| { |
| "_name_or_path": "/data/zhangyan/data/pretrained_model/videot5-mod/t5-instruct-qformer-vision", |
| "am_adapter_channels": 384, |
| "am_adapter_kernel_size_l": 3, |
| "am_adapter_kernel_size_t": 3, |
| "am_adapter_mask_zero": false, |
| "am_adapter_scope": "all", |
| "am_bbox_scale": 100, |
| "am_ocr_multi_gran": true, |
| "am_space_attn": "bros-level", |
| "am_time_mlp_freq": 1, |
| "am_time_mlp_pretrained": false, |
| "am_time_type": "adapter-conv", |
| "am_type": "divided-space-time", |
| "architectures": [ |
| "VideoQformerT5" |
| ], |
| "classifier_dropout": 0.0, |
| "d_ff": 3072, |
| "d_kv": 64, |
| "d_model": 768, |
| "decoder_start_token_id": 0, |
| "dense_act_fn": "relu", |
| "dropout_rate": 0.1, |
| "encoder_hidden_states_scope": "all", |
| "eos_token_id": 1, |
| "feed_forward_proj": "relu", |
| "initializer_factor": 1.0, |
| "initializer_range": 0.02, |
| "is_encoder_decoder": true, |
| "is_gated_act": false, |
| "key_frame_module_batch_attn": true, |
| "keyframe_module_model_name": "/data/zhangyan/data/pretrained_model/clip-vit-base-patch32", |
| "layer_norm_epsilon": 1e-06, |
| "max_length": 65, |
| "model_type": "multimodalt5", |
| "n_positions": 512, |
| "num_decoder_layers": 12, |
| "num_frames": 10, |
| "num_heads": 12, |
| "num_images": null, |
| "num_key_frames": 2, |
| "num_layers": 12, |
| "obj_len": 60, |
| "ocr_len": 250, |
| "output_past": true, |
| "pad_token_id": 0, |
| "qformer_module_model_name": "/data/zhangyan/data/pretrained_model/Salesforce/instructblip-vicuna-7b", |
| "qm_decoder_input": "all", |
| "qm_global_query_init": true, |
| "qm_interact_type": "self-cross", |
| "qm_loss": "none", |
| "qm_loss_scale": 0.5, |
| "qm_num_layers": 12, |
| "qm_num_query_tokens": 32, |
| "qm_query_decoder_qkv": "kv", |
| "qm_scope": "ocr", |
| "qm_use_ques_guide": true, |
| "ques_len": 45, |
| "relative_attention_max_distance": 128, |
| "relative_attention_num_buckets": 32, |
| "task_specific_params": { |
| "summarization": { |
| "early_stopping": true, |
| "length_penalty": 2.0, |
| "max_length": 200, |
| "min_length": 30, |
| "no_repeat_ngram_size": 3, |
| "num_beams": 4, |
| "prefix": "summarize: " |
| }, |
| "translation_en_to_de": { |
| "early_stopping": true, |
| "max_length": 300, |
| "num_beams": 4, |
| "prefix": "translate English to German: " |
| }, |
| "translation_en_to_fr": { |
| "early_stopping": true, |
| "max_length": 300, |
| "num_beams": 4, |
| "prefix": "translate English to French: " |
| }, |
| "translation_en_to_ro": { |
| "early_stopping": true, |
| "max_length": 300, |
| "num_beams": 4, |
| "prefix": "translate English to Romanian: " |
| } |
| }, |
| "torch_dtype": "float32", |
| "transformers_version": "4.42.3", |
| "use_bbox": true, |
| "use_cache": true, |
| "use_feature": false, |
| "use_frame_type": true, |
| "use_frcn": false, |
| "use_key_frame_module": false, |
| "use_qformer_module": true, |
| "video_feature_dim": 1024, |
| "vocab_size": 32128 |
| } |
|
|
