Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- .ipynb_checkpoints/config-checkpoint.json +382 -0
- README.md +156 -0
- _config.json +393 -0
- chat_template.jinja +141 -0
- config.json +406 -0
- generation_config.json +14 -0
- model-00001-of-00024.safetensors +3 -0
- model-00002-of-00024.safetensors +3 -0
- model-00003-of-00024.safetensors +3 -0
- model-00004-of-00024.safetensors +3 -0
- model-00005-of-00024.safetensors +3 -0
- model-00006-of-00024.safetensors +3 -0
- model-00007-of-00024.safetensors +3 -0
- model-00008-of-00024.safetensors +3 -0
- model-00009-of-00024.safetensors +3 -0
- model-00010-of-00024.safetensors +3 -0
- model-00011-of-00024.safetensors +3 -0
- model-00012-of-00024.safetensors +3 -0
- model-00013-of-00024.safetensors +3 -0
- model-00014-of-00024.safetensors +3 -0
- model-00015-of-00024.safetensors +3 -0
- model-00016-of-00024.safetensors +3 -0
- model-00017-of-00024.safetensors +3 -0
- model-00018-of-00024.safetensors +3 -0
- model-00019-of-00024.safetensors +3 -0
- model-00020-of-00024.safetensors +3 -0
- model-00021-of-00024.safetensors +3 -0
- model-00022-of-00024.safetensors +3 -0
- model-00023-of-00024.safetensors +3 -0
- model-00024-of-00024.safetensors +3 -0
- model.safetensors.index.json +0 -0
- preprocessor_config.json +11 -0
- recipe.yaml +38 -0
- special_tokens_map.json +42 -0
- tokenizer.json +3 -0
- tokenizer_config.json +327 -0
- video_preprocessor_config.json +11 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
.ipynb_checkpoints/config-checkpoint.json
ADDED
|
@@ -0,0 +1,382 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Glm4vMoeForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "bfloat16",
|
| 6 |
+
"image_end_token_id": 151340,
|
| 7 |
+
"image_start_token_id": 151339,
|
| 8 |
+
"image_token_id": 151363,
|
| 9 |
+
"model_type": "glm4v_moe",
|
| 10 |
+
"quantization_config": {
|
| 11 |
+
"config_groups": {
|
| 12 |
+
"group_0": {
|
| 13 |
+
"format": "pack-quantized",
|
| 14 |
+
"input_activations": null,
|
| 15 |
+
"output_activations": null,
|
| 16 |
+
"targets": [
|
| 17 |
+
"Linear"
|
| 18 |
+
],
|
| 19 |
+
"weights": {
|
| 20 |
+
"actorder": null,
|
| 21 |
+
"block_structure": null,
|
| 22 |
+
"dynamic": false,
|
| 23 |
+
"group_size": 32,
|
| 24 |
+
"num_bits": 8,
|
| 25 |
+
"observer": "mse",
|
| 26 |
+
"observer_kwargs": {},
|
| 27 |
+
"strategy": "group",
|
| 28 |
+
"symmetric": true,
|
| 29 |
+
"type": "int"
|
| 30 |
+
}
|
| 31 |
+
}
|
| 32 |
+
},
|
| 33 |
+
"format": "pack-quantized",
|
| 34 |
+
"global_compression_ratio": null,
|
| 35 |
+
"ignore": [
|
| 36 |
+
"model.visual.blocks.0.attn.qkv",
|
| 37 |
+
"model.visual.blocks.0.attn.proj",
|
| 38 |
+
"model.visual.blocks.0.mlp.gate_proj",
|
| 39 |
+
"model.visual.blocks.0.mlp.up_proj",
|
| 40 |
+
"model.visual.blocks.0.mlp.down_proj",
|
| 41 |
+
"model.visual.blocks.1.attn.qkv",
|
| 42 |
+
"model.visual.blocks.1.attn.proj",
|
| 43 |
+
"model.visual.blocks.1.mlp.gate_proj",
|
| 44 |
+
"model.visual.blocks.1.mlp.up_proj",
|
| 45 |
+
"model.visual.blocks.1.mlp.down_proj",
|
| 46 |
+
"model.visual.blocks.2.attn.qkv",
|
| 47 |
+
"model.visual.blocks.2.attn.proj",
|
| 48 |
+
"model.visual.blocks.2.mlp.gate_proj",
|
| 49 |
+
"model.visual.blocks.2.mlp.up_proj",
|
| 50 |
+
"model.visual.blocks.2.mlp.down_proj",
|
| 51 |
+
"model.visual.blocks.3.attn.qkv",
|
| 52 |
+
"model.visual.blocks.3.attn.proj",
|
| 53 |
+
"model.visual.blocks.3.mlp.gate_proj",
|
| 54 |
+
"model.visual.blocks.3.mlp.up_proj",
|
| 55 |
+
"model.visual.blocks.3.mlp.down_proj",
|
| 56 |
+
"model.visual.blocks.4.attn.qkv",
|
| 57 |
+
"model.visual.blocks.4.attn.proj",
|
| 58 |
+
"model.visual.blocks.4.mlp.gate_proj",
|
| 59 |
+
"model.visual.blocks.4.mlp.up_proj",
|
| 60 |
+
"model.visual.blocks.4.mlp.down_proj",
|
| 61 |
+
"model.visual.blocks.5.attn.qkv",
|
| 62 |
+
"model.visual.blocks.5.attn.proj",
|
| 63 |
+
"model.visual.blocks.5.mlp.gate_proj",
|
| 64 |
+
"model.visual.blocks.5.mlp.up_proj",
|
| 65 |
+
"model.visual.blocks.5.mlp.down_proj",
|
| 66 |
+
"model.visual.blocks.6.attn.qkv",
|
| 67 |
+
"model.visual.blocks.6.attn.proj",
|
| 68 |
+
"model.visual.blocks.6.mlp.gate_proj",
|
| 69 |
+
"model.visual.blocks.6.mlp.up_proj",
|
| 70 |
+
"model.visual.blocks.6.mlp.down_proj",
|
| 71 |
+
"model.visual.blocks.7.attn.qkv",
|
| 72 |
+
"model.visual.blocks.7.attn.proj",
|
| 73 |
+
"model.visual.blocks.7.mlp.gate_proj",
|
| 74 |
+
"model.visual.blocks.7.mlp.up_proj",
|
| 75 |
+
"model.visual.blocks.7.mlp.down_proj",
|
| 76 |
+
"model.visual.blocks.8.attn.qkv",
|
| 77 |
+
"model.visual.blocks.8.attn.proj",
|
| 78 |
+
"model.visual.blocks.8.mlp.gate_proj",
|
| 79 |
+
"model.visual.blocks.8.mlp.up_proj",
|
| 80 |
+
"model.visual.blocks.8.mlp.down_proj",
|
| 81 |
+
"model.visual.blocks.9.attn.qkv",
|
| 82 |
+
"model.visual.blocks.9.attn.proj",
|
| 83 |
+
"model.visual.blocks.9.mlp.gate_proj",
|
| 84 |
+
"model.visual.blocks.9.mlp.up_proj",
|
| 85 |
+
"model.visual.blocks.9.mlp.down_proj",
|
| 86 |
+
"model.visual.blocks.10.attn.qkv",
|
| 87 |
+
"model.visual.blocks.10.attn.proj",
|
| 88 |
+
"model.visual.blocks.10.mlp.gate_proj",
|
| 89 |
+
"model.visual.blocks.10.mlp.up_proj",
|
| 90 |
+
"model.visual.blocks.10.mlp.down_proj",
|
| 91 |
+
"model.visual.blocks.11.attn.qkv",
|
| 92 |
+
"model.visual.blocks.11.attn.proj",
|
| 93 |
+
"model.visual.blocks.11.mlp.gate_proj",
|
| 94 |
+
"model.visual.blocks.11.mlp.up_proj",
|
| 95 |
+
"model.visual.blocks.11.mlp.down_proj",
|
| 96 |
+
"model.visual.blocks.12.attn.qkv",
|
| 97 |
+
"model.visual.blocks.12.attn.proj",
|
| 98 |
+
"model.visual.blocks.12.mlp.gate_proj",
|
| 99 |
+
"model.visual.blocks.12.mlp.up_proj",
|
| 100 |
+
"model.visual.blocks.12.mlp.down_proj",
|
| 101 |
+
"model.visual.blocks.13.attn.qkv",
|
| 102 |
+
"model.visual.blocks.13.attn.proj",
|
| 103 |
+
"model.visual.blocks.13.mlp.gate_proj",
|
| 104 |
+
"model.visual.blocks.13.mlp.up_proj",
|
| 105 |
+
"model.visual.blocks.13.mlp.down_proj",
|
| 106 |
+
"model.visual.blocks.14.attn.qkv",
|
| 107 |
+
"model.visual.blocks.14.attn.proj",
|
| 108 |
+
"model.visual.blocks.14.mlp.gate_proj",
|
| 109 |
+
"model.visual.blocks.14.mlp.up_proj",
|
| 110 |
+
"model.visual.blocks.14.mlp.down_proj",
|
| 111 |
+
"model.visual.blocks.15.attn.qkv",
|
| 112 |
+
"model.visual.blocks.15.attn.proj",
|
| 113 |
+
"model.visual.blocks.15.mlp.gate_proj",
|
| 114 |
+
"model.visual.blocks.15.mlp.up_proj",
|
| 115 |
+
"model.visual.blocks.15.mlp.down_proj",
|
| 116 |
+
"model.visual.blocks.16.attn.qkv",
|
| 117 |
+
"model.visual.blocks.16.attn.proj",
|
| 118 |
+
"model.visual.blocks.16.mlp.gate_proj",
|
| 119 |
+
"model.visual.blocks.16.mlp.up_proj",
|
| 120 |
+
"model.visual.blocks.16.mlp.down_proj",
|
| 121 |
+
"model.visual.blocks.17.attn.qkv",
|
| 122 |
+
"model.visual.blocks.17.attn.proj",
|
| 123 |
+
"model.visual.blocks.17.mlp.gate_proj",
|
| 124 |
+
"model.visual.blocks.17.mlp.up_proj",
|
| 125 |
+
"model.visual.blocks.17.mlp.down_proj",
|
| 126 |
+
"model.visual.blocks.18.attn.qkv",
|
| 127 |
+
"model.visual.blocks.18.attn.proj",
|
| 128 |
+
"model.visual.blocks.18.mlp.gate_proj",
|
| 129 |
+
"model.visual.blocks.18.mlp.up_proj",
|
| 130 |
+
"model.visual.blocks.18.mlp.down_proj",
|
| 131 |
+
"model.visual.blocks.19.attn.qkv",
|
| 132 |
+
"model.visual.blocks.19.attn.proj",
|
| 133 |
+
"model.visual.blocks.19.mlp.gate_proj",
|
| 134 |
+
"model.visual.blocks.19.mlp.up_proj",
|
| 135 |
+
"model.visual.blocks.19.mlp.down_proj",
|
| 136 |
+
"model.visual.blocks.20.attn.qkv",
|
| 137 |
+
"model.visual.blocks.20.attn.proj",
|
| 138 |
+
"model.visual.blocks.20.mlp.gate_proj",
|
| 139 |
+
"model.visual.blocks.20.mlp.up_proj",
|
| 140 |
+
"model.visual.blocks.20.mlp.down_proj",
|
| 141 |
+
"model.visual.blocks.21.attn.qkv",
|
| 142 |
+
"model.visual.blocks.21.attn.proj",
|
| 143 |
+
"model.visual.blocks.21.mlp.gate_proj",
|
| 144 |
+
"model.visual.blocks.21.mlp.up_proj",
|
| 145 |
+
"model.visual.blocks.21.mlp.down_proj",
|
| 146 |
+
"model.visual.blocks.22.attn.qkv",
|
| 147 |
+
"model.visual.blocks.22.attn.proj",
|
| 148 |
+
"model.visual.blocks.22.mlp.gate_proj",
|
| 149 |
+
"model.visual.blocks.22.mlp.up_proj",
|
| 150 |
+
"model.visual.blocks.22.mlp.down_proj",
|
| 151 |
+
"model.visual.blocks.23.attn.qkv",
|
| 152 |
+
"model.visual.blocks.23.attn.proj",
|
| 153 |
+
"model.visual.blocks.23.mlp.gate_proj",
|
| 154 |
+
"model.visual.blocks.23.mlp.up_proj",
|
| 155 |
+
"model.visual.blocks.23.mlp.down_proj",
|
| 156 |
+
"model.visual.merger.proj",
|
| 157 |
+
"model.visual.merger.gate_proj",
|
| 158 |
+
"model.visual.merger.up_proj",
|
| 159 |
+
"model.visual.merger.down_proj",
|
| 160 |
+
"model.language_model.layers.0.self_attn.q_proj",
|
| 161 |
+
"model.language_model.layers.0.self_attn.k_proj",
|
| 162 |
+
"model.language_model.layers.0.self_attn.v_proj",
|
| 163 |
+
"model.language_model.layers.0.self_attn.o_proj",
|
| 164 |
+
"model.language_model.layers.0.mlp.gate_proj",
|
| 165 |
+
"model.language_model.layers.0.mlp.up_proj",
|
| 166 |
+
"model.language_model.layers.0.mlp.down_proj",
|
| 167 |
+
"model.language_model.layers.1.mlp.shared_experts.gate_proj",
|
| 168 |
+
"model.language_model.layers.1.mlp.shared_experts.up_proj",
|
| 169 |
+
"model.language_model.layers.1.mlp.shared_experts.down_proj",
|
| 170 |
+
"model.language_model.layers.2.mlp.shared_experts.gate_proj",
|
| 171 |
+
"model.language_model.layers.2.mlp.shared_experts.up_proj",
|
| 172 |
+
"model.language_model.layers.2.mlp.shared_experts.down_proj",
|
| 173 |
+
"model.language_model.layers.3.mlp.shared_experts.gate_proj",
|
| 174 |
+
"model.language_model.layers.3.mlp.shared_experts.up_proj",
|
| 175 |
+
"model.language_model.layers.3.mlp.shared_experts.down_proj",
|
| 176 |
+
"model.language_model.layers.4.mlp.shared_experts.gate_proj",
|
| 177 |
+
"model.language_model.layers.4.mlp.shared_experts.up_proj",
|
| 178 |
+
"model.language_model.layers.4.mlp.shared_experts.down_proj",
|
| 179 |
+
"model.language_model.layers.5.mlp.shared_experts.gate_proj",
|
| 180 |
+
"model.language_model.layers.5.mlp.shared_experts.up_proj",
|
| 181 |
+
"model.language_model.layers.5.mlp.shared_experts.down_proj",
|
| 182 |
+
"model.language_model.layers.6.mlp.shared_experts.gate_proj",
|
| 183 |
+
"model.language_model.layers.6.mlp.shared_experts.up_proj",
|
| 184 |
+
"model.language_model.layers.6.mlp.shared_experts.down_proj",
|
| 185 |
+
"model.language_model.layers.7.mlp.shared_experts.gate_proj",
|
| 186 |
+
"model.language_model.layers.7.mlp.shared_experts.up_proj",
|
| 187 |
+
"model.language_model.layers.7.mlp.shared_experts.down_proj",
|
| 188 |
+
"model.language_model.layers.8.mlp.shared_experts.gate_proj",
|
| 189 |
+
"model.language_model.layers.8.mlp.shared_experts.up_proj",
|
| 190 |
+
"model.language_model.layers.8.mlp.shared_experts.down_proj",
|
| 191 |
+
"model.language_model.layers.9.mlp.shared_experts.gate_proj",
|
| 192 |
+
"model.language_model.layers.9.mlp.shared_experts.up_proj",
|
| 193 |
+
"model.language_model.layers.9.mlp.shared_experts.down_proj",
|
| 194 |
+
"model.language_model.layers.10.mlp.shared_experts.gate_proj",
|
| 195 |
+
"model.language_model.layers.10.mlp.shared_experts.up_proj",
|
| 196 |
+
"model.language_model.layers.10.mlp.shared_experts.down_proj",
|
| 197 |
+
"model.language_model.layers.11.mlp.shared_experts.gate_proj",
|
| 198 |
+
"model.language_model.layers.11.mlp.shared_experts.up_proj",
|
| 199 |
+
"model.language_model.layers.11.mlp.shared_experts.down_proj",
|
| 200 |
+
"model.language_model.layers.12.mlp.shared_experts.gate_proj",
|
| 201 |
+
"model.language_model.layers.12.mlp.shared_experts.up_proj",
|
| 202 |
+
"model.language_model.layers.12.mlp.shared_experts.down_proj",
|
| 203 |
+
"model.language_model.layers.13.mlp.shared_experts.gate_proj",
|
| 204 |
+
"model.language_model.layers.13.mlp.shared_experts.up_proj",
|
| 205 |
+
"model.language_model.layers.13.mlp.shared_experts.down_proj",
|
| 206 |
+
"model.language_model.layers.14.mlp.shared_experts.gate_proj",
|
| 207 |
+
"model.language_model.layers.14.mlp.shared_experts.up_proj",
|
| 208 |
+
"model.language_model.layers.14.mlp.shared_experts.down_proj",
|
| 209 |
+
"model.language_model.layers.15.mlp.shared_experts.gate_proj",
|
| 210 |
+
"model.language_model.layers.15.mlp.shared_experts.up_proj",
|
| 211 |
+
"model.language_model.layers.15.mlp.shared_experts.down_proj",
|
| 212 |
+
"model.language_model.layers.16.mlp.shared_experts.gate_proj",
|
| 213 |
+
"model.language_model.layers.16.mlp.shared_experts.up_proj",
|
| 214 |
+
"model.language_model.layers.16.mlp.shared_experts.down_proj",
|
| 215 |
+
"model.language_model.layers.17.mlp.shared_experts.gate_proj",
|
| 216 |
+
"model.language_model.layers.17.mlp.shared_experts.up_proj",
|
| 217 |
+
"model.language_model.layers.17.mlp.shared_experts.down_proj",
|
| 218 |
+
"model.language_model.layers.18.mlp.shared_experts.gate_proj",
|
| 219 |
+
"model.language_model.layers.18.mlp.shared_experts.up_proj",
|
| 220 |
+
"model.language_model.layers.18.mlp.shared_experts.down_proj",
|
| 221 |
+
"model.language_model.layers.19.mlp.shared_experts.gate_proj",
|
| 222 |
+
"model.language_model.layers.19.mlp.shared_experts.up_proj",
|
| 223 |
+
"model.language_model.layers.19.mlp.shared_experts.down_proj",
|
| 224 |
+
"model.language_model.layers.20.mlp.shared_experts.gate_proj",
|
| 225 |
+
"model.language_model.layers.20.mlp.shared_experts.up_proj",
|
| 226 |
+
"model.language_model.layers.20.mlp.shared_experts.down_proj",
|
| 227 |
+
"model.language_model.layers.21.mlp.shared_experts.gate_proj",
|
| 228 |
+
"model.language_model.layers.21.mlp.shared_experts.up_proj",
|
| 229 |
+
"model.language_model.layers.21.mlp.shared_experts.down_proj",
|
| 230 |
+
"model.language_model.layers.22.mlp.shared_experts.gate_proj",
|
| 231 |
+
"model.language_model.layers.22.mlp.shared_experts.up_proj",
|
| 232 |
+
"model.language_model.layers.22.mlp.shared_experts.down_proj",
|
| 233 |
+
"model.language_model.layers.23.mlp.shared_experts.gate_proj",
|
| 234 |
+
"model.language_model.layers.23.mlp.shared_experts.up_proj",
|
| 235 |
+
"model.language_model.layers.23.mlp.shared_experts.down_proj",
|
| 236 |
+
"model.language_model.layers.24.mlp.shared_experts.gate_proj",
|
| 237 |
+
"model.language_model.layers.24.mlp.shared_experts.up_proj",
|
| 238 |
+
"model.language_model.layers.24.mlp.shared_experts.down_proj",
|
| 239 |
+
"model.language_model.layers.25.mlp.shared_experts.gate_proj",
|
| 240 |
+
"model.language_model.layers.25.mlp.shared_experts.up_proj",
|
| 241 |
+
"model.language_model.layers.25.mlp.shared_experts.down_proj",
|
| 242 |
+
"model.language_model.layers.26.mlp.shared_experts.gate_proj",
|
| 243 |
+
"model.language_model.layers.26.mlp.shared_experts.up_proj",
|
| 244 |
+
"model.language_model.layers.26.mlp.shared_experts.down_proj",
|
| 245 |
+
"model.language_model.layers.27.mlp.shared_experts.gate_proj",
|
| 246 |
+
"model.language_model.layers.27.mlp.shared_experts.up_proj",
|
| 247 |
+
"model.language_model.layers.27.mlp.shared_experts.down_proj",
|
| 248 |
+
"model.language_model.layers.28.mlp.shared_experts.gate_proj",
|
| 249 |
+
"model.language_model.layers.28.mlp.shared_experts.up_proj",
|
| 250 |
+
"model.language_model.layers.28.mlp.shared_experts.down_proj",
|
| 251 |
+
"model.language_model.layers.29.mlp.shared_experts.gate_proj",
|
| 252 |
+
"model.language_model.layers.29.mlp.shared_experts.up_proj",
|
| 253 |
+
"model.language_model.layers.29.mlp.shared_experts.down_proj",
|
| 254 |
+
"model.language_model.layers.30.mlp.shared_experts.gate_proj",
|
| 255 |
+
"model.language_model.layers.30.mlp.shared_experts.up_proj",
|
| 256 |
+
"model.language_model.layers.30.mlp.shared_experts.down_proj",
|
| 257 |
+
"model.language_model.layers.31.mlp.shared_experts.gate_proj",
|
| 258 |
+
"model.language_model.layers.31.mlp.shared_experts.up_proj",
|
| 259 |
+
"model.language_model.layers.31.mlp.shared_experts.down_proj",
|
| 260 |
+
"model.language_model.layers.32.mlp.shared_experts.gate_proj",
|
| 261 |
+
"model.language_model.layers.32.mlp.shared_experts.up_proj",
|
| 262 |
+
"model.language_model.layers.32.mlp.shared_experts.down_proj",
|
| 263 |
+
"model.language_model.layers.33.mlp.shared_experts.gate_proj",
|
| 264 |
+
"model.language_model.layers.33.mlp.shared_experts.up_proj",
|
| 265 |
+
"model.language_model.layers.33.mlp.shared_experts.down_proj",
|
| 266 |
+
"model.language_model.layers.34.mlp.shared_experts.gate_proj",
|
| 267 |
+
"model.language_model.layers.34.mlp.shared_experts.up_proj",
|
| 268 |
+
"model.language_model.layers.34.mlp.shared_experts.down_proj",
|
| 269 |
+
"model.language_model.layers.35.mlp.shared_experts.gate_proj",
|
| 270 |
+
"model.language_model.layers.35.mlp.shared_experts.up_proj",
|
| 271 |
+
"model.language_model.layers.35.mlp.shared_experts.down_proj",
|
| 272 |
+
"model.language_model.layers.36.mlp.shared_experts.gate_proj",
|
| 273 |
+
"model.language_model.layers.36.mlp.shared_experts.up_proj",
|
| 274 |
+
"model.language_model.layers.36.mlp.shared_experts.down_proj",
|
| 275 |
+
"model.language_model.layers.37.mlp.shared_experts.gate_proj",
|
| 276 |
+
"model.language_model.layers.37.mlp.shared_experts.up_proj",
|
| 277 |
+
"model.language_model.layers.37.mlp.shared_experts.down_proj",
|
| 278 |
+
"model.language_model.layers.38.mlp.shared_experts.gate_proj",
|
| 279 |
+
"model.language_model.layers.38.mlp.shared_experts.up_proj",
|
| 280 |
+
"model.language_model.layers.38.mlp.shared_experts.down_proj",
|
| 281 |
+
"model.language_model.layers.39.mlp.shared_experts.gate_proj",
|
| 282 |
+
"model.language_model.layers.39.mlp.shared_experts.up_proj",
|
| 283 |
+
"model.language_model.layers.39.mlp.shared_experts.down_proj",
|
| 284 |
+
"model.language_model.layers.40.mlp.shared_experts.gate_proj",
|
| 285 |
+
"model.language_model.layers.40.mlp.shared_experts.up_proj",
|
| 286 |
+
"model.language_model.layers.40.mlp.shared_experts.down_proj",
|
| 287 |
+
"model.language_model.layers.41.mlp.shared_experts.gate_proj",
|
| 288 |
+
"model.language_model.layers.41.mlp.shared_experts.up_proj",
|
| 289 |
+
"model.language_model.layers.41.mlp.shared_experts.down_proj",
|
| 290 |
+
"model.language_model.layers.42.mlp.shared_experts.gate_proj",
|
| 291 |
+
"model.language_model.layers.42.mlp.shared_experts.up_proj",
|
| 292 |
+
"model.language_model.layers.42.mlp.shared_experts.down_proj",
|
| 293 |
+
"model.language_model.layers.43.mlp.shared_experts.gate_proj",
|
| 294 |
+
"model.language_model.layers.43.mlp.shared_experts.up_proj",
|
| 295 |
+
"model.language_model.layers.43.mlp.shared_experts.down_proj",
|
| 296 |
+
"model.language_model.layers.44.mlp.shared_experts.gate_proj",
|
| 297 |
+
"model.language_model.layers.44.mlp.shared_experts.up_proj",
|
| 298 |
+
"model.language_model.layers.44.mlp.shared_experts.down_proj",
|
| 299 |
+
"model.language_model.layers.45.mlp.shared_experts.gate_proj",
|
| 300 |
+
"model.language_model.layers.45.mlp.shared_experts.up_proj",
|
| 301 |
+
"model.language_model.layers.45.mlp.shared_experts.down_proj",
|
| 302 |
+
"lm_head"
|
| 303 |
+
],
|
| 304 |
+
"kv_cache_scheme": null,
|
| 305 |
+
"quant_method": "compressed-tensors",
|
| 306 |
+
"quantization_status": "compressed",
|
| 307 |
+
"sparsity_config": {},
|
| 308 |
+
"transform_config": {},
|
| 309 |
+
"version": "0.12.3.a20251203"
|
| 310 |
+
},
|
| 311 |
+
"text_config": {
|
| 312 |
+
"attention_bias": true,
|
| 313 |
+
"attention_dropout": 0.0,
|
| 314 |
+
"dtype": "bfloat16",
|
| 315 |
+
"eos_token_id": [
|
| 316 |
+
151329,
|
| 317 |
+
151336,
|
| 318 |
+
151338
|
| 319 |
+
],
|
| 320 |
+
"first_k_dense_replace": 1,
|
| 321 |
+
"head_dim": 128,
|
| 322 |
+
"hidden_act": "silu",
|
| 323 |
+
"hidden_size": 4096,
|
| 324 |
+
"initializer_range": 0.02,
|
| 325 |
+
"intermediate_size": 10944,
|
| 326 |
+
"max_position_embeddings": 131072,
|
| 327 |
+
"model_type": "Glm4vMoe_text",
|
| 328 |
+
"moe_intermediate_size": 1408,
|
| 329 |
+
"n_group": 1,
|
| 330 |
+
"n_routed_experts": 128,
|
| 331 |
+
"n_shared_experts": 1,
|
| 332 |
+
"norm_topk_prob": true,
|
| 333 |
+
"num_attention_heads": 96,
|
| 334 |
+
"num_experts_per_tok": 8,
|
| 335 |
+
"num_hidden_layers": 46,
|
| 336 |
+
"num_key_value_heads": 8,
|
| 337 |
+
"num_nextn_predict_layers": 0,
|
| 338 |
+
"pad_token_id": 151329,
|
| 339 |
+
"partial_rotary_factor": 0.5,
|
| 340 |
+
"qk_layernorm": false,
|
| 341 |
+
"rms_norm_eps": 1e-05,
|
| 342 |
+
"rope_parameters": {
|
| 343 |
+
"mrope_section": [
|
| 344 |
+
8,
|
| 345 |
+
12,
|
| 346 |
+
12
|
| 347 |
+
],
|
| 348 |
+
"partial_rotary_factor": 0.5,
|
| 349 |
+
"rope_theta": 500000,
|
| 350 |
+
"rope_type": "default"
|
| 351 |
+
},
|
| 352 |
+
"routed_scaling_factor": 1.0,
|
| 353 |
+
"topk_group": 1,
|
| 354 |
+
"use_cache": true,
|
| 355 |
+
"use_qk_norm": false,
|
| 356 |
+
"vocab_size": 151552
|
| 357 |
+
},
|
| 358 |
+
"tie_word_embeddings": false,
|
| 359 |
+
"transformers_version": "4.57.3",
|
| 360 |
+
"video_end_token_id": 151342,
|
| 361 |
+
"video_start_token_id": 151341,
|
| 362 |
+
"video_token_id": 151364,
|
| 363 |
+
"vision_config": {
|
| 364 |
+
"attention_bias": false,
|
| 365 |
+
"attention_dropout": 0.0,
|
| 366 |
+
"depth": 24,
|
| 367 |
+
"hidden_act": "silu",
|
| 368 |
+
"hidden_dropout_prob": 0.0,
|
| 369 |
+
"hidden_size": 1536,
|
| 370 |
+
"image_size": 336,
|
| 371 |
+
"in_channels": 3,
|
| 372 |
+
"initializer_range": 0.02,
|
| 373 |
+
"intermediate_size": 10944,
|
| 374 |
+
"model_type": "glm4v_moe",
|
| 375 |
+
"num_heads": 12,
|
| 376 |
+
"out_hidden_size": 4096,
|
| 377 |
+
"patch_size": 14,
|
| 378 |
+
"rms_norm_eps": 1e-05,
|
| 379 |
+
"spatial_merge_size": 2,
|
| 380 |
+
"temporal_patch_size": 2
|
| 381 |
+
}
|
| 382 |
+
}
|
README.md
ADDED
|
@@ -0,0 +1,156 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- zh
|
| 4 |
+
- en
|
| 5 |
+
library_name: transformers
|
| 6 |
+
license: mit
|
| 7 |
+
pipeline_tag: image-text-to-text
|
| 8 |
+
base_model: zai-org/GLM-4.6V
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# GLM-4.6V
|
| 12 |
+
|
| 13 |
+
<div align="center">
|
| 14 |
+
<img src=https://raw.githubusercontent.com/zai-org/GLM-V/refs/heads/main/resources/logo.svg width="40%"/>
|
| 15 |
+
</div>
|
| 16 |
+
|
| 17 |
+
This model is part of the GLM-V family of models, introduced in the paper [GLM-4.1V-Thinking and GLM-4.5V: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning](https://huggingface.co/papers/2507.01006).
|
| 18 |
+
|
| 19 |
+
- **GLM-4.6V Blog**: [https://z.ai/blog/glm-4.6v](https://z.ai/blog/glm-4.6v)
|
| 20 |
+
- **Paper**: [https://huggingface.co/papers/2507.01006](https://huggingface.co/papers/2507.01006)
|
| 21 |
+
- **GitHub Repository**: [https://github.com/zai-org/GLM-V](https://github.com/zai-org/GLM-V)
|
| 22 |
+
- **Online Demo**: [https://chat.z.ai/](https://chat.z.ai/)
|
| 23 |
+
- **API Access**: [Z.ai Open Platform](https://docs.z.ai/guides/vlm/glm-4.6v)
|
| 24 |
+
- **Desktop Assistant App**: [https://huggingface.co/spaces/zai-org/GLM-4.5V-Demo-App](https://huggingface.co/spaces/zai-org/GLM-4.5V-Demo-App)
|
| 25 |
+
|
| 26 |
+
## Introduction
|
| 27 |
+
|
| 28 |
+
GLM-4.6V series model includes two versions: GLM-4.6V (106B), a foundation model designed for cloud and high-performance
|
| 29 |
+
cluster scenarios,
|
| 30 |
+
and GLM-4.6V-Flash (9B), a lightweight model optimized for local deployment and low-latency applications.
|
| 31 |
+
GLM-4.6V scales its context window to 128k tokens in training,
|
| 32 |
+
and achieves SoTA performance in visual understanding among models of similar parameter scales.
|
| 33 |
+
Crucially, we integrate native Function Calling capabilities for the first time.
|
| 34 |
+
This effectively bridges the gap between "visual perception" and "executable action"
|
| 35 |
+
providing a unified technical foundation for multimodal agents in real-world business scenarios.
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
Beyond achieves SoTA performance across major multimodal benchmarks at comparable model scales. GLM-4.6V introduces
|
| 40 |
+
several key features:
|
| 41 |
+
|
| 42 |
+
- **Native Multimodal Function Calling**
|
| 43 |
+
Enables native vision-driven tool use. Images, screenshots, and document pages can be passed directly as tool inputs without text conversion, while visual outputs (charts, search images, rendered pages) are interpreted and integrated into the reasoning chain. This closes the loop from perception to understanding to execution.
|
| 44 |
+
|
| 45 |
+
- **Interleaved Image-Text Content Generation**
|
| 46 |
+
Supports high-quality mixed media creation from complex multimodal inputs. GLM-4.6V takes a multimodal context—spanning documents, user inputs, and tool-retrieved images—and synthesizes coherent, interleaved image-text content tailored to the task. During generation it can actively call search and retrieval tools to gather and curate additional text and visuals, producing rich, visually grounded content.
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
- **Multimodal Document Understanding**
|
| 50 |
+
GLM-4.6V can process up to 128K tokens of multi-document or long-document input, directly interpreting richly formatted pages as images. It understands text, layout, charts, tables, and figures jointly, enabling accurate comprehension of complex, image-heavy documents without requiring prior conversion to plain text.
|
| 51 |
+
|
| 52 |
+
- **Frontend Replication & Visual Editing**
|
| 53 |
+
Reconstructs pixel-accurate HTML/CSS from UI screenshots and supports natural-language-driven edits. It detects layout, components, and styles visually, generates clean code, and applies iterative visual modifications through simple user instructions.
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
**This Hugging Face repository hosts the `GLM-4.6V` model, part of the `GLM-V` series.**
|
| 57 |
+
|
| 58 |
+
## Usage
|
| 59 |
+
|
| 60 |
+
### Environment Installation
|
| 61 |
+
|
| 62 |
+
For `SGLang`:
|
| 63 |
+
|
| 64 |
+
```bash
|
| 65 |
+
pip install sglang>=0.5.6post1
|
| 66 |
+
pip install transformers>=5.0.0rc0
|
| 67 |
+
```
|
| 68 |
+
|
| 69 |
+
For `vLLM`:
|
| 70 |
+
|
| 71 |
+
```bash
|
| 72 |
+
pip install vllm>=0.12.0
|
| 73 |
+
pip install transformers>=5.0.0rc0
|
| 74 |
+
```
|
| 75 |
+
|
| 76 |
+
### Quick Start with Transformers
|
| 77 |
+
|
| 78 |
+
```python
|
| 79 |
+
from transformers import AutoProcessor, Glm4vMoeForConditionalGeneration
|
| 80 |
+
import torch
|
| 81 |
+
|
| 82 |
+
MODEL_PATH = "zai-org/GLM-4.6V"
|
| 83 |
+
messages = [
|
| 84 |
+
{
|
| 85 |
+
"role": "user",
|
| 86 |
+
"content": [
|
| 87 |
+
{
|
| 88 |
+
"type": "image",
|
| 89 |
+
"url": "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png"
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"type": "text",
|
| 93 |
+
"text": "describe this image"
|
| 94 |
+
}
|
| 95 |
+
],
|
| 96 |
+
}
|
| 97 |
+
]
|
| 98 |
+
processor = AutoProcessor.from_pretrained(MODEL_PATH)
|
| 99 |
+
model = Glm4vMoeForConditionalGeneration.from_pretrained(
|
| 100 |
+
pretrained_model_name_or_path=MODEL_PATH,
|
| 101 |
+
torch_dtype="auto",
|
| 102 |
+
device_map="auto",
|
| 103 |
+
)
|
| 104 |
+
inputs = processor.apply_chat_template(
|
| 105 |
+
messages,
|
| 106 |
+
tokenize=True,
|
| 107 |
+
add_generation_prompt=True,
|
| 108 |
+
return_dict=True,
|
| 109 |
+
return_tensors="pt"
|
| 110 |
+
).to(model.device)
|
| 111 |
+
inputs.pop("token_type_ids", None)
|
| 112 |
+
generated_ids = model.generate(**inputs, max_new_tokens=8192)
|
| 113 |
+
output_text = processor.decode(generated_ids[0][inputs["input_ids"].shape[1]:], skip_special_tokens=False)
|
| 114 |
+
print(output_text)
|
| 115 |
+
```
|
| 116 |
+
|
| 117 |
+
## Evaluation Settings
|
| 118 |
+
|
| 119 |
+
We primarily use vLLM as the backend for model inference. For faster and more reliable performance on video tasks, we employ SGLang. To reproduce our leaderboard results, we recommend the following decoding parameters:
|
| 120 |
+
|
| 121 |
+
+ top_p: 0.6
|
| 122 |
+
+ top_k: 2
|
| 123 |
+
+ temperature: 0.8
|
| 124 |
+
+ repetition_penalty: 1.1
|
| 125 |
+
+ max_generate_tokens: 16K
|
| 126 |
+
|
| 127 |
+
For more usage details, please refer to Our [Github](https://github.com/zai-org/GLM-V).
|
| 128 |
+
|
| 129 |
+
## Fixed and Remaining Issues
|
| 130 |
+
|
| 131 |
+
Since the open-sourcing of GLM-4.1V, we have received extensive feedback from the community and are well aware that the model still has many shortcomings. In subsequent iterations, we attempted to address several common issues — such as repetitive thinking outputs and formatting errors — which have been mitigated to some extent in this new version.
|
| 132 |
+
|
| 133 |
+
However, the model still has several limitations and issues that we will fix as soon as possible:
|
| 134 |
+
|
| 135 |
+
1. Pure text QA capabilities still have significant room for improvement. In this development cycle, our primary focus was on visual multimodal scenarios, and we will enhance pure text abilities in upcoming updates.
|
| 136 |
+
2. The model may still overthink or even repeat itself in certain cases, especially when dealing with complex prompts.
|
| 137 |
+
3. In some situations, the model may restate the answer again at the end.
|
| 138 |
+
4. There remain certain perception limitations, such as counting accuracy and identifying specific individuals, which still require improvement.
|
| 139 |
+
|
| 140 |
+
Thank you for your patience and understanding. We also welcome feedback and suggestions in the issue section — we will respond and improve as much as we can!
|
| 141 |
+
|
| 142 |
+
## Citation
|
| 143 |
+
|
| 144 |
+
If you use this model, please cite the following paper:
|
| 145 |
+
|
| 146 |
+
```bibtex
|
| 147 |
+
@misc{vteam2025glm45vglm41vthinkingversatilemultimodal,
|
| 148 |
+
title={GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning},
|
| 149 |
+
author={V Team and Wenyi Hong and Wenmeng Yu and Xiaotao Gu and Guo Wang and Guobing Gan and Haomiao Tang and Jiale Cheng and Ji Qi and Junhui Ji and Lihang Pan and Shuaiqi Duan and Weihan Wang and Yan Wang and Yean Cheng and Zehai He and Zhe Su and Zhen Yang and Ziyang Pan and Aohan Zeng and Baoxu Wang and Bin Chen and Boyan Shi and Changyu Pang and Chenhui Zhang and Da Yin and Fan Yang and Guoqing Chen and Jiazheng Xu and Jiale Zhu and Jiali Chen and Jing Chen and Jinhao Chen and Jinghao Lin and Jinjiang Wang and Junjie Chen and Leqi Lei and Letian Gong and Leyi Pan and Mingdao Liu and Mingde Xu and Mingzhi Zhang and Qinkai Zheng and Sheng Yang and Shi Zhong and Shiyu Huang and Shuyuan Zhao and Siyan Xue and Shangqin Tu and Shengbiao Meng and Tianshu Zhang and Tianwei Luo and Tianxiang Hao and Tianyu Tong and Wenkai Li and Wei Jia and Xiao Liu and Xiaohan Zhang and Xin Lyu and Xinyue Fan and Xuancheng Huang and Yanling Wang and Yadong Xue and Yanfeng Wang and Yanzi Wang and Yifan An and Yifan Du and Yiming Shi and Yiheng Huang and Yilin Niu and Yuan Wang and Yuanchang Yue and Yuchen Li and Yutao Zhang and Yuting Wang and Yu Wang and Yuxuan Zhang and Zhao Xue and Zhenyu Hou and Zhengxiao Du and Zihan Wang and Peng Zhang and Debing Liu and Bin Xu and Juanzi Li and Minlie Huang and Yuxiao Dong and Jie Tang},
|
| 150 |
+
year={2025},
|
| 151 |
+
eprint={2507.01006},
|
| 152 |
+
archivePrefix={arXiv},
|
| 153 |
+
primaryClass={cs.CV},
|
| 154 |
+
url={https://arxiv.org/abs/2507.01006},
|
| 155 |
+
}
|
| 156 |
+
```
|
_config.json
ADDED
|
@@ -0,0 +1,393 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Glm4vMoeForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "bfloat16",
|
| 6 |
+
"image_end_token_id": 151340,
|
| 7 |
+
"image_start_token_id": 151339,
|
| 8 |
+
"image_token_id": 151363,
|
| 9 |
+
"model_type": "glm4v_moe",
|
| 10 |
+
"quantization_config": {
|
| 11 |
+
"config_groups": {
|
| 12 |
+
"group_0": {
|
| 13 |
+
"format": "pack-quantized",
|
| 14 |
+
"input_activations": null,
|
| 15 |
+
"output_activations": null,
|
| 16 |
+
"targets": [
|
| 17 |
+
"Linear"
|
| 18 |
+
],
|
| 19 |
+
"weights": {
|
| 20 |
+
"actorder": null,
|
| 21 |
+
"block_structure": null,
|
| 22 |
+
"dynamic": false,
|
| 23 |
+
"group_size": 32,
|
| 24 |
+
"num_bits": 8,
|
| 25 |
+
"observer": "mse",
|
| 26 |
+
"observer_kwargs": {},
|
| 27 |
+
"strategy": "group",
|
| 28 |
+
"symmetric": true,
|
| 29 |
+
"type": "int"
|
| 30 |
+
}
|
| 31 |
+
}
|
| 32 |
+
},
|
| 33 |
+
"format": "pack-quantized",
|
| 34 |
+
"global_compression_ratio": null,
|
| 35 |
+
"ignore": [
|
| 36 |
+
"model.visual.blocks.0.attn.qkv",
|
| 37 |
+
"model.visual.blocks.0.attn.proj",
|
| 38 |
+
"model.visual.blocks.0.mlp.gate_proj",
|
| 39 |
+
"model.visual.blocks.0.mlp.up_proj",
|
| 40 |
+
"model.visual.blocks.0.mlp.down_proj",
|
| 41 |
+
"model.visual.blocks.1.attn.qkv",
|
| 42 |
+
"model.visual.blocks.1.attn.proj",
|
| 43 |
+
"model.visual.blocks.1.mlp.gate_proj",
|
| 44 |
+
"model.visual.blocks.1.mlp.up_proj",
|
| 45 |
+
"model.visual.blocks.1.mlp.down_proj",
|
| 46 |
+
"model.visual.blocks.2.attn.qkv",
|
| 47 |
+
"model.visual.blocks.2.attn.proj",
|
| 48 |
+
"model.visual.blocks.2.mlp.gate_proj",
|
| 49 |
+
"model.visual.blocks.2.mlp.up_proj",
|
| 50 |
+
"model.visual.blocks.2.mlp.down_proj",
|
| 51 |
+
"model.visual.blocks.3.attn.qkv",
|
| 52 |
+
"model.visual.blocks.3.attn.proj",
|
| 53 |
+
"model.visual.blocks.3.mlp.gate_proj",
|
| 54 |
+
"model.visual.blocks.3.mlp.up_proj",
|
| 55 |
+
"model.visual.blocks.3.mlp.down_proj",
|
| 56 |
+
"model.visual.blocks.4.attn.qkv",
|
| 57 |
+
"model.visual.blocks.4.attn.proj",
|
| 58 |
+
"model.visual.blocks.4.mlp.gate_proj",
|
| 59 |
+
"model.visual.blocks.4.mlp.up_proj",
|
| 60 |
+
"model.visual.blocks.4.mlp.down_proj",
|
| 61 |
+
"model.visual.blocks.5.attn.qkv",
|
| 62 |
+
"model.visual.blocks.5.attn.proj",
|
| 63 |
+
"model.visual.blocks.5.mlp.gate_proj",
|
| 64 |
+
"model.visual.blocks.5.mlp.up_proj",
|
| 65 |
+
"model.visual.blocks.5.mlp.down_proj",
|
| 66 |
+
"model.visual.blocks.6.attn.qkv",
|
| 67 |
+
"model.visual.blocks.6.attn.proj",
|
| 68 |
+
"model.visual.blocks.6.mlp.gate_proj",
|
| 69 |
+
"model.visual.blocks.6.mlp.up_proj",
|
| 70 |
+
"model.visual.blocks.6.mlp.down_proj",
|
| 71 |
+
"model.visual.blocks.7.attn.qkv",
|
| 72 |
+
"model.visual.blocks.7.attn.proj",
|
| 73 |
+
"model.visual.blocks.7.mlp.gate_proj",
|
| 74 |
+
"model.visual.blocks.7.mlp.up_proj",
|
| 75 |
+
"model.visual.blocks.7.mlp.down_proj",
|
| 76 |
+
"model.visual.blocks.8.attn.qkv",
|
| 77 |
+
"model.visual.blocks.8.attn.proj",
|
| 78 |
+
"model.visual.blocks.8.mlp.gate_proj",
|
| 79 |
+
"model.visual.blocks.8.mlp.up_proj",
|
| 80 |
+
"model.visual.blocks.8.mlp.down_proj",
|
| 81 |
+
"model.visual.blocks.9.attn.qkv",
|
| 82 |
+
"model.visual.blocks.9.attn.proj",
|
| 83 |
+
"model.visual.blocks.9.mlp.gate_proj",
|
| 84 |
+
"model.visual.blocks.9.mlp.up_proj",
|
| 85 |
+
"model.visual.blocks.9.mlp.down_proj",
|
| 86 |
+
"model.visual.blocks.10.attn.qkv",
|
| 87 |
+
"model.visual.blocks.10.attn.proj",
|
| 88 |
+
"model.visual.blocks.10.mlp.gate_proj",
|
| 89 |
+
"model.visual.blocks.10.mlp.up_proj",
|
| 90 |
+
"model.visual.blocks.10.mlp.down_proj",
|
| 91 |
+
"model.visual.blocks.11.attn.qkv",
|
| 92 |
+
"model.visual.blocks.11.attn.proj",
|
| 93 |
+
"model.visual.blocks.11.mlp.gate_proj",
|
| 94 |
+
"model.visual.blocks.11.mlp.up_proj",
|
| 95 |
+
"model.visual.blocks.11.mlp.down_proj",
|
| 96 |
+
"model.visual.blocks.12.attn.qkv",
|
| 97 |
+
"model.visual.blocks.12.attn.proj",
|
| 98 |
+
"model.visual.blocks.12.mlp.gate_proj",
|
| 99 |
+
"model.visual.blocks.12.mlp.up_proj",
|
| 100 |
+
"model.visual.blocks.12.mlp.down_proj",
|
| 101 |
+
"model.visual.blocks.13.attn.qkv",
|
| 102 |
+
"model.visual.blocks.13.attn.proj",
|
| 103 |
+
"model.visual.blocks.13.mlp.gate_proj",
|
| 104 |
+
"model.visual.blocks.13.mlp.up_proj",
|
| 105 |
+
"model.visual.blocks.13.mlp.down_proj",
|
| 106 |
+
"model.visual.blocks.14.attn.qkv",
|
| 107 |
+
"model.visual.blocks.14.attn.proj",
|
| 108 |
+
"model.visual.blocks.14.mlp.gate_proj",
|
| 109 |
+
"model.visual.blocks.14.mlp.up_proj",
|
| 110 |
+
"model.visual.blocks.14.mlp.down_proj",
|
| 111 |
+
"model.visual.blocks.15.attn.qkv",
|
| 112 |
+
"model.visual.blocks.15.attn.proj",
|
| 113 |
+
"model.visual.blocks.15.mlp.gate_proj",
|
| 114 |
+
"model.visual.blocks.15.mlp.up_proj",
|
| 115 |
+
"model.visual.blocks.15.mlp.down_proj",
|
| 116 |
+
"model.visual.blocks.16.attn.qkv",
|
| 117 |
+
"model.visual.blocks.16.attn.proj",
|
| 118 |
+
"model.visual.blocks.16.mlp.gate_proj",
|
| 119 |
+
"model.visual.blocks.16.mlp.up_proj",
|
| 120 |
+
"model.visual.blocks.16.mlp.down_proj",
|
| 121 |
+
"model.visual.blocks.17.attn.qkv",
|
| 122 |
+
"model.visual.blocks.17.attn.proj",
|
| 123 |
+
"model.visual.blocks.17.mlp.gate_proj",
|
| 124 |
+
"model.visual.blocks.17.mlp.up_proj",
|
| 125 |
+
"model.visual.blocks.17.mlp.down_proj",
|
| 126 |
+
"model.visual.blocks.18.attn.qkv",
|
| 127 |
+
"model.visual.blocks.18.attn.proj",
|
| 128 |
+
"model.visual.blocks.18.mlp.gate_proj",
|
| 129 |
+
"model.visual.blocks.18.mlp.up_proj",
|
| 130 |
+
"model.visual.blocks.18.mlp.down_proj",
|
| 131 |
+
"model.visual.blocks.19.attn.qkv",
|
| 132 |
+
"model.visual.blocks.19.attn.proj",
|
| 133 |
+
"model.visual.blocks.19.mlp.gate_proj",
|
| 134 |
+
"model.visual.blocks.19.mlp.up_proj",
|
| 135 |
+
"model.visual.blocks.19.mlp.down_proj",
|
| 136 |
+
"model.visual.blocks.20.attn.qkv",
|
| 137 |
+
"model.visual.blocks.20.attn.proj",
|
| 138 |
+
"model.visual.blocks.20.mlp.gate_proj",
|
| 139 |
+
"model.visual.blocks.20.mlp.up_proj",
|
| 140 |
+
"model.visual.blocks.20.mlp.down_proj",
|
| 141 |
+
"model.visual.blocks.21.attn.qkv",
|
| 142 |
+
"model.visual.blocks.21.attn.proj",
|
| 143 |
+
"model.visual.blocks.21.mlp.gate_proj",
|
| 144 |
+
"model.visual.blocks.21.mlp.up_proj",
|
| 145 |
+
"model.visual.blocks.21.mlp.down_proj",
|
| 146 |
+
"model.visual.blocks.22.attn.qkv",
|
| 147 |
+
"model.visual.blocks.22.attn.proj",
|
| 148 |
+
"model.visual.blocks.22.mlp.gate_proj",
|
| 149 |
+
"model.visual.blocks.22.mlp.up_proj",
|
| 150 |
+
"model.visual.blocks.22.mlp.down_proj",
|
| 151 |
+
"model.visual.blocks.23.attn.qkv",
|
| 152 |
+
"model.visual.blocks.23.attn.proj",
|
| 153 |
+
"model.visual.blocks.23.mlp.gate_proj",
|
| 154 |
+
"model.visual.blocks.23.mlp.up_proj",
|
| 155 |
+
"model.visual.blocks.23.mlp.down_proj",
|
| 156 |
+
"model.visual.merger.proj",
|
| 157 |
+
"model.visual.merger.gate_proj",
|
| 158 |
+
"model.visual.merger.up_proj",
|
| 159 |
+
"model.visual.merger.down_proj",
|
| 160 |
+
"model.language_model.layers.0.self_attn.q_proj",
|
| 161 |
+
"model.language_model.layers.0.self_attn.k_proj",
|
| 162 |
+
"model.language_model.layers.0.self_attn.v_proj",
|
| 163 |
+
"model.language_model.layers.0.self_attn.o_proj",
|
| 164 |
+
"model.language_model.layers.0.mlp.gate_proj",
|
| 165 |
+
"model.language_model.layers.0.mlp.up_proj",
|
| 166 |
+
"model.language_model.layers.0.mlp.down_proj",
|
| 167 |
+
"model.language_model.layers.1.mlp.shared_experts.gate_proj",
|
| 168 |
+
"model.language_model.layers.1.mlp.shared_experts.up_proj",
|
| 169 |
+
"model.language_model.layers.1.mlp.shared_experts.down_proj",
|
| 170 |
+
"model.language_model.layers.2.mlp.shared_experts.gate_proj",
|
| 171 |
+
"model.language_model.layers.2.mlp.shared_experts.up_proj",
|
| 172 |
+
"model.language_model.layers.2.mlp.shared_experts.down_proj",
|
| 173 |
+
"model.language_model.layers.3.mlp.shared_experts.gate_proj",
|
| 174 |
+
"model.language_model.layers.3.mlp.shared_experts.up_proj",
|
| 175 |
+
"model.language_model.layers.3.mlp.shared_experts.down_proj",
|
| 176 |
+
"model.language_model.layers.4.mlp.shared_experts.gate_proj",
|
| 177 |
+
"model.language_model.layers.4.mlp.shared_experts.up_proj",
|
| 178 |
+
"model.language_model.layers.4.mlp.shared_experts.down_proj",
|
| 179 |
+
"model.language_model.layers.5.mlp.shared_experts.gate_proj",
|
| 180 |
+
"model.language_model.layers.5.mlp.shared_experts.up_proj",
|
| 181 |
+
"model.language_model.layers.5.mlp.shared_experts.down_proj",
|
| 182 |
+
"model.language_model.layers.6.mlp.shared_experts.gate_proj",
|
| 183 |
+
"model.language_model.layers.6.mlp.shared_experts.up_proj",
|
| 184 |
+
"model.language_model.layers.6.mlp.shared_experts.down_proj",
|
| 185 |
+
"model.language_model.layers.7.mlp.shared_experts.gate_proj",
|
| 186 |
+
"model.language_model.layers.7.mlp.shared_experts.up_proj",
|
| 187 |
+
"model.language_model.layers.7.mlp.shared_experts.down_proj",
|
| 188 |
+
"model.language_model.layers.8.mlp.shared_experts.gate_proj",
|
| 189 |
+
"model.language_model.layers.8.mlp.shared_experts.up_proj",
|
| 190 |
+
"model.language_model.layers.8.mlp.shared_experts.down_proj",
|
| 191 |
+
"model.language_model.layers.9.mlp.shared_experts.gate_proj",
|
| 192 |
+
"model.language_model.layers.9.mlp.shared_experts.up_proj",
|
| 193 |
+
"model.language_model.layers.9.mlp.shared_experts.down_proj",
|
| 194 |
+
"model.language_model.layers.10.mlp.shared_experts.gate_proj",
|
| 195 |
+
"model.language_model.layers.10.mlp.shared_experts.up_proj",
|
| 196 |
+
"model.language_model.layers.10.mlp.shared_experts.down_proj",
|
| 197 |
+
"model.language_model.layers.11.mlp.shared_experts.gate_proj",
|
| 198 |
+
"model.language_model.layers.11.mlp.shared_experts.up_proj",
|
| 199 |
+
"model.language_model.layers.11.mlp.shared_experts.down_proj",
|
| 200 |
+
"model.language_model.layers.12.mlp.shared_experts.gate_proj",
|
| 201 |
+
"model.language_model.layers.12.mlp.shared_experts.up_proj",
|
| 202 |
+
"model.language_model.layers.12.mlp.shared_experts.down_proj",
|
| 203 |
+
"model.language_model.layers.13.mlp.shared_experts.gate_proj",
|
| 204 |
+
"model.language_model.layers.13.mlp.shared_experts.up_proj",
|
| 205 |
+
"model.language_model.layers.13.mlp.shared_experts.down_proj",
|
| 206 |
+
"model.language_model.layers.14.mlp.shared_experts.gate_proj",
|
| 207 |
+
"model.language_model.layers.14.mlp.shared_experts.up_proj",
|
| 208 |
+
"model.language_model.layers.14.mlp.shared_experts.down_proj",
|
| 209 |
+
"model.language_model.layers.15.mlp.shared_experts.gate_proj",
|
| 210 |
+
"model.language_model.layers.15.mlp.shared_experts.up_proj",
|
| 211 |
+
"model.language_model.layers.15.mlp.shared_experts.down_proj",
|
| 212 |
+
"model.language_model.layers.16.mlp.shared_experts.gate_proj",
|
| 213 |
+
"model.language_model.layers.16.mlp.shared_experts.up_proj",
|
| 214 |
+
"model.language_model.layers.16.mlp.shared_experts.down_proj",
|
| 215 |
+
"model.language_model.layers.17.mlp.shared_experts.gate_proj",
|
| 216 |
+
"model.language_model.layers.17.mlp.shared_experts.up_proj",
|
| 217 |
+
"model.language_model.layers.17.mlp.shared_experts.down_proj",
|
| 218 |
+
"model.language_model.layers.18.mlp.shared_experts.gate_proj",
|
| 219 |
+
"model.language_model.layers.18.mlp.shared_experts.up_proj",
|
| 220 |
+
"model.language_model.layers.18.mlp.shared_experts.down_proj",
|
| 221 |
+
"model.language_model.layers.19.mlp.shared_experts.gate_proj",
|
| 222 |
+
"model.language_model.layers.19.mlp.shared_experts.up_proj",
|
| 223 |
+
"model.language_model.layers.19.mlp.shared_experts.down_proj",
|
| 224 |
+
"model.language_model.layers.20.mlp.shared_experts.gate_proj",
|
| 225 |
+
"model.language_model.layers.20.mlp.shared_experts.up_proj",
|
| 226 |
+
"model.language_model.layers.20.mlp.shared_experts.down_proj",
|
| 227 |
+
"model.language_model.layers.21.mlp.shared_experts.gate_proj",
|
| 228 |
+
"model.language_model.layers.21.mlp.shared_experts.up_proj",
|
| 229 |
+
"model.language_model.layers.21.mlp.shared_experts.down_proj",
|
| 230 |
+
"model.language_model.layers.22.mlp.shared_experts.gate_proj",
|
| 231 |
+
"model.language_model.layers.22.mlp.shared_experts.up_proj",
|
| 232 |
+
"model.language_model.layers.22.mlp.shared_experts.down_proj",
|
| 233 |
+
"model.language_model.layers.23.mlp.shared_experts.gate_proj",
|
| 234 |
+
"model.language_model.layers.23.mlp.shared_experts.up_proj",
|
| 235 |
+
"model.language_model.layers.23.mlp.shared_experts.down_proj",
|
| 236 |
+
"model.language_model.layers.24.mlp.shared_experts.gate_proj",
|
| 237 |
+
"model.language_model.layers.24.mlp.shared_experts.up_proj",
|
| 238 |
+
"model.language_model.layers.24.mlp.shared_experts.down_proj",
|
| 239 |
+
"model.language_model.layers.25.mlp.shared_experts.gate_proj",
|
| 240 |
+
"model.language_model.layers.25.mlp.shared_experts.up_proj",
|
| 241 |
+
"model.language_model.layers.25.mlp.shared_experts.down_proj",
|
| 242 |
+
"model.language_model.layers.26.mlp.shared_experts.gate_proj",
|
| 243 |
+
"model.language_model.layers.26.mlp.shared_experts.up_proj",
|
| 244 |
+
"model.language_model.layers.26.mlp.shared_experts.down_proj",
|
| 245 |
+
"model.language_model.layers.27.mlp.shared_experts.gate_proj",
|
| 246 |
+
"model.language_model.layers.27.mlp.shared_experts.up_proj",
|
| 247 |
+
"model.language_model.layers.27.mlp.shared_experts.down_proj",
|
| 248 |
+
"model.language_model.layers.28.mlp.shared_experts.gate_proj",
|
| 249 |
+
"model.language_model.layers.28.mlp.shared_experts.up_proj",
|
| 250 |
+
"model.language_model.layers.28.mlp.shared_experts.down_proj",
|
| 251 |
+
"model.language_model.layers.29.mlp.shared_experts.gate_proj",
|
| 252 |
+
"model.language_model.layers.29.mlp.shared_experts.up_proj",
|
| 253 |
+
"model.language_model.layers.29.mlp.shared_experts.down_proj",
|
| 254 |
+
"model.language_model.layers.30.mlp.shared_experts.gate_proj",
|
| 255 |
+
"model.language_model.layers.30.mlp.shared_experts.up_proj",
|
| 256 |
+
"model.language_model.layers.30.mlp.shared_experts.down_proj",
|
| 257 |
+
"model.language_model.layers.31.mlp.shared_experts.gate_proj",
|
| 258 |
+
"model.language_model.layers.31.mlp.shared_experts.up_proj",
|
| 259 |
+
"model.language_model.layers.31.mlp.shared_experts.down_proj",
|
| 260 |
+
"model.language_model.layers.32.mlp.shared_experts.gate_proj",
|
| 261 |
+
"model.language_model.layers.32.mlp.shared_experts.up_proj",
|
| 262 |
+
"model.language_model.layers.32.mlp.shared_experts.down_proj",
|
| 263 |
+
"model.language_model.layers.33.mlp.shared_experts.gate_proj",
|
| 264 |
+
"model.language_model.layers.33.mlp.shared_experts.up_proj",
|
| 265 |
+
"model.language_model.layers.33.mlp.shared_experts.down_proj",
|
| 266 |
+
"model.language_model.layers.34.mlp.shared_experts.gate_proj",
|
| 267 |
+
"model.language_model.layers.34.mlp.shared_experts.up_proj",
|
| 268 |
+
"model.language_model.layers.34.mlp.shared_experts.down_proj",
|
| 269 |
+
"model.language_model.layers.35.mlp.shared_experts.gate_proj",
|
| 270 |
+
"model.language_model.layers.35.mlp.shared_experts.up_proj",
|
| 271 |
+
"model.language_model.layers.35.mlp.shared_experts.down_proj",
|
| 272 |
+
"model.language_model.layers.36.mlp.shared_experts.gate_proj",
|
| 273 |
+
"model.language_model.layers.36.mlp.shared_experts.up_proj",
|
| 274 |
+
"model.language_model.layers.36.mlp.shared_experts.down_proj",
|
| 275 |
+
"model.language_model.layers.37.mlp.shared_experts.gate_proj",
|
| 276 |
+
"model.language_model.layers.37.mlp.shared_experts.up_proj",
|
| 277 |
+
"model.language_model.layers.37.mlp.shared_experts.down_proj",
|
| 278 |
+
"model.language_model.layers.38.mlp.shared_experts.gate_proj",
|
| 279 |
+
"model.language_model.layers.38.mlp.shared_experts.up_proj",
|
| 280 |
+
"model.language_model.layers.38.mlp.shared_experts.down_proj",
|
| 281 |
+
"model.language_model.layers.39.mlp.shared_experts.gate_proj",
|
| 282 |
+
"model.language_model.layers.39.mlp.shared_experts.up_proj",
|
| 283 |
+
"model.language_model.layers.39.mlp.shared_experts.down_proj",
|
| 284 |
+
"model.language_model.layers.40.mlp.shared_experts.gate_proj",
|
| 285 |
+
"model.language_model.layers.40.mlp.shared_experts.up_proj",
|
| 286 |
+
"model.language_model.layers.40.mlp.shared_experts.down_proj",
|
| 287 |
+
"model.language_model.layers.41.mlp.shared_experts.gate_proj",
|
| 288 |
+
"model.language_model.layers.41.mlp.shared_experts.up_proj",
|
| 289 |
+
"model.language_model.layers.41.mlp.shared_experts.down_proj",
|
| 290 |
+
"model.language_model.layers.42.mlp.shared_experts.gate_proj",
|
| 291 |
+
"model.language_model.layers.42.mlp.shared_experts.up_proj",
|
| 292 |
+
"model.language_model.layers.42.mlp.shared_experts.down_proj",
|
| 293 |
+
"model.language_model.layers.43.mlp.shared_experts.gate_proj",
|
| 294 |
+
"model.language_model.layers.43.mlp.shared_experts.up_proj",
|
| 295 |
+
"model.language_model.layers.43.mlp.shared_experts.down_proj",
|
| 296 |
+
"model.language_model.layers.44.mlp.shared_experts.gate_proj",
|
| 297 |
+
"model.language_model.layers.44.mlp.shared_experts.up_proj",
|
| 298 |
+
"model.language_model.layers.44.mlp.shared_experts.down_proj",
|
| 299 |
+
"model.language_model.layers.45.mlp.shared_experts.gate_proj",
|
| 300 |
+
"model.language_model.layers.45.mlp.shared_experts.up_proj",
|
| 301 |
+
"model.language_model.layers.45.mlp.shared_experts.down_proj",
|
| 302 |
+
"lm_head"
|
| 303 |
+
],
|
| 304 |
+
"kv_cache_scheme": null,
|
| 305 |
+
"quant_method": "compressed-tensors",
|
| 306 |
+
"quantization_status": "compressed",
|
| 307 |
+
"sparsity_config": {},
|
| 308 |
+
"transform_config": {},
|
| 309 |
+
"version": "0.12.3.a20251203"
|
| 310 |
+
},
|
| 311 |
+
"text_config": {
|
| 312 |
+
"attention_bias": true,
|
| 313 |
+
"attention_dropout": 0.0,
|
| 314 |
+
"dtype": "bfloat16",
|
| 315 |
+
"eos_token_id": [
|
| 316 |
+
151329,
|
| 317 |
+
151336,
|
| 318 |
+
151338
|
| 319 |
+
],
|
| 320 |
+
"first_k_dense_replace": 1,
|
| 321 |
+
"head_dim": 128,
|
| 322 |
+
"hidden_act": "silu",
|
| 323 |
+
"hidden_size": 4096,
|
| 324 |
+
"initializer_range": 0.02,
|
| 325 |
+
"intermediate_size": 10944,
|
| 326 |
+
"max_position_embeddings": 131072,
|
| 327 |
+
"model_type": "Glm4vMoe_text",
|
| 328 |
+
"moe_intermediate_size": 1408,
|
| 329 |
+
"n_group": 1,
|
| 330 |
+
"n_routed_experts": 128,
|
| 331 |
+
"n_shared_experts": 1,
|
| 332 |
+
"norm_topk_prob": true,
|
| 333 |
+
"num_attention_heads": 96,
|
| 334 |
+
"num_experts_per_tok": 8,
|
| 335 |
+
"num_hidden_layers": 46,
|
| 336 |
+
"num_key_value_heads": 8,
|
| 337 |
+
"num_nextn_predict_layers": 0,
|
| 338 |
+
"pad_token_id": 151329,
|
| 339 |
+
"partial_rotary_factor": 0.5,
|
| 340 |
+
"qk_layernorm": false,
|
| 341 |
+
"rms_norm_eps": 1e-05,
|
| 342 |
+
"rope_parameters": {
|
| 343 |
+
"mrope_section": [
|
| 344 |
+
8,
|
| 345 |
+
12,
|
| 346 |
+
12
|
| 347 |
+
],
|
| 348 |
+
"partial_rotary_factor": 0.5,
|
| 349 |
+
"rope_theta": 500000,
|
| 350 |
+
"rope_type": "default"
|
| 351 |
+
},
|
| 352 |
+
"rope_scaling": {
|
| 353 |
+
"mrope_section": [
|
| 354 |
+
8,
|
| 355 |
+
12,
|
| 356 |
+
12
|
| 357 |
+
],
|
| 358 |
+
"partial_rotary_factor": 0.5,
|
| 359 |
+
"rope_theta": 500000,
|
| 360 |
+
"rope_type": "default"
|
| 361 |
+
},
|
| 362 |
+
"rope_theta": 10000.0,
|
| 363 |
+
"routed_scaling_factor": 1.0,
|
| 364 |
+
"topk_group": 1,
|
| 365 |
+
"use_cache": true,
|
| 366 |
+
"use_qk_norm": false,
|
| 367 |
+
"vocab_size": 151552
|
| 368 |
+
},
|
| 369 |
+
"tie_word_embeddings": false,
|
| 370 |
+
"transformers_version": "4.57.3",
|
| 371 |
+
"video_end_token_id": 151342,
|
| 372 |
+
"video_start_token_id": 151341,
|
| 373 |
+
"video_token_id": 151364,
|
| 374 |
+
"vision_config": {
|
| 375 |
+
"attention_bias": false,
|
| 376 |
+
"attention_dropout": 0.0,
|
| 377 |
+
"depth": 24,
|
| 378 |
+
"hidden_act": "silu",
|
| 379 |
+
"hidden_dropout_prob": 0.0,
|
| 380 |
+
"hidden_size": 1536,
|
| 381 |
+
"image_size": 336,
|
| 382 |
+
"in_channels": 3,
|
| 383 |
+
"initializer_range": 0.02,
|
| 384 |
+
"intermediate_size": 10944,
|
| 385 |
+
"model_type": "glm4v_moe",
|
| 386 |
+
"num_heads": 12,
|
| 387 |
+
"out_hidden_size": 4096,
|
| 388 |
+
"patch_size": 14,
|
| 389 |
+
"rms_norm_eps": 1e-05,
|
| 390 |
+
"spatial_merge_size": 2,
|
| 391 |
+
"temporal_patch_size": 2
|
| 392 |
+
}
|
| 393 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,141 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[gMASK]<sop>
|
| 2 |
+
{%- if tools -%}
|
| 3 |
+
<|system|>
|
| 4 |
+
# Tools
|
| 5 |
+
|
| 6 |
+
You may call one or more functions to assist with the user query.
|
| 7 |
+
|
| 8 |
+
You are provided with function signatures within <tools></tools> XML tags:
|
| 9 |
+
<tools>
|
| 10 |
+
{% for tool in tools %}
|
| 11 |
+
{{ tool | tojson(ensure_ascii=False) }}
|
| 12 |
+
{% endfor %}
|
| 13 |
+
</tools>
|
| 14 |
+
|
| 15 |
+
For each function call, output the function name and arguments within the following XML format:
|
| 16 |
+
<tool_call>{function-name}
|
| 17 |
+
<arg_key>{arg-key-1}</arg_key>
|
| 18 |
+
<arg_value>{arg-value-1}</arg_value>
|
| 19 |
+
<arg_key>{arg-key-2}</arg_key>
|
| 20 |
+
<arg_value>{arg-value-2}</arg_value>
|
| 21 |
+
...
|
| 22 |
+
</tool_call>{%- endif -%}
|
| 23 |
+
{%- macro visible_text(content) -%}
|
| 24 |
+
{%- if content is string -%}
|
| 25 |
+
{{- content }}
|
| 26 |
+
{%- elif content is iterable and content is not mapping -%}
|
| 27 |
+
{%- for item in content -%}
|
| 28 |
+
{%- if item is mapping and item.type == 'text' -%}
|
| 29 |
+
{{- item.text }}
|
| 30 |
+
{%- elif item is mapping and (item.type == 'image' or 'image' in item) -%}
|
| 31 |
+
<|begin_of_image|><|image|><|end_of_image|>
|
| 32 |
+
{%- elif item is mapping and (item.type == 'video' or 'video' in item) -%}
|
| 33 |
+
<|begin_of_video|><|video|><|end_of_video|>
|
| 34 |
+
{%- elif item is string -%}
|
| 35 |
+
{{- item }}
|
| 36 |
+
{%- endif -%}
|
| 37 |
+
{%- endfor -%}
|
| 38 |
+
{%- else -%}
|
| 39 |
+
{{- content }}
|
| 40 |
+
{%- endif -%}
|
| 41 |
+
{%- endmacro -%}
|
| 42 |
+
{%- set ns = namespace(last_user_index=-1) %}
|
| 43 |
+
{%- for m in messages %}
|
| 44 |
+
{%- if m.role == 'user' %}
|
| 45 |
+
{% set ns.last_user_index = loop.index0 -%}
|
| 46 |
+
{%- endif %}
|
| 47 |
+
{%- endfor %}
|
| 48 |
+
{% for m in messages %}
|
| 49 |
+
{%- if m.role == 'user' -%}<|user|>
|
| 50 |
+
{% if m.content is string %}
|
| 51 |
+
{{ m.content }}
|
| 52 |
+
{%- else %}
|
| 53 |
+
{%- for item in m.content %}
|
| 54 |
+
{% if item.type == 'video' or 'video' in item %}
|
| 55 |
+
<|begin_of_video|><|video|><|end_of_video|>{% elif item.type == 'image' or 'image' in item %}
|
| 56 |
+
<|begin_of_image|><|image|><|end_of_image|>{% elif item.type == 'text' %}
|
| 57 |
+
{{ item.text }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endfor %}
|
| 60 |
+
{%- endif %}
|
| 61 |
+
{{- '/nothink' if (enable_thinking is defined and not enable_thinking and not visible_text(m.content).endswith("/nothink")) else '' -}}
|
| 62 |
+
{%- elif m.role == 'assistant' -%}
|
| 63 |
+
<|assistant|>
|
| 64 |
+
{%- set reasoning_content = '' %}
|
| 65 |
+
{%- set content = visible_text(m.content) %}
|
| 66 |
+
{%- if m.reasoning_content is string %}
|
| 67 |
+
{%- set reasoning_content = m.reasoning_content %}
|
| 68 |
+
{%- else %}
|
| 69 |
+
{%- if '</think>' in content %}
|
| 70 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 71 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 72 |
+
{%- endif %}
|
| 73 |
+
{%- endif %}
|
| 74 |
+
{%- if loop.index0 > ns.last_user_index and reasoning_content -%}
|
| 75 |
+
{{ '\n<think>' + reasoning_content.strip() + '</think>'}}
|
| 76 |
+
{%- else -%}
|
| 77 |
+
{{ '\n<think></think>' }}
|
| 78 |
+
{%- endif -%}
|
| 79 |
+
{%- if content.strip() -%}
|
| 80 |
+
{{ '\n' + content.strip() }}
|
| 81 |
+
{%- endif -%}
|
| 82 |
+
{% if m.tool_calls %}
|
| 83 |
+
{% for tc in m.tool_calls %}
|
| 84 |
+
{%- if tc.function %}
|
| 85 |
+
{%- set tc = tc.function %}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{{ '\n<tool_call>' + tc.name }}
|
| 88 |
+
{% set _args = tc.arguments %}
|
| 89 |
+
{% for k, v in _args.items() %}
|
| 90 |
+
<arg_key>{{ k }}</arg_key>
|
| 91 |
+
<arg_value>{{ v | tojson(ensure_ascii=False) if v is not string else v }}</arg_value>
|
| 92 |
+
{% endfor %}
|
| 93 |
+
</tool_call>{% endfor %}
|
| 94 |
+
{% endif %}
|
| 95 |
+
{%- elif m.role == 'tool' -%}
|
| 96 |
+
{%- if m.content is string -%}
|
| 97 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 98 |
+
{{- '<|observation|>' }}
|
| 99 |
+
{%- endif %}
|
| 100 |
+
{{- '\n<tool_response>\n' }}
|
| 101 |
+
{{- m.content }}
|
| 102 |
+
{{- '\n</tool_response>' }}
|
| 103 |
+
{% elif m.content is iterable and m.content is not mapping %}
|
| 104 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 105 |
+
{{- '<|observation|>' }}
|
| 106 |
+
{%- endif %}
|
| 107 |
+
{{- '\n<tool_response>\n' }}
|
| 108 |
+
{%- for tr in m.content -%}
|
| 109 |
+
{%- if tr is mapping and tr.type is defined -%}
|
| 110 |
+
{%- set t = tr.type | lower -%}
|
| 111 |
+
{%- if t == 'text' and tr.text is defined -%}
|
| 112 |
+
{{ tr.text }}
|
| 113 |
+
{%- elif t in ['image', 'image_url'] -%}
|
| 114 |
+
<|begin_of_image|><|image|><|end_of_image|>
|
| 115 |
+
{%- elif t in ['video', 'video_url'] -%}
|
| 116 |
+
<|begin_of_video|><|video|><|end_of_video|>
|
| 117 |
+
{%- else -%}
|
| 118 |
+
{{ tr | tojson(ensure_ascii=False) }}
|
| 119 |
+
{%- endif -%}
|
| 120 |
+
{%- else -%}
|
| 121 |
+
{{ tr.output if tr.output is defined else tr }}
|
| 122 |
+
{%- endif -%}
|
| 123 |
+
{%- endfor -%}
|
| 124 |
+
{{- '\n</tool_response>' }}
|
| 125 |
+
{%- else -%}
|
| 126 |
+
<|observation|>{% for tr in m.content %}
|
| 127 |
+
|
| 128 |
+
<tool_response>
|
| 129 |
+
{{ tr.output if tr.output is defined else tr }}
|
| 130 |
+
</tool_response>{% endfor -%}
|
| 131 |
+
{% endif -%}
|
| 132 |
+
{# ====== 逻辑结束 ====== #}
|
| 133 |
+
{%- elif m.role == 'system' -%}
|
| 134 |
+
<|system|>
|
| 135 |
+
{{ visible_text(m.content) }}
|
| 136 |
+
{%- endif -%}
|
| 137 |
+
{%- endfor -%}
|
| 138 |
+
{%- if add_generation_prompt -%}
|
| 139 |
+
<|assistant|>
|
| 140 |
+
{{'<think></think>\n' if (enable_thinking is defined and not enable_thinking) else ''}}
|
| 141 |
+
{%- endif -%}
|
config.json
ADDED
|
@@ -0,0 +1,406 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Glm4vMoeForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "bfloat16",
|
| 6 |
+
"image_end_token_id": 151340,
|
| 7 |
+
"image_start_token_id": 151339,
|
| 8 |
+
"image_token_id": 151363,
|
| 9 |
+
"model_type": "glm4v_moe",
|
| 10 |
+
"quantization_config": {
|
| 11 |
+
"config_groups": {
|
| 12 |
+
"group_0": {
|
| 13 |
+
"format": "pack-quantized",
|
| 14 |
+
"input_activations": null,
|
| 15 |
+
"output_activations": null,
|
| 16 |
+
"targets": [
|
| 17 |
+
"Linear"
|
| 18 |
+
],
|
| 19 |
+
"weights": {
|
| 20 |
+
"actorder": null,
|
| 21 |
+
"block_structure": null,
|
| 22 |
+
"dynamic": false,
|
| 23 |
+
"group_size": 32,
|
| 24 |
+
"num_bits": 8,
|
| 25 |
+
"observer": "mse",
|
| 26 |
+
"observer_kwargs": {},
|
| 27 |
+
"strategy": "group",
|
| 28 |
+
"symmetric": true,
|
| 29 |
+
"type": "int"
|
| 30 |
+
}
|
| 31 |
+
}
|
| 32 |
+
},
|
| 33 |
+
"format": "pack-quantized",
|
| 34 |
+
"global_compression_ratio": null,
|
| 35 |
+
"ignore": [
|
| 36 |
+
"model.visual.blocks.0.attn.qkv_proj",
|
| 37 |
+
"model.visual.blocks.0.attn.qkv",
|
| 38 |
+
"model.visual.blocks.0.attn.proj",
|
| 39 |
+
"model.visual.blocks.0.mlp.gate_proj",
|
| 40 |
+
"model.visual.blocks.0.mlp.up_proj",
|
| 41 |
+
"model.visual.blocks.0.mlp.down_proj",
|
| 42 |
+
"model.visual.blocks.1.attn.qkv_proj",
|
| 43 |
+
"model.visual.blocks.1.attn.qkv",
|
| 44 |
+
"model.visual.blocks.1.attn.proj",
|
| 45 |
+
"model.visual.blocks.1.mlp.gate_proj",
|
| 46 |
+
"model.visual.blocks.1.mlp.up_proj",
|
| 47 |
+
"model.visual.blocks.1.mlp.down_proj",
|
| 48 |
+
"model.visual.blocks.2.attn.qkv_proj",
|
| 49 |
+
"model.visual.blocks.2.attn.qkv",
|
| 50 |
+
"model.visual.blocks.2.attn.proj",
|
| 51 |
+
"model.visual.blocks.2.mlp.gate_proj",
|
| 52 |
+
"model.visual.blocks.2.mlp.up_proj",
|
| 53 |
+
"model.visual.blocks.2.mlp.down_proj",
|
| 54 |
+
"model.visual.blocks.3.attn.qkv_proj",
|
| 55 |
+
"model.visual.blocks.3.attn.qkv",
|
| 56 |
+
"model.visual.blocks.3.attn.proj",
|
| 57 |
+
"model.visual.blocks.3.mlp.gate_proj",
|
| 58 |
+
"model.visual.blocks.3.mlp.up_proj",
|
| 59 |
+
"model.visual.blocks.3.mlp.down_proj",
|
| 60 |
+
"model.visual.blocks.4.attn.qkv_proj",
|
| 61 |
+
"model.visual.blocks.4.attn.qkv",
|
| 62 |
+
"model.visual.blocks.4.attn.proj",
|
| 63 |
+
"model.visual.blocks.4.mlp.gate_proj",
|
| 64 |
+
"model.visual.blocks.4.mlp.up_proj",
|
| 65 |
+
"model.visual.blocks.4.mlp.down_proj",
|
| 66 |
+
"model.visual.blocks.5.attn.qkv_proj",
|
| 67 |
+
"model.visual.blocks.5.attn.qkv",
|
| 68 |
+
"model.visual.blocks.5.attn.proj",
|
| 69 |
+
"model.visual.blocks.5.mlp.gate_proj",
|
| 70 |
+
"model.visual.blocks.5.mlp.up_proj",
|
| 71 |
+
"model.visual.blocks.5.mlp.down_proj",
|
| 72 |
+
"model.visual.blocks.6.attn.qkv_proj",
|
| 73 |
+
"model.visual.blocks.6.attn.qkv",
|
| 74 |
+
"model.visual.blocks.6.attn.proj",
|
| 75 |
+
"model.visual.blocks.6.mlp.gate_proj",
|
| 76 |
+
"model.visual.blocks.6.mlp.up_proj",
|
| 77 |
+
"model.visual.blocks.6.mlp.down_proj",
|
| 78 |
+
"model.visual.blocks.7.attn.qkv_proj",
|
| 79 |
+
"model.visual.blocks.7.attn.qkv",
|
| 80 |
+
"model.visual.blocks.7.attn.proj",
|
| 81 |
+
"model.visual.blocks.7.mlp.gate_proj",
|
| 82 |
+
"model.visual.blocks.7.mlp.up_proj",
|
| 83 |
+
"model.visual.blocks.7.mlp.down_proj",
|
| 84 |
+
"model.visual.blocks.8.attn.qkv_proj",
|
| 85 |
+
"model.visual.blocks.8.attn.qkv",
|
| 86 |
+
"model.visual.blocks.8.attn.proj",
|
| 87 |
+
"model.visual.blocks.8.mlp.gate_proj",
|
| 88 |
+
"model.visual.blocks.8.mlp.up_proj",
|
| 89 |
+
"model.visual.blocks.8.mlp.down_proj",
|
| 90 |
+
"model.visual.blocks.9.attn.qkv_proj",
|
| 91 |
+
"model.visual.blocks.9.attn.qkv",
|
| 92 |
+
"model.visual.blocks.9.attn.proj",
|
| 93 |
+
"model.visual.blocks.9.mlp.gate_proj",
|
| 94 |
+
"model.visual.blocks.9.mlp.up_proj",
|
| 95 |
+
"model.visual.blocks.9.mlp.down_proj",
|
| 96 |
+
"model.visual.blocks.10.attn.qkv_proj",
|
| 97 |
+
"model.visual.blocks.10.attn.qkv",
|
| 98 |
+
"model.visual.blocks.10.attn.proj",
|
| 99 |
+
"model.visual.blocks.10.mlp.gate_proj",
|
| 100 |
+
"model.visual.blocks.10.mlp.up_proj",
|
| 101 |
+
"model.visual.blocks.10.mlp.down_proj",
|
| 102 |
+
"model.visual.blocks.11.attn.qkv_proj",
|
| 103 |
+
"model.visual.blocks.11.attn.qkv",
|
| 104 |
+
"model.visual.blocks.11.attn.proj",
|
| 105 |
+
"model.visual.blocks.11.mlp.gate_proj",
|
| 106 |
+
"model.visual.blocks.11.mlp.up_proj",
|
| 107 |
+
"model.visual.blocks.11.mlp.down_proj",
|
| 108 |
+
"model.visual.blocks.12.attn.qkv_proj",
|
| 109 |
+
"model.visual.blocks.12.attn.qkv",
|
| 110 |
+
"model.visual.blocks.12.attn.proj",
|
| 111 |
+
"model.visual.blocks.12.mlp.gate_proj",
|
| 112 |
+
"model.visual.blocks.12.mlp.up_proj",
|
| 113 |
+
"model.visual.blocks.12.mlp.down_proj",
|
| 114 |
+
"model.visual.blocks.13.attn.qkv_proj",
|
| 115 |
+
"model.visual.blocks.13.attn.qkv",
|
| 116 |
+
"model.visual.blocks.13.attn.proj",
|
| 117 |
+
"model.visual.blocks.13.mlp.gate_proj",
|
| 118 |
+
"model.visual.blocks.13.mlp.up_proj",
|
| 119 |
+
"model.visual.blocks.13.mlp.down_proj",
|
| 120 |
+
"model.visual.blocks.14.attn.qkv_proj",
|
| 121 |
+
"model.visual.blocks.14.attn.qkv",
|
| 122 |
+
"model.visual.blocks.14.attn.proj",
|
| 123 |
+
"model.visual.blocks.14.mlp.gate_proj",
|
| 124 |
+
"model.visual.blocks.14.mlp.up_proj",
|
| 125 |
+
"model.visual.blocks.14.mlp.down_proj",
|
| 126 |
+
"model.visual.blocks.15.attn.qkv_proj",
|
| 127 |
+
"model.visual.blocks.15.attn.qkv",
|
| 128 |
+
"model.visual.blocks.15.attn.proj",
|
| 129 |
+
"model.visual.blocks.15.mlp.gate_proj",
|
| 130 |
+
"model.visual.blocks.15.mlp.up_proj",
|
| 131 |
+
"model.visual.blocks.15.mlp.down_proj",
|
| 132 |
+
"model.visual.blocks.16.attn.qkv_proj",
|
| 133 |
+
"model.visual.blocks.16.attn.qkv",
|
| 134 |
+
"model.visual.blocks.16.attn.proj",
|
| 135 |
+
"model.visual.blocks.16.mlp.gate_proj",
|
| 136 |
+
"model.visual.blocks.16.mlp.up_proj",
|
| 137 |
+
"model.visual.blocks.16.mlp.down_proj",
|
| 138 |
+
"model.visual.blocks.17.attn.qkv_proj",
|
| 139 |
+
"model.visual.blocks.17.attn.qkv",
|
| 140 |
+
"model.visual.blocks.17.attn.proj",
|
| 141 |
+
"model.visual.blocks.17.mlp.gate_proj",
|
| 142 |
+
"model.visual.blocks.17.mlp.up_proj",
|
| 143 |
+
"model.visual.blocks.17.mlp.down_proj",
|
| 144 |
+
"model.visual.blocks.18.attn.qkv_proj",
|
| 145 |
+
"model.visual.blocks.18.attn.qkv",
|
| 146 |
+
"model.visual.blocks.18.attn.proj",
|
| 147 |
+
"model.visual.blocks.18.mlp.gate_proj",
|
| 148 |
+
"model.visual.blocks.18.mlp.up_proj",
|
| 149 |
+
"model.visual.blocks.18.mlp.down_proj",
|
| 150 |
+
"model.visual.blocks.19.attn.qkv_proj",
|
| 151 |
+
"model.visual.blocks.19.attn.qkv",
|
| 152 |
+
"model.visual.blocks.19.attn.proj",
|
| 153 |
+
"model.visual.blocks.19.mlp.gate_proj",
|
| 154 |
+
"model.visual.blocks.19.mlp.up_proj",
|
| 155 |
+
"model.visual.blocks.19.mlp.down_proj",
|
| 156 |
+
"model.visual.blocks.20.attn.qkv_proj",
|
| 157 |
+
"model.visual.blocks.20.attn.qkv",
|
| 158 |
+
"model.visual.blocks.20.attn.proj",
|
| 159 |
+
"model.visual.blocks.20.mlp.gate_proj",
|
| 160 |
+
"model.visual.blocks.20.mlp.up_proj",
|
| 161 |
+
"model.visual.blocks.20.mlp.down_proj",
|
| 162 |
+
"model.visual.blocks.21.attn.qkv_proj",
|
| 163 |
+
"model.visual.blocks.21.attn.qkv",
|
| 164 |
+
"model.visual.blocks.21.attn.proj",
|
| 165 |
+
"model.visual.blocks.21.mlp.gate_proj",
|
| 166 |
+
"model.visual.blocks.21.mlp.up_proj",
|
| 167 |
+
"model.visual.blocks.21.mlp.down_proj",
|
| 168 |
+
"model.visual.blocks.22.attn.qkv_proj",
|
| 169 |
+
"model.visual.blocks.22.attn.qkv",
|
| 170 |
+
"model.visual.blocks.22.attn.proj",
|
| 171 |
+
"model.visual.blocks.22.mlp.gate_proj",
|
| 172 |
+
"model.visual.blocks.22.mlp.up_proj",
|
| 173 |
+
"model.visual.blocks.22.mlp.down_proj",
|
| 174 |
+
"model.visual.blocks.23.attn.qkv_proj",
|
| 175 |
+
"model.visual.blocks.23.attn.qkv",
|
| 176 |
+
"model.visual.blocks.23.attn.proj",
|
| 177 |
+
"model.visual.blocks.23.mlp.gate_proj",
|
| 178 |
+
"model.visual.blocks.23.mlp.up_proj",
|
| 179 |
+
"model.visual.blocks.23.mlp.down_proj",
|
| 180 |
+
"model.visual.merger.proj",
|
| 181 |
+
"model.visual.merger.gate_proj",
|
| 182 |
+
"model.visual.merger.up_proj",
|
| 183 |
+
"model.visual.merger.down_proj",
|
| 184 |
+
"model.language_model.layers.0.self_attn.q_proj",
|
| 185 |
+
"model.language_model.layers.0.self_attn.k_proj",
|
| 186 |
+
"model.language_model.layers.0.self_attn.v_proj",
|
| 187 |
+
"model.language_model.layers.0.self_attn.o_proj",
|
| 188 |
+
"model.language_model.layers.0.mlp.gate_proj",
|
| 189 |
+
"model.language_model.layers.0.mlp.up_proj",
|
| 190 |
+
"model.language_model.layers.0.mlp.down_proj",
|
| 191 |
+
"model.language_model.layers.1.mlp.shared_experts.gate_proj",
|
| 192 |
+
"model.language_model.layers.1.mlp.shared_experts.up_proj",
|
| 193 |
+
"model.language_model.layers.1.mlp.shared_experts.down_proj",
|
| 194 |
+
"model.language_model.layers.2.mlp.shared_experts.gate_proj",
|
| 195 |
+
"model.language_model.layers.2.mlp.shared_experts.up_proj",
|
| 196 |
+
"model.language_model.layers.2.mlp.shared_experts.down_proj",
|
| 197 |
+
"model.language_model.layers.3.mlp.shared_experts.gate_proj",
|
| 198 |
+
"model.language_model.layers.3.mlp.shared_experts.up_proj",
|
| 199 |
+
"model.language_model.layers.3.mlp.shared_experts.down_proj",
|
| 200 |
+
"model.language_model.layers.4.mlp.shared_experts.gate_proj",
|
| 201 |
+
"model.language_model.layers.4.mlp.shared_experts.up_proj",
|
| 202 |
+
"model.language_model.layers.4.mlp.shared_experts.down_proj",
|
| 203 |
+
"model.language_model.layers.5.mlp.shared_experts.gate_proj",
|
| 204 |
+
"model.language_model.layers.5.mlp.shared_experts.up_proj",
|
| 205 |
+
"model.language_model.layers.5.mlp.shared_experts.down_proj",
|
| 206 |
+
"model.language_model.layers.6.mlp.shared_experts.gate_proj",
|
| 207 |
+
"model.language_model.layers.6.mlp.shared_experts.up_proj",
|
| 208 |
+
"model.language_model.layers.6.mlp.shared_experts.down_proj",
|
| 209 |
+
"model.language_model.layers.7.mlp.shared_experts.gate_proj",
|
| 210 |
+
"model.language_model.layers.7.mlp.shared_experts.up_proj",
|
| 211 |
+
"model.language_model.layers.7.mlp.shared_experts.down_proj",
|
| 212 |
+
"model.language_model.layers.8.mlp.shared_experts.gate_proj",
|
| 213 |
+
"model.language_model.layers.8.mlp.shared_experts.up_proj",
|
| 214 |
+
"model.language_model.layers.8.mlp.shared_experts.down_proj",
|
| 215 |
+
"model.language_model.layers.9.mlp.shared_experts.gate_proj",
|
| 216 |
+
"model.language_model.layers.9.mlp.shared_experts.up_proj",
|
| 217 |
+
"model.language_model.layers.9.mlp.shared_experts.down_proj",
|
| 218 |
+
"model.language_model.layers.10.mlp.shared_experts.gate_proj",
|
| 219 |
+
"model.language_model.layers.10.mlp.shared_experts.up_proj",
|
| 220 |
+
"model.language_model.layers.10.mlp.shared_experts.down_proj",
|
| 221 |
+
"model.language_model.layers.11.mlp.shared_experts.gate_proj",
|
| 222 |
+
"model.language_model.layers.11.mlp.shared_experts.up_proj",
|
| 223 |
+
"model.language_model.layers.11.mlp.shared_experts.down_proj",
|
| 224 |
+
"model.language_model.layers.12.mlp.shared_experts.gate_proj",
|
| 225 |
+
"model.language_model.layers.12.mlp.shared_experts.up_proj",
|
| 226 |
+
"model.language_model.layers.12.mlp.shared_experts.down_proj",
|
| 227 |
+
"model.language_model.layers.13.mlp.shared_experts.gate_proj",
|
| 228 |
+
"model.language_model.layers.13.mlp.shared_experts.up_proj",
|
| 229 |
+
"model.language_model.layers.13.mlp.shared_experts.down_proj",
|
| 230 |
+
"model.language_model.layers.14.mlp.shared_experts.gate_proj",
|
| 231 |
+
"model.language_model.layers.14.mlp.shared_experts.up_proj",
|
| 232 |
+
"model.language_model.layers.14.mlp.shared_experts.down_proj",
|
| 233 |
+
"model.language_model.layers.15.mlp.shared_experts.gate_proj",
|
| 234 |
+
"model.language_model.layers.15.mlp.shared_experts.up_proj",
|
| 235 |
+
"model.language_model.layers.15.mlp.shared_experts.down_proj",
|
| 236 |
+
"model.language_model.layers.16.mlp.shared_experts.gate_proj",
|
| 237 |
+
"model.language_model.layers.16.mlp.shared_experts.up_proj",
|
| 238 |
+
"model.language_model.layers.16.mlp.shared_experts.down_proj",
|
| 239 |
+
"model.language_model.layers.17.mlp.shared_experts.gate_proj",
|
| 240 |
+
"model.language_model.layers.17.mlp.shared_experts.up_proj",
|
| 241 |
+
"model.language_model.layers.17.mlp.shared_experts.down_proj",
|
| 242 |
+
"model.language_model.layers.18.mlp.shared_experts.gate_proj",
|
| 243 |
+
"model.language_model.layers.18.mlp.shared_experts.up_proj",
|
| 244 |
+
"model.language_model.layers.18.mlp.shared_experts.down_proj",
|
| 245 |
+
"model.language_model.layers.19.mlp.shared_experts.gate_proj",
|
| 246 |
+
"model.language_model.layers.19.mlp.shared_experts.up_proj",
|
| 247 |
+
"model.language_model.layers.19.mlp.shared_experts.down_proj",
|
| 248 |
+
"model.language_model.layers.20.mlp.shared_experts.gate_proj",
|
| 249 |
+
"model.language_model.layers.20.mlp.shared_experts.up_proj",
|
| 250 |
+
"model.language_model.layers.20.mlp.shared_experts.down_proj",
|
| 251 |
+
"model.language_model.layers.21.mlp.shared_experts.gate_proj",
|
| 252 |
+
"model.language_model.layers.21.mlp.shared_experts.up_proj",
|
| 253 |
+
"model.language_model.layers.21.mlp.shared_experts.down_proj",
|
| 254 |
+
"model.language_model.layers.22.mlp.shared_experts.gate_proj",
|
| 255 |
+
"model.language_model.layers.22.mlp.shared_experts.up_proj",
|
| 256 |
+
"model.language_model.layers.22.mlp.shared_experts.down_proj",
|
| 257 |
+
"model.language_model.layers.23.mlp.shared_experts.gate_proj",
|
| 258 |
+
"model.language_model.layers.23.mlp.shared_experts.up_proj",
|
| 259 |
+
"model.language_model.layers.23.mlp.shared_experts.down_proj",
|
| 260 |
+
"model.language_model.layers.24.mlp.shared_experts.gate_proj",
|
| 261 |
+
"model.language_model.layers.24.mlp.shared_experts.up_proj",
|
| 262 |
+
"model.language_model.layers.24.mlp.shared_experts.down_proj",
|
| 263 |
+
"model.language_model.layers.25.mlp.shared_experts.gate_proj",
|
| 264 |
+
"model.language_model.layers.25.mlp.shared_experts.up_proj",
|
| 265 |
+
"model.language_model.layers.25.mlp.shared_experts.down_proj",
|
| 266 |
+
"model.language_model.layers.26.mlp.shared_experts.gate_proj",
|
| 267 |
+
"model.language_model.layers.26.mlp.shared_experts.up_proj",
|
| 268 |
+
"model.language_model.layers.26.mlp.shared_experts.down_proj",
|
| 269 |
+
"model.language_model.layers.27.mlp.shared_experts.gate_proj",
|
| 270 |
+
"model.language_model.layers.27.mlp.shared_experts.up_proj",
|
| 271 |
+
"model.language_model.layers.27.mlp.shared_experts.down_proj",
|
| 272 |
+
"model.language_model.layers.28.mlp.shared_experts.gate_proj",
|
| 273 |
+
"model.language_model.layers.28.mlp.shared_experts.up_proj",
|
| 274 |
+
"model.language_model.layers.28.mlp.shared_experts.down_proj",
|
| 275 |
+
"model.language_model.layers.29.mlp.shared_experts.gate_proj",
|
| 276 |
+
"model.language_model.layers.29.mlp.shared_experts.up_proj",
|
| 277 |
+
"model.language_model.layers.29.mlp.shared_experts.down_proj",
|
| 278 |
+
"model.language_model.layers.30.mlp.shared_experts.gate_proj",
|
| 279 |
+
"model.language_model.layers.30.mlp.shared_experts.up_proj",
|
| 280 |
+
"model.language_model.layers.30.mlp.shared_experts.down_proj",
|
| 281 |
+
"model.language_model.layers.31.mlp.shared_experts.gate_proj",
|
| 282 |
+
"model.language_model.layers.31.mlp.shared_experts.up_proj",
|
| 283 |
+
"model.language_model.layers.31.mlp.shared_experts.down_proj",
|
| 284 |
+
"model.language_model.layers.32.mlp.shared_experts.gate_proj",
|
| 285 |
+
"model.language_model.layers.32.mlp.shared_experts.up_proj",
|
| 286 |
+
"model.language_model.layers.32.mlp.shared_experts.down_proj",
|
| 287 |
+
"model.language_model.layers.33.mlp.shared_experts.gate_proj",
|
| 288 |
+
"model.language_model.layers.33.mlp.shared_experts.up_proj",
|
| 289 |
+
"model.language_model.layers.33.mlp.shared_experts.down_proj",
|
| 290 |
+
"model.language_model.layers.34.mlp.shared_experts.gate_proj",
|
| 291 |
+
"model.language_model.layers.34.mlp.shared_experts.up_proj",
|
| 292 |
+
"model.language_model.layers.34.mlp.shared_experts.down_proj",
|
| 293 |
+
"model.language_model.layers.35.mlp.shared_experts.gate_proj",
|
| 294 |
+
"model.language_model.layers.35.mlp.shared_experts.up_proj",
|
| 295 |
+
"model.language_model.layers.35.mlp.shared_experts.down_proj",
|
| 296 |
+
"model.language_model.layers.36.mlp.shared_experts.gate_proj",
|
| 297 |
+
"model.language_model.layers.36.mlp.shared_experts.up_proj",
|
| 298 |
+
"model.language_model.layers.36.mlp.shared_experts.down_proj",
|
| 299 |
+
"model.language_model.layers.37.mlp.shared_experts.gate_proj",
|
| 300 |
+
"model.language_model.layers.37.mlp.shared_experts.up_proj",
|
| 301 |
+
"model.language_model.layers.37.mlp.shared_experts.down_proj",
|
| 302 |
+
"model.language_model.layers.38.mlp.shared_experts.gate_proj",
|
| 303 |
+
"model.language_model.layers.38.mlp.shared_experts.up_proj",
|
| 304 |
+
"model.language_model.layers.38.mlp.shared_experts.down_proj",
|
| 305 |
+
"model.language_model.layers.39.mlp.shared_experts.gate_proj",
|
| 306 |
+
"model.language_model.layers.39.mlp.shared_experts.up_proj",
|
| 307 |
+
"model.language_model.layers.39.mlp.shared_experts.down_proj",
|
| 308 |
+
"model.language_model.layers.40.mlp.shared_experts.gate_proj",
|
| 309 |
+
"model.language_model.layers.40.mlp.shared_experts.up_proj",
|
| 310 |
+
"model.language_model.layers.40.mlp.shared_experts.down_proj",
|
| 311 |
+
"model.language_model.layers.41.mlp.shared_experts.gate_proj",
|
| 312 |
+
"model.language_model.layers.41.mlp.shared_experts.up_proj",
|
| 313 |
+
"model.language_model.layers.41.mlp.shared_experts.down_proj",
|
| 314 |
+
"model.language_model.layers.42.mlp.shared_experts.gate_proj",
|
| 315 |
+
"model.language_model.layers.42.mlp.shared_experts.up_proj",
|
| 316 |
+
"model.language_model.layers.42.mlp.shared_experts.down_proj",
|
| 317 |
+
"model.language_model.layers.43.mlp.shared_experts.gate_proj",
|
| 318 |
+
"model.language_model.layers.43.mlp.shared_experts.up_proj",
|
| 319 |
+
"model.language_model.layers.43.mlp.shared_experts.down_proj",
|
| 320 |
+
"model.language_model.layers.44.mlp.shared_experts.gate_proj",
|
| 321 |
+
"model.language_model.layers.44.mlp.shared_experts.up_proj",
|
| 322 |
+
"model.language_model.layers.44.mlp.shared_experts.down_proj",
|
| 323 |
+
"model.language_model.layers.45.mlp.shared_experts.gate_proj",
|
| 324 |
+
"model.language_model.layers.45.mlp.shared_experts.up_proj",
|
| 325 |
+
"model.language_model.layers.45.mlp.shared_experts.down_proj",
|
| 326 |
+
"lm_head"
|
| 327 |
+
],
|
| 328 |
+
"kv_cache_scheme": null,
|
| 329 |
+
"quant_method": "compressed-tensors",
|
| 330 |
+
"quantization_status": "compressed",
|
| 331 |
+
"sparsity_config": {},
|
| 332 |
+
"transform_config": {},
|
| 333 |
+
"version": "0.12.3.a20251203"
|
| 334 |
+
},
|
| 335 |
+
"text_config": {
|
| 336 |
+
"attention_bias": true,
|
| 337 |
+
"attention_dropout": 0.0,
|
| 338 |
+
"dtype": "bfloat16",
|
| 339 |
+
"eos_token_id": [
|
| 340 |
+
151329,
|
| 341 |
+
151336,
|
| 342 |
+
151338
|
| 343 |
+
],
|
| 344 |
+
"first_k_dense_replace": 1,
|
| 345 |
+
"head_dim": 128,
|
| 346 |
+
"hidden_act": "silu",
|
| 347 |
+
"hidden_size": 4096,
|
| 348 |
+
"initializer_range": 0.02,
|
| 349 |
+
"intermediate_size": 10944,
|
| 350 |
+
"max_position_embeddings": 131072,
|
| 351 |
+
"model_type": "Glm4vMoe_text",
|
| 352 |
+
"moe_intermediate_size": 1408,
|
| 353 |
+
"n_group": 1,
|
| 354 |
+
"n_routed_experts": 128,
|
| 355 |
+
"n_shared_experts": 1,
|
| 356 |
+
"norm_topk_prob": true,
|
| 357 |
+
"num_attention_heads": 96,
|
| 358 |
+
"num_experts_per_tok": 8,
|
| 359 |
+
"num_hidden_layers": 46,
|
| 360 |
+
"num_key_value_heads": 8,
|
| 361 |
+
"num_nextn_predict_layers": 0,
|
| 362 |
+
"pad_token_id": 151329,
|
| 363 |
+
"partial_rotary_factor": 0.5,
|
| 364 |
+
"qk_layernorm": false,
|
| 365 |
+
"rms_norm_eps": 1e-05,
|
| 366 |
+
"rope_parameters": {
|
| 367 |
+
"mrope_section": [
|
| 368 |
+
8,
|
| 369 |
+
12,
|
| 370 |
+
12
|
| 371 |
+
],
|
| 372 |
+
"partial_rotary_factor": 0.5,
|
| 373 |
+
"rope_theta": 500000,
|
| 374 |
+
"rope_type": "default"
|
| 375 |
+
},
|
| 376 |
+
"routed_scaling_factor": 1.0,
|
| 377 |
+
"topk_group": 1,
|
| 378 |
+
"use_cache": true,
|
| 379 |
+
"use_qk_norm": false,
|
| 380 |
+
"vocab_size": 151552
|
| 381 |
+
},
|
| 382 |
+
"tie_word_embeddings": false,
|
| 383 |
+
"transformers_version": "4.57.3",
|
| 384 |
+
"video_end_token_id": 151342,
|
| 385 |
+
"video_start_token_id": 151341,
|
| 386 |
+
"video_token_id": 151364,
|
| 387 |
+
"vision_config": {
|
| 388 |
+
"attention_bias": false,
|
| 389 |
+
"attention_dropout": 0.0,
|
| 390 |
+
"depth": 24,
|
| 391 |
+
"hidden_act": "silu",
|
| 392 |
+
"hidden_dropout_prob": 0.0,
|
| 393 |
+
"hidden_size": 1536,
|
| 394 |
+
"image_size": 336,
|
| 395 |
+
"in_channels": 3,
|
| 396 |
+
"initializer_range": 0.02,
|
| 397 |
+
"intermediate_size": 10944,
|
| 398 |
+
"model_type": "glm4v_moe",
|
| 399 |
+
"num_heads": 12,
|
| 400 |
+
"out_hidden_size": 4096,
|
| 401 |
+
"patch_size": 14,
|
| 402 |
+
"rms_norm_eps": 1e-05,
|
| 403 |
+
"spatial_merge_size": 2,
|
| 404 |
+
"temporal_patch_size": 2
|
| 405 |
+
}
|
| 406 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151329,
|
| 6 |
+
151336,
|
| 7 |
+
151338
|
| 8 |
+
],
|
| 9 |
+
"pad_token_id": 151329,
|
| 10 |
+
"temperature": 0.8,
|
| 11 |
+
"top_k": 2,
|
| 12 |
+
"top_p": 0.6,
|
| 13 |
+
"transformers_version": "4.57.3"
|
| 14 |
+
}
|
model-00001-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b76785d15404e44115d23dfd13761a693b6b60db9e7d544d5a0104577b7b0b0
|
| 3 |
+
size 4995869512
|
model-00002-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a6d2322e05ec2fddef1feb300ab0f645360c16c95c8bbc7666c0ca7fc84844b9
|
| 3 |
+
size 4997218360
|
model-00003-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0eb003638651ed1e922a2afe47c5e80f745a9781e294bc2a827d2eec02c929ee
|
| 3 |
+
size 4997218360
|
model-00004-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d171d617389c5f6c3c6e3cc43bdb8b6d0be341e5bb9a78b0562a88d3107f179
|
| 3 |
+
size 4997218360
|
model-00005-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5f91be7d486cdc4ecf8079c0f2f9b0218191a65269f2fd08978bd8dbe2b10c8
|
| 3 |
+
size 4997218360
|
model-00006-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbe7bdf664320f8c2896e053075df342ee99ce15a7ae35c94b827085aa887510
|
| 3 |
+
size 4997220224
|
model-00007-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:591e9444df514ac78d340d884ad42bc4a4b0649187518b8b2c3cff14ab5fd202
|
| 3 |
+
size 4997220704
|
model-00008-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dca23b2f5f4526ab430d1a27c9e05876882d8d1393d541dcfb634dd839aab4be
|
| 3 |
+
size 4997220704
|
model-00009-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e59d404e98aa728be9ae74e571d5480014a3a7457d3382722bd2c618f6a262cf
|
| 3 |
+
size 4997220704
|
model-00010-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd6b0a902a1cf37145b21892ba2becdf60b51fdb21d60252c7fb057303a74228
|
| 3 |
+
size 4997220704
|
model-00011-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f344abd3bdccb427f2999777d69f0995c5ab9f974316179b67a13183244e7696
|
| 3 |
+
size 4997220704
|
model-00012-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a1ed2476b15192228ebd8677e5726218eacc9d3aac2c80314b8779f79c55db8
|
| 3 |
+
size 4997220704
|
model-00013-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:13411a32931265694cd9cc322f44586fb3cdae0355d372c667ec4065033b256a
|
| 3 |
+
size 4997220704
|
model-00014-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:272a22356daad527a21f8f17b50ef4bdd1e22ab9f4b29b0ed94eacee2970e86b
|
| 3 |
+
size 4997220704
|
model-00015-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cee33e4aec15d942b4df27eee51170d73d27eeaa4d7250aa772416b7ee4bd695
|
| 3 |
+
size 4997220704
|
model-00016-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fdad2d95c9c440008964f5c95fc95f0690b7a754dc20e964273e340bac8a4048
|
| 3 |
+
size 4997220704
|
model-00017-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c40cb047537aa2ae3dabd53fcfb1e2ea34e797d56a3cce616207c620bb49f0cd
|
| 3 |
+
size 4997220704
|
model-00018-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:398600cad6ec48f4a8a48a88915d5ffc45235bcc6595d26b855ff440100d5c9b
|
| 3 |
+
size 4997220704
|
model-00019-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef7ff1814a489a1aa72a17d12900ecac57d8186db74932e892ca474190e82881
|
| 3 |
+
size 4997220704
|
model-00020-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b3004c0fa30db6db7b3cdd9d325a777f0b71c70d081fae2b83d8a98c25bd6d4
|
| 3 |
+
size 4997220704
|
model-00021-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf8053c7bfb98ae0d281d2cab5aa438bb8e65f645ab1b7d5600ac088233f5647
|
| 3 |
+
size 4997220704
|
model-00022-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c5f294cd02f0ef65a00020b4ed136dad68343641600fae3e8fe51798281e1f8
|
| 3 |
+
size 4997220704
|
model-00023-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c112dd8a1c5f54af28aaf2b374296674d2013c888a33ec503627c5c9d55ce12
|
| 3 |
+
size 4997220704
|
model-00024-of-00024.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ebe1f4f7790247f6f2db1e947e86c704e70808db59a0cffd5205163645c6c628
|
| 3 |
+
size 2466032800
|
model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"size": {"shortest_edge": 12544, "longest_edge": 9633792},
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"patch_size": 14,
|
| 5 |
+
"temporal_patch_size": 2,
|
| 6 |
+
"merge_size": 2,
|
| 7 |
+
"image_mean": [0.48145466, 0.4578275, 0.40821073],
|
| 8 |
+
"image_std": [0.26862954, 0.26130258, 0.27577711],
|
| 9 |
+
"image_processor_type": "Glm46VImageProcessor",
|
| 10 |
+
"processor_class": "Glm46VProcessor"
|
| 11 |
+
}
|
recipe.yaml
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
default_stage:
|
| 2 |
+
default_modifiers:
|
| 3 |
+
AWQModifier:
|
| 4 |
+
config_groups:
|
| 5 |
+
group_0:
|
| 6 |
+
targets: [Linear]
|
| 7 |
+
weights:
|
| 8 |
+
num_bits: 8
|
| 9 |
+
type: int
|
| 10 |
+
symmetric: true
|
| 11 |
+
group_size: 32
|
| 12 |
+
strategy: group
|
| 13 |
+
block_structure: null
|
| 14 |
+
dynamic: false
|
| 15 |
+
actorder: null
|
| 16 |
+
scale_dtype: null
|
| 17 |
+
zp_dtype: null
|
| 18 |
+
observer: mse
|
| 19 |
+
observer_kwargs: {}
|
| 20 |
+
input_activations: null
|
| 21 |
+
output_activations: null
|
| 22 |
+
format: null
|
| 23 |
+
targets: [Linear]
|
| 24 |
+
ignore: [lm_head, 're:.*embed_tokens', 're:.*input_layernorm', 're:.*post_attention_layernorm',
|
| 25 |
+
model.language_model.norm, 're:.*shared_experts.*', 're:model[.]language_model[.]layers[.]0[.].*',
|
| 26 |
+
're:.*mlp[.]gate$', 're:model[.]visual.*']
|
| 27 |
+
mappings:
|
| 28 |
+
- smooth_layer: re:.*input_layernorm$
|
| 29 |
+
balance_layers: ['re:.*q_proj$', 're:.*k_proj$', 're:.*v_proj$']
|
| 30 |
+
- smooth_layer: re:.*v_proj$
|
| 31 |
+
balance_layers: ['re:.*o_proj$']
|
| 32 |
+
- smooth_layer: re:.*post_attention_layernorm$
|
| 33 |
+
balance_layers: ['re:.*gate_proj$', 're:.*up_proj$']
|
| 34 |
+
- smooth_layer: re:.*up_proj$
|
| 35 |
+
balance_layers: ['re:.*down_proj$']
|
| 36 |
+
offload_device: !!python/object/apply:torch.device [cpu]
|
| 37 |
+
duo_scaling: true
|
| 38 |
+
n_grid: 20
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|endoftext|>",
|
| 4 |
+
"[MASK]",
|
| 5 |
+
"[gMASK]",
|
| 6 |
+
"[sMASK]",
|
| 7 |
+
"<sop>",
|
| 8 |
+
"<eop>",
|
| 9 |
+
"<|system|>",
|
| 10 |
+
"<|user|>",
|
| 11 |
+
"<|assistant|>",
|
| 12 |
+
"<|observation|>",
|
| 13 |
+
"<|begin_of_image|>",
|
| 14 |
+
"<|end_of_image|>",
|
| 15 |
+
"<|begin_of_video|>",
|
| 16 |
+
"<|end_of_video|>",
|
| 17 |
+
"<|begin_of_audio|>",
|
| 18 |
+
"<|end_of_audio|>",
|
| 19 |
+
"<|image|>",
|
| 20 |
+
"<|video|>",
|
| 21 |
+
"<|begin_of_transcription|>",
|
| 22 |
+
"<|end_of_transcription|>",
|
| 23 |
+
"<|code_prefix|>",
|
| 24 |
+
"<|code_middle|>",
|
| 25 |
+
"<|code_suffix|>",
|
| 26 |
+
"/nothink"
|
| 27 |
+
],
|
| 28 |
+
"eos_token": {
|
| 29 |
+
"content": "<|endoftext|>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false
|
| 34 |
+
},
|
| 35 |
+
"pad_token": {
|
| 36 |
+
"content": "<|endoftext|>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false
|
| 41 |
+
}
|
| 42 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2ff52959093921034528ecd6a59926e5fd543f56f94f2a0034ed4ba458c0a86
|
| 3 |
+
size 19970698
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,327 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"151329": {
|
| 4 |
+
"content": "<|endoftext|>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"151330": {
|
| 12 |
+
"content": "[MASK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"151331": {
|
| 20 |
+
"content": "[gMASK]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"151332": {
|
| 28 |
+
"content": "[sMASK]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"151333": {
|
| 36 |
+
"content": "<sop>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"151334": {
|
| 44 |
+
"content": "<eop>",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"151335": {
|
| 52 |
+
"content": "<|system|>",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": false,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": true
|
| 58 |
+
},
|
| 59 |
+
"151336": {
|
| 60 |
+
"content": "<|user|>",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": false,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false,
|
| 65 |
+
"special": true
|
| 66 |
+
},
|
| 67 |
+
"151337": {
|
| 68 |
+
"content": "<|assistant|>",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": false,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": true
|
| 74 |
+
},
|
| 75 |
+
"151338": {
|
| 76 |
+
"content": "<|observation|>",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": false,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": true
|
| 82 |
+
},
|
| 83 |
+
"151339": {
|
| 84 |
+
"content": "<|begin_of_image|>",
|
| 85 |
+
"lstrip": false,
|
| 86 |
+
"normalized": false,
|
| 87 |
+
"rstrip": false,
|
| 88 |
+
"single_word": false,
|
| 89 |
+
"special": true
|
| 90 |
+
},
|
| 91 |
+
"151340": {
|
| 92 |
+
"content": "<|end_of_image|>",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": false,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false,
|
| 97 |
+
"special": true
|
| 98 |
+
},
|
| 99 |
+
"151341": {
|
| 100 |
+
"content": "<|begin_of_video|>",
|
| 101 |
+
"lstrip": false,
|
| 102 |
+
"normalized": false,
|
| 103 |
+
"rstrip": false,
|
| 104 |
+
"single_word": false,
|
| 105 |
+
"special": true
|
| 106 |
+
},
|
| 107 |
+
"151342": {
|
| 108 |
+
"content": "<|end_of_video|>",
|
| 109 |
+
"lstrip": false,
|
| 110 |
+
"normalized": false,
|
| 111 |
+
"rstrip": false,
|
| 112 |
+
"single_word": false,
|
| 113 |
+
"special": true
|
| 114 |
+
},
|
| 115 |
+
"151343": {
|
| 116 |
+
"content": "<|begin_of_audio|>",
|
| 117 |
+
"lstrip": false,
|
| 118 |
+
"normalized": false,
|
| 119 |
+
"rstrip": false,
|
| 120 |
+
"single_word": false,
|
| 121 |
+
"special": true
|
| 122 |
+
},
|
| 123 |
+
"151344": {
|
| 124 |
+
"content": "<|end_of_audio|>",
|
| 125 |
+
"lstrip": false,
|
| 126 |
+
"normalized": false,
|
| 127 |
+
"rstrip": false,
|
| 128 |
+
"single_word": false,
|
| 129 |
+
"special": true
|
| 130 |
+
},
|
| 131 |
+
"151345": {
|
| 132 |
+
"content": "<|begin_of_transcription|>",
|
| 133 |
+
"lstrip": false,
|
| 134 |
+
"normalized": false,
|
| 135 |
+
"rstrip": false,
|
| 136 |
+
"single_word": false,
|
| 137 |
+
"special": true
|
| 138 |
+
},
|
| 139 |
+
"151346": {
|
| 140 |
+
"content": "<|end_of_transcription|>",
|
| 141 |
+
"lstrip": false,
|
| 142 |
+
"normalized": false,
|
| 143 |
+
"rstrip": false,
|
| 144 |
+
"single_word": false,
|
| 145 |
+
"special": true
|
| 146 |
+
},
|
| 147 |
+
"151347": {
|
| 148 |
+
"content": "<|code_prefix|>",
|
| 149 |
+
"lstrip": false,
|
| 150 |
+
"normalized": false,
|
| 151 |
+
"rstrip": false,
|
| 152 |
+
"single_word": false,
|
| 153 |
+
"special": true
|
| 154 |
+
},
|
| 155 |
+
"151348": {
|
| 156 |
+
"content": "<|code_middle|>",
|
| 157 |
+
"lstrip": false,
|
| 158 |
+
"normalized": false,
|
| 159 |
+
"rstrip": false,
|
| 160 |
+
"single_word": false,
|
| 161 |
+
"special": true
|
| 162 |
+
},
|
| 163 |
+
"151349": {
|
| 164 |
+
"content": "<|code_suffix|>",
|
| 165 |
+
"lstrip": false,
|
| 166 |
+
"normalized": false,
|
| 167 |
+
"rstrip": false,
|
| 168 |
+
"single_word": false,
|
| 169 |
+
"special": true
|
| 170 |
+
},
|
| 171 |
+
"151350": {
|
| 172 |
+
"content": "<think>",
|
| 173 |
+
"lstrip": false,
|
| 174 |
+
"normalized": false,
|
| 175 |
+
"rstrip": false,
|
| 176 |
+
"single_word": false,
|
| 177 |
+
"special": false
|
| 178 |
+
},
|
| 179 |
+
"151351": {
|
| 180 |
+
"content": "</think>",
|
| 181 |
+
"lstrip": false,
|
| 182 |
+
"normalized": false,
|
| 183 |
+
"rstrip": false,
|
| 184 |
+
"single_word": false,
|
| 185 |
+
"special": false
|
| 186 |
+
},
|
| 187 |
+
"151352": {
|
| 188 |
+
"content": "<tool_call>",
|
| 189 |
+
"lstrip": false,
|
| 190 |
+
"normalized": false,
|
| 191 |
+
"rstrip": false,
|
| 192 |
+
"single_word": false,
|
| 193 |
+
"special": false
|
| 194 |
+
},
|
| 195 |
+
"151353": {
|
| 196 |
+
"content": "</tool_call>",
|
| 197 |
+
"lstrip": false,
|
| 198 |
+
"normalized": false,
|
| 199 |
+
"rstrip": false,
|
| 200 |
+
"single_word": false,
|
| 201 |
+
"special": false
|
| 202 |
+
},
|
| 203 |
+
"151354": {
|
| 204 |
+
"content": "<tool_response>",
|
| 205 |
+
"lstrip": false,
|
| 206 |
+
"normalized": false,
|
| 207 |
+
"rstrip": false,
|
| 208 |
+
"single_word": false,
|
| 209 |
+
"special": false
|
| 210 |
+
},
|
| 211 |
+
"151355": {
|
| 212 |
+
"content": "</tool_response>",
|
| 213 |
+
"lstrip": false,
|
| 214 |
+
"normalized": false,
|
| 215 |
+
"rstrip": false,
|
| 216 |
+
"single_word": false,
|
| 217 |
+
"special": false
|
| 218 |
+
},
|
| 219 |
+
"151356": {
|
| 220 |
+
"content": "<arg_key>",
|
| 221 |
+
"lstrip": false,
|
| 222 |
+
"normalized": false,
|
| 223 |
+
"rstrip": false,
|
| 224 |
+
"single_word": false,
|
| 225 |
+
"special": false
|
| 226 |
+
},
|
| 227 |
+
"151357": {
|
| 228 |
+
"content": "</arg_key>",
|
| 229 |
+
"lstrip": false,
|
| 230 |
+
"normalized": false,
|
| 231 |
+
"rstrip": false,
|
| 232 |
+
"single_word": false,
|
| 233 |
+
"special": false
|
| 234 |
+
},
|
| 235 |
+
"151358": {
|
| 236 |
+
"content": "<arg_value>",
|
| 237 |
+
"lstrip": false,
|
| 238 |
+
"normalized": false,
|
| 239 |
+
"rstrip": false,
|
| 240 |
+
"single_word": false,
|
| 241 |
+
"special": false
|
| 242 |
+
},
|
| 243 |
+
"151359": {
|
| 244 |
+
"content": "</arg_value>",
|
| 245 |
+
"lstrip": false,
|
| 246 |
+
"normalized": false,
|
| 247 |
+
"rstrip": false,
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"special": false
|
| 250 |
+
},
|
| 251 |
+
"151360": {
|
| 252 |
+
"content": "/nothink",
|
| 253 |
+
"lstrip": false,
|
| 254 |
+
"normalized": false,
|
| 255 |
+
"rstrip": false,
|
| 256 |
+
"single_word": false,
|
| 257 |
+
"special": true
|
| 258 |
+
},
|
| 259 |
+
"151361": {
|
| 260 |
+
"content": "<|begin_of_box|>",
|
| 261 |
+
"lstrip": false,
|
| 262 |
+
"normalized": false,
|
| 263 |
+
"rstrip": false,
|
| 264 |
+
"single_word": false,
|
| 265 |
+
"special": false
|
| 266 |
+
},
|
| 267 |
+
"151362": {
|
| 268 |
+
"content": "<|end_of_box|>",
|
| 269 |
+
"lstrip": false,
|
| 270 |
+
"normalized": false,
|
| 271 |
+
"rstrip": false,
|
| 272 |
+
"single_word": false,
|
| 273 |
+
"special": false
|
| 274 |
+
},
|
| 275 |
+
"151363": {
|
| 276 |
+
"content": "<|image|>",
|
| 277 |
+
"lstrip": false,
|
| 278 |
+
"normalized": false,
|
| 279 |
+
"rstrip": false,
|
| 280 |
+
"single_word": false,
|
| 281 |
+
"special": true
|
| 282 |
+
},
|
| 283 |
+
"151364": {
|
| 284 |
+
"content": "<|video|>",
|
| 285 |
+
"lstrip": false,
|
| 286 |
+
"normalized": false,
|
| 287 |
+
"rstrip": false,
|
| 288 |
+
"single_word": false,
|
| 289 |
+
"special": true
|
| 290 |
+
}
|
| 291 |
+
},
|
| 292 |
+
"additional_special_tokens": [
|
| 293 |
+
"<|endoftext|>",
|
| 294 |
+
"[MASK]",
|
| 295 |
+
"[gMASK]",
|
| 296 |
+
"[sMASK]",
|
| 297 |
+
"<sop>",
|
| 298 |
+
"<eop>",
|
| 299 |
+
"<|system|>",
|
| 300 |
+
"<|user|>",
|
| 301 |
+
"<|assistant|>",
|
| 302 |
+
"<|observation|>",
|
| 303 |
+
"<|begin_of_image|>",
|
| 304 |
+
"<|end_of_image|>",
|
| 305 |
+
"<|begin_of_video|>",
|
| 306 |
+
"<|end_of_video|>",
|
| 307 |
+
"<|begin_of_audio|>",
|
| 308 |
+
"<|end_of_audio|>",
|
| 309 |
+
"<|image|>",
|
| 310 |
+
"<|video|>",
|
| 311 |
+
"<|begin_of_transcription|>",
|
| 312 |
+
"<|end_of_transcription|>",
|
| 313 |
+
"<|code_prefix|>",
|
| 314 |
+
"<|code_middle|>",
|
| 315 |
+
"<|code_suffix|>",
|
| 316 |
+
"/nothink"
|
| 317 |
+
],
|
| 318 |
+
"clean_up_tokenization_spaces": false,
|
| 319 |
+
"do_lower_case": false,
|
| 320 |
+
"eos_token": "<|endoftext|>",
|
| 321 |
+
"extra_special_tokens": {},
|
| 322 |
+
"model_max_length": 128000,
|
| 323 |
+
"pad_token": "<|endoftext|>",
|
| 324 |
+
"padding_side": "left",
|
| 325 |
+
"remove_space": false,
|
| 326 |
+
"tokenizer_class": "PreTrainedTokenizerFast"
|
| 327 |
+
}
|
video_preprocessor_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"size": {"shortest_edge": 12544, "longest_edge": 47040000},
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"patch_size": 14,
|
| 5 |
+
"temporal_patch_size": 2,
|
| 6 |
+
"merge_size": 2,
|
| 7 |
+
"image_mean": [0.48145466, 0.4578275, 0.40821073],
|
| 8 |
+
"image_std": [0.26862954, 0.26130258, 0.27577711],
|
| 9 |
+
"video_processor_type": "Glm46VVideoProcessor",
|
| 10 |
+
"processor_class": "Glm46VProcessor"
|
| 11 |
+
}
|