Upload folder using huggingface_hub
Browse files- chat_template.jinja +7 -0
- config.json +236 -165
- generation_config.json +1 -1
- model.safetensors.index.json +1 -0
- processor_config.json +86 -0
- recipe.yaml +1 -1
- tokenizer_config.json +0 -1
chat_template.jinja
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
|
| 2 |
+
You are a helpful assistant.<|im_end|>
|
| 3 |
+
{% endif %}<|im_start|>{{ message['role'] }}
|
| 4 |
+
{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
|
| 5 |
+
{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
|
| 6 |
+
{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
|
| 7 |
+
{% endif %}
|
config.json
CHANGED
|
@@ -4,6 +4,7 @@
|
|
| 4 |
],
|
| 5 |
"attention_dropout": 0.0,
|
| 6 |
"bos_token_id": 151643,
|
|
|
|
| 7 |
"eos_token_id": 151645,
|
| 8 |
"hidden_act": "silu",
|
| 9 |
"hidden_size": 3584,
|
|
@@ -19,6 +20,7 @@
|
|
| 19 |
"quantization_config": {
|
| 20 |
"config_groups": {
|
| 21 |
"group_0": {
|
|
|
|
| 22 |
"input_activations": {
|
| 23 |
"actorder": null,
|
| 24 |
"block_structure": null,
|
|
@@ -52,173 +54,176 @@
|
|
| 52 |
"format": "float-quantized",
|
| 53 |
"global_compression_ratio": null,
|
| 54 |
"ignore": [
|
| 55 |
-
"visual.blocks.0.attn.qkv",
|
| 56 |
-
"visual.blocks.0.attn.proj",
|
| 57 |
-
"visual.blocks.0.mlp.gate_proj",
|
| 58 |
-
"visual.blocks.0.mlp.up_proj",
|
| 59 |
-
"visual.blocks.0.mlp.down_proj",
|
| 60 |
-
"visual.blocks.1.attn.qkv",
|
| 61 |
-
"visual.blocks.1.attn.proj",
|
| 62 |
-
"visual.blocks.1.mlp.gate_proj",
|
| 63 |
-
"visual.blocks.1.mlp.up_proj",
|
| 64 |
-
"visual.blocks.1.mlp.down_proj",
|
| 65 |
-
"visual.blocks.2.attn.qkv",
|
| 66 |
-
"visual.blocks.2.attn.proj",
|
| 67 |
-
"visual.blocks.2.mlp.gate_proj",
|
| 68 |
-
"visual.blocks.2.mlp.up_proj",
|
| 69 |
-
"visual.blocks.2.mlp.down_proj",
|
| 70 |
-
"visual.blocks.3.attn.qkv",
|
| 71 |
-
"visual.blocks.3.attn.proj",
|
| 72 |
-
"visual.blocks.3.mlp.gate_proj",
|
| 73 |
-
"visual.blocks.3.mlp.up_proj",
|
| 74 |
-
"visual.blocks.3.mlp.down_proj",
|
| 75 |
-
"visual.blocks.4.attn.qkv",
|
| 76 |
-
"visual.blocks.4.attn.proj",
|
| 77 |
-
"visual.blocks.4.mlp.gate_proj",
|
| 78 |
-
"visual.blocks.4.mlp.up_proj",
|
| 79 |
-
"visual.blocks.4.mlp.down_proj",
|
| 80 |
-
"visual.blocks.5.attn.qkv",
|
| 81 |
-
"visual.blocks.5.attn.proj",
|
| 82 |
-
"visual.blocks.5.mlp.gate_proj",
|
| 83 |
-
"visual.blocks.5.mlp.up_proj",
|
| 84 |
-
"visual.blocks.5.mlp.down_proj",
|
| 85 |
-
"visual.blocks.6.attn.qkv",
|
| 86 |
-
"visual.blocks.6.attn.proj",
|
| 87 |
-
"visual.blocks.6.mlp.gate_proj",
|
| 88 |
-
"visual.blocks.6.mlp.up_proj",
|
| 89 |
-
"visual.blocks.6.mlp.down_proj",
|
| 90 |
-
"visual.blocks.7.attn.qkv",
|
| 91 |
-
"visual.blocks.7.attn.proj",
|
| 92 |
-
"visual.blocks.7.mlp.gate_proj",
|
| 93 |
-
"visual.blocks.7.mlp.up_proj",
|
| 94 |
-
"visual.blocks.7.mlp.down_proj",
|
| 95 |
-
"visual.blocks.8.attn.qkv",
|
| 96 |
-
"visual.blocks.8.attn.proj",
|
| 97 |
-
"visual.blocks.8.mlp.gate_proj",
|
| 98 |
-
"visual.blocks.8.mlp.up_proj",
|
| 99 |
-
"visual.blocks.8.mlp.down_proj",
|
| 100 |
-
"visual.blocks.9.attn.qkv",
|
| 101 |
-
"visual.blocks.9.attn.proj",
|
| 102 |
-
"visual.blocks.9.mlp.gate_proj",
|
| 103 |
-
"visual.blocks.9.mlp.up_proj",
|
| 104 |
-
"visual.blocks.9.mlp.down_proj",
|
| 105 |
-
"visual.blocks.10.attn.qkv",
|
| 106 |
-
"visual.blocks.10.attn.proj",
|
| 107 |
-
"visual.blocks.10.mlp.gate_proj",
|
| 108 |
-
"visual.blocks.10.mlp.up_proj",
|
| 109 |
-
"visual.blocks.10.mlp.down_proj",
|
| 110 |
-
"visual.blocks.11.attn.qkv",
|
| 111 |
-
"visual.blocks.11.attn.proj",
|
| 112 |
-
"visual.blocks.11.mlp.gate_proj",
|
| 113 |
-
"visual.blocks.11.mlp.up_proj",
|
| 114 |
-
"visual.blocks.11.mlp.down_proj",
|
| 115 |
-
"visual.blocks.12.attn.qkv",
|
| 116 |
-
"visual.blocks.12.attn.proj",
|
| 117 |
-
"visual.blocks.12.mlp.gate_proj",
|
| 118 |
-
"visual.blocks.12.mlp.up_proj",
|
| 119 |
-
"visual.blocks.12.mlp.down_proj",
|
| 120 |
-
"visual.blocks.13.attn.qkv",
|
| 121 |
-
"visual.blocks.13.attn.proj",
|
| 122 |
-
"visual.blocks.13.mlp.gate_proj",
|
| 123 |
-
"visual.blocks.13.mlp.up_proj",
|
| 124 |
-
"visual.blocks.13.mlp.down_proj",
|
| 125 |
-
"visual.blocks.14.attn.qkv",
|
| 126 |
-
"visual.blocks.14.attn.proj",
|
| 127 |
-
"visual.blocks.14.mlp.gate_proj",
|
| 128 |
-
"visual.blocks.14.mlp.up_proj",
|
| 129 |
-
"visual.blocks.14.mlp.down_proj",
|
| 130 |
-
"visual.blocks.15.attn.qkv",
|
| 131 |
-
"visual.blocks.15.attn.proj",
|
| 132 |
-
"visual.blocks.15.mlp.gate_proj",
|
| 133 |
-
"visual.blocks.15.mlp.up_proj",
|
| 134 |
-
"visual.blocks.15.mlp.down_proj",
|
| 135 |
-
"visual.blocks.16.attn.qkv",
|
| 136 |
-
"visual.blocks.16.attn.proj",
|
| 137 |
-
"visual.blocks.16.mlp.gate_proj",
|
| 138 |
-
"visual.blocks.16.mlp.up_proj",
|
| 139 |
-
"visual.blocks.16.mlp.down_proj",
|
| 140 |
-
"visual.blocks.17.attn.qkv",
|
| 141 |
-
"visual.blocks.17.attn.proj",
|
| 142 |
-
"visual.blocks.17.mlp.gate_proj",
|
| 143 |
-
"visual.blocks.17.mlp.up_proj",
|
| 144 |
-
"visual.blocks.17.mlp.down_proj",
|
| 145 |
-
"visual.blocks.18.attn.qkv",
|
| 146 |
-
"visual.blocks.18.attn.proj",
|
| 147 |
-
"visual.blocks.18.mlp.gate_proj",
|
| 148 |
-
"visual.blocks.18.mlp.up_proj",
|
| 149 |
-
"visual.blocks.18.mlp.down_proj",
|
| 150 |
-
"visual.blocks.19.attn.qkv",
|
| 151 |
-
"visual.blocks.19.attn.proj",
|
| 152 |
-
"visual.blocks.19.mlp.gate_proj",
|
| 153 |
-
"visual.blocks.19.mlp.up_proj",
|
| 154 |
-
"visual.blocks.19.mlp.down_proj",
|
| 155 |
-
"visual.blocks.20.attn.qkv",
|
| 156 |
-
"visual.blocks.20.attn.proj",
|
| 157 |
-
"visual.blocks.20.mlp.gate_proj",
|
| 158 |
-
"visual.blocks.20.mlp.up_proj",
|
| 159 |
-
"visual.blocks.20.mlp.down_proj",
|
| 160 |
-
"visual.blocks.21.attn.qkv",
|
| 161 |
-
"visual.blocks.21.attn.proj",
|
| 162 |
-
"visual.blocks.21.mlp.gate_proj",
|
| 163 |
-
"visual.blocks.21.mlp.up_proj",
|
| 164 |
-
"visual.blocks.21.mlp.down_proj",
|
| 165 |
-
"visual.blocks.22.attn.qkv",
|
| 166 |
-
"visual.blocks.22.attn.proj",
|
| 167 |
-
"visual.blocks.22.mlp.gate_proj",
|
| 168 |
-
"visual.blocks.22.mlp.up_proj",
|
| 169 |
-
"visual.blocks.22.mlp.down_proj",
|
| 170 |
-
"visual.blocks.23.attn.qkv",
|
| 171 |
-
"visual.blocks.23.attn.proj",
|
| 172 |
-
"visual.blocks.23.mlp.gate_proj",
|
| 173 |
-
"visual.blocks.23.mlp.up_proj",
|
| 174 |
-
"visual.blocks.23.mlp.down_proj",
|
| 175 |
-
"visual.blocks.24.attn.qkv",
|
| 176 |
-
"visual.blocks.24.attn.proj",
|
| 177 |
-
"visual.blocks.24.mlp.gate_proj",
|
| 178 |
-
"visual.blocks.24.mlp.up_proj",
|
| 179 |
-
"visual.blocks.24.mlp.down_proj",
|
| 180 |
-
"visual.blocks.25.attn.qkv",
|
| 181 |
-
"visual.blocks.25.attn.proj",
|
| 182 |
-
"visual.blocks.25.mlp.gate_proj",
|
| 183 |
-
"visual.blocks.25.mlp.up_proj",
|
| 184 |
-
"visual.blocks.25.mlp.down_proj",
|
| 185 |
-
"visual.blocks.26.attn.qkv",
|
| 186 |
-
"visual.blocks.26.attn.proj",
|
| 187 |
-
"visual.blocks.26.mlp.gate_proj",
|
| 188 |
-
"visual.blocks.26.mlp.up_proj",
|
| 189 |
-
"visual.blocks.26.mlp.down_proj",
|
| 190 |
-
"visual.blocks.27.attn.qkv",
|
| 191 |
-
"visual.blocks.27.attn.proj",
|
| 192 |
-
"visual.blocks.27.mlp.gate_proj",
|
| 193 |
-
"visual.blocks.27.mlp.up_proj",
|
| 194 |
-
"visual.blocks.27.mlp.down_proj",
|
| 195 |
-
"visual.blocks.28.attn.qkv",
|
| 196 |
-
"visual.blocks.28.attn.proj",
|
| 197 |
-
"visual.blocks.28.mlp.gate_proj",
|
| 198 |
-
"visual.blocks.28.mlp.up_proj",
|
| 199 |
-
"visual.blocks.28.mlp.down_proj",
|
| 200 |
-
"visual.blocks.29.attn.qkv",
|
| 201 |
-
"visual.blocks.29.attn.proj",
|
| 202 |
-
"visual.blocks.29.mlp.gate_proj",
|
| 203 |
-
"visual.blocks.29.mlp.up_proj",
|
| 204 |
-
"visual.blocks.29.mlp.down_proj",
|
| 205 |
-
"visual.blocks.30.attn.qkv",
|
| 206 |
-
"visual.blocks.30.attn.proj",
|
| 207 |
-
"visual.blocks.30.mlp.gate_proj",
|
| 208 |
-
"visual.blocks.30.mlp.up_proj",
|
| 209 |
-
"visual.blocks.30.mlp.down_proj",
|
| 210 |
-
"visual.blocks.31.attn.qkv",
|
| 211 |
-
"visual.blocks.31.attn.proj",
|
| 212 |
-
"visual.blocks.31.mlp.gate_proj",
|
| 213 |
-
"visual.blocks.31.mlp.up_proj",
|
| 214 |
-
"visual.blocks.31.mlp.down_proj",
|
| 215 |
-
"visual.merger.mlp.0",
|
| 216 |
-
"visual.merger.mlp.2",
|
| 217 |
"lm_head"
|
| 218 |
],
|
| 219 |
"kv_cache_scheme": null,
|
| 220 |
"quant_method": "compressed-tensors",
|
| 221 |
-
"quantization_status": "compressed"
|
|
|
|
|
|
|
|
|
|
| 222 |
},
|
| 223 |
"rms_norm_eps": 1e-06,
|
| 224 |
"rope_scaling": {
|
|
@@ -232,9 +237,74 @@
|
|
| 232 |
},
|
| 233 |
"rope_theta": 1000000.0,
|
| 234 |
"sliding_window": 32768,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 235 |
"tie_word_embeddings": false,
|
| 236 |
-
"
|
| 237 |
-
"transformers_version": "4.51.3",
|
| 238 |
"use_cache": true,
|
| 239 |
"use_sliding_window": false,
|
| 240 |
"video_token_id": 151656,
|
|
@@ -250,6 +320,7 @@
|
|
| 250 |
"hidden_size": 1280,
|
| 251 |
"in_channels": 3,
|
| 252 |
"in_chans": 3,
|
|
|
|
| 253 |
"intermediate_size": 3420,
|
| 254 |
"model_type": "qwen2_5_vl",
|
| 255 |
"num_heads": 16,
|
|
|
|
| 4 |
],
|
| 5 |
"attention_dropout": 0.0,
|
| 6 |
"bos_token_id": 151643,
|
| 7 |
+
"dtype": "bfloat16",
|
| 8 |
"eos_token_id": 151645,
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 3584,
|
|
|
|
| 20 |
"quantization_config": {
|
| 21 |
"config_groups": {
|
| 22 |
"group_0": {
|
| 23 |
+
"format": "float-quantized",
|
| 24 |
"input_activations": {
|
| 25 |
"actorder": null,
|
| 26 |
"block_structure": null,
|
|
|
|
| 54 |
"format": "float-quantized",
|
| 55 |
"global_compression_ratio": null,
|
| 56 |
"ignore": [
|
| 57 |
+
"model.visual.blocks.0.attn.qkv",
|
| 58 |
+
"model.visual.blocks.0.attn.proj",
|
| 59 |
+
"model.visual.blocks.0.mlp.gate_proj",
|
| 60 |
+
"model.visual.blocks.0.mlp.up_proj",
|
| 61 |
+
"model.visual.blocks.0.mlp.down_proj",
|
| 62 |
+
"model.visual.blocks.1.attn.qkv",
|
| 63 |
+
"model.visual.blocks.1.attn.proj",
|
| 64 |
+
"model.visual.blocks.1.mlp.gate_proj",
|
| 65 |
+
"model.visual.blocks.1.mlp.up_proj",
|
| 66 |
+
"model.visual.blocks.1.mlp.down_proj",
|
| 67 |
+
"model.visual.blocks.2.attn.qkv",
|
| 68 |
+
"model.visual.blocks.2.attn.proj",
|
| 69 |
+
"model.visual.blocks.2.mlp.gate_proj",
|
| 70 |
+
"model.visual.blocks.2.mlp.up_proj",
|
| 71 |
+
"model.visual.blocks.2.mlp.down_proj",
|
| 72 |
+
"model.visual.blocks.3.attn.qkv",
|
| 73 |
+
"model.visual.blocks.3.attn.proj",
|
| 74 |
+
"model.visual.blocks.3.mlp.gate_proj",
|
| 75 |
+
"model.visual.blocks.3.mlp.up_proj",
|
| 76 |
+
"model.visual.blocks.3.mlp.down_proj",
|
| 77 |
+
"model.visual.blocks.4.attn.qkv",
|
| 78 |
+
"model.visual.blocks.4.attn.proj",
|
| 79 |
+
"model.visual.blocks.4.mlp.gate_proj",
|
| 80 |
+
"model.visual.blocks.4.mlp.up_proj",
|
| 81 |
+
"model.visual.blocks.4.mlp.down_proj",
|
| 82 |
+
"model.visual.blocks.5.attn.qkv",
|
| 83 |
+
"model.visual.blocks.5.attn.proj",
|
| 84 |
+
"model.visual.blocks.5.mlp.gate_proj",
|
| 85 |
+
"model.visual.blocks.5.mlp.up_proj",
|
| 86 |
+
"model.visual.blocks.5.mlp.down_proj",
|
| 87 |
+
"model.visual.blocks.6.attn.qkv",
|
| 88 |
+
"model.visual.blocks.6.attn.proj",
|
| 89 |
+
"model.visual.blocks.6.mlp.gate_proj",
|
| 90 |
+
"model.visual.blocks.6.mlp.up_proj",
|
| 91 |
+
"model.visual.blocks.6.mlp.down_proj",
|
| 92 |
+
"model.visual.blocks.7.attn.qkv",
|
| 93 |
+
"model.visual.blocks.7.attn.proj",
|
| 94 |
+
"model.visual.blocks.7.mlp.gate_proj",
|
| 95 |
+
"model.visual.blocks.7.mlp.up_proj",
|
| 96 |
+
"model.visual.blocks.7.mlp.down_proj",
|
| 97 |
+
"model.visual.blocks.8.attn.qkv",
|
| 98 |
+
"model.visual.blocks.8.attn.proj",
|
| 99 |
+
"model.visual.blocks.8.mlp.gate_proj",
|
| 100 |
+
"model.visual.blocks.8.mlp.up_proj",
|
| 101 |
+
"model.visual.blocks.8.mlp.down_proj",
|
| 102 |
+
"model.visual.blocks.9.attn.qkv",
|
| 103 |
+
"model.visual.blocks.9.attn.proj",
|
| 104 |
+
"model.visual.blocks.9.mlp.gate_proj",
|
| 105 |
+
"model.visual.blocks.9.mlp.up_proj",
|
| 106 |
+
"model.visual.blocks.9.mlp.down_proj",
|
| 107 |
+
"model.visual.blocks.10.attn.qkv",
|
| 108 |
+
"model.visual.blocks.10.attn.proj",
|
| 109 |
+
"model.visual.blocks.10.mlp.gate_proj",
|
| 110 |
+
"model.visual.blocks.10.mlp.up_proj",
|
| 111 |
+
"model.visual.blocks.10.mlp.down_proj",
|
| 112 |
+
"model.visual.blocks.11.attn.qkv",
|
| 113 |
+
"model.visual.blocks.11.attn.proj",
|
| 114 |
+
"model.visual.blocks.11.mlp.gate_proj",
|
| 115 |
+
"model.visual.blocks.11.mlp.up_proj",
|
| 116 |
+
"model.visual.blocks.11.mlp.down_proj",
|
| 117 |
+
"model.visual.blocks.12.attn.qkv",
|
| 118 |
+
"model.visual.blocks.12.attn.proj",
|
| 119 |
+
"model.visual.blocks.12.mlp.gate_proj",
|
| 120 |
+
"model.visual.blocks.12.mlp.up_proj",
|
| 121 |
+
"model.visual.blocks.12.mlp.down_proj",
|
| 122 |
+
"model.visual.blocks.13.attn.qkv",
|
| 123 |
+
"model.visual.blocks.13.attn.proj",
|
| 124 |
+
"model.visual.blocks.13.mlp.gate_proj",
|
| 125 |
+
"model.visual.blocks.13.mlp.up_proj",
|
| 126 |
+
"model.visual.blocks.13.mlp.down_proj",
|
| 127 |
+
"model.visual.blocks.14.attn.qkv",
|
| 128 |
+
"model.visual.blocks.14.attn.proj",
|
| 129 |
+
"model.visual.blocks.14.mlp.gate_proj",
|
| 130 |
+
"model.visual.blocks.14.mlp.up_proj",
|
| 131 |
+
"model.visual.blocks.14.mlp.down_proj",
|
| 132 |
+
"model.visual.blocks.15.attn.qkv",
|
| 133 |
+
"model.visual.blocks.15.attn.proj",
|
| 134 |
+
"model.visual.blocks.15.mlp.gate_proj",
|
| 135 |
+
"model.visual.blocks.15.mlp.up_proj",
|
| 136 |
+
"model.visual.blocks.15.mlp.down_proj",
|
| 137 |
+
"model.visual.blocks.16.attn.qkv",
|
| 138 |
+
"model.visual.blocks.16.attn.proj",
|
| 139 |
+
"model.visual.blocks.16.mlp.gate_proj",
|
| 140 |
+
"model.visual.blocks.16.mlp.up_proj",
|
| 141 |
+
"model.visual.blocks.16.mlp.down_proj",
|
| 142 |
+
"model.visual.blocks.17.attn.qkv",
|
| 143 |
+
"model.visual.blocks.17.attn.proj",
|
| 144 |
+
"model.visual.blocks.17.mlp.gate_proj",
|
| 145 |
+
"model.visual.blocks.17.mlp.up_proj",
|
| 146 |
+
"model.visual.blocks.17.mlp.down_proj",
|
| 147 |
+
"model.visual.blocks.18.attn.qkv",
|
| 148 |
+
"model.visual.blocks.18.attn.proj",
|
| 149 |
+
"model.visual.blocks.18.mlp.gate_proj",
|
| 150 |
+
"model.visual.blocks.18.mlp.up_proj",
|
| 151 |
+
"model.visual.blocks.18.mlp.down_proj",
|
| 152 |
+
"model.visual.blocks.19.attn.qkv",
|
| 153 |
+
"model.visual.blocks.19.attn.proj",
|
| 154 |
+
"model.visual.blocks.19.mlp.gate_proj",
|
| 155 |
+
"model.visual.blocks.19.mlp.up_proj",
|
| 156 |
+
"model.visual.blocks.19.mlp.down_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",
|
| 163 |
+
"model.visual.blocks.21.attn.proj",
|
| 164 |
+
"model.visual.blocks.21.mlp.gate_proj",
|
| 165 |
+
"model.visual.blocks.21.mlp.up_proj",
|
| 166 |
+
"model.visual.blocks.21.mlp.down_proj",
|
| 167 |
+
"model.visual.blocks.22.attn.qkv",
|
| 168 |
+
"model.visual.blocks.22.attn.proj",
|
| 169 |
+
"model.visual.blocks.22.mlp.gate_proj",
|
| 170 |
+
"model.visual.blocks.22.mlp.up_proj",
|
| 171 |
+
"model.visual.blocks.22.mlp.down_proj",
|
| 172 |
+
"model.visual.blocks.23.attn.qkv",
|
| 173 |
+
"model.visual.blocks.23.attn.proj",
|
| 174 |
+
"model.visual.blocks.23.mlp.gate_proj",
|
| 175 |
+
"model.visual.blocks.23.mlp.up_proj",
|
| 176 |
+
"model.visual.blocks.23.mlp.down_proj",
|
| 177 |
+
"model.visual.blocks.24.attn.qkv",
|
| 178 |
+
"model.visual.blocks.24.attn.proj",
|
| 179 |
+
"model.visual.blocks.24.mlp.gate_proj",
|
| 180 |
+
"model.visual.blocks.24.mlp.up_proj",
|
| 181 |
+
"model.visual.blocks.24.mlp.down_proj",
|
| 182 |
+
"model.visual.blocks.25.attn.qkv",
|
| 183 |
+
"model.visual.blocks.25.attn.proj",
|
| 184 |
+
"model.visual.blocks.25.mlp.gate_proj",
|
| 185 |
+
"model.visual.blocks.25.mlp.up_proj",
|
| 186 |
+
"model.visual.blocks.25.mlp.down_proj",
|
| 187 |
+
"model.visual.blocks.26.attn.qkv",
|
| 188 |
+
"model.visual.blocks.26.attn.proj",
|
| 189 |
+
"model.visual.blocks.26.mlp.gate_proj",
|
| 190 |
+
"model.visual.blocks.26.mlp.up_proj",
|
| 191 |
+
"model.visual.blocks.26.mlp.down_proj",
|
| 192 |
+
"model.visual.blocks.27.attn.qkv",
|
| 193 |
+
"model.visual.blocks.27.attn.proj",
|
| 194 |
+
"model.visual.blocks.27.mlp.gate_proj",
|
| 195 |
+
"model.visual.blocks.27.mlp.up_proj",
|
| 196 |
+
"model.visual.blocks.27.mlp.down_proj",
|
| 197 |
+
"model.visual.blocks.28.attn.qkv",
|
| 198 |
+
"model.visual.blocks.28.attn.proj",
|
| 199 |
+
"model.visual.blocks.28.mlp.gate_proj",
|
| 200 |
+
"model.visual.blocks.28.mlp.up_proj",
|
| 201 |
+
"model.visual.blocks.28.mlp.down_proj",
|
| 202 |
+
"model.visual.blocks.29.attn.qkv",
|
| 203 |
+
"model.visual.blocks.29.attn.proj",
|
| 204 |
+
"model.visual.blocks.29.mlp.gate_proj",
|
| 205 |
+
"model.visual.blocks.29.mlp.up_proj",
|
| 206 |
+
"model.visual.blocks.29.mlp.down_proj",
|
| 207 |
+
"model.visual.blocks.30.attn.qkv",
|
| 208 |
+
"model.visual.blocks.30.attn.proj",
|
| 209 |
+
"model.visual.blocks.30.mlp.gate_proj",
|
| 210 |
+
"model.visual.blocks.30.mlp.up_proj",
|
| 211 |
+
"model.visual.blocks.30.mlp.down_proj",
|
| 212 |
+
"model.visual.blocks.31.attn.qkv",
|
| 213 |
+
"model.visual.blocks.31.attn.proj",
|
| 214 |
+
"model.visual.blocks.31.mlp.gate_proj",
|
| 215 |
+
"model.visual.blocks.31.mlp.up_proj",
|
| 216 |
+
"model.visual.blocks.31.mlp.down_proj",
|
| 217 |
+
"model.visual.merger.mlp.0",
|
| 218 |
+
"model.visual.merger.mlp.2",
|
| 219 |
"lm_head"
|
| 220 |
],
|
| 221 |
"kv_cache_scheme": null,
|
| 222 |
"quant_method": "compressed-tensors",
|
| 223 |
+
"quantization_status": "compressed",
|
| 224 |
+
"sparsity_config": {},
|
| 225 |
+
"transform_config": {},
|
| 226 |
+
"version": "0.12.2"
|
| 227 |
},
|
| 228 |
"rms_norm_eps": 1e-06,
|
| 229 |
"rope_scaling": {
|
|
|
|
| 237 |
},
|
| 238 |
"rope_theta": 1000000.0,
|
| 239 |
"sliding_window": 32768,
|
| 240 |
+
"text_config": {
|
| 241 |
+
"_name_or_path": "Qwen/Qwen2.5-VL-7B-Instruct",
|
| 242 |
+
"architectures": [
|
| 243 |
+
"Qwen2_5_VLForConditionalGeneration"
|
| 244 |
+
],
|
| 245 |
+
"attention_dropout": 0.0,
|
| 246 |
+
"bos_token_id": 151643,
|
| 247 |
+
"dtype": "bfloat16",
|
| 248 |
+
"eos_token_id": 151645,
|
| 249 |
+
"hidden_act": "silu",
|
| 250 |
+
"hidden_size": 3584,
|
| 251 |
+
"initializer_range": 0.02,
|
| 252 |
+
"intermediate_size": 18944,
|
| 253 |
+
"layer_types": [
|
| 254 |
+
"full_attention",
|
| 255 |
+
"full_attention",
|
| 256 |
+
"full_attention",
|
| 257 |
+
"full_attention",
|
| 258 |
+
"full_attention",
|
| 259 |
+
"full_attention",
|
| 260 |
+
"full_attention",
|
| 261 |
+
"full_attention",
|
| 262 |
+
"full_attention",
|
| 263 |
+
"full_attention",
|
| 264 |
+
"full_attention",
|
| 265 |
+
"full_attention",
|
| 266 |
+
"full_attention",
|
| 267 |
+
"full_attention",
|
| 268 |
+
"full_attention",
|
| 269 |
+
"full_attention",
|
| 270 |
+
"full_attention",
|
| 271 |
+
"full_attention",
|
| 272 |
+
"full_attention",
|
| 273 |
+
"full_attention",
|
| 274 |
+
"full_attention",
|
| 275 |
+
"full_attention",
|
| 276 |
+
"full_attention",
|
| 277 |
+
"full_attention",
|
| 278 |
+
"full_attention",
|
| 279 |
+
"full_attention",
|
| 280 |
+
"full_attention",
|
| 281 |
+
"full_attention"
|
| 282 |
+
],
|
| 283 |
+
"max_position_embeddings": 128000,
|
| 284 |
+
"max_window_layers": 28,
|
| 285 |
+
"model_type": "qwen2_5_vl_text",
|
| 286 |
+
"num_attention_heads": 28,
|
| 287 |
+
"num_hidden_layers": 28,
|
| 288 |
+
"num_key_value_heads": 4,
|
| 289 |
+
"rms_norm_eps": 1e-06,
|
| 290 |
+
"rope_scaling": {
|
| 291 |
+
"mrope_section": [
|
| 292 |
+
16,
|
| 293 |
+
24,
|
| 294 |
+
24
|
| 295 |
+
],
|
| 296 |
+
"rope_type": "default",
|
| 297 |
+
"type": "default"
|
| 298 |
+
},
|
| 299 |
+
"rope_theta": 1000000.0,
|
| 300 |
+
"sliding_window": null,
|
| 301 |
+
"use_cache": true,
|
| 302 |
+
"use_sliding_window": false,
|
| 303 |
+
"vision_token_id": 151654,
|
| 304 |
+
"vocab_size": 152064
|
| 305 |
+
},
|
| 306 |
"tie_word_embeddings": false,
|
| 307 |
+
"transformers_version": "5.0.0.dev0",
|
|
|
|
| 308 |
"use_cache": true,
|
| 309 |
"use_sliding_window": false,
|
| 310 |
"video_token_id": 151656,
|
|
|
|
| 320 |
"hidden_size": 1280,
|
| 321 |
"in_channels": 3,
|
| 322 |
"in_chans": 3,
|
| 323 |
+
"initializer_range": 0.02,
|
| 324 |
"intermediate_size": 3420,
|
| 325 |
"model_type": "qwen2_5_vl",
|
| 326 |
"num_heads": 16,
|
generation_config.json
CHANGED
|
@@ -8,5 +8,5 @@
|
|
| 8 |
"pad_token_id": 151643,
|
| 9 |
"repetition_penalty": 1.05,
|
| 10 |
"temperature": 1e-06,
|
| 11 |
-
"transformers_version": "
|
| 12 |
}
|
|
|
|
| 8 |
"pad_token_id": 151643,
|
| 9 |
"repetition_penalty": 1.05,
|
| 10 |
"temperature": 1e-06,
|
| 11 |
+
"transformers_version": "5.0.0.dev0"
|
| 12 |
}
|
model.safetensors.index.json
CHANGED
|
@@ -1,5 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
|
|
|
| 3 |
"total_size": 10061826048
|
| 4 |
},
|
| 5 |
"weight_map": {
|
|
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
| 3 |
+
"total_parameters": 8293557248,
|
| 4 |
"total_size": 10061826048
|
| 5 |
},
|
| 6 |
"weight_map": {
|
processor_config.json
ADDED
|
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"image_processor": {
|
| 3 |
+
"crop_size": null,
|
| 4 |
+
"data_format": "channels_first",
|
| 5 |
+
"device": null,
|
| 6 |
+
"disable_grouping": null,
|
| 7 |
+
"do_center_crop": null,
|
| 8 |
+
"do_convert_rgb": true,
|
| 9 |
+
"do_normalize": true,
|
| 10 |
+
"do_pad": null,
|
| 11 |
+
"do_rescale": true,
|
| 12 |
+
"do_resize": true,
|
| 13 |
+
"image_mean": [
|
| 14 |
+
0.48145466,
|
| 15 |
+
0.4578275,
|
| 16 |
+
0.40821073
|
| 17 |
+
],
|
| 18 |
+
"image_processor_type": "Qwen2VLImageProcessorFast",
|
| 19 |
+
"image_std": [
|
| 20 |
+
0.26862954,
|
| 21 |
+
0.26130258,
|
| 22 |
+
0.27577711
|
| 23 |
+
],
|
| 24 |
+
"input_data_format": null,
|
| 25 |
+
"max_pixels": 12845056,
|
| 26 |
+
"merge_size": 2,
|
| 27 |
+
"min_pixels": 3136,
|
| 28 |
+
"pad_size": null,
|
| 29 |
+
"patch_size": 14,
|
| 30 |
+
"processor_class": "Qwen2_5_VLProcessor",
|
| 31 |
+
"resample": 3,
|
| 32 |
+
"rescale_factor": 0.00392156862745098,
|
| 33 |
+
"return_tensors": null,
|
| 34 |
+
"size": {
|
| 35 |
+
"longest_edge": 12845056,
|
| 36 |
+
"shortest_edge": 3136
|
| 37 |
+
},
|
| 38 |
+
"temporal_patch_size": 2
|
| 39 |
+
},
|
| 40 |
+
"processor_class": "Qwen2_5_VLProcessor",
|
| 41 |
+
"video_processor": {
|
| 42 |
+
"crop_size": null,
|
| 43 |
+
"data_format": "channels_first",
|
| 44 |
+
"default_to_square": true,
|
| 45 |
+
"device": null,
|
| 46 |
+
"do_center_crop": null,
|
| 47 |
+
"do_convert_rgb": true,
|
| 48 |
+
"do_normalize": true,
|
| 49 |
+
"do_pad": null,
|
| 50 |
+
"do_rescale": true,
|
| 51 |
+
"do_resize": true,
|
| 52 |
+
"do_sample_frames": false,
|
| 53 |
+
"fps": null,
|
| 54 |
+
"image_mean": [
|
| 55 |
+
0.48145466,
|
| 56 |
+
0.4578275,
|
| 57 |
+
0.40821073
|
| 58 |
+
],
|
| 59 |
+
"image_std": [
|
| 60 |
+
0.26862954,
|
| 61 |
+
0.26130258,
|
| 62 |
+
0.27577711
|
| 63 |
+
],
|
| 64 |
+
"input_data_format": null,
|
| 65 |
+
"max_frames": 768,
|
| 66 |
+
"max_pixels": 12845056,
|
| 67 |
+
"merge_size": 2,
|
| 68 |
+
"min_frames": 4,
|
| 69 |
+
"min_pixels": 3136,
|
| 70 |
+
"num_frames": null,
|
| 71 |
+
"pad_size": null,
|
| 72 |
+
"patch_size": 14,
|
| 73 |
+
"processor_class": "Qwen2_5_VLProcessor",
|
| 74 |
+
"resample": 3,
|
| 75 |
+
"rescale_factor": 0.00392156862745098,
|
| 76 |
+
"return_metadata": false,
|
| 77 |
+
"return_tensors": null,
|
| 78 |
+
"size": {
|
| 79 |
+
"longest_edge": 12845056,
|
| 80 |
+
"shortest_edge": 3136
|
| 81 |
+
},
|
| 82 |
+
"temporal_patch_size": 2,
|
| 83 |
+
"video_metadata": null,
|
| 84 |
+
"video_processor_type": "Qwen2VLVideoProcessor"
|
| 85 |
+
}
|
| 86 |
+
}
|
recipe.yaml
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
default_stage:
|
| 2 |
default_modifiers:
|
| 3 |
QuantizationModifier:
|
| 4 |
-
ignore: ['re:.*lm_head', 're:visual.*']
|
| 5 |
targets: [Linear]
|
|
|
|
| 6 |
scheme: FP8_DYNAMIC
|
|
|
|
| 1 |
default_stage:
|
| 2 |
default_modifiers:
|
| 3 |
QuantizationModifier:
|
|
|
|
| 4 |
targets: [Linear]
|
| 5 |
+
ignore: [lm_head, 're:visual.*', 're:model.visual.*']
|
| 6 |
scheme: FP8_DYNAMIC
|
tokenizer_config.json
CHANGED
|
@@ -195,7 +195,6 @@
|
|
| 195 |
"<|video_pad|>"
|
| 196 |
],
|
| 197 |
"bos_token": null,
|
| 198 |
-
"chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
|
| 199 |
"clean_up_tokenization_spaces": false,
|
| 200 |
"eos_token": "<|im_end|>",
|
| 201 |
"errors": "replace",
|
|
|
|
| 195 |
"<|video_pad|>"
|
| 196 |
],
|
| 197 |
"bos_token": null,
|
|
|
|
| 198 |
"clean_up_tokenization_spaces": false,
|
| 199 |
"eos_token": "<|im_end|>",
|
| 200 |
"errors": "replace",
|