| { | |
| "architectures": [ | |
| "Qwen3VLForConditionalGeneration" | |
| ], | |
| "torch_dtype": "bfloat16", | |
| "eos_token_id": 151645, | |
| "image_token_id": 151655, | |
| "model_type": "qwen3_vl", | |
| "pad_token_id": 151654, | |
| "quantization_config": { | |
| "_load_in_4bit": true, | |
| "_load_in_8bit": false, | |
| "bnb_4bit_compute_dtype": "bfloat16", | |
| "bnb_4bit_quant_storage": "uint8", | |
| "bnb_4bit_quant_type": "nf4", | |
| "bnb_4bit_use_double_quant": true, | |
| "llm_int8_enable_fp32_cpu_offload": false, | |
| "llm_int8_has_fp16_weight": false, | |
| "llm_int8_skip_modules": [ | |
| "embed_tokens", | |
| "embedding", | |
| "lm_head", | |
| "multi_modal_projector", | |
| "merger", | |
| "modality_projection", | |
| "router", | |
| "visual", | |
| "vision_tower", | |
| "model.language_model.layers.32.mlp", | |
| "model.language_model.layers.11.mlp", | |
| "model.language_model.layers.30.mlp", | |
| "model.language_model.layers.7.self_attn", | |
| "model.language_model.layers.3.mlp", | |
| "model.visual.deepstack_merger_list.2", | |
| "model.language_model.layers.8.mlp", | |
| "model.language_model.layers.4.mlp", | |
| "model.visual.deepstack_merger_list.0", | |
| "model.language_model.layers.7.mlp", | |
| "model.language_model.layers.37.mlp", | |
| "model.language_model.layers.12.self_attn", | |
| "model.visual.blocks.21.attn", | |
| "model.language_model.layers.3.self_attn", | |
| "model.visual.blocks.26.attn", | |
| "model.visual.deepstack_merger_list.1", | |
| "model.visual.blocks.24.attn", | |
| "model.language_model.layers.6.mlp", | |
| "model.language_model.layers.34.mlp", | |
| "model.language_model.layers.43.mlp", | |
| "model.visual.blocks.19.attn", | |
| "model.language_model.layers.62.mlp", | |
| "model.visual.blocks.20.attn", | |
| "model.visual.blocks.25.mlp", | |
| "model.visual.blocks.17.mlp", | |
| "model.visual.blocks.11.mlp", | |
| "model.visual.blocks.4.attn", | |
| "model.visual.merger", | |
| "model.visual.blocks.21.mlp", | |
| "model.visual.blocks.6.mlp", | |
| "model.visual.blocks.15.attn", | |
| "model.visual.blocks.25.attn", | |
| "model.visual.blocks.10.mlp", | |
| "model.visual.blocks.14.mlp", | |
| "model.visual.blocks.9.attn", | |
| "model.visual.blocks.18.attn", | |
| "model.visual.blocks.23.attn", | |
| "model.visual.blocks.9.mlp", | |
| "model.visual.blocks.7.mlp", | |
| "model.visual.blocks.22.mlp", | |
| "model.visual.blocks.7.attn", | |
| "model.visual.blocks.13.mlp", | |
| "model.visual.blocks.20.mlp", | |
| "model.visual.blocks.16.attn", | |
| "model.visual.blocks.1.attn", | |
| "model.visual.blocks.8.mlp", | |
| "model.visual.blocks.17.attn", | |
| "model.visual.blocks.14.attn", | |
| "model.visual.blocks.5.mlp", | |
| "model.visual.blocks.24.mlp", | |
| "model.visual.blocks.11.attn", | |
| "model.visual.blocks.22.attn", | |
| "model.visual.blocks.2.attn", | |
| "model.visual.blocks.1.mlp", | |
| "model.visual.blocks.26.mlp", | |
| "model.visual.blocks.16.mlp", | |
| "model.visual.blocks.13.attn", | |
| "model.visual.blocks.2.mlp", | |
| "model.visual.blocks.18.mlp", | |
| "model.visual.blocks.3.mlp", | |
| "model.visual.blocks.0.mlp", | |
| "model.visual.blocks.12.mlp", | |
| "model.visual.blocks.12.attn", | |
| "model.visual.blocks.15.mlp", | |
| "model.visual.blocks.5.attn", | |
| "model.visual.blocks.3.attn", | |
| "model.visual.blocks.10.attn", | |
| "model.visual.blocks.0.attn", | |
| "model.visual.blocks.4.mlp", | |
| "model.visual.blocks.19.mlp", | |
| "model.visual.blocks.23.mlp", | |
| "model.visual.blocks.8.attn", | |
| "model.visual.blocks.6.attn", | |
| "model.language_model.layers.31.mlp.down_proj", | |
| "model.language_model.layers.43.mlp.up_proj", | |
| "model.language_model.layers.45.self_attn.o_proj" | |
| ], | |
| "llm_int8_threshold": 6.0, | |
| "load_in_4bit": true, | |
| "load_in_8bit": false, | |
| "quant_method": "bitsandbytes" | |
| }, | |
| "text_config": { | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 151643, | |
| "torch_dtype": "bfloat16", | |
| "eos_token_id": 151645, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 5120, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 25600, | |
| "max_position_embeddings": 262144, | |
| "model_type": "qwen3_vl_text", | |
| "num_attention_heads": 64, | |
| "num_hidden_layers": 64, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": { | |
| "mrope_interleaved": true, | |
| "mrope_section": [ | |
| 24, | |
| 20, | |
| 20 | |
| ], | |
| "rope_type": "default" | |
| }, | |
| "rope_theta": 5000000, | |
| "use_cache": true, | |
| "vocab_size": 151936 | |
| }, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "4.57.1", | |
| "unsloth_fixed": true, | |
| "video_token_id": 151656, | |
| "vision_config": { | |
| "deepstack_visual_indexes": [ | |
| 8, | |
| 16, | |
| 24 | |
| ], | |
| "depth": 27, | |
| "torch_dtype": "bfloat16", | |
| "hidden_act": "gelu_pytorch_tanh", | |
| "hidden_size": 1152, | |
| "in_channels": 3, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4304, | |
| "model_type": "qwen3_vl", | |
| "num_heads": 16, | |
| "num_position_embeddings": 2304, | |
| "out_hidden_size": 5120, | |
| "patch_size": 16, | |
| "spatial_merge_size": 2, | |
| "temporal_patch_size": 2 | |
| }, | |
| "vision_end_token_id": 151653, | |
| "vision_start_token_id": 151652 | |
| } | |