Update index after adding input_scale params
Browse files- model.safetensors.index.json +194 -1
model.safetensors.index.json
CHANGED
|
@@ -1,9 +1,10 @@
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
| 3 |
"total_parameters": 18164649200,
|
| 4 |
-
"total_size":
|
| 5 |
},
|
| 6 |
"weight_map": {
|
|
|
|
| 7 |
"lm_head.weight": "model-00003-of-00003.safetensors",
|
| 8 |
"lm_head.weight_scale": "model-00003-of-00003.safetensors",
|
| 9 |
"lm_head.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
@@ -24,12 +25,15 @@
|
|
| 24 |
"model.language_model.layers.0.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 25 |
"model.language_model.layers.0.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 26 |
"model.language_model.layers.0.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 27 |
"model.language_model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 28 |
"model.language_model.layers.0.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 29 |
"model.language_model.layers.0.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 30 |
"model.language_model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 31 |
"model.language_model.layers.0.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 32 |
"model.language_model.layers.0.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 33 |
"model.language_model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 34 |
"model.language_model.layers.0.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 35 |
"model.language_model.layers.0.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -50,12 +54,15 @@
|
|
| 50 |
"model.language_model.layers.1.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 51 |
"model.language_model.layers.1.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 52 |
"model.language_model.layers.1.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 53 |
"model.language_model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 54 |
"model.language_model.layers.1.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 55 |
"model.language_model.layers.1.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 56 |
"model.language_model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 57 |
"model.language_model.layers.1.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 58 |
"model.language_model.layers.1.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 59 |
"model.language_model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 60 |
"model.language_model.layers.1.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 61 |
"model.language_model.layers.1.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -76,23 +83,29 @@
|
|
| 76 |
"model.language_model.layers.10.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 77 |
"model.language_model.layers.10.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 78 |
"model.language_model.layers.10.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 79 |
"model.language_model.layers.10.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 80 |
"model.language_model.layers.10.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 81 |
"model.language_model.layers.10.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 82 |
"model.language_model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 83 |
"model.language_model.layers.10.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 84 |
"model.language_model.layers.10.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 85 |
"model.language_model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 86 |
"model.language_model.layers.10.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 87 |
"model.language_model.layers.10.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 88 |
"model.language_model.layers.10.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 89 |
"model.language_model.layers.11.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 90 |
"model.language_model.layers.11.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 91 |
"model.language_model.layers.11.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 92 |
"model.language_model.layers.11.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 93 |
"model.language_model.layers.11.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 94 |
"model.language_model.layers.11.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 95 |
"model.language_model.layers.11.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 96 |
"model.language_model.layers.11.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 97 |
"model.language_model.layers.11.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 98 |
"model.language_model.layers.11.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -127,12 +140,15 @@
|
|
| 127 |
"model.language_model.layers.12.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 128 |
"model.language_model.layers.12.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 129 |
"model.language_model.layers.12.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 130 |
"model.language_model.layers.12.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 131 |
"model.language_model.layers.12.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 132 |
"model.language_model.layers.12.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 133 |
"model.language_model.layers.12.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 134 |
"model.language_model.layers.12.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 135 |
"model.language_model.layers.12.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 136 |
"model.language_model.layers.12.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 137 |
"model.language_model.layers.12.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 138 |
"model.language_model.layers.12.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -153,12 +169,15 @@
|
|
| 153 |
"model.language_model.layers.13.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 154 |
"model.language_model.layers.13.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 155 |
"model.language_model.layers.13.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 156 |
"model.language_model.layers.13.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 157 |
"model.language_model.layers.13.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 158 |
"model.language_model.layers.13.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 159 |
"model.language_model.layers.13.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 160 |
"model.language_model.layers.13.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 161 |
"model.language_model.layers.13.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 162 |
"model.language_model.layers.13.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 163 |
"model.language_model.layers.13.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 164 |
"model.language_model.layers.13.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -179,23 +198,29 @@
|
|
| 179 |
"model.language_model.layers.14.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 180 |
"model.language_model.layers.14.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 181 |
"model.language_model.layers.14.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 182 |
"model.language_model.layers.14.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 183 |
"model.language_model.layers.14.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 184 |
"model.language_model.layers.14.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 185 |
"model.language_model.layers.14.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 186 |
"model.language_model.layers.14.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 187 |
"model.language_model.layers.14.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 188 |
"model.language_model.layers.14.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 189 |
"model.language_model.layers.14.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 190 |
"model.language_model.layers.14.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 191 |
"model.language_model.layers.14.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 192 |
"model.language_model.layers.15.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 193 |
"model.language_model.layers.15.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 194 |
"model.language_model.layers.15.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 195 |
"model.language_model.layers.15.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 196 |
"model.language_model.layers.15.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 197 |
"model.language_model.layers.15.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 198 |
"model.language_model.layers.15.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 199 |
"model.language_model.layers.15.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 200 |
"model.language_model.layers.15.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 201 |
"model.language_model.layers.15.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -230,12 +255,15 @@
|
|
| 230 |
"model.language_model.layers.16.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 231 |
"model.language_model.layers.16.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 232 |
"model.language_model.layers.16.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 233 |
"model.language_model.layers.16.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 234 |
"model.language_model.layers.16.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 235 |
"model.language_model.layers.16.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 236 |
"model.language_model.layers.16.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 237 |
"model.language_model.layers.16.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 238 |
"model.language_model.layers.16.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 239 |
"model.language_model.layers.16.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 240 |
"model.language_model.layers.16.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 241 |
"model.language_model.layers.16.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -256,12 +284,15 @@
|
|
| 256 |
"model.language_model.layers.17.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 257 |
"model.language_model.layers.17.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 258 |
"model.language_model.layers.17.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 259 |
"model.language_model.layers.17.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 260 |
"model.language_model.layers.17.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 261 |
"model.language_model.layers.17.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 262 |
"model.language_model.layers.17.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 263 |
"model.language_model.layers.17.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 264 |
"model.language_model.layers.17.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 265 |
"model.language_model.layers.17.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 266 |
"model.language_model.layers.17.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 267 |
"model.language_model.layers.17.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -282,23 +313,29 @@
|
|
| 282 |
"model.language_model.layers.18.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 283 |
"model.language_model.layers.18.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 284 |
"model.language_model.layers.18.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 285 |
"model.language_model.layers.18.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 286 |
"model.language_model.layers.18.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 287 |
"model.language_model.layers.18.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 288 |
"model.language_model.layers.18.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 289 |
"model.language_model.layers.18.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 290 |
"model.language_model.layers.18.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 291 |
"model.language_model.layers.18.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 292 |
"model.language_model.layers.18.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 293 |
"model.language_model.layers.18.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 294 |
"model.language_model.layers.18.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 295 |
"model.language_model.layers.19.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 296 |
"model.language_model.layers.19.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 297 |
"model.language_model.layers.19.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 298 |
"model.language_model.layers.19.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 299 |
"model.language_model.layers.19.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 300 |
"model.language_model.layers.19.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 301 |
"model.language_model.layers.19.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 302 |
"model.language_model.layers.19.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 303 |
"model.language_model.layers.19.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 304 |
"model.language_model.layers.19.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -333,12 +370,15 @@
|
|
| 333 |
"model.language_model.layers.2.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 334 |
"model.language_model.layers.2.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 335 |
"model.language_model.layers.2.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 336 |
"model.language_model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 337 |
"model.language_model.layers.2.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 338 |
"model.language_model.layers.2.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 339 |
"model.language_model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 340 |
"model.language_model.layers.2.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 341 |
"model.language_model.layers.2.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 342 |
"model.language_model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 343 |
"model.language_model.layers.2.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 344 |
"model.language_model.layers.2.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -359,12 +399,15 @@
|
|
| 359 |
"model.language_model.layers.20.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 360 |
"model.language_model.layers.20.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 361 |
"model.language_model.layers.20.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 362 |
"model.language_model.layers.20.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 363 |
"model.language_model.layers.20.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 364 |
"model.language_model.layers.20.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 365 |
"model.language_model.layers.20.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 366 |
"model.language_model.layers.20.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 367 |
"model.language_model.layers.20.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 368 |
"model.language_model.layers.20.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 369 |
"model.language_model.layers.20.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 370 |
"model.language_model.layers.20.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -385,12 +428,15 @@
|
|
| 385 |
"model.language_model.layers.21.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 386 |
"model.language_model.layers.21.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 387 |
"model.language_model.layers.21.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 388 |
"model.language_model.layers.21.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 389 |
"model.language_model.layers.21.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 390 |
"model.language_model.layers.21.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 391 |
"model.language_model.layers.21.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 392 |
"model.language_model.layers.21.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 393 |
"model.language_model.layers.21.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 394 |
"model.language_model.layers.21.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 395 |
"model.language_model.layers.21.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 396 |
"model.language_model.layers.21.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -411,23 +457,29 @@
|
|
| 411 |
"model.language_model.layers.22.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 412 |
"model.language_model.layers.22.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 413 |
"model.language_model.layers.22.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 414 |
"model.language_model.layers.22.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 415 |
"model.language_model.layers.22.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 416 |
"model.language_model.layers.22.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 417 |
"model.language_model.layers.22.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 418 |
"model.language_model.layers.22.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 419 |
"model.language_model.layers.22.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 420 |
"model.language_model.layers.22.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 421 |
"model.language_model.layers.22.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 422 |
"model.language_model.layers.22.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 423 |
"model.language_model.layers.22.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 424 |
"model.language_model.layers.23.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 425 |
"model.language_model.layers.23.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 426 |
"model.language_model.layers.23.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 427 |
"model.language_model.layers.23.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 428 |
"model.language_model.layers.23.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 429 |
"model.language_model.layers.23.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 430 |
"model.language_model.layers.23.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 431 |
"model.language_model.layers.23.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 432 |
"model.language_model.layers.23.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 433 |
"model.language_model.layers.23.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -462,12 +514,15 @@
|
|
| 462 |
"model.language_model.layers.24.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 463 |
"model.language_model.layers.24.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 464 |
"model.language_model.layers.24.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 465 |
"model.language_model.layers.24.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 466 |
"model.language_model.layers.24.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 467 |
"model.language_model.layers.24.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 468 |
"model.language_model.layers.24.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 469 |
"model.language_model.layers.24.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 470 |
"model.language_model.layers.24.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 471 |
"model.language_model.layers.24.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 472 |
"model.language_model.layers.24.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 473 |
"model.language_model.layers.24.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -488,12 +543,15 @@
|
|
| 488 |
"model.language_model.layers.25.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 489 |
"model.language_model.layers.25.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 490 |
"model.language_model.layers.25.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 491 |
"model.language_model.layers.25.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 492 |
"model.language_model.layers.25.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 493 |
"model.language_model.layers.25.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 494 |
"model.language_model.layers.25.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 495 |
"model.language_model.layers.25.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 496 |
"model.language_model.layers.25.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 497 |
"model.language_model.layers.25.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 498 |
"model.language_model.layers.25.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 499 |
"model.language_model.layers.25.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -514,23 +572,29 @@
|
|
| 514 |
"model.language_model.layers.26.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 515 |
"model.language_model.layers.26.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 516 |
"model.language_model.layers.26.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 517 |
"model.language_model.layers.26.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 518 |
"model.language_model.layers.26.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 519 |
"model.language_model.layers.26.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 520 |
"model.language_model.layers.26.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 521 |
"model.language_model.layers.26.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 522 |
"model.language_model.layers.26.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 523 |
"model.language_model.layers.26.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 524 |
"model.language_model.layers.26.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 525 |
"model.language_model.layers.26.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 526 |
"model.language_model.layers.26.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 527 |
"model.language_model.layers.27.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 528 |
"model.language_model.layers.27.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 529 |
"model.language_model.layers.27.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 530 |
"model.language_model.layers.27.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 531 |
"model.language_model.layers.27.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 532 |
"model.language_model.layers.27.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 533 |
"model.language_model.layers.27.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 534 |
"model.language_model.layers.27.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 535 |
"model.language_model.layers.27.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 536 |
"model.language_model.layers.27.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -565,12 +629,15 @@
|
|
| 565 |
"model.language_model.layers.28.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 566 |
"model.language_model.layers.28.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 567 |
"model.language_model.layers.28.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 568 |
"model.language_model.layers.28.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 569 |
"model.language_model.layers.28.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 570 |
"model.language_model.layers.28.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 571 |
"model.language_model.layers.28.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 572 |
"model.language_model.layers.28.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 573 |
"model.language_model.layers.28.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 574 |
"model.language_model.layers.28.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 575 |
"model.language_model.layers.28.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 576 |
"model.language_model.layers.28.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -591,23 +658,29 @@
|
|
| 591 |
"model.language_model.layers.29.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 592 |
"model.language_model.layers.29.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 593 |
"model.language_model.layers.29.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 594 |
"model.language_model.layers.29.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 595 |
"model.language_model.layers.29.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 596 |
"model.language_model.layers.29.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 597 |
"model.language_model.layers.29.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 598 |
"model.language_model.layers.29.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 599 |
"model.language_model.layers.29.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 600 |
"model.language_model.layers.29.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 601 |
"model.language_model.layers.29.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 602 |
"model.language_model.layers.29.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 603 |
"model.language_model.layers.29.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 604 |
"model.language_model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 605 |
"model.language_model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 606 |
"model.language_model.layers.3.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 607 |
"model.language_model.layers.3.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 608 |
"model.language_model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 609 |
"model.language_model.layers.3.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 610 |
"model.language_model.layers.3.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 611 |
"model.language_model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 612 |
"model.language_model.layers.3.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 613 |
"model.language_model.layers.3.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -642,23 +715,29 @@
|
|
| 642 |
"model.language_model.layers.30.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 643 |
"model.language_model.layers.30.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 644 |
"model.language_model.layers.30.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 645 |
"model.language_model.layers.30.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 646 |
"model.language_model.layers.30.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 647 |
"model.language_model.layers.30.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 648 |
"model.language_model.layers.30.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 649 |
"model.language_model.layers.30.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 650 |
"model.language_model.layers.30.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 651 |
"model.language_model.layers.30.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 652 |
"model.language_model.layers.30.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 653 |
"model.language_model.layers.30.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 654 |
"model.language_model.layers.30.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 655 |
"model.language_model.layers.31.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 656 |
"model.language_model.layers.31.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 657 |
"model.language_model.layers.31.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 658 |
"model.language_model.layers.31.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 659 |
"model.language_model.layers.31.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 660 |
"model.language_model.layers.31.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 661 |
"model.language_model.layers.31.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 662 |
"model.language_model.layers.31.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 663 |
"model.language_model.layers.31.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 664 |
"model.language_model.layers.31.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -693,12 +772,15 @@
|
|
| 693 |
"model.language_model.layers.32.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 694 |
"model.language_model.layers.32.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 695 |
"model.language_model.layers.32.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 696 |
"model.language_model.layers.32.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 697 |
"model.language_model.layers.32.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 698 |
"model.language_model.layers.32.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 699 |
"model.language_model.layers.32.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 700 |
"model.language_model.layers.32.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 701 |
"model.language_model.layers.32.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 702 |
"model.language_model.layers.32.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 703 |
"model.language_model.layers.32.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 704 |
"model.language_model.layers.32.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -719,12 +801,15 @@
|
|
| 719 |
"model.language_model.layers.33.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 720 |
"model.language_model.layers.33.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 721 |
"model.language_model.layers.33.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 722 |
"model.language_model.layers.33.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 723 |
"model.language_model.layers.33.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 724 |
"model.language_model.layers.33.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 725 |
"model.language_model.layers.33.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 726 |
"model.language_model.layers.33.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 727 |
"model.language_model.layers.33.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 728 |
"model.language_model.layers.33.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 729 |
"model.language_model.layers.33.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 730 |
"model.language_model.layers.33.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -745,23 +830,29 @@
|
|
| 745 |
"model.language_model.layers.34.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 746 |
"model.language_model.layers.34.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 747 |
"model.language_model.layers.34.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 748 |
"model.language_model.layers.34.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 749 |
"model.language_model.layers.34.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 750 |
"model.language_model.layers.34.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 751 |
"model.language_model.layers.34.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 752 |
"model.language_model.layers.34.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 753 |
"model.language_model.layers.34.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 754 |
"model.language_model.layers.34.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 755 |
"model.language_model.layers.34.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 756 |
"model.language_model.layers.34.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 757 |
"model.language_model.layers.34.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 758 |
"model.language_model.layers.35.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 759 |
"model.language_model.layers.35.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 760 |
"model.language_model.layers.35.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 761 |
"model.language_model.layers.35.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 762 |
"model.language_model.layers.35.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 763 |
"model.language_model.layers.35.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 764 |
"model.language_model.layers.35.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 765 |
"model.language_model.layers.35.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 766 |
"model.language_model.layers.35.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 767 |
"model.language_model.layers.35.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -796,12 +887,15 @@
|
|
| 796 |
"model.language_model.layers.36.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 797 |
"model.language_model.layers.36.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 798 |
"model.language_model.layers.36.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 799 |
"model.language_model.layers.36.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 800 |
"model.language_model.layers.36.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 801 |
"model.language_model.layers.36.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 802 |
"model.language_model.layers.36.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 803 |
"model.language_model.layers.36.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 804 |
"model.language_model.layers.36.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 805 |
"model.language_model.layers.36.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 806 |
"model.language_model.layers.36.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 807 |
"model.language_model.layers.36.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -822,12 +916,15 @@
|
|
| 822 |
"model.language_model.layers.37.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 823 |
"model.language_model.layers.37.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 824 |
"model.language_model.layers.37.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 825 |
"model.language_model.layers.37.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 826 |
"model.language_model.layers.37.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 827 |
"model.language_model.layers.37.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 828 |
"model.language_model.layers.37.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 829 |
"model.language_model.layers.37.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 830 |
"model.language_model.layers.37.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 831 |
"model.language_model.layers.37.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 832 |
"model.language_model.layers.37.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 833 |
"model.language_model.layers.37.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -848,23 +945,29 @@
|
|
| 848 |
"model.language_model.layers.38.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 849 |
"model.language_model.layers.38.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 850 |
"model.language_model.layers.38.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 851 |
"model.language_model.layers.38.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 852 |
"model.language_model.layers.38.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 853 |
"model.language_model.layers.38.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 854 |
"model.language_model.layers.38.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 855 |
"model.language_model.layers.38.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 856 |
"model.language_model.layers.38.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 857 |
"model.language_model.layers.38.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 858 |
"model.language_model.layers.38.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 859 |
"model.language_model.layers.38.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 860 |
"model.language_model.layers.38.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 861 |
"model.language_model.layers.39.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 862 |
"model.language_model.layers.39.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 863 |
"model.language_model.layers.39.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 864 |
"model.language_model.layers.39.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 865 |
"model.language_model.layers.39.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 866 |
"model.language_model.layers.39.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 867 |
"model.language_model.layers.39.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 868 |
"model.language_model.layers.39.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 869 |
"model.language_model.layers.39.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 870 |
"model.language_model.layers.39.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -899,12 +1002,15 @@
|
|
| 899 |
"model.language_model.layers.4.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 900 |
"model.language_model.layers.4.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 901 |
"model.language_model.layers.4.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 902 |
"model.language_model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 903 |
"model.language_model.layers.4.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 904 |
"model.language_model.layers.4.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 905 |
"model.language_model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 906 |
"model.language_model.layers.4.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 907 |
"model.language_model.layers.4.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 908 |
"model.language_model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 909 |
"model.language_model.layers.4.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 910 |
"model.language_model.layers.4.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -925,12 +1031,15 @@
|
|
| 925 |
"model.language_model.layers.40.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 926 |
"model.language_model.layers.40.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 927 |
"model.language_model.layers.40.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 928 |
"model.language_model.layers.40.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 929 |
"model.language_model.layers.40.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 930 |
"model.language_model.layers.40.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 931 |
"model.language_model.layers.40.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 932 |
"model.language_model.layers.40.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 933 |
"model.language_model.layers.40.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 934 |
"model.language_model.layers.40.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 935 |
"model.language_model.layers.40.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 936 |
"model.language_model.layers.40.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -951,12 +1060,15 @@
|
|
| 951 |
"model.language_model.layers.41.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 952 |
"model.language_model.layers.41.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 953 |
"model.language_model.layers.41.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 954 |
"model.language_model.layers.41.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 955 |
"model.language_model.layers.41.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 956 |
"model.language_model.layers.41.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 957 |
"model.language_model.layers.41.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 958 |
"model.language_model.layers.41.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 959 |
"model.language_model.layers.41.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 960 |
"model.language_model.layers.41.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 961 |
"model.language_model.layers.41.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 962 |
"model.language_model.layers.41.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -977,23 +1089,29 @@
|
|
| 977 |
"model.language_model.layers.42.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 978 |
"model.language_model.layers.42.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 979 |
"model.language_model.layers.42.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 980 |
"model.language_model.layers.42.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 981 |
"model.language_model.layers.42.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 982 |
"model.language_model.layers.42.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 983 |
"model.language_model.layers.42.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 984 |
"model.language_model.layers.42.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 985 |
"model.language_model.layers.42.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 986 |
"model.language_model.layers.42.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 987 |
"model.language_model.layers.42.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 988 |
"model.language_model.layers.42.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 989 |
"model.language_model.layers.42.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 990 |
"model.language_model.layers.43.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 991 |
"model.language_model.layers.43.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 992 |
"model.language_model.layers.43.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 993 |
"model.language_model.layers.43.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 994 |
"model.language_model.layers.43.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 995 |
"model.language_model.layers.43.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 996 |
"model.language_model.layers.43.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 997 |
"model.language_model.layers.43.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 998 |
"model.language_model.layers.43.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 999 |
"model.language_model.layers.43.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1028,12 +1146,15 @@
|
|
| 1028 |
"model.language_model.layers.44.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1029 |
"model.language_model.layers.44.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1030 |
"model.language_model.layers.44.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1031 |
"model.language_model.layers.44.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1032 |
"model.language_model.layers.44.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1033 |
"model.language_model.layers.44.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1034 |
"model.language_model.layers.44.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1035 |
"model.language_model.layers.44.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1036 |
"model.language_model.layers.44.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1037 |
"model.language_model.layers.44.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1038 |
"model.language_model.layers.44.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1039 |
"model.language_model.layers.44.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1054,12 +1175,15 @@
|
|
| 1054 |
"model.language_model.layers.45.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1055 |
"model.language_model.layers.45.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1056 |
"model.language_model.layers.45.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1057 |
"model.language_model.layers.45.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1058 |
"model.language_model.layers.45.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1059 |
"model.language_model.layers.45.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1060 |
"model.language_model.layers.45.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1061 |
"model.language_model.layers.45.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1062 |
"model.language_model.layers.45.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1063 |
"model.language_model.layers.45.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1064 |
"model.language_model.layers.45.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1065 |
"model.language_model.layers.45.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1080,23 +1204,29 @@
|
|
| 1080 |
"model.language_model.layers.46.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1081 |
"model.language_model.layers.46.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1082 |
"model.language_model.layers.46.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1083 |
"model.language_model.layers.46.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1084 |
"model.language_model.layers.46.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1085 |
"model.language_model.layers.46.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1086 |
"model.language_model.layers.46.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1087 |
"model.language_model.layers.46.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1088 |
"model.language_model.layers.46.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1089 |
"model.language_model.layers.46.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1090 |
"model.language_model.layers.46.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1091 |
"model.language_model.layers.46.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1092 |
"model.language_model.layers.46.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1093 |
"model.language_model.layers.47.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 1094 |
"model.language_model.layers.47.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1095 |
"model.language_model.layers.47.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1096 |
"model.language_model.layers.47.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1097 |
"model.language_model.layers.47.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1098 |
"model.language_model.layers.47.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1099 |
"model.language_model.layers.47.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1100 |
"model.language_model.layers.47.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1101 |
"model.language_model.layers.47.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1102 |
"model.language_model.layers.47.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1131,12 +1261,15 @@
|
|
| 1131 |
"model.language_model.layers.48.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1132 |
"model.language_model.layers.48.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1133 |
"model.language_model.layers.48.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1134 |
"model.language_model.layers.48.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1135 |
"model.language_model.layers.48.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1136 |
"model.language_model.layers.48.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1137 |
"model.language_model.layers.48.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1138 |
"model.language_model.layers.48.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1139 |
"model.language_model.layers.48.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1140 |
"model.language_model.layers.48.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1141 |
"model.language_model.layers.48.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1142 |
"model.language_model.layers.48.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1157,12 +1290,15 @@
|
|
| 1157 |
"model.language_model.layers.49.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1158 |
"model.language_model.layers.49.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1159 |
"model.language_model.layers.49.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1160 |
"model.language_model.layers.49.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1161 |
"model.language_model.layers.49.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1162 |
"model.language_model.layers.49.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1163 |
"model.language_model.layers.49.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1164 |
"model.language_model.layers.49.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1165 |
"model.language_model.layers.49.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1166 |
"model.language_model.layers.49.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1167 |
"model.language_model.layers.49.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1168 |
"model.language_model.layers.49.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1183,12 +1319,15 @@
|
|
| 1183 |
"model.language_model.layers.5.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1184 |
"model.language_model.layers.5.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1185 |
"model.language_model.layers.5.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 1186 |
"model.language_model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1187 |
"model.language_model.layers.5.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1188 |
"model.language_model.layers.5.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1189 |
"model.language_model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1190 |
"model.language_model.layers.5.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1191 |
"model.language_model.layers.5.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1192 |
"model.language_model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1193 |
"model.language_model.layers.5.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1194 |
"model.language_model.layers.5.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -1209,23 +1348,29 @@
|
|
| 1209 |
"model.language_model.layers.50.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1210 |
"model.language_model.layers.50.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1211 |
"model.language_model.layers.50.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1212 |
"model.language_model.layers.50.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1213 |
"model.language_model.layers.50.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1214 |
"model.language_model.layers.50.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1215 |
"model.language_model.layers.50.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1216 |
"model.language_model.layers.50.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1217 |
"model.language_model.layers.50.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1218 |
"model.language_model.layers.50.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1219 |
"model.language_model.layers.50.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1220 |
"model.language_model.layers.50.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1221 |
"model.language_model.layers.50.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1222 |
"model.language_model.layers.51.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 1223 |
"model.language_model.layers.51.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1224 |
"model.language_model.layers.51.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1225 |
"model.language_model.layers.51.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1226 |
"model.language_model.layers.51.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1227 |
"model.language_model.layers.51.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1228 |
"model.language_model.layers.51.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1229 |
"model.language_model.layers.51.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1230 |
"model.language_model.layers.51.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1231 |
"model.language_model.layers.51.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1260,12 +1405,15 @@
|
|
| 1260 |
"model.language_model.layers.52.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1261 |
"model.language_model.layers.52.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1262 |
"model.language_model.layers.52.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1263 |
"model.language_model.layers.52.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1264 |
"model.language_model.layers.52.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1265 |
"model.language_model.layers.52.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1266 |
"model.language_model.layers.52.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1267 |
"model.language_model.layers.52.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1268 |
"model.language_model.layers.52.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1269 |
"model.language_model.layers.52.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1270 |
"model.language_model.layers.52.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1271 |
"model.language_model.layers.52.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1286,12 +1434,15 @@
|
|
| 1286 |
"model.language_model.layers.53.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1287 |
"model.language_model.layers.53.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1288 |
"model.language_model.layers.53.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1289 |
"model.language_model.layers.53.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1290 |
"model.language_model.layers.53.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1291 |
"model.language_model.layers.53.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1292 |
"model.language_model.layers.53.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1293 |
"model.language_model.layers.53.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1294 |
"model.language_model.layers.53.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1295 |
"model.language_model.layers.53.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1296 |
"model.language_model.layers.53.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1297 |
"model.language_model.layers.53.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1312,23 +1463,29 @@
|
|
| 1312 |
"model.language_model.layers.54.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1313 |
"model.language_model.layers.54.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1314 |
"model.language_model.layers.54.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1315 |
"model.language_model.layers.54.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1316 |
"model.language_model.layers.54.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1317 |
"model.language_model.layers.54.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1318 |
"model.language_model.layers.54.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1319 |
"model.language_model.layers.54.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1320 |
"model.language_model.layers.54.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1321 |
"model.language_model.layers.54.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1322 |
"model.language_model.layers.54.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1323 |
"model.language_model.layers.54.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1324 |
"model.language_model.layers.54.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1325 |
"model.language_model.layers.55.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 1326 |
"model.language_model.layers.55.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1327 |
"model.language_model.layers.55.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1328 |
"model.language_model.layers.55.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1329 |
"model.language_model.layers.55.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1330 |
"model.language_model.layers.55.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1331 |
"model.language_model.layers.55.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1332 |
"model.language_model.layers.55.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1333 |
"model.language_model.layers.55.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1334 |
"model.language_model.layers.55.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1363,12 +1520,15 @@
|
|
| 1363 |
"model.language_model.layers.56.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1364 |
"model.language_model.layers.56.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1365 |
"model.language_model.layers.56.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1366 |
"model.language_model.layers.56.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1367 |
"model.language_model.layers.56.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1368 |
"model.language_model.layers.56.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1369 |
"model.language_model.layers.56.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1370 |
"model.language_model.layers.56.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1371 |
"model.language_model.layers.56.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1372 |
"model.language_model.layers.56.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1373 |
"model.language_model.layers.56.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1374 |
"model.language_model.layers.56.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1389,12 +1549,15 @@
|
|
| 1389 |
"model.language_model.layers.57.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1390 |
"model.language_model.layers.57.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1391 |
"model.language_model.layers.57.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1392 |
"model.language_model.layers.57.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1393 |
"model.language_model.layers.57.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1394 |
"model.language_model.layers.57.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1395 |
"model.language_model.layers.57.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1396 |
"model.language_model.layers.57.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1397 |
"model.language_model.layers.57.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1398 |
"model.language_model.layers.57.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1399 |
"model.language_model.layers.57.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1400 |
"model.language_model.layers.57.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1415,23 +1578,29 @@
|
|
| 1415 |
"model.language_model.layers.58.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1416 |
"model.language_model.layers.58.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1417 |
"model.language_model.layers.58.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1418 |
"model.language_model.layers.58.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1419 |
"model.language_model.layers.58.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1420 |
"model.language_model.layers.58.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1421 |
"model.language_model.layers.58.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1422 |
"model.language_model.layers.58.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1423 |
"model.language_model.layers.58.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1424 |
"model.language_model.layers.58.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1425 |
"model.language_model.layers.58.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1426 |
"model.language_model.layers.58.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1427 |
"model.language_model.layers.58.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1428 |
"model.language_model.layers.59.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
|
|
|
| 1429 |
"model.language_model.layers.59.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1430 |
"model.language_model.layers.59.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1431 |
"model.language_model.layers.59.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1432 |
"model.language_model.layers.59.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1433 |
"model.language_model.layers.59.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1434 |
"model.language_model.layers.59.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1435 |
"model.language_model.layers.59.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1436 |
"model.language_model.layers.59.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1437 |
"model.language_model.layers.59.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1466,12 +1635,15 @@
|
|
| 1466 |
"model.language_model.layers.6.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1467 |
"model.language_model.layers.6.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1468 |
"model.language_model.layers.6.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 1469 |
"model.language_model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1470 |
"model.language_model.layers.6.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1471 |
"model.language_model.layers.6.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1472 |
"model.language_model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1473 |
"model.language_model.layers.6.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1474 |
"model.language_model.layers.6.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1475 |
"model.language_model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1476 |
"model.language_model.layers.6.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1477 |
"model.language_model.layers.6.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -1492,12 +1664,15 @@
|
|
| 1492 |
"model.language_model.layers.60.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1493 |
"model.language_model.layers.60.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1494 |
"model.language_model.layers.60.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1495 |
"model.language_model.layers.60.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1496 |
"model.language_model.layers.60.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1497 |
"model.language_model.layers.60.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1498 |
"model.language_model.layers.60.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1499 |
"model.language_model.layers.60.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1500 |
"model.language_model.layers.60.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1501 |
"model.language_model.layers.60.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1502 |
"model.language_model.layers.60.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1503 |
"model.language_model.layers.60.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1518,12 +1693,15 @@
|
|
| 1518 |
"model.language_model.layers.61.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1519 |
"model.language_model.layers.61.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1520 |
"model.language_model.layers.61.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1521 |
"model.language_model.layers.61.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1522 |
"model.language_model.layers.61.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1523 |
"model.language_model.layers.61.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1524 |
"model.language_model.layers.61.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1525 |
"model.language_model.layers.61.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1526 |
"model.language_model.layers.61.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1527 |
"model.language_model.layers.61.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1528 |
"model.language_model.layers.61.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1529 |
"model.language_model.layers.61.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
@@ -1544,23 +1722,29 @@
|
|
| 1544 |
"model.language_model.layers.62.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1545 |
"model.language_model.layers.62.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1546 |
"model.language_model.layers.62.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
|
|
|
| 1547 |
"model.language_model.layers.62.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
|
| 1548 |
"model.language_model.layers.62.mlp.down_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1549 |
"model.language_model.layers.62.mlp.down_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 1550 |
"model.language_model.layers.62.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
|
| 1551 |
"model.language_model.layers.62.mlp.gate_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1552 |
"model.language_model.layers.62.mlp.gate_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 1553 |
"model.language_model.layers.62.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
|
| 1554 |
"model.language_model.layers.62.mlp.up_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1555 |
"model.language_model.layers.62.mlp.up_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1556 |
"model.language_model.layers.62.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
|
| 1557 |
"model.language_model.layers.63.input_layernorm.weight": "model-00003-of-00003.safetensors",
|
|
|
|
| 1558 |
"model.language_model.layers.63.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
|
| 1559 |
"model.language_model.layers.63.mlp.down_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1560 |
"model.language_model.layers.63.mlp.down_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 1561 |
"model.language_model.layers.63.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
|
| 1562 |
"model.language_model.layers.63.mlp.gate_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1563 |
"model.language_model.layers.63.mlp.gate_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 1564 |
"model.language_model.layers.63.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
|
| 1565 |
"model.language_model.layers.63.mlp.up_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1566 |
"model.language_model.layers.63.mlp.up_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
@@ -1580,12 +1764,15 @@
|
|
| 1580 |
"model.language_model.layers.63.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
|
| 1581 |
"model.language_model.layers.63.self_attn.v_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1582 |
"model.language_model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
|
|
|
| 1583 |
"model.language_model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1584 |
"model.language_model.layers.7.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1585 |
"model.language_model.layers.7.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1586 |
"model.language_model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1587 |
"model.language_model.layers.7.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1588 |
"model.language_model.layers.7.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1589 |
"model.language_model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1590 |
"model.language_model.layers.7.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1591 |
"model.language_model.layers.7.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -1620,12 +1807,15 @@
|
|
| 1620 |
"model.language_model.layers.8.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1621 |
"model.language_model.layers.8.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1622 |
"model.language_model.layers.8.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 1623 |
"model.language_model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1624 |
"model.language_model.layers.8.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1625 |
"model.language_model.layers.8.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1626 |
"model.language_model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1627 |
"model.language_model.layers.8.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1628 |
"model.language_model.layers.8.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1629 |
"model.language_model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1630 |
"model.language_model.layers.8.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1631 |
"model.language_model.layers.8.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
@@ -1646,12 +1836,15 @@
|
|
| 1646 |
"model.language_model.layers.9.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1647 |
"model.language_model.layers.9.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1648 |
"model.language_model.layers.9.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
|
|
|
| 1649 |
"model.language_model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1650 |
"model.language_model.layers.9.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1651 |
"model.language_model.layers.9.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1652 |
"model.language_model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1653 |
"model.language_model.layers.9.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1654 |
"model.language_model.layers.9.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1655 |
"model.language_model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1656 |
"model.language_model.layers.9.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1657 |
"model.language_model.layers.9.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
| 3 |
"total_parameters": 18164649200,
|
| 4 |
+
"total_size": 21921428072
|
| 5 |
},
|
| 6 |
"weight_map": {
|
| 7 |
+
"lm_head.input_scale": "model-00003-of-00003.safetensors",
|
| 8 |
"lm_head.weight": "model-00003-of-00003.safetensors",
|
| 9 |
"lm_head.weight_scale": "model-00003-of-00003.safetensors",
|
| 10 |
"lm_head.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 25 |
"model.language_model.layers.0.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 26 |
"model.language_model.layers.0.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 27 |
"model.language_model.layers.0.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 28 |
+
"model.language_model.layers.0.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 29 |
"model.language_model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 30 |
"model.language_model.layers.0.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 31 |
"model.language_model.layers.0.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 32 |
+
"model.language_model.layers.0.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 33 |
"model.language_model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 34 |
"model.language_model.layers.0.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 35 |
"model.language_model.layers.0.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 36 |
+
"model.language_model.layers.0.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 37 |
"model.language_model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 38 |
"model.language_model.layers.0.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 39 |
"model.language_model.layers.0.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 54 |
"model.language_model.layers.1.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 55 |
"model.language_model.layers.1.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 56 |
"model.language_model.layers.1.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 57 |
+
"model.language_model.layers.1.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 58 |
"model.language_model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 59 |
"model.language_model.layers.1.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 60 |
"model.language_model.layers.1.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 61 |
+
"model.language_model.layers.1.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 62 |
"model.language_model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 63 |
"model.language_model.layers.1.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 64 |
"model.language_model.layers.1.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 65 |
+
"model.language_model.layers.1.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 66 |
"model.language_model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 67 |
"model.language_model.layers.1.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 68 |
"model.language_model.layers.1.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 83 |
"model.language_model.layers.10.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 84 |
"model.language_model.layers.10.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 85 |
"model.language_model.layers.10.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 86 |
+
"model.language_model.layers.10.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 87 |
"model.language_model.layers.10.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 88 |
"model.language_model.layers.10.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 89 |
"model.language_model.layers.10.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 90 |
+
"model.language_model.layers.10.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 91 |
"model.language_model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 92 |
"model.language_model.layers.10.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 93 |
"model.language_model.layers.10.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 94 |
+
"model.language_model.layers.10.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 95 |
"model.language_model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 96 |
"model.language_model.layers.10.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 97 |
"model.language_model.layers.10.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 98 |
"model.language_model.layers.10.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 99 |
"model.language_model.layers.11.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 100 |
+
"model.language_model.layers.11.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 101 |
"model.language_model.layers.11.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 102 |
"model.language_model.layers.11.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 103 |
"model.language_model.layers.11.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 104 |
+
"model.language_model.layers.11.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 105 |
"model.language_model.layers.11.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 106 |
"model.language_model.layers.11.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 107 |
"model.language_model.layers.11.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 108 |
+
"model.language_model.layers.11.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 109 |
"model.language_model.layers.11.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 110 |
"model.language_model.layers.11.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 111 |
"model.language_model.layers.11.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 140 |
"model.language_model.layers.12.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 141 |
"model.language_model.layers.12.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 142 |
"model.language_model.layers.12.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 143 |
+
"model.language_model.layers.12.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 144 |
"model.language_model.layers.12.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 145 |
"model.language_model.layers.12.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 146 |
"model.language_model.layers.12.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 147 |
+
"model.language_model.layers.12.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 148 |
"model.language_model.layers.12.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 149 |
"model.language_model.layers.12.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 150 |
"model.language_model.layers.12.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 151 |
+
"model.language_model.layers.12.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 152 |
"model.language_model.layers.12.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 153 |
"model.language_model.layers.12.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 154 |
"model.language_model.layers.12.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 169 |
"model.language_model.layers.13.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 170 |
"model.language_model.layers.13.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 171 |
"model.language_model.layers.13.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 172 |
+
"model.language_model.layers.13.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 173 |
"model.language_model.layers.13.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 174 |
"model.language_model.layers.13.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 175 |
"model.language_model.layers.13.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 176 |
+
"model.language_model.layers.13.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 177 |
"model.language_model.layers.13.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 178 |
"model.language_model.layers.13.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 179 |
"model.language_model.layers.13.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 180 |
+
"model.language_model.layers.13.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 181 |
"model.language_model.layers.13.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 182 |
"model.language_model.layers.13.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 183 |
"model.language_model.layers.13.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 198 |
"model.language_model.layers.14.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 199 |
"model.language_model.layers.14.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 200 |
"model.language_model.layers.14.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 201 |
+
"model.language_model.layers.14.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 202 |
"model.language_model.layers.14.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 203 |
"model.language_model.layers.14.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 204 |
"model.language_model.layers.14.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 205 |
+
"model.language_model.layers.14.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 206 |
"model.language_model.layers.14.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 207 |
"model.language_model.layers.14.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 208 |
"model.language_model.layers.14.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 209 |
+
"model.language_model.layers.14.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 210 |
"model.language_model.layers.14.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 211 |
"model.language_model.layers.14.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 212 |
"model.language_model.layers.14.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 213 |
"model.language_model.layers.14.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 214 |
"model.language_model.layers.15.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 215 |
+
"model.language_model.layers.15.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 216 |
"model.language_model.layers.15.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 217 |
"model.language_model.layers.15.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 218 |
"model.language_model.layers.15.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 219 |
+
"model.language_model.layers.15.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 220 |
"model.language_model.layers.15.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 221 |
"model.language_model.layers.15.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 222 |
"model.language_model.layers.15.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 223 |
+
"model.language_model.layers.15.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 224 |
"model.language_model.layers.15.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 225 |
"model.language_model.layers.15.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 226 |
"model.language_model.layers.15.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 255 |
"model.language_model.layers.16.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 256 |
"model.language_model.layers.16.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 257 |
"model.language_model.layers.16.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 258 |
+
"model.language_model.layers.16.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 259 |
"model.language_model.layers.16.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 260 |
"model.language_model.layers.16.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 261 |
"model.language_model.layers.16.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 262 |
+
"model.language_model.layers.16.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 263 |
"model.language_model.layers.16.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 264 |
"model.language_model.layers.16.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 265 |
"model.language_model.layers.16.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 266 |
+
"model.language_model.layers.16.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 267 |
"model.language_model.layers.16.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 268 |
"model.language_model.layers.16.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 269 |
"model.language_model.layers.16.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 284 |
"model.language_model.layers.17.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 285 |
"model.language_model.layers.17.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 286 |
"model.language_model.layers.17.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 287 |
+
"model.language_model.layers.17.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 288 |
"model.language_model.layers.17.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 289 |
"model.language_model.layers.17.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 290 |
"model.language_model.layers.17.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 291 |
+
"model.language_model.layers.17.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 292 |
"model.language_model.layers.17.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 293 |
"model.language_model.layers.17.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 294 |
"model.language_model.layers.17.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 295 |
+
"model.language_model.layers.17.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 296 |
"model.language_model.layers.17.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 297 |
"model.language_model.layers.17.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 298 |
"model.language_model.layers.17.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 313 |
"model.language_model.layers.18.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 314 |
"model.language_model.layers.18.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 315 |
"model.language_model.layers.18.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 316 |
+
"model.language_model.layers.18.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 317 |
"model.language_model.layers.18.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 318 |
"model.language_model.layers.18.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 319 |
"model.language_model.layers.18.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 320 |
+
"model.language_model.layers.18.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 321 |
"model.language_model.layers.18.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 322 |
"model.language_model.layers.18.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 323 |
"model.language_model.layers.18.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 324 |
+
"model.language_model.layers.18.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 325 |
"model.language_model.layers.18.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 326 |
"model.language_model.layers.18.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 327 |
"model.language_model.layers.18.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 328 |
"model.language_model.layers.18.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 329 |
"model.language_model.layers.19.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 330 |
+
"model.language_model.layers.19.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 331 |
"model.language_model.layers.19.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 332 |
"model.language_model.layers.19.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 333 |
"model.language_model.layers.19.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 334 |
+
"model.language_model.layers.19.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 335 |
"model.language_model.layers.19.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 336 |
"model.language_model.layers.19.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 337 |
"model.language_model.layers.19.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 338 |
+
"model.language_model.layers.19.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 339 |
"model.language_model.layers.19.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 340 |
"model.language_model.layers.19.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 341 |
"model.language_model.layers.19.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 370 |
"model.language_model.layers.2.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 371 |
"model.language_model.layers.2.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 372 |
"model.language_model.layers.2.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 373 |
+
"model.language_model.layers.2.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 374 |
"model.language_model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 375 |
"model.language_model.layers.2.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 376 |
"model.language_model.layers.2.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 377 |
+
"model.language_model.layers.2.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 378 |
"model.language_model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 379 |
"model.language_model.layers.2.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 380 |
"model.language_model.layers.2.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 381 |
+
"model.language_model.layers.2.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 382 |
"model.language_model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 383 |
"model.language_model.layers.2.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 384 |
"model.language_model.layers.2.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 399 |
"model.language_model.layers.20.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 400 |
"model.language_model.layers.20.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 401 |
"model.language_model.layers.20.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 402 |
+
"model.language_model.layers.20.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 403 |
"model.language_model.layers.20.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 404 |
"model.language_model.layers.20.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 405 |
"model.language_model.layers.20.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 406 |
+
"model.language_model.layers.20.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 407 |
"model.language_model.layers.20.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 408 |
"model.language_model.layers.20.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 409 |
"model.language_model.layers.20.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 410 |
+
"model.language_model.layers.20.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 411 |
"model.language_model.layers.20.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 412 |
"model.language_model.layers.20.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 413 |
"model.language_model.layers.20.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 428 |
"model.language_model.layers.21.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 429 |
"model.language_model.layers.21.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 430 |
"model.language_model.layers.21.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 431 |
+
"model.language_model.layers.21.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 432 |
"model.language_model.layers.21.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 433 |
"model.language_model.layers.21.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 434 |
"model.language_model.layers.21.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 435 |
+
"model.language_model.layers.21.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 436 |
"model.language_model.layers.21.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 437 |
"model.language_model.layers.21.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 438 |
"model.language_model.layers.21.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 439 |
+
"model.language_model.layers.21.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 440 |
"model.language_model.layers.21.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 441 |
"model.language_model.layers.21.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 442 |
"model.language_model.layers.21.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 457 |
"model.language_model.layers.22.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 458 |
"model.language_model.layers.22.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 459 |
"model.language_model.layers.22.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 460 |
+
"model.language_model.layers.22.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 461 |
"model.language_model.layers.22.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 462 |
"model.language_model.layers.22.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 463 |
"model.language_model.layers.22.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 464 |
+
"model.language_model.layers.22.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 465 |
"model.language_model.layers.22.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 466 |
"model.language_model.layers.22.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 467 |
"model.language_model.layers.22.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 468 |
+
"model.language_model.layers.22.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 469 |
"model.language_model.layers.22.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 470 |
"model.language_model.layers.22.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 471 |
"model.language_model.layers.22.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 472 |
"model.language_model.layers.22.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 473 |
"model.language_model.layers.23.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 474 |
+
"model.language_model.layers.23.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 475 |
"model.language_model.layers.23.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 476 |
"model.language_model.layers.23.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 477 |
"model.language_model.layers.23.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 478 |
+
"model.language_model.layers.23.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 479 |
"model.language_model.layers.23.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 480 |
"model.language_model.layers.23.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 481 |
"model.language_model.layers.23.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 482 |
+
"model.language_model.layers.23.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 483 |
"model.language_model.layers.23.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 484 |
"model.language_model.layers.23.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 485 |
"model.language_model.layers.23.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 514 |
"model.language_model.layers.24.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 515 |
"model.language_model.layers.24.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 516 |
"model.language_model.layers.24.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 517 |
+
"model.language_model.layers.24.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 518 |
"model.language_model.layers.24.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 519 |
"model.language_model.layers.24.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 520 |
"model.language_model.layers.24.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 521 |
+
"model.language_model.layers.24.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 522 |
"model.language_model.layers.24.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 523 |
"model.language_model.layers.24.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 524 |
"model.language_model.layers.24.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 525 |
+
"model.language_model.layers.24.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 526 |
"model.language_model.layers.24.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 527 |
"model.language_model.layers.24.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 528 |
"model.language_model.layers.24.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 543 |
"model.language_model.layers.25.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 544 |
"model.language_model.layers.25.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 545 |
"model.language_model.layers.25.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 546 |
+
"model.language_model.layers.25.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 547 |
"model.language_model.layers.25.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 548 |
"model.language_model.layers.25.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 549 |
"model.language_model.layers.25.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 550 |
+
"model.language_model.layers.25.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 551 |
"model.language_model.layers.25.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 552 |
"model.language_model.layers.25.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 553 |
"model.language_model.layers.25.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 554 |
+
"model.language_model.layers.25.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 555 |
"model.language_model.layers.25.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 556 |
"model.language_model.layers.25.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 557 |
"model.language_model.layers.25.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 572 |
"model.language_model.layers.26.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 573 |
"model.language_model.layers.26.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 574 |
"model.language_model.layers.26.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 575 |
+
"model.language_model.layers.26.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 576 |
"model.language_model.layers.26.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 577 |
"model.language_model.layers.26.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 578 |
"model.language_model.layers.26.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 579 |
+
"model.language_model.layers.26.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 580 |
"model.language_model.layers.26.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 581 |
"model.language_model.layers.26.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 582 |
"model.language_model.layers.26.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 583 |
+
"model.language_model.layers.26.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 584 |
"model.language_model.layers.26.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 585 |
"model.language_model.layers.26.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 586 |
"model.language_model.layers.26.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 587 |
"model.language_model.layers.26.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 588 |
"model.language_model.layers.27.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 589 |
+
"model.language_model.layers.27.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 590 |
"model.language_model.layers.27.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 591 |
"model.language_model.layers.27.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 592 |
"model.language_model.layers.27.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 593 |
+
"model.language_model.layers.27.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 594 |
"model.language_model.layers.27.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 595 |
"model.language_model.layers.27.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 596 |
"model.language_model.layers.27.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 597 |
+
"model.language_model.layers.27.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 598 |
"model.language_model.layers.27.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 599 |
"model.language_model.layers.27.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 600 |
"model.language_model.layers.27.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 629 |
"model.language_model.layers.28.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 630 |
"model.language_model.layers.28.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 631 |
"model.language_model.layers.28.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 632 |
+
"model.language_model.layers.28.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 633 |
"model.language_model.layers.28.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 634 |
"model.language_model.layers.28.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 635 |
"model.language_model.layers.28.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 636 |
+
"model.language_model.layers.28.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 637 |
"model.language_model.layers.28.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 638 |
"model.language_model.layers.28.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 639 |
"model.language_model.layers.28.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 640 |
+
"model.language_model.layers.28.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 641 |
"model.language_model.layers.28.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 642 |
"model.language_model.layers.28.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 643 |
"model.language_model.layers.28.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 658 |
"model.language_model.layers.29.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 659 |
"model.language_model.layers.29.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 660 |
"model.language_model.layers.29.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 661 |
+
"model.language_model.layers.29.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 662 |
"model.language_model.layers.29.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 663 |
"model.language_model.layers.29.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 664 |
"model.language_model.layers.29.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 665 |
+
"model.language_model.layers.29.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 666 |
"model.language_model.layers.29.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 667 |
"model.language_model.layers.29.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 668 |
"model.language_model.layers.29.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 669 |
+
"model.language_model.layers.29.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 670 |
"model.language_model.layers.29.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 671 |
"model.language_model.layers.29.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 672 |
"model.language_model.layers.29.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 673 |
"model.language_model.layers.29.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 674 |
"model.language_model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 675 |
+
"model.language_model.layers.3.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 676 |
"model.language_model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 677 |
"model.language_model.layers.3.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 678 |
"model.language_model.layers.3.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 679 |
+
"model.language_model.layers.3.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 680 |
"model.language_model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 681 |
"model.language_model.layers.3.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 682 |
"model.language_model.layers.3.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 683 |
+
"model.language_model.layers.3.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 684 |
"model.language_model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 685 |
"model.language_model.layers.3.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 686 |
"model.language_model.layers.3.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 715 |
"model.language_model.layers.30.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 716 |
"model.language_model.layers.30.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 717 |
"model.language_model.layers.30.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 718 |
+
"model.language_model.layers.30.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 719 |
"model.language_model.layers.30.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 720 |
"model.language_model.layers.30.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 721 |
"model.language_model.layers.30.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 722 |
+
"model.language_model.layers.30.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 723 |
"model.language_model.layers.30.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 724 |
"model.language_model.layers.30.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 725 |
"model.language_model.layers.30.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 726 |
+
"model.language_model.layers.30.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 727 |
"model.language_model.layers.30.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 728 |
"model.language_model.layers.30.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 729 |
"model.language_model.layers.30.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 730 |
"model.language_model.layers.30.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 731 |
"model.language_model.layers.31.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 732 |
+
"model.language_model.layers.31.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 733 |
"model.language_model.layers.31.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 734 |
"model.language_model.layers.31.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 735 |
"model.language_model.layers.31.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 736 |
+
"model.language_model.layers.31.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 737 |
"model.language_model.layers.31.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 738 |
"model.language_model.layers.31.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 739 |
"model.language_model.layers.31.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 740 |
+
"model.language_model.layers.31.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 741 |
"model.language_model.layers.31.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 742 |
"model.language_model.layers.31.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 743 |
"model.language_model.layers.31.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 772 |
"model.language_model.layers.32.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 773 |
"model.language_model.layers.32.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 774 |
"model.language_model.layers.32.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 775 |
+
"model.language_model.layers.32.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 776 |
"model.language_model.layers.32.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 777 |
"model.language_model.layers.32.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 778 |
"model.language_model.layers.32.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 779 |
+
"model.language_model.layers.32.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 780 |
"model.language_model.layers.32.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 781 |
"model.language_model.layers.32.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 782 |
"model.language_model.layers.32.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 783 |
+
"model.language_model.layers.32.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 784 |
"model.language_model.layers.32.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 785 |
"model.language_model.layers.32.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 786 |
"model.language_model.layers.32.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 801 |
"model.language_model.layers.33.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 802 |
"model.language_model.layers.33.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 803 |
"model.language_model.layers.33.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 804 |
+
"model.language_model.layers.33.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 805 |
"model.language_model.layers.33.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 806 |
"model.language_model.layers.33.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 807 |
"model.language_model.layers.33.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 808 |
+
"model.language_model.layers.33.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 809 |
"model.language_model.layers.33.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 810 |
"model.language_model.layers.33.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 811 |
"model.language_model.layers.33.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 812 |
+
"model.language_model.layers.33.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 813 |
"model.language_model.layers.33.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 814 |
"model.language_model.layers.33.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 815 |
"model.language_model.layers.33.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 830 |
"model.language_model.layers.34.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 831 |
"model.language_model.layers.34.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 832 |
"model.language_model.layers.34.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 833 |
+
"model.language_model.layers.34.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 834 |
"model.language_model.layers.34.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 835 |
"model.language_model.layers.34.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 836 |
"model.language_model.layers.34.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 837 |
+
"model.language_model.layers.34.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 838 |
"model.language_model.layers.34.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 839 |
"model.language_model.layers.34.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 840 |
"model.language_model.layers.34.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 841 |
+
"model.language_model.layers.34.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 842 |
"model.language_model.layers.34.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 843 |
"model.language_model.layers.34.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 844 |
"model.language_model.layers.34.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 845 |
"model.language_model.layers.34.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 846 |
"model.language_model.layers.35.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 847 |
+
"model.language_model.layers.35.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 848 |
"model.language_model.layers.35.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 849 |
"model.language_model.layers.35.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 850 |
"model.language_model.layers.35.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 851 |
+
"model.language_model.layers.35.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 852 |
"model.language_model.layers.35.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 853 |
"model.language_model.layers.35.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 854 |
"model.language_model.layers.35.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 855 |
+
"model.language_model.layers.35.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 856 |
"model.language_model.layers.35.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 857 |
"model.language_model.layers.35.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 858 |
"model.language_model.layers.35.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 887 |
"model.language_model.layers.36.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 888 |
"model.language_model.layers.36.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 889 |
"model.language_model.layers.36.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 890 |
+
"model.language_model.layers.36.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 891 |
"model.language_model.layers.36.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 892 |
"model.language_model.layers.36.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 893 |
"model.language_model.layers.36.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 894 |
+
"model.language_model.layers.36.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 895 |
"model.language_model.layers.36.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 896 |
"model.language_model.layers.36.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 897 |
"model.language_model.layers.36.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 898 |
+
"model.language_model.layers.36.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 899 |
"model.language_model.layers.36.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 900 |
"model.language_model.layers.36.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 901 |
"model.language_model.layers.36.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 916 |
"model.language_model.layers.37.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 917 |
"model.language_model.layers.37.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 918 |
"model.language_model.layers.37.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 919 |
+
"model.language_model.layers.37.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 920 |
"model.language_model.layers.37.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 921 |
"model.language_model.layers.37.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 922 |
"model.language_model.layers.37.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 923 |
+
"model.language_model.layers.37.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 924 |
"model.language_model.layers.37.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 925 |
"model.language_model.layers.37.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 926 |
"model.language_model.layers.37.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 927 |
+
"model.language_model.layers.37.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 928 |
"model.language_model.layers.37.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 929 |
"model.language_model.layers.37.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 930 |
"model.language_model.layers.37.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 945 |
"model.language_model.layers.38.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 946 |
"model.language_model.layers.38.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 947 |
"model.language_model.layers.38.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 948 |
+
"model.language_model.layers.38.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 949 |
"model.language_model.layers.38.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 950 |
"model.language_model.layers.38.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 951 |
"model.language_model.layers.38.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 952 |
+
"model.language_model.layers.38.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 953 |
"model.language_model.layers.38.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 954 |
"model.language_model.layers.38.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 955 |
"model.language_model.layers.38.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 956 |
+
"model.language_model.layers.38.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 957 |
"model.language_model.layers.38.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 958 |
"model.language_model.layers.38.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 959 |
"model.language_model.layers.38.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 960 |
"model.language_model.layers.38.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 961 |
"model.language_model.layers.39.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 962 |
+
"model.language_model.layers.39.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 963 |
"model.language_model.layers.39.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 964 |
"model.language_model.layers.39.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 965 |
"model.language_model.layers.39.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 966 |
+
"model.language_model.layers.39.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 967 |
"model.language_model.layers.39.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 968 |
"model.language_model.layers.39.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 969 |
"model.language_model.layers.39.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 970 |
+
"model.language_model.layers.39.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 971 |
"model.language_model.layers.39.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 972 |
"model.language_model.layers.39.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 973 |
"model.language_model.layers.39.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1002 |
"model.language_model.layers.4.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1003 |
"model.language_model.layers.4.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1004 |
"model.language_model.layers.4.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1005 |
+
"model.language_model.layers.4.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1006 |
"model.language_model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1007 |
"model.language_model.layers.4.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1008 |
"model.language_model.layers.4.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1009 |
+
"model.language_model.layers.4.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1010 |
"model.language_model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1011 |
"model.language_model.layers.4.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1012 |
"model.language_model.layers.4.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1013 |
+
"model.language_model.layers.4.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1014 |
"model.language_model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1015 |
"model.language_model.layers.4.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1016 |
"model.language_model.layers.4.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1031 |
"model.language_model.layers.40.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1032 |
"model.language_model.layers.40.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1033 |
"model.language_model.layers.40.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1034 |
+
"model.language_model.layers.40.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1035 |
"model.language_model.layers.40.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1036 |
"model.language_model.layers.40.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1037 |
"model.language_model.layers.40.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1038 |
+
"model.language_model.layers.40.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1039 |
"model.language_model.layers.40.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1040 |
"model.language_model.layers.40.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1041 |
"model.language_model.layers.40.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1042 |
+
"model.language_model.layers.40.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1043 |
"model.language_model.layers.40.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1044 |
"model.language_model.layers.40.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1045 |
"model.language_model.layers.40.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1060 |
"model.language_model.layers.41.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1061 |
"model.language_model.layers.41.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1062 |
"model.language_model.layers.41.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1063 |
+
"model.language_model.layers.41.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1064 |
"model.language_model.layers.41.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1065 |
"model.language_model.layers.41.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1066 |
"model.language_model.layers.41.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1067 |
+
"model.language_model.layers.41.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1068 |
"model.language_model.layers.41.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1069 |
"model.language_model.layers.41.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1070 |
"model.language_model.layers.41.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1071 |
+
"model.language_model.layers.41.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1072 |
"model.language_model.layers.41.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1073 |
"model.language_model.layers.41.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1074 |
"model.language_model.layers.41.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1089 |
"model.language_model.layers.42.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1090 |
"model.language_model.layers.42.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1091 |
"model.language_model.layers.42.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1092 |
+
"model.language_model.layers.42.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1093 |
"model.language_model.layers.42.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1094 |
"model.language_model.layers.42.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1095 |
"model.language_model.layers.42.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1096 |
+
"model.language_model.layers.42.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1097 |
"model.language_model.layers.42.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1098 |
"model.language_model.layers.42.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1099 |
"model.language_model.layers.42.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1100 |
+
"model.language_model.layers.42.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1101 |
"model.language_model.layers.42.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1102 |
"model.language_model.layers.42.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1103 |
"model.language_model.layers.42.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1104 |
"model.language_model.layers.42.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1105 |
"model.language_model.layers.43.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1106 |
+
"model.language_model.layers.43.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1107 |
"model.language_model.layers.43.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1108 |
"model.language_model.layers.43.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1109 |
"model.language_model.layers.43.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1110 |
+
"model.language_model.layers.43.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1111 |
"model.language_model.layers.43.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1112 |
"model.language_model.layers.43.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1113 |
"model.language_model.layers.43.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1114 |
+
"model.language_model.layers.43.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1115 |
"model.language_model.layers.43.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1116 |
"model.language_model.layers.43.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1117 |
"model.language_model.layers.43.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1146 |
"model.language_model.layers.44.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1147 |
"model.language_model.layers.44.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1148 |
"model.language_model.layers.44.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1149 |
+
"model.language_model.layers.44.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1150 |
"model.language_model.layers.44.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1151 |
"model.language_model.layers.44.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1152 |
"model.language_model.layers.44.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1153 |
+
"model.language_model.layers.44.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1154 |
"model.language_model.layers.44.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1155 |
"model.language_model.layers.44.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1156 |
"model.language_model.layers.44.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1157 |
+
"model.language_model.layers.44.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1158 |
"model.language_model.layers.44.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1159 |
"model.language_model.layers.44.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1160 |
"model.language_model.layers.44.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1175 |
"model.language_model.layers.45.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1176 |
"model.language_model.layers.45.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1177 |
"model.language_model.layers.45.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1178 |
+
"model.language_model.layers.45.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1179 |
"model.language_model.layers.45.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1180 |
"model.language_model.layers.45.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1181 |
"model.language_model.layers.45.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1182 |
+
"model.language_model.layers.45.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1183 |
"model.language_model.layers.45.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1184 |
"model.language_model.layers.45.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1185 |
"model.language_model.layers.45.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1186 |
+
"model.language_model.layers.45.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1187 |
"model.language_model.layers.45.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1188 |
"model.language_model.layers.45.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1189 |
"model.language_model.layers.45.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1204 |
"model.language_model.layers.46.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1205 |
"model.language_model.layers.46.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1206 |
"model.language_model.layers.46.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1207 |
+
"model.language_model.layers.46.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1208 |
"model.language_model.layers.46.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1209 |
"model.language_model.layers.46.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1210 |
"model.language_model.layers.46.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1211 |
+
"model.language_model.layers.46.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1212 |
"model.language_model.layers.46.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1213 |
"model.language_model.layers.46.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1214 |
"model.language_model.layers.46.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1215 |
+
"model.language_model.layers.46.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1216 |
"model.language_model.layers.46.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1217 |
"model.language_model.layers.46.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1218 |
"model.language_model.layers.46.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1219 |
"model.language_model.layers.46.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1220 |
"model.language_model.layers.47.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1221 |
+
"model.language_model.layers.47.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1222 |
"model.language_model.layers.47.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1223 |
"model.language_model.layers.47.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1224 |
"model.language_model.layers.47.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1225 |
+
"model.language_model.layers.47.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1226 |
"model.language_model.layers.47.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1227 |
"model.language_model.layers.47.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1228 |
"model.language_model.layers.47.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1229 |
+
"model.language_model.layers.47.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1230 |
"model.language_model.layers.47.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1231 |
"model.language_model.layers.47.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1232 |
"model.language_model.layers.47.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1261 |
"model.language_model.layers.48.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1262 |
"model.language_model.layers.48.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1263 |
"model.language_model.layers.48.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1264 |
+
"model.language_model.layers.48.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1265 |
"model.language_model.layers.48.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1266 |
"model.language_model.layers.48.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1267 |
"model.language_model.layers.48.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1268 |
+
"model.language_model.layers.48.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1269 |
"model.language_model.layers.48.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1270 |
"model.language_model.layers.48.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1271 |
"model.language_model.layers.48.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1272 |
+
"model.language_model.layers.48.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1273 |
"model.language_model.layers.48.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1274 |
"model.language_model.layers.48.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1275 |
"model.language_model.layers.48.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1290 |
"model.language_model.layers.49.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1291 |
"model.language_model.layers.49.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1292 |
"model.language_model.layers.49.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1293 |
+
"model.language_model.layers.49.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1294 |
"model.language_model.layers.49.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1295 |
"model.language_model.layers.49.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1296 |
"model.language_model.layers.49.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1297 |
+
"model.language_model.layers.49.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1298 |
"model.language_model.layers.49.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1299 |
"model.language_model.layers.49.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1300 |
"model.language_model.layers.49.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1301 |
+
"model.language_model.layers.49.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1302 |
"model.language_model.layers.49.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1303 |
"model.language_model.layers.49.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1304 |
"model.language_model.layers.49.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1319 |
"model.language_model.layers.5.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1320 |
"model.language_model.layers.5.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1321 |
"model.language_model.layers.5.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1322 |
+
"model.language_model.layers.5.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1323 |
"model.language_model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1324 |
"model.language_model.layers.5.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1325 |
"model.language_model.layers.5.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1326 |
+
"model.language_model.layers.5.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1327 |
"model.language_model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1328 |
"model.language_model.layers.5.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1329 |
"model.language_model.layers.5.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1330 |
+
"model.language_model.layers.5.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1331 |
"model.language_model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1332 |
"model.language_model.layers.5.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1333 |
"model.language_model.layers.5.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1348 |
"model.language_model.layers.50.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1349 |
"model.language_model.layers.50.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1350 |
"model.language_model.layers.50.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1351 |
+
"model.language_model.layers.50.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1352 |
"model.language_model.layers.50.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1353 |
"model.language_model.layers.50.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1354 |
"model.language_model.layers.50.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1355 |
+
"model.language_model.layers.50.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1356 |
"model.language_model.layers.50.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1357 |
"model.language_model.layers.50.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1358 |
"model.language_model.layers.50.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1359 |
+
"model.language_model.layers.50.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1360 |
"model.language_model.layers.50.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1361 |
"model.language_model.layers.50.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1362 |
"model.language_model.layers.50.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1363 |
"model.language_model.layers.50.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1364 |
"model.language_model.layers.51.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1365 |
+
"model.language_model.layers.51.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1366 |
"model.language_model.layers.51.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1367 |
"model.language_model.layers.51.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1368 |
"model.language_model.layers.51.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1369 |
+
"model.language_model.layers.51.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1370 |
"model.language_model.layers.51.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1371 |
"model.language_model.layers.51.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1372 |
"model.language_model.layers.51.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1373 |
+
"model.language_model.layers.51.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1374 |
"model.language_model.layers.51.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1375 |
"model.language_model.layers.51.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1376 |
"model.language_model.layers.51.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1405 |
"model.language_model.layers.52.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1406 |
"model.language_model.layers.52.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1407 |
"model.language_model.layers.52.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1408 |
+
"model.language_model.layers.52.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1409 |
"model.language_model.layers.52.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1410 |
"model.language_model.layers.52.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1411 |
"model.language_model.layers.52.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1412 |
+
"model.language_model.layers.52.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1413 |
"model.language_model.layers.52.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1414 |
"model.language_model.layers.52.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1415 |
"model.language_model.layers.52.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1416 |
+
"model.language_model.layers.52.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1417 |
"model.language_model.layers.52.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1418 |
"model.language_model.layers.52.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1419 |
"model.language_model.layers.52.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1434 |
"model.language_model.layers.53.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1435 |
"model.language_model.layers.53.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1436 |
"model.language_model.layers.53.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1437 |
+
"model.language_model.layers.53.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1438 |
"model.language_model.layers.53.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1439 |
"model.language_model.layers.53.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1440 |
"model.language_model.layers.53.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1441 |
+
"model.language_model.layers.53.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1442 |
"model.language_model.layers.53.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1443 |
"model.language_model.layers.53.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1444 |
"model.language_model.layers.53.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1445 |
+
"model.language_model.layers.53.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1446 |
"model.language_model.layers.53.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1447 |
"model.language_model.layers.53.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1448 |
"model.language_model.layers.53.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1463 |
"model.language_model.layers.54.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1464 |
"model.language_model.layers.54.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1465 |
"model.language_model.layers.54.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1466 |
+
"model.language_model.layers.54.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1467 |
"model.language_model.layers.54.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1468 |
"model.language_model.layers.54.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1469 |
"model.language_model.layers.54.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1470 |
+
"model.language_model.layers.54.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1471 |
"model.language_model.layers.54.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1472 |
"model.language_model.layers.54.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1473 |
"model.language_model.layers.54.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1474 |
+
"model.language_model.layers.54.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1475 |
"model.language_model.layers.54.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1476 |
"model.language_model.layers.54.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1477 |
"model.language_model.layers.54.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1478 |
"model.language_model.layers.54.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1479 |
"model.language_model.layers.55.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1480 |
+
"model.language_model.layers.55.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1481 |
"model.language_model.layers.55.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1482 |
"model.language_model.layers.55.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1483 |
"model.language_model.layers.55.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1484 |
+
"model.language_model.layers.55.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1485 |
"model.language_model.layers.55.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1486 |
"model.language_model.layers.55.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1487 |
"model.language_model.layers.55.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1488 |
+
"model.language_model.layers.55.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1489 |
"model.language_model.layers.55.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1490 |
"model.language_model.layers.55.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1491 |
"model.language_model.layers.55.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1520 |
"model.language_model.layers.56.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1521 |
"model.language_model.layers.56.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1522 |
"model.language_model.layers.56.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1523 |
+
"model.language_model.layers.56.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1524 |
"model.language_model.layers.56.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1525 |
"model.language_model.layers.56.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1526 |
"model.language_model.layers.56.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1527 |
+
"model.language_model.layers.56.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1528 |
"model.language_model.layers.56.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1529 |
"model.language_model.layers.56.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1530 |
"model.language_model.layers.56.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1531 |
+
"model.language_model.layers.56.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1532 |
"model.language_model.layers.56.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1533 |
"model.language_model.layers.56.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1534 |
"model.language_model.layers.56.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1549 |
"model.language_model.layers.57.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1550 |
"model.language_model.layers.57.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1551 |
"model.language_model.layers.57.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1552 |
+
"model.language_model.layers.57.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1553 |
"model.language_model.layers.57.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1554 |
"model.language_model.layers.57.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1555 |
"model.language_model.layers.57.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1556 |
+
"model.language_model.layers.57.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1557 |
"model.language_model.layers.57.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1558 |
"model.language_model.layers.57.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1559 |
"model.language_model.layers.57.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1560 |
+
"model.language_model.layers.57.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1561 |
"model.language_model.layers.57.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1562 |
"model.language_model.layers.57.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1563 |
"model.language_model.layers.57.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1578 |
"model.language_model.layers.58.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1579 |
"model.language_model.layers.58.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1580 |
"model.language_model.layers.58.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1581 |
+
"model.language_model.layers.58.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1582 |
"model.language_model.layers.58.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1583 |
"model.language_model.layers.58.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1584 |
"model.language_model.layers.58.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1585 |
+
"model.language_model.layers.58.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1586 |
"model.language_model.layers.58.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1587 |
"model.language_model.layers.58.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1588 |
"model.language_model.layers.58.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1589 |
+
"model.language_model.layers.58.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1590 |
"model.language_model.layers.58.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1591 |
"model.language_model.layers.58.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1592 |
"model.language_model.layers.58.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1593 |
"model.language_model.layers.58.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1594 |
"model.language_model.layers.59.input_layernorm.weight": "model-00002-of-00003.safetensors",
|
| 1595 |
+
"model.language_model.layers.59.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1596 |
"model.language_model.layers.59.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1597 |
"model.language_model.layers.59.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1598 |
"model.language_model.layers.59.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1599 |
+
"model.language_model.layers.59.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1600 |
"model.language_model.layers.59.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1601 |
"model.language_model.layers.59.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1602 |
"model.language_model.layers.59.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1603 |
+
"model.language_model.layers.59.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1604 |
"model.language_model.layers.59.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1605 |
"model.language_model.layers.59.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1606 |
"model.language_model.layers.59.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1635 |
"model.language_model.layers.6.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1636 |
"model.language_model.layers.6.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1637 |
"model.language_model.layers.6.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1638 |
+
"model.language_model.layers.6.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1639 |
"model.language_model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1640 |
"model.language_model.layers.6.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1641 |
"model.language_model.layers.6.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1642 |
+
"model.language_model.layers.6.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1643 |
"model.language_model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1644 |
"model.language_model.layers.6.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1645 |
"model.language_model.layers.6.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1646 |
+
"model.language_model.layers.6.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1647 |
"model.language_model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1648 |
"model.language_model.layers.6.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1649 |
"model.language_model.layers.6.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1664 |
"model.language_model.layers.60.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1665 |
"model.language_model.layers.60.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1666 |
"model.language_model.layers.60.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1667 |
+
"model.language_model.layers.60.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1668 |
"model.language_model.layers.60.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1669 |
"model.language_model.layers.60.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1670 |
"model.language_model.layers.60.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1671 |
+
"model.language_model.layers.60.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1672 |
"model.language_model.layers.60.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1673 |
"model.language_model.layers.60.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1674 |
"model.language_model.layers.60.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1675 |
+
"model.language_model.layers.60.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1676 |
"model.language_model.layers.60.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1677 |
"model.language_model.layers.60.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1678 |
"model.language_model.layers.60.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1693 |
"model.language_model.layers.61.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1694 |
"model.language_model.layers.61.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1695 |
"model.language_model.layers.61.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1696 |
+
"model.language_model.layers.61.mlp.down_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1697 |
"model.language_model.layers.61.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
|
| 1698 |
"model.language_model.layers.61.mlp.down_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1699 |
"model.language_model.layers.61.mlp.down_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1700 |
+
"model.language_model.layers.61.mlp.gate_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1701 |
"model.language_model.layers.61.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
|
| 1702 |
"model.language_model.layers.61.mlp.gate_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1703 |
"model.language_model.layers.61.mlp.gate_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
| 1704 |
+
"model.language_model.layers.61.mlp.up_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1705 |
"model.language_model.layers.61.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
|
| 1706 |
"model.language_model.layers.61.mlp.up_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1707 |
"model.language_model.layers.61.mlp.up_proj.weight_scale_2": "model-00002-of-00003.safetensors",
|
|
|
|
| 1722 |
"model.language_model.layers.62.linear_attn.out_proj.input_scale": "model-00002-of-00003.safetensors",
|
| 1723 |
"model.language_model.layers.62.linear_attn.out_proj.weight": "model-00002-of-00003.safetensors",
|
| 1724 |
"model.language_model.layers.62.linear_attn.out_proj.weight_scale": "model-00002-of-00003.safetensors",
|
| 1725 |
+
"model.language_model.layers.62.mlp.down_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1726 |
"model.language_model.layers.62.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
|
| 1727 |
"model.language_model.layers.62.mlp.down_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1728 |
"model.language_model.layers.62.mlp.down_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1729 |
+
"model.language_model.layers.62.mlp.gate_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1730 |
"model.language_model.layers.62.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
|
| 1731 |
"model.language_model.layers.62.mlp.gate_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1732 |
"model.language_model.layers.62.mlp.gate_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1733 |
+
"model.language_model.layers.62.mlp.up_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1734 |
"model.language_model.layers.62.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
|
| 1735 |
"model.language_model.layers.62.mlp.up_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1736 |
"model.language_model.layers.62.mlp.up_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1737 |
"model.language_model.layers.62.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
|
| 1738 |
"model.language_model.layers.63.input_layernorm.weight": "model-00003-of-00003.safetensors",
|
| 1739 |
+
"model.language_model.layers.63.mlp.down_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1740 |
"model.language_model.layers.63.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
|
| 1741 |
"model.language_model.layers.63.mlp.down_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1742 |
"model.language_model.layers.63.mlp.down_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1743 |
+
"model.language_model.layers.63.mlp.gate_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1744 |
"model.language_model.layers.63.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
|
| 1745 |
"model.language_model.layers.63.mlp.gate_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1746 |
"model.language_model.layers.63.mlp.gate_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
| 1747 |
+
"model.language_model.layers.63.mlp.up_proj.input_scale": "model-00003-of-00003.safetensors",
|
| 1748 |
"model.language_model.layers.63.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
|
| 1749 |
"model.language_model.layers.63.mlp.up_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1750 |
"model.language_model.layers.63.mlp.up_proj.weight_scale_2": "model-00003-of-00003.safetensors",
|
|
|
|
| 1764 |
"model.language_model.layers.63.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
|
| 1765 |
"model.language_model.layers.63.self_attn.v_proj.weight_scale": "model-00003-of-00003.safetensors",
|
| 1766 |
"model.language_model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
|
| 1767 |
+
"model.language_model.layers.7.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1768 |
"model.language_model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1769 |
"model.language_model.layers.7.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1770 |
"model.language_model.layers.7.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1771 |
+
"model.language_model.layers.7.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1772 |
"model.language_model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1773 |
"model.language_model.layers.7.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1774 |
"model.language_model.layers.7.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1775 |
+
"model.language_model.layers.7.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1776 |
"model.language_model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1777 |
"model.language_model.layers.7.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1778 |
"model.language_model.layers.7.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1807 |
"model.language_model.layers.8.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1808 |
"model.language_model.layers.8.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1809 |
"model.language_model.layers.8.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1810 |
+
"model.language_model.layers.8.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1811 |
"model.language_model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1812 |
"model.language_model.layers.8.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1813 |
"model.language_model.layers.8.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1814 |
+
"model.language_model.layers.8.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1815 |
"model.language_model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1816 |
"model.language_model.layers.8.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1817 |
"model.language_model.layers.8.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1818 |
+
"model.language_model.layers.8.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1819 |
"model.language_model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1820 |
"model.language_model.layers.8.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1821 |
"model.language_model.layers.8.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
|
|
|
| 1836 |
"model.language_model.layers.9.linear_attn.out_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1837 |
"model.language_model.layers.9.linear_attn.out_proj.weight": "model-00001-of-00003.safetensors",
|
| 1838 |
"model.language_model.layers.9.linear_attn.out_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1839 |
+
"model.language_model.layers.9.mlp.down_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1840 |
"model.language_model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
|
| 1841 |
"model.language_model.layers.9.mlp.down_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1842 |
"model.language_model.layers.9.mlp.down_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1843 |
+
"model.language_model.layers.9.mlp.gate_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1844 |
"model.language_model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
|
| 1845 |
"model.language_model.layers.9.mlp.gate_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1846 |
"model.language_model.layers.9.mlp.gate_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|
| 1847 |
+
"model.language_model.layers.9.mlp.up_proj.input_scale": "model-00001-of-00003.safetensors",
|
| 1848 |
"model.language_model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
|
| 1849 |
"model.language_model.layers.9.mlp.up_proj.weight_scale": "model-00001-of-00003.safetensors",
|
| 1850 |
"model.language_model.layers.9.mlp.up_proj.weight_scale_2": "model-00001-of-00003.safetensors",
|