| { |
| "_name_or_path": "/mnt/data/jiaxing.zjx/code/HumanOmni/work_dirs/humanomniqwen2_siglip/finetune_HumanOmni_1B_Omni", |
| "architectures": [ |
| "HumanOmniQwen2ForCausalLM" |
| ], |
| "attention_dropout": 0.0, |
| "audio_hidden_size": 1280, |
| "bos_token_id": 151643, |
| "eos_token_id": 151645, |
| "freeze_mm_mlp_adapter": false, |
| "hidden_act": "silu", |
| "hidden_size": 896, |
| "hyper_layers": [ |
| 15, |
| 22 |
| ], |
| "image_aspect_ratio": "pad", |
| "initializer_range": 0.02, |
| "intermediate_size": 4864, |
| "max_position_embeddings": 32768, |
| "max_window_layers": 21, |
| "mm_audio_projector_type": "mlp2x_gelu", |
| "mm_audio_tower": "openai/whisper-large-v3", |
| "mm_hidden_size": 768, |
| "mm_projector_lr": null, |
| "mm_projector_type": "all_in_one_small", |
| "mm_tunable_parts": "mm_mlp_adapter,audio_projector,mm_language_model", |
| "mm_use_x_start_end": true, |
| "mm_vision_select_feature": "patch", |
| "mm_vision_select_layer": -2, |
| "mm_vision_tower": "google/siglip-base-patch16-224", |
| "model_type": "HumanOmni_qwen2", |
| "num_attention_heads": 14, |
| "num_frames": 8, |
| "num_hidden_layers": 24, |
| "num_key_value_heads": 2, |
| "rms_norm_eps": 1e-06, |
| "rope_theta": 1000000.0, |
| "sliding_window": null, |
| "tie_word_embeddings": true, |
| "tokenizer_model_max_length": 2048, |
| "tokenizer_padding_side": "right", |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.45.0.dev0", |
| "tune_mm_mlp_adapter": false, |
| "use_cache": true, |
| "use_mm_proj": true, |
| "use_sliding_window": false, |
| "vocab_size": 151936 |
| } |
|
|