| { |
| "_name_or_path": "LinWeizheDragon/PreFLMR_ViT-L_ENCN", |
| "architectures": [ |
| "FLMRModelForRetrieval" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_flmr.FLMRConfig", |
| "AutoModel": "modeling_flmr.FLMRModelForRetrieval" |
| }, |
| "context_concat_output_from_text_encoder": true, |
| "context_concat_output_from_vision_encoder": false, |
| "dim": 128, |
| "initializer_range": 0.02, |
| "load_cpu_extension": true, |
| "mapping_network_prefix_length": 32, |
| "mask_instruction_token": "\n:", |
| "mask_punctuation": true, |
| "model_type": "flmr", |
| "query_concat_output_from_text_encoder": true, |
| "query_concat_output_from_vision_encoder": true, |
| "query_mask_input_ids_skip_list": [ |
| 1 |
| ], |
| "separate_query_and_context_text_encoder": false, |
| "separate_query_and_context_vision_encoder": false, |
| "text_config": { |
| "architectures": [ |
| "XLMRobertaModel" |
| ], |
| "gradient_checkpointing": false, |
| "hidden_size": 1024, |
| "model_type": "flmr_text_model", |
| "projection_dim": 1024, |
| "text_encoder_base_model": "BAAI/bge-m3", |
| "query_maxlen": 64, |
| "use_cache": true |
| }, |
| "torch_dtype": "float32", |
| "transformer_mapping_config_base": "google-bert/bert-large-uncased", |
| "transformer_mapping_cross_attention_length": 32, |
| "transformer_mapping_num_hidden_layers": 1, |
| "transformers_version": "4.44.2", |
| "use_transformer_mapping_network": false, |
| "use_vision_encoder": true, |
| "vision_config": { |
| "dropout": 0.0, |
| "hidden_size": 1024, |
| "intermediate_size": 4096, |
| "model_type": "flmr_vision_model", |
| "num_attention_heads": 16, |
| "num_hidden_layers": 24, |
| "patch_size": 14, |
| "projection_dim": 768 |
| }, |
| "vision_model_version": "openai/clip-vit-large-patch14" |
| } |
|
|
