2.000000000000001,161 2.000000000000001,162 2.000000000000001,163 2.000000000000001,164 2.000000000000001,165 2.0500000000000007,166 2.0500000000000007,167 2.0500000000000007,168 2.1000000000000005,169 2.1000000000000005,170 2.1000000000000005,171 2.1500000000000004,172 2.1500000000000004,173 2.3499999999999996,174 2.549999999999999,175 2.5999999999999988,176 2.5999999999999988,177 2.7499999999999982,178 2.849999999999998,179 2.849999999999998,180 2.9499999999999975181 ],182 "type": "su"183 },184 "rope_theta": 10000.0,185 "sep_token_id": null,186 "sliding_window": 262144,187 "suppress_tokens": null,188 "task_specific_params": null,189 "temperature": 1.0,190 "tf_legacy_loss": false,191 "tie_encoder_decoder": false,192 "tie_word_embeddings": false,193 "tokenizer_class": null,194 "top_k": 50,195 "top_p": 1.0,196 "torch_dtype": "bfloat16",197 "torchscript": false,198 "transformers_version": "4.37.2",199 "typical_p": 1.0,200 "use_bfloat16": true,201 "use_cache": false,202 "vocab_size": 32020203 },204 "max_dynamic_patch": 12,205 "min_dynamic_patch": 1,206 "model_type": "internvl_chat",207 "pad2square": false,208 "ps_version": "v2",209 "select_layer": -1,210 "template": "phi3-chat",211 "torch_dtype": "bfloat16",212 "transformers_version": null,213 "use_backbone_lora": 0,214 "use_llm_lora": 0,215 "use_thumbnail": true,216 "vision_config": {217 "_name_or_path": "",218 "add_cross_attention": false,219 "architectures": [220 "InternVisionModel"221 ],222 "attention_dropout": 0.0,223 "bad_words_ids": null,224 "begin_suppress_tokens": null,225 "bos_token_id": null,226 "chunk_size_feed_forward": 0,227 "cross_attention_hidden_size": null,228 "decoder_start_token_id": null,229 "diversity_penalty": 0.0,230 "do_sample": false,231 "drop_path_rate": 0.1,232 "dropout": 0.0,233 "early_stopping": false,234 "encoder_no_repeat_ngram_size": 0,235 "eos_token_id": null,236 "exponential_decay_length_penalty": null,237 "finetuning_task": null,238 "forced_bos_token_id": null,239 "forced_eos_token_id": null,240 "hidden_act": "gelu",241 "hidden_size": 1024,242 "id2label": {243 "0": "LABEL_0",244 "1": "LABEL_1"245 },246 "image_size": 448,247 "initializer_factor": 1.0,248 "initializer_range": 0.02,249 "intermediate_size": 4096,250 "is_decoder": false,251 "is_encoder_decoder": false,252 "label2id": {253 "LABEL_0": 0,254 "LABEL_1": 1255 },256 "layer_norm_eps": 1e-06,257 "length_penalty": 1.0,258 "max_length": 20,259 "min_length": 0,260 "model_type": "intern_vit_6b",261 "no_repeat_ngram_size": 0,262 "norm_type": "layer_norm",263 "num_attention_heads": 16,264 "num_beam_groups": 1,265 "num_beams": 1,266 "num_channels": 3,267 "num_hidden_layers": 24,268 "num_return_sequences": 1,269 "output_attentions": false,270 "output_hidden_states": false,271 "output_scores": false,272 "pad_token_id": null,273 "patch_size": 14,274 "prefix": null,275 "problem_type": null,276 "pruned_heads": {},277 "qk_normalization": false,278 "qkv_bias": true,279 "remove_invalid_values": false,280 "repetition_penalty": 1.0,281 "return_dict": true,282 "return_dict_in_generate": false,283 "sep_token_id": null,284 "suppress_tokens": null,285 "task_specific_params": null,286 "temperature": 1.0,287 "tf_legacy_loss": false,288 "tie_encoder_decoder": false,289 "tie_word_embeddings": true,290 "tokenizer_class": null,291 "top_k": 50,292 "top_p": 1.0,293 "torch_dtype": "bfloat16",294 "torchscript": false,295 "transformers_version": "4.37.2",296 "typical_p": 1.0,297 "use_bfloat16": true,298 "use_flash_attn": true299 }300}301