ByteDance/Sa2VA-1B

1{2  "_commit_hash": null,3  "architectures": [4    "Sa2VAChatModel"5  ],6  "auto_map": {7    "AutoConfig": "configuration_sa2va_chat.Sa2VAChatConfig",8    "AutoModel": "modeling_sa2va_chat.Sa2VAChatModel",9    "AutoModelForCausalLM": "modeling_sa2va_chat.Sa2VAChatModel"10  },11  "downsample_ratio": 0.5,12  "dynamic_image_size": true,13  "force_image_size": 448,14  "hidden_size": 896,15  "llm_config": {16    "_name_or_path": "Qwen/Qwen2.5-0.5B-Instruct",17    "add_cross_attention": false,18    "architectures": [19      "Qwen2ForCausalLM"20    ],21    "attention_dropout": 0.0,22    "bad_words_ids": null,23    "begin_suppress_tokens": null,24    "bos_token_id": 151643,25    "chunk_size_feed_forward": 0,26    "cross_attention_hidden_size": null,27    "decoder_start_token_id": null,28    "diversity_penalty": 0.0,29    "do_sample": false,30    "early_stopping": false,31    "encoder_no_repeat_ngram_size": 0,32    "eos_token_id": 151645,33    "exponential_decay_length_penalty": null,34    "finetuning_task": null,35    "forced_bos_token_id": null,36    "forced_eos_token_id": null,37    "hidden_act": "silu",38    "hidden_size": 896,39    "id2label": {40      "0": "LABEL_0",41      "1": "LABEL_1"42    },43    "initializer_range": 0.02,44    "intermediate_size": 4864,45    "is_decoder": false,46    "is_encoder_decoder": false,47    "label2id": {48      "LABEL_0": 0,49      "LABEL_1": 150    },51    "length_penalty": 1.0,52    "max_length": 20,53    "max_position_embeddings": 32768,54    "max_window_layers": 21,55    "min_length": 0,56    "model_type": "qwen2",57    "no_repeat_ngram_size": 0,58    "num_attention_heads": 14,59    "num_beam_groups": 1,60    "num_beams": 1,61    "num_hidden_layers": 24,62    "num_key_value_heads": 2,63    "num_return_sequences": 1,64    "output_attentions": false,65    "output_hidden_states": false,66    "output_scores": false,67    "pad_token_id": null,68    "prefix": null,69    "problem_type": null,70    "pruned_heads": {},71    "remove_invalid_values": false,72    "repetition_penalty": 1.0,73    "return_dict": true,74    "return_dict_in_generate": false,75    "rms_norm_eps": 1e-06,76    "rope_theta": 1000000.0,77    "sep_token_id": null,78    "sliding_window": null,79    "suppress_tokens": null,80    "task_specific_params": null,81    "temperature": 1.0,82    "tf_legacy_loss": false,83    "tie_encoder_decoder": false,84    "tie_word_embeddings": false,85    "tokenizer_class": null,86    "top_k": 50,87    "top_p": 1.0,88    "torch_dtype": "bfloat16",89    "torchscript": false,90    "transformers_version": "4.44.0",91    "typical_p": 1.0,92    "use_bfloat16": true,93    "use_cache": true,94    "use_sliding_window": false,95    "vocab_size": 15167996  },97  "max_dynamic_patch": 12,98  "min_dynamic_patch": 1,99  "model_type": "sa2va_chat",100  "pad2square": false,101  "ps_version": "v2",102  "select_layer": -1,103  "template": "qwen_chat",104  "tie_word_embeddings": false,105  "torch_dtype": "bfloat16",106  "transformers_version": null,107  "use_backbone_lora": 0,108  "use_llm_lora": 0,109  "use_thumbnail": true,110  "vision_config": {111    "_name_or_path": "",112    "add_cross_attention": false,113    "architectures": [114      "InternVisionModel"115    ],116    "attention_dropout": 0.0,117    "bad_words_ids": null,118    "begin_suppress_tokens": null,119    "bos_token_id": null,120    "chunk_size_feed_forward": 0,121    "cross_attention_hidden_size": null,122    "decoder_start_token_id": null,123    "diversity_penalty": 0.0,124    "do_sample": false,125    "drop_path_rate": 0.0,126    "dropout": 0.0,127    "early_stopping": false,128    "encoder_no_repeat_ngram_size": 0,129    "eos_token_id": null,130    "exponential_decay_length_penalty": null,131    "finetuning_task": null,132    "forced_bos_token_id": null,133    "forced_eos_token_id": null,134    "hidden_act": "gelu",135    "hidden_size": 1024,136    "id2label": {137      "0": "LABEL_0",138      "1": "LABEL_1"139    },140    "image_size": 448,141    "initializer_factor": 1.0,142    "initializer_range": 0.02,143    "intermediate_size": 4096,144    "is_decoder": false,145    "is_encoder_decoder": false,146    "label2id": {147      "LABEL_0": 0,148      "LABEL_1": 1149    },150    "layer_norm_eps": 1e-06,151    "length_penalty": 1.0,152    "max_length": 20,153    "min_length": 0,154    "model_type": "intern_vit_6b",155    "no_repeat_ngram_size": 0,156    "norm_type": "layer_norm",157    "num_attention_heads": 16,158    "num_beam_groups": 1,159    "num_beams": 1,160    "num_channels": 3,161    "num_hidden_layers": 24,162    "num_return_sequences": 1,163    "output_attentions": false,164    "output_hidden_states": false,165    "output_scores": false,166    "pad_token_id": null,167    "patch_size": 14,168    "prefix": null,169    "problem_type": null,170    "pruned_heads": {},171    "qk_normalization": false,172    "qkv_bias": true,173    "remove_invalid_values": false,174    "repetition_penalty": 1.0,175    "return_dict": true,176    "return_dict_in_generate": false,177    "sep_token_id": null,178    "suppress_tokens": null,179    "task_specific_params": null,180    "temperature": 1.0,181    "tf_legacy_loss": false,182    "tie_encoder_decoder": false,183    "tie_word_embeddings": true,184    "tokenizer_class": null,185    "top_k": 50,186    "top_p": 1.0,187    "torch_dtype": "bfloat16",188    "torchscript": false,189    "transformers_version": "4.44.0",190    "typical_p": 1.0,191    "use_bfloat16": true,192    "use_flash_attn": true193  }194}195