CoolFace
Modelpublic

KalvinPhan/MathCoder-VL-34bit

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face
0likes5downloads
config.json148 linesDownload Raw Back to root
1{2  "architectures": [3    "Qwen3VLForConditionalGeneration"4  ],5  "torch_dtype": "bfloat16",6  "eos_token_id": 151645,7  "image_token_id": 151655,8  "model_type": "qwen3_vl",9  "pad_token_id": 151654,10  "quantization_config": {11    "_load_in_4bit": true,12    "_load_in_8bit": false,13    "bnb_4bit_compute_dtype": "bfloat16",14    "bnb_4bit_quant_storage": "uint8",15    "bnb_4bit_quant_type": "nf4",16    "bnb_4bit_use_double_quant": true,17    "llm_int8_enable_fp32_cpu_offload": false,18    "llm_int8_has_fp16_weight": false,19    "llm_int8_skip_modules": [20      "embed_tokens",21      "embedding",22      "lm_head",23      "multi_modal_projector",24      "merger",25      "modality_projection",26      "router",27      "visual",28      "vision_tower",29      "model.language_model.layers.3.mlp",30      "model.visual.blocks.23.attn",31      "model.language_model.layers.0.self_attn",32      "model.visual.blocks.22.mlp",33      "model.visual.blocks.19.attn",34      "model.language_model.layers.0.mlp",35      "model.visual.blocks.20.attn",36      "model.visual.merger",37      "model.visual.blocks.15.attn",38      "model.visual.blocks.22.attn",39      "model.language_model.layers.27.mlp",40      "model.visual.blocks.21.attn",41      "model.visual.blocks.18.attn",42      "model.visual.blocks.17.mlp",43      "model.visual.blocks.17.attn",44      "model.visual.blocks.19.mlp",45      "model.visual.blocks.14.mlp",46      "model.visual.blocks.16.attn",47      "model.visual.blocks.16.mlp",48      "model.visual.blocks.21.mlp",49      "model.visual.blocks.15.mlp",50      "model.visual.blocks.14.attn",51      "model.visual.blocks.18.mlp",52      "model.visual.blocks.20.mlp",53      "model.language_model.layers.2.mlp",54      "model.visual.blocks.13.mlp",55      "model.visual.blocks.23.mlp",56      "model.visual.blocks.13.attn",57      "model.visual.blocks.11.mlp",58      "model.visual.blocks.12.attn",59      "model.visual.blocks.12.mlp",60      "model.visual.blocks.11.attn",61      "model.visual.blocks.10.attn",62      "model.visual.blocks.10.mlp",63      "model.visual.blocks.0.attn",64      "model.visual.blocks.7.mlp",65      "model.visual.blocks.9.mlp",66      "model.visual.blocks.9.attn",67      "model.visual.blocks.6.attn",68      "model.visual.blocks.4.mlp",69      "model.visual.blocks.3.mlp",70      "model.visual.blocks.4.attn",71      "model.visual.blocks.2.attn",72      "model.visual.blocks.3.attn",73      "model.visual.blocks.8.mlp",74      "model.visual.blocks.8.attn",75      "model.visual.blocks.5.mlp",76      "model.visual.blocks.7.attn",77      "model.visual.blocks.6.mlp",78      "model.visual.blocks.5.attn",79      "model.visual.blocks.1.mlp",80      "model.visual.blocks.1.attn",81      "model.visual.blocks.2.mlp",82      "model.visual.blocks.0.mlp"83    ],84    "llm_int8_threshold": 6.0,85    "load_in_4bit": true,86    "load_in_8bit": false,87    "quant_method": "bitsandbytes"88  },89  "text_config": {90    "attention_bias": false,91    "attention_dropout": 0.0,92    "bos_token_id": 151643,93    "torch_dtype": "bfloat16",94    "eos_token_id": 151645,95    "head_dim": 128,96    "hidden_act": "silu",97    "hidden_size": 2048,98    "initializer_range": 0.02,99    "intermediate_size": 6144,100    "max_position_embeddings": 262144,101    "model_type": "qwen3_vl_text",102    "num_attention_heads": 16,103    "num_hidden_layers": 28,104    "num_key_value_heads": 8,105    "rms_norm_eps": 1e-06,106    "rope_scaling": {107      "mrope_interleaved": true,108      "mrope_section": [109        24,110        20,111        20112      ],113      "rope_type": "default"114    },115    "rope_theta": 5000000,116    "tie_word_embeddings": true,117    "use_cache": true,118    "vocab_size": 151936119  },120  "tie_word_embeddings": true,121  "transformers_version": "4.57.1",122  "unsloth_fixed": true,123  "video_token_id": 151656,124  "vision_config": {125    "deepstack_visual_indexes": [126      5,127      11,128      17129    ],130    "depth": 24,131    "torch_dtype": "bfloat16",132    "hidden_act": "gelu_pytorch_tanh",133    "hidden_size": 1024,134    "in_channels": 3,135    "initializer_range": 0.02,136    "intermediate_size": 4096,137    "model_type": "qwen3_vl",138    "num_heads": 16,139    "num_position_embeddings": 2304,140    "out_hidden_size": 2048,141    "patch_size": 16,142    "spatial_merge_size": 2,143    "temporal_patch_size": 2144  },145  "vision_end_token_id": 151653,146  "vision_start_token_id": 151652147}148