| { | |
| "X": [ | |
| "Image", | |
| "Video" | |
| ], | |
| "_attn_implementation_autoset": true, | |
| "_name_or_path": "DAMO-NLP-SG/VideoLLaMA3-7B", | |
| "architectures": [ | |
| "Videollama3Qwen2ForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "DAMO-NLP-SG/VideoLLaMA3-7B--configuration_videollama3.Videollama3Qwen2Config", | |
| "AutoModelForCausalLM": "DAMO-NLP-SG/VideoLLaMA3-7B--modeling_videollama3.Videollama3Qwen2ForCausalLM" | |
| }, | |
| "bos_token_id": 151643, | |
| "eos_token_id": 151645, | |
| "freeze_mm_mlp_adapter": true, | |
| "hidden_act": "silu", | |
| "hidden_size": 3584, | |
| "image_aspect_ratio": "pad", | |
| "image_grid_pinpoints": null, | |
| "image_token_index": 151665, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 18944, | |
| "max_position_embeddings": 32768, | |
| "max_window_layers": 28, | |
| "mm_projector_type": "mlp2x_gelu", | |
| "mm_use_x_patch_token": false, | |
| "mm_use_x_start_end": false, | |
| "model_type": "videollama3_qwen2", | |
| "num_attention_heads": 28, | |
| "num_hidden_layers": 28, | |
| "num_key_value_heads": 4, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.3", | |
| "tune_mm_mlp_adapter": false, | |
| "use_cache": true, | |
| "use_sliding_window": false, | |
| "use_token_compression": false, | |
| "vision_encoder": null, | |
| "vision_encoder_config": { | |
| "hidden_size": 1152, | |
| "intermediate_size": 4304, | |
| "model_type": "videollama3_vision_encoder", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 27, | |
| "patch_size": 14 | |
| }, | |
| "vocab_size": 152064 | |
| } | |