Upload folder using huggingface_hub
Browse files- README.md +4 -4
- config.json +1 -1
- model.safetensors +2 -2
- quant_strategy.json +214 -214
- tokenizer.json +6 -24
- tokenizer_config.json +4 -20
README.md
CHANGED
|
@@ -4,9 +4,9 @@ tags:
|
|
| 4 |
- mlx
|
| 5 |
---
|
| 6 |
|
| 7 |
-
# GreenBitAI/01-Yi-6B-
|
| 8 |
-
This quantized low-bit model was converted to MLX format from [`GreenBitAI/01-Yi-6B-
|
| 9 |
-
Refer to the [original model card](https://huggingface.co/GreenBitAI/01-Yi-6B-
|
| 10 |
## Use with mlx
|
| 11 |
|
| 12 |
```bash
|
|
@@ -16,6 +16,6 @@ pip install gbx-lm
|
|
| 16 |
```python
|
| 17 |
from gbx_lm import load, generate
|
| 18 |
|
| 19 |
-
model, tokenizer = load("GreenBitAI/01-Yi-6B-
|
| 20 |
response = generate(model, tokenizer, prompt="hello", verbose=True)
|
| 21 |
```
|
|
|
|
| 4 |
- mlx
|
| 5 |
---
|
| 6 |
|
| 7 |
+
# GreenBitAI/01-Yi-6B-Chat-layer-mix-bpw-3.0-mlx
|
| 8 |
+
This quantized low-bit model was converted to MLX format from [`GreenBitAI/01-Yi-6B-Chat-layer-mix-bpw-3.0`]().
|
| 9 |
+
Refer to the [original model card](https://huggingface.co/GreenBitAI/01-Yi-6B-Chat-layer-mix-bpw-3.0) for more details on the model.
|
| 10 |
## Use with mlx
|
| 11 |
|
| 12 |
```bash
|
|
|
|
| 16 |
```python
|
| 17 |
from gbx_lm import load, generate
|
| 18 |
|
| 19 |
+
model, tokenizer = load("GreenBitAI/01-Yi-6B-Chat-layer-mix-bpw-3.0-mlx")
|
| 20 |
response = generate(model, tokenizer, prompt="hello", verbose=True)
|
| 21 |
```
|
config.json
CHANGED
|
@@ -77,7 +77,7 @@
|
|
| 77 |
"top_p": 1.0,
|
| 78 |
"torch_dtype": "float16",
|
| 79 |
"torchscript": false,
|
| 80 |
-
"transformers_version": "4.39.
|
| 81 |
"typical_p": 1.0,
|
| 82 |
"use_bfloat16": false,
|
| 83 |
"use_cache": true,
|
|
|
|
| 77 |
"top_p": 1.0,
|
| 78 |
"torch_dtype": "float16",
|
| 79 |
"torchscript": false,
|
| 80 |
+
"transformers_version": "4.39.3",
|
| 81 |
"typical_p": 1.0,
|
| 82 |
"use_bfloat16": false,
|
| 83 |
"use_cache": true,
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21a8d50bdf8fbd1540698c8f71db74b9de56d6a375c5d676122c490a05b15432
|
| 3 |
+
size 3358077486
|
quant_strategy.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"measurement": {
|
| 3 |
"model.layers.0": {
|
| 4 |
-
"accuracy": 0.
|
| 5 |
-
"total_bits":
|
| 6 |
"q_proj": {
|
| 7 |
"group_size": {
|
| 8 |
"2": 64
|
|
@@ -17,10 +17,10 @@
|
|
| 17 |
},
|
| 18 |
"k_proj": {
|
| 19 |
"group_size": {
|
| 20 |
-
"
|
| 21 |
},
|
| 22 |
"bits": [
|
| 23 |
-
|
| 24 |
],
|
| 25 |
"bits_prop": [
|
| 26 |
1
|
|
@@ -89,8 +89,8 @@
|
|
| 89 |
}
|
| 90 |
},
|
| 91 |
"model.layers.1": {
|
| 92 |
-
"accuracy": 0.
|
| 93 |
-
"total_bits":
|
| 94 |
"q_proj": {
|
| 95 |
"group_size": {
|
| 96 |
"2": 64
|
|
@@ -105,10 +105,10 @@
|
|
| 105 |
},
|
| 106 |
"k_proj": {
|
| 107 |
"group_size": {
|
| 108 |
-
"
|
| 109 |
},
|
| 110 |
"bits": [
|
| 111 |
-
|
| 112 |
],
|
| 113 |
"bits_prop": [
|
| 114 |
1
|
|
@@ -129,10 +129,10 @@
|
|
| 129 |
},
|
| 130 |
"o_proj": {
|
| 131 |
"group_size": {
|
| 132 |
-
"
|
| 133 |
},
|
| 134 |
"bits": [
|
| 135 |
-
|
| 136 |
],
|
| 137 |
"bits_prop": [
|
| 138 |
1
|
|
@@ -177,14 +177,14 @@
|
|
| 177 |
}
|
| 178 |
},
|
| 179 |
"model.layers.2": {
|
| 180 |
-
"accuracy": 0.
|
| 181 |
-
"total_bits":
|
| 182 |
"q_proj": {
|
| 183 |
"group_size": {
|
| 184 |
-
"
|
| 185 |
},
|
| 186 |
"bits": [
|
| 187 |
-
|
| 188 |
],
|
| 189 |
"bits_prop": [
|
| 190 |
1
|
|
@@ -193,10 +193,10 @@
|
|
| 193 |
},
|
| 194 |
"k_proj": {
|
| 195 |
"group_size": {
|
| 196 |
-
"
|
| 197 |
},
|
| 198 |
"bits": [
|
| 199 |
-
|
| 200 |
],
|
| 201 |
"bits_prop": [
|
| 202 |
1
|
|
@@ -217,10 +217,10 @@
|
|
| 217 |
},
|
| 218 |
"o_proj": {
|
| 219 |
"group_size": {
|
| 220 |
-
"
|
| 221 |
},
|
| 222 |
"bits": [
|
| 223 |
-
|
| 224 |
],
|
| 225 |
"bits_prop": [
|
| 226 |
1
|
|
@@ -253,10 +253,10 @@
|
|
| 253 |
},
|
| 254 |
"down_proj": {
|
| 255 |
"group_size": {
|
| 256 |
-
"
|
| 257 |
},
|
| 258 |
"bits": [
|
| 259 |
-
|
| 260 |
],
|
| 261 |
"bits_prop": [
|
| 262 |
1
|
|
@@ -265,14 +265,14 @@
|
|
| 265 |
}
|
| 266 |
},
|
| 267 |
"model.layers.3": {
|
| 268 |
-
"accuracy": 0.
|
| 269 |
-
"total_bits":
|
| 270 |
"q_proj": {
|
| 271 |
"group_size": {
|
| 272 |
-
"
|
| 273 |
},
|
| 274 |
"bits": [
|
| 275 |
-
|
| 276 |
],
|
| 277 |
"bits_prop": [
|
| 278 |
1
|
|
@@ -305,10 +305,10 @@
|
|
| 305 |
},
|
| 306 |
"o_proj": {
|
| 307 |
"group_size": {
|
| 308 |
-
"
|
| 309 |
},
|
| 310 |
"bits": [
|
| 311 |
-
|
| 312 |
],
|
| 313 |
"bits_prop": [
|
| 314 |
1
|
|
@@ -353,14 +353,14 @@
|
|
| 353 |
}
|
| 354 |
},
|
| 355 |
"model.layers.4": {
|
| 356 |
-
"accuracy": 0.
|
| 357 |
-
"total_bits":
|
| 358 |
"q_proj": {
|
| 359 |
"group_size": {
|
| 360 |
-
"
|
| 361 |
},
|
| 362 |
"bits": [
|
| 363 |
-
|
| 364 |
],
|
| 365 |
"bits_prop": [
|
| 366 |
1
|
|
@@ -369,10 +369,10 @@
|
|
| 369 |
},
|
| 370 |
"k_proj": {
|
| 371 |
"group_size": {
|
| 372 |
-
"
|
| 373 |
},
|
| 374 |
"bits": [
|
| 375 |
-
|
| 376 |
],
|
| 377 |
"bits_prop": [
|
| 378 |
1
|
|
@@ -429,10 +429,10 @@
|
|
| 429 |
},
|
| 430 |
"down_proj": {
|
| 431 |
"group_size": {
|
| 432 |
-
"
|
| 433 |
},
|
| 434 |
"bits": [
|
| 435 |
-
|
| 436 |
],
|
| 437 |
"bits_prop": [
|
| 438 |
1
|
|
@@ -441,14 +441,14 @@
|
|
| 441 |
}
|
| 442 |
},
|
| 443 |
"model.layers.5": {
|
| 444 |
-
"accuracy": 0.
|
| 445 |
-
"total_bits":
|
| 446 |
"q_proj": {
|
| 447 |
"group_size": {
|
| 448 |
-
"
|
| 449 |
},
|
| 450 |
"bits": [
|
| 451 |
-
|
| 452 |
],
|
| 453 |
"bits_prop": [
|
| 454 |
1
|
|
@@ -481,10 +481,10 @@
|
|
| 481 |
},
|
| 482 |
"o_proj": {
|
| 483 |
"group_size": {
|
| 484 |
-
"
|
| 485 |
},
|
| 486 |
"bits": [
|
| 487 |
-
|
| 488 |
],
|
| 489 |
"bits_prop": [
|
| 490 |
1
|
|
@@ -493,10 +493,10 @@
|
|
| 493 |
},
|
| 494 |
"up_proj": {
|
| 495 |
"group_size": {
|
| 496 |
-
"
|
| 497 |
},
|
| 498 |
"bits": [
|
| 499 |
-
|
| 500 |
],
|
| 501 |
"bits_prop": [
|
| 502 |
1
|
|
@@ -505,10 +505,10 @@
|
|
| 505 |
},
|
| 506 |
"gate_proj": {
|
| 507 |
"group_size": {
|
| 508 |
-
"
|
| 509 |
},
|
| 510 |
"bits": [
|
| 511 |
-
|
| 512 |
],
|
| 513 |
"bits_prop": [
|
| 514 |
1
|
|
@@ -517,10 +517,10 @@
|
|
| 517 |
},
|
| 518 |
"down_proj": {
|
| 519 |
"group_size": {
|
| 520 |
-
"
|
| 521 |
},
|
| 522 |
"bits": [
|
| 523 |
-
|
| 524 |
],
|
| 525 |
"bits_prop": [
|
| 526 |
1
|
|
@@ -529,14 +529,14 @@
|
|
| 529 |
}
|
| 530 |
},
|
| 531 |
"model.layers.6": {
|
| 532 |
-
"accuracy": 0.
|
| 533 |
-
"total_bits":
|
| 534 |
"q_proj": {
|
| 535 |
"group_size": {
|
| 536 |
-
"
|
| 537 |
},
|
| 538 |
"bits": [
|
| 539 |
-
|
| 540 |
],
|
| 541 |
"bits_prop": [
|
| 542 |
1
|
|
@@ -545,10 +545,10 @@
|
|
| 545 |
},
|
| 546 |
"k_proj": {
|
| 547 |
"group_size": {
|
| 548 |
-
"
|
| 549 |
},
|
| 550 |
"bits": [
|
| 551 |
-
|
| 552 |
],
|
| 553 |
"bits_prop": [
|
| 554 |
1
|
|
@@ -569,10 +569,10 @@
|
|
| 569 |
},
|
| 570 |
"o_proj": {
|
| 571 |
"group_size": {
|
| 572 |
-
"
|
| 573 |
},
|
| 574 |
"bits": [
|
| 575 |
-
|
| 576 |
],
|
| 577 |
"bits_prop": [
|
| 578 |
1
|
|
@@ -581,10 +581,10 @@
|
|
| 581 |
},
|
| 582 |
"up_proj": {
|
| 583 |
"group_size": {
|
| 584 |
-
"
|
| 585 |
},
|
| 586 |
"bits": [
|
| 587 |
-
|
| 588 |
],
|
| 589 |
"bits_prop": [
|
| 590 |
1
|
|
@@ -593,10 +593,10 @@
|
|
| 593 |
},
|
| 594 |
"gate_proj": {
|
| 595 |
"group_size": {
|
| 596 |
-
"
|
| 597 |
},
|
| 598 |
"bits": [
|
| 599 |
-
|
| 600 |
],
|
| 601 |
"bits_prop": [
|
| 602 |
1
|
|
@@ -617,14 +617,14 @@
|
|
| 617 |
}
|
| 618 |
},
|
| 619 |
"model.layers.7": {
|
| 620 |
-
"accuracy": 0.
|
| 621 |
-
"total_bits":
|
| 622 |
"q_proj": {
|
| 623 |
"group_size": {
|
| 624 |
-
"
|
| 625 |
},
|
| 626 |
"bits": [
|
| 627 |
-
|
| 628 |
],
|
| 629 |
"bits_prop": [
|
| 630 |
1
|
|
@@ -633,10 +633,10 @@
|
|
| 633 |
},
|
| 634 |
"k_proj": {
|
| 635 |
"group_size": {
|
| 636 |
-
"
|
| 637 |
},
|
| 638 |
"bits": [
|
| 639 |
-
|
| 640 |
],
|
| 641 |
"bits_prop": [
|
| 642 |
1
|
|
@@ -669,10 +669,10 @@
|
|
| 669 |
},
|
| 670 |
"up_proj": {
|
| 671 |
"group_size": {
|
| 672 |
-
"
|
| 673 |
},
|
| 674 |
"bits": [
|
| 675 |
-
|
| 676 |
],
|
| 677 |
"bits_prop": [
|
| 678 |
1
|
|
@@ -681,10 +681,10 @@
|
|
| 681 |
},
|
| 682 |
"gate_proj": {
|
| 683 |
"group_size": {
|
| 684 |
-
"
|
| 685 |
},
|
| 686 |
"bits": [
|
| 687 |
-
|
| 688 |
],
|
| 689 |
"bits_prop": [
|
| 690 |
1
|
|
@@ -705,8 +705,8 @@
|
|
| 705 |
}
|
| 706 |
},
|
| 707 |
"model.layers.8": {
|
| 708 |
-
"accuracy": 0.
|
| 709 |
-
"total_bits":
|
| 710 |
"q_proj": {
|
| 711 |
"group_size": {
|
| 712 |
"2": 64
|
|
@@ -721,10 +721,10 @@
|
|
| 721 |
},
|
| 722 |
"k_proj": {
|
| 723 |
"group_size": {
|
| 724 |
-
"
|
| 725 |
},
|
| 726 |
"bits": [
|
| 727 |
-
|
| 728 |
],
|
| 729 |
"bits_prop": [
|
| 730 |
1
|
|
@@ -745,10 +745,10 @@
|
|
| 745 |
},
|
| 746 |
"o_proj": {
|
| 747 |
"group_size": {
|
| 748 |
-
"
|
| 749 |
},
|
| 750 |
"bits": [
|
| 751 |
-
|
| 752 |
],
|
| 753 |
"bits_prop": [
|
| 754 |
1
|
|
@@ -793,14 +793,14 @@
|
|
| 793 |
}
|
| 794 |
},
|
| 795 |
"model.layers.9": {
|
| 796 |
-
"accuracy": 0.
|
| 797 |
-
"total_bits":
|
| 798 |
"q_proj": {
|
| 799 |
"group_size": {
|
| 800 |
-
"
|
| 801 |
},
|
| 802 |
"bits": [
|
| 803 |
-
|
| 804 |
],
|
| 805 |
"bits_prop": [
|
| 806 |
1
|
|
@@ -809,10 +809,10 @@
|
|
| 809 |
},
|
| 810 |
"k_proj": {
|
| 811 |
"group_size": {
|
| 812 |
-
"
|
| 813 |
},
|
| 814 |
"bits": [
|
| 815 |
-
|
| 816 |
],
|
| 817 |
"bits_prop": [
|
| 818 |
1
|
|
@@ -881,8 +881,8 @@
|
|
| 881 |
}
|
| 882 |
},
|
| 883 |
"model.layers.10": {
|
| 884 |
-
"accuracy": 0.
|
| 885 |
-
"total_bits":
|
| 886 |
"q_proj": {
|
| 887 |
"group_size": {
|
| 888 |
"2": 64
|
|
@@ -921,10 +921,10 @@
|
|
| 921 |
},
|
| 922 |
"o_proj": {
|
| 923 |
"group_size": {
|
| 924 |
-
"
|
| 925 |
},
|
| 926 |
"bits": [
|
| 927 |
-
|
| 928 |
],
|
| 929 |
"bits_prop": [
|
| 930 |
1
|
|
@@ -969,14 +969,14 @@
|
|
| 969 |
}
|
| 970 |
},
|
| 971 |
"model.layers.11": {
|
| 972 |
-
"accuracy": 0.
|
| 973 |
-
"total_bits":
|
| 974 |
"q_proj": {
|
| 975 |
"group_size": {
|
| 976 |
-
"
|
| 977 |
},
|
| 978 |
"bits": [
|
| 979 |
-
|
| 980 |
],
|
| 981 |
"bits_prop": [
|
| 982 |
1
|
|
@@ -1057,8 +1057,8 @@
|
|
| 1057 |
}
|
| 1058 |
},
|
| 1059 |
"model.layers.12": {
|
| 1060 |
-
"accuracy": 0.
|
| 1061 |
-
"total_bits":
|
| 1062 |
"q_proj": {
|
| 1063 |
"group_size": {
|
| 1064 |
"2": 64
|
|
@@ -1073,10 +1073,10 @@
|
|
| 1073 |
},
|
| 1074 |
"k_proj": {
|
| 1075 |
"group_size": {
|
| 1076 |
-
"
|
| 1077 |
},
|
| 1078 |
"bits": [
|
| 1079 |
-
|
| 1080 |
],
|
| 1081 |
"bits_prop": [
|
| 1082 |
1
|
|
@@ -1097,10 +1097,10 @@
|
|
| 1097 |
},
|
| 1098 |
"o_proj": {
|
| 1099 |
"group_size": {
|
| 1100 |
-
"
|
| 1101 |
},
|
| 1102 |
"bits": [
|
| 1103 |
-
|
| 1104 |
],
|
| 1105 |
"bits_prop": [
|
| 1106 |
1
|
|
@@ -1145,8 +1145,8 @@
|
|
| 1145 |
}
|
| 1146 |
},
|
| 1147 |
"model.layers.13": {
|
| 1148 |
-
"accuracy": 0.
|
| 1149 |
-
"total_bits":
|
| 1150 |
"q_proj": {
|
| 1151 |
"group_size": {
|
| 1152 |
"2": 64
|
|
@@ -1161,10 +1161,10 @@
|
|
| 1161 |
},
|
| 1162 |
"k_proj": {
|
| 1163 |
"group_size": {
|
| 1164 |
-
"
|
| 1165 |
},
|
| 1166 |
"bits": [
|
| 1167 |
-
|
| 1168 |
],
|
| 1169 |
"bits_prop": [
|
| 1170 |
1
|
|
@@ -1185,10 +1185,10 @@
|
|
| 1185 |
},
|
| 1186 |
"o_proj": {
|
| 1187 |
"group_size": {
|
| 1188 |
-
"
|
| 1189 |
},
|
| 1190 |
"bits": [
|
| 1191 |
-
|
| 1192 |
],
|
| 1193 |
"bits_prop": [
|
| 1194 |
1
|
|
@@ -1233,8 +1233,8 @@
|
|
| 1233 |
}
|
| 1234 |
},
|
| 1235 |
"model.layers.14": {
|
| 1236 |
-
"accuracy": 0.
|
| 1237 |
-
"total_bits":
|
| 1238 |
"q_proj": {
|
| 1239 |
"group_size": {
|
| 1240 |
"2": 64
|
|
@@ -1249,10 +1249,10 @@
|
|
| 1249 |
},
|
| 1250 |
"k_proj": {
|
| 1251 |
"group_size": {
|
| 1252 |
-
"
|
| 1253 |
},
|
| 1254 |
"bits": [
|
| 1255 |
-
|
| 1256 |
],
|
| 1257 |
"bits_prop": [
|
| 1258 |
1
|
|
@@ -1273,10 +1273,10 @@
|
|
| 1273 |
},
|
| 1274 |
"o_proj": {
|
| 1275 |
"group_size": {
|
| 1276 |
-
"
|
| 1277 |
},
|
| 1278 |
"bits": [
|
| 1279 |
-
|
| 1280 |
],
|
| 1281 |
"bits_prop": [
|
| 1282 |
1
|
|
@@ -1321,8 +1321,8 @@
|
|
| 1321 |
}
|
| 1322 |
},
|
| 1323 |
"model.layers.15": {
|
| 1324 |
-
"accuracy": 0.
|
| 1325 |
-
"total_bits":
|
| 1326 |
"q_proj": {
|
| 1327 |
"group_size": {
|
| 1328 |
"2": 64
|
|
@@ -1361,10 +1361,10 @@
|
|
| 1361 |
},
|
| 1362 |
"o_proj": {
|
| 1363 |
"group_size": {
|
| 1364 |
-
"
|
| 1365 |
},
|
| 1366 |
"bits": [
|
| 1367 |
-
|
| 1368 |
],
|
| 1369 |
"bits_prop": [
|
| 1370 |
1
|
|
@@ -1397,10 +1397,10 @@
|
|
| 1397 |
},
|
| 1398 |
"down_proj": {
|
| 1399 |
"group_size": {
|
| 1400 |
-
"
|
| 1401 |
},
|
| 1402 |
"bits": [
|
| 1403 |
-
|
| 1404 |
],
|
| 1405 |
"bits_prop": [
|
| 1406 |
1
|
|
@@ -1409,8 +1409,8 @@
|
|
| 1409 |
}
|
| 1410 |
},
|
| 1411 |
"model.layers.16": {
|
| 1412 |
-
"accuracy": 0.
|
| 1413 |
-
"total_bits":
|
| 1414 |
"q_proj": {
|
| 1415 |
"group_size": {
|
| 1416 |
"2": 64
|
|
@@ -1425,10 +1425,10 @@
|
|
| 1425 |
},
|
| 1426 |
"k_proj": {
|
| 1427 |
"group_size": {
|
| 1428 |
-
"
|
| 1429 |
},
|
| 1430 |
"bits": [
|
| 1431 |
-
|
| 1432 |
],
|
| 1433 |
"bits_prop": [
|
| 1434 |
1
|
|
@@ -1449,10 +1449,10 @@
|
|
| 1449 |
},
|
| 1450 |
"o_proj": {
|
| 1451 |
"group_size": {
|
| 1452 |
-
"
|
| 1453 |
},
|
| 1454 |
"bits": [
|
| 1455 |
-
|
| 1456 |
],
|
| 1457 |
"bits_prop": [
|
| 1458 |
1
|
|
@@ -1497,14 +1497,14 @@
|
|
| 1497 |
}
|
| 1498 |
},
|
| 1499 |
"model.layers.17": {
|
| 1500 |
-
"accuracy": 0.
|
| 1501 |
-
"total_bits":
|
| 1502 |
"q_proj": {
|
| 1503 |
"group_size": {
|
| 1504 |
-
"
|
| 1505 |
},
|
| 1506 |
"bits": [
|
| 1507 |
-
|
| 1508 |
],
|
| 1509 |
"bits_prop": [
|
| 1510 |
1
|
|
@@ -1513,10 +1513,10 @@
|
|
| 1513 |
},
|
| 1514 |
"k_proj": {
|
| 1515 |
"group_size": {
|
| 1516 |
-
"
|
| 1517 |
},
|
| 1518 |
"bits": [
|
| 1519 |
-
|
| 1520 |
],
|
| 1521 |
"bits_prop": [
|
| 1522 |
1
|
|
@@ -1585,8 +1585,8 @@
|
|
| 1585 |
}
|
| 1586 |
},
|
| 1587 |
"model.layers.18": {
|
| 1588 |
-
"accuracy": 0.
|
| 1589 |
-
"total_bits":
|
| 1590 |
"q_proj": {
|
| 1591 |
"group_size": {
|
| 1592 |
"4": 128
|
|
@@ -1601,10 +1601,10 @@
|
|
| 1601 |
},
|
| 1602 |
"k_proj": {
|
| 1603 |
"group_size": {
|
| 1604 |
-
"
|
| 1605 |
},
|
| 1606 |
"bits": [
|
| 1607 |
-
|
| 1608 |
],
|
| 1609 |
"bits_prop": [
|
| 1610 |
1
|
|
@@ -1673,14 +1673,14 @@
|
|
| 1673 |
}
|
| 1674 |
},
|
| 1675 |
"model.layers.19": {
|
| 1676 |
-
"accuracy": 0.
|
| 1677 |
-
"total_bits":
|
| 1678 |
"q_proj": {
|
| 1679 |
"group_size": {
|
| 1680 |
-
"
|
| 1681 |
},
|
| 1682 |
"bits": [
|
| 1683 |
-
|
| 1684 |
],
|
| 1685 |
"bits_prop": [
|
| 1686 |
1
|
|
@@ -1725,10 +1725,10 @@
|
|
| 1725 |
},
|
| 1726 |
"up_proj": {
|
| 1727 |
"group_size": {
|
| 1728 |
-
"
|
| 1729 |
},
|
| 1730 |
"bits": [
|
| 1731 |
-
|
| 1732 |
],
|
| 1733 |
"bits_prop": [
|
| 1734 |
1
|
|
@@ -1761,8 +1761,8 @@
|
|
| 1761 |
}
|
| 1762 |
},
|
| 1763 |
"model.layers.20": {
|
| 1764 |
-
"accuracy": 0.
|
| 1765 |
-
"total_bits":
|
| 1766 |
"q_proj": {
|
| 1767 |
"group_size": {
|
| 1768 |
"4": 128
|
|
@@ -1777,10 +1777,10 @@
|
|
| 1777 |
},
|
| 1778 |
"k_proj": {
|
| 1779 |
"group_size": {
|
| 1780 |
-
"
|
| 1781 |
},
|
| 1782 |
"bits": [
|
| 1783 |
-
|
| 1784 |
],
|
| 1785 |
"bits_prop": [
|
| 1786 |
1
|
|
@@ -1813,10 +1813,10 @@
|
|
| 1813 |
},
|
| 1814 |
"up_proj": {
|
| 1815 |
"group_size": {
|
| 1816 |
-
"
|
| 1817 |
},
|
| 1818 |
"bits": [
|
| 1819 |
-
|
| 1820 |
],
|
| 1821 |
"bits_prop": [
|
| 1822 |
1
|
|
@@ -1849,14 +1849,14 @@
|
|
| 1849 |
}
|
| 1850 |
},
|
| 1851 |
"model.layers.21": {
|
| 1852 |
-
"accuracy": 0.
|
| 1853 |
-
"total_bits":
|
| 1854 |
"q_proj": {
|
| 1855 |
"group_size": {
|
| 1856 |
-
"
|
| 1857 |
},
|
| 1858 |
"bits": [
|
| 1859 |
-
|
| 1860 |
],
|
| 1861 |
"bits_prop": [
|
| 1862 |
1
|
|
@@ -1901,10 +1901,10 @@
|
|
| 1901 |
},
|
| 1902 |
"up_proj": {
|
| 1903 |
"group_size": {
|
| 1904 |
-
"
|
| 1905 |
},
|
| 1906 |
"bits": [
|
| 1907 |
-
|
| 1908 |
],
|
| 1909 |
"bits_prop": [
|
| 1910 |
1
|
|
@@ -1913,10 +1913,10 @@
|
|
| 1913 |
},
|
| 1914 |
"gate_proj": {
|
| 1915 |
"group_size": {
|
| 1916 |
-
"
|
| 1917 |
},
|
| 1918 |
"bits": [
|
| 1919 |
-
|
| 1920 |
],
|
| 1921 |
"bits_prop": [
|
| 1922 |
1
|
|
@@ -1937,8 +1937,8 @@
|
|
| 1937 |
}
|
| 1938 |
},
|
| 1939 |
"model.layers.22": {
|
| 1940 |
-
"accuracy": 0.
|
| 1941 |
-
"total_bits":
|
| 1942 |
"q_proj": {
|
| 1943 |
"group_size": {
|
| 1944 |
"4": 128
|
|
@@ -1953,10 +1953,10 @@
|
|
| 1953 |
},
|
| 1954 |
"k_proj": {
|
| 1955 |
"group_size": {
|
| 1956 |
-
"
|
| 1957 |
},
|
| 1958 |
"bits": [
|
| 1959 |
-
|
| 1960 |
],
|
| 1961 |
"bits_prop": [
|
| 1962 |
1
|
|
@@ -2001,10 +2001,10 @@
|
|
| 2001 |
},
|
| 2002 |
"gate_proj": {
|
| 2003 |
"group_size": {
|
| 2004 |
-
"
|
| 2005 |
},
|
| 2006 |
"bits": [
|
| 2007 |
-
|
| 2008 |
],
|
| 2009 |
"bits_prop": [
|
| 2010 |
1
|
|
@@ -2025,8 +2025,8 @@
|
|
| 2025 |
}
|
| 2026 |
},
|
| 2027 |
"model.layers.23": {
|
| 2028 |
-
"accuracy": 0.
|
| 2029 |
-
"total_bits":
|
| 2030 |
"q_proj": {
|
| 2031 |
"group_size": {
|
| 2032 |
"4": 128
|
|
@@ -2041,10 +2041,10 @@
|
|
| 2041 |
},
|
| 2042 |
"k_proj": {
|
| 2043 |
"group_size": {
|
| 2044 |
-
"
|
| 2045 |
},
|
| 2046 |
"bits": [
|
| 2047 |
-
|
| 2048 |
],
|
| 2049 |
"bits_prop": [
|
| 2050 |
1
|
|
@@ -2089,10 +2089,10 @@
|
|
| 2089 |
},
|
| 2090 |
"gate_proj": {
|
| 2091 |
"group_size": {
|
| 2092 |
-
"
|
| 2093 |
},
|
| 2094 |
"bits": [
|
| 2095 |
-
|
| 2096 |
],
|
| 2097 |
"bits_prop": [
|
| 2098 |
1
|
|
@@ -2113,8 +2113,8 @@
|
|
| 2113 |
}
|
| 2114 |
},
|
| 2115 |
"model.layers.24": {
|
| 2116 |
-
"accuracy": 0.
|
| 2117 |
-
"total_bits":
|
| 2118 |
"q_proj": {
|
| 2119 |
"group_size": {
|
| 2120 |
"4": 128
|
|
@@ -2129,10 +2129,10 @@
|
|
| 2129 |
},
|
| 2130 |
"k_proj": {
|
| 2131 |
"group_size": {
|
| 2132 |
-
"
|
| 2133 |
},
|
| 2134 |
"bits": [
|
| 2135 |
-
|
| 2136 |
],
|
| 2137 |
"bits_prop": [
|
| 2138 |
1
|
|
@@ -2177,10 +2177,10 @@
|
|
| 2177 |
},
|
| 2178 |
"gate_proj": {
|
| 2179 |
"group_size": {
|
| 2180 |
-
"
|
| 2181 |
},
|
| 2182 |
"bits": [
|
| 2183 |
-
|
| 2184 |
],
|
| 2185 |
"bits_prop": [
|
| 2186 |
1
|
|
@@ -2201,8 +2201,8 @@
|
|
| 2201 |
}
|
| 2202 |
},
|
| 2203 |
"model.layers.25": {
|
| 2204 |
-
"accuracy": 0.
|
| 2205 |
-
"total_bits":
|
| 2206 |
"q_proj": {
|
| 2207 |
"group_size": {
|
| 2208 |
"4": 128
|
|
@@ -2217,10 +2217,10 @@
|
|
| 2217 |
},
|
| 2218 |
"k_proj": {
|
| 2219 |
"group_size": {
|
| 2220 |
-
"
|
| 2221 |
},
|
| 2222 |
"bits": [
|
| 2223 |
-
|
| 2224 |
],
|
| 2225 |
"bits_prop": [
|
| 2226 |
1
|
|
@@ -2265,10 +2265,10 @@
|
|
| 2265 |
},
|
| 2266 |
"gate_proj": {
|
| 2267 |
"group_size": {
|
| 2268 |
-
"
|
| 2269 |
},
|
| 2270 |
"bits": [
|
| 2271 |
-
|
| 2272 |
],
|
| 2273 |
"bits_prop": [
|
| 2274 |
1
|
|
@@ -2289,14 +2289,14 @@
|
|
| 2289 |
}
|
| 2290 |
},
|
| 2291 |
"model.layers.26": {
|
| 2292 |
-
"accuracy": 0.
|
| 2293 |
-
"total_bits":
|
| 2294 |
"q_proj": {
|
| 2295 |
"group_size": {
|
| 2296 |
-
"
|
| 2297 |
},
|
| 2298 |
"bits": [
|
| 2299 |
-
|
| 2300 |
],
|
| 2301 |
"bits_prop": [
|
| 2302 |
1
|
|
@@ -2305,10 +2305,10 @@
|
|
| 2305 |
},
|
| 2306 |
"k_proj": {
|
| 2307 |
"group_size": {
|
| 2308 |
-
"
|
| 2309 |
},
|
| 2310 |
"bits": [
|
| 2311 |
-
|
| 2312 |
],
|
| 2313 |
"bits_prop": [
|
| 2314 |
1
|
|
@@ -2341,10 +2341,10 @@
|
|
| 2341 |
},
|
| 2342 |
"up_proj": {
|
| 2343 |
"group_size": {
|
| 2344 |
-
"
|
| 2345 |
},
|
| 2346 |
"bits": [
|
| 2347 |
-
|
| 2348 |
],
|
| 2349 |
"bits_prop": [
|
| 2350 |
1
|
|
@@ -2377,14 +2377,14 @@
|
|
| 2377 |
}
|
| 2378 |
},
|
| 2379 |
"model.layers.27": {
|
| 2380 |
-
"accuracy": 0.
|
| 2381 |
-
"total_bits":
|
| 2382 |
"q_proj": {
|
| 2383 |
"group_size": {
|
| 2384 |
-
"
|
| 2385 |
},
|
| 2386 |
"bits": [
|
| 2387 |
-
|
| 2388 |
],
|
| 2389 |
"bits_prop": [
|
| 2390 |
1
|
|
@@ -2393,10 +2393,10 @@
|
|
| 2393 |
},
|
| 2394 |
"k_proj": {
|
| 2395 |
"group_size": {
|
| 2396 |
-
"
|
| 2397 |
},
|
| 2398 |
"bits": [
|
| 2399 |
-
|
| 2400 |
],
|
| 2401 |
"bits_prop": [
|
| 2402 |
1
|
|
@@ -2441,10 +2441,10 @@
|
|
| 2441 |
},
|
| 2442 |
"gate_proj": {
|
| 2443 |
"group_size": {
|
| 2444 |
-
"
|
| 2445 |
},
|
| 2446 |
"bits": [
|
| 2447 |
-
|
| 2448 |
],
|
| 2449 |
"bits_prop": [
|
| 2450 |
1
|
|
@@ -2465,8 +2465,8 @@
|
|
| 2465 |
}
|
| 2466 |
},
|
| 2467 |
"model.layers.28": {
|
| 2468 |
-
"accuracy": 0.
|
| 2469 |
-
"total_bits":
|
| 2470 |
"q_proj": {
|
| 2471 |
"group_size": {
|
| 2472 |
"4": 128
|
|
@@ -2517,10 +2517,10 @@
|
|
| 2517 |
},
|
| 2518 |
"up_proj": {
|
| 2519 |
"group_size": {
|
| 2520 |
-
"
|
| 2521 |
},
|
| 2522 |
"bits": [
|
| 2523 |
-
|
| 2524 |
],
|
| 2525 |
"bits_prop": [
|
| 2526 |
1
|
|
@@ -2529,10 +2529,10 @@
|
|
| 2529 |
},
|
| 2530 |
"gate_proj": {
|
| 2531 |
"group_size": {
|
| 2532 |
-
"
|
| 2533 |
},
|
| 2534 |
"bits": [
|
| 2535 |
-
|
| 2536 |
],
|
| 2537 |
"bits_prop": [
|
| 2538 |
1
|
|
@@ -2553,14 +2553,14 @@
|
|
| 2553 |
}
|
| 2554 |
},
|
| 2555 |
"model.layers.29": {
|
| 2556 |
-
"accuracy": 0.
|
| 2557 |
-
"total_bits":
|
| 2558 |
"q_proj": {
|
| 2559 |
"group_size": {
|
| 2560 |
-
"
|
| 2561 |
},
|
| 2562 |
"bits": [
|
| 2563 |
-
|
| 2564 |
],
|
| 2565 |
"bits_prop": [
|
| 2566 |
1
|
|
@@ -2569,10 +2569,10 @@
|
|
| 2569 |
},
|
| 2570 |
"k_proj": {
|
| 2571 |
"group_size": {
|
| 2572 |
-
"
|
| 2573 |
},
|
| 2574 |
"bits": [
|
| 2575 |
-
|
| 2576 |
],
|
| 2577 |
"bits_prop": [
|
| 2578 |
1
|
|
@@ -2605,10 +2605,10 @@
|
|
| 2605 |
},
|
| 2606 |
"up_proj": {
|
| 2607 |
"group_size": {
|
| 2608 |
-
"
|
| 2609 |
},
|
| 2610 |
"bits": [
|
| 2611 |
-
|
| 2612 |
],
|
| 2613 |
"bits_prop": [
|
| 2614 |
1
|
|
@@ -2617,10 +2617,10 @@
|
|
| 2617 |
},
|
| 2618 |
"gate_proj": {
|
| 2619 |
"group_size": {
|
| 2620 |
-
"
|
| 2621 |
},
|
| 2622 |
"bits": [
|
| 2623 |
-
|
| 2624 |
],
|
| 2625 |
"bits_prop": [
|
| 2626 |
1
|
|
@@ -2641,14 +2641,14 @@
|
|
| 2641 |
}
|
| 2642 |
},
|
| 2643 |
"model.layers.30": {
|
| 2644 |
-
"accuracy": 0.
|
| 2645 |
-
"total_bits":
|
| 2646 |
"q_proj": {
|
| 2647 |
"group_size": {
|
| 2648 |
-
"
|
| 2649 |
},
|
| 2650 |
"bits": [
|
| 2651 |
-
|
| 2652 |
],
|
| 2653 |
"bits_prop": [
|
| 2654 |
1
|
|
@@ -2657,10 +2657,10 @@
|
|
| 2657 |
},
|
| 2658 |
"k_proj": {
|
| 2659 |
"group_size": {
|
| 2660 |
-
"
|
| 2661 |
},
|
| 2662 |
"bits": [
|
| 2663 |
-
|
| 2664 |
],
|
| 2665 |
"bits_prop": [
|
| 2666 |
1
|
|
@@ -2693,10 +2693,10 @@
|
|
| 2693 |
},
|
| 2694 |
"up_proj": {
|
| 2695 |
"group_size": {
|
| 2696 |
-
"
|
| 2697 |
},
|
| 2698 |
"bits": [
|
| 2699 |
-
|
| 2700 |
],
|
| 2701 |
"bits_prop": [
|
| 2702 |
1
|
|
@@ -2705,10 +2705,10 @@
|
|
| 2705 |
},
|
| 2706 |
"gate_proj": {
|
| 2707 |
"group_size": {
|
| 2708 |
-
"
|
| 2709 |
},
|
| 2710 |
"bits": [
|
| 2711 |
-
|
| 2712 |
],
|
| 2713 |
"bits_prop": [
|
| 2714 |
1
|
|
|
|
| 1 |
{
|
| 2 |
"measurement": {
|
| 3 |
"model.layers.0": {
|
| 4 |
+
"accuracy": 0.9242749214172363,
|
| 5 |
+
"total_bits": 360997440,
|
| 6 |
"q_proj": {
|
| 7 |
"group_size": {
|
| 8 |
"2": 64
|
|
|
|
| 17 |
},
|
| 18 |
"k_proj": {
|
| 19 |
"group_size": {
|
| 20 |
+
"2": 64
|
| 21 |
},
|
| 22 |
"bits": [
|
| 23 |
+
2
|
| 24 |
],
|
| 25 |
"bits_prop": [
|
| 26 |
1
|
|
|
|
| 89 |
}
|
| 90 |
},
|
| 91 |
"model.layers.1": {
|
| 92 |
+
"accuracy": 0.9216856956481934,
|
| 93 |
+
"total_bits": 486917664,
|
| 94 |
"q_proj": {
|
| 95 |
"group_size": {
|
| 96 |
"2": 64
|
|
|
|
| 105 |
},
|
| 106 |
"k_proj": {
|
| 107 |
"group_size": {
|
| 108 |
+
"4": 128
|
| 109 |
},
|
| 110 |
"bits": [
|
| 111 |
+
4
|
| 112 |
],
|
| 113 |
"bits_prop": [
|
| 114 |
1
|
|
|
|
| 129 |
},
|
| 130 |
"o_proj": {
|
| 131 |
"group_size": {
|
| 132 |
+
"4": 128
|
| 133 |
},
|
| 134 |
"bits": [
|
| 135 |
+
4
|
| 136 |
],
|
| 137 |
"bits_prop": [
|
| 138 |
1
|
|
|
|
| 177 |
}
|
| 178 |
},
|
| 179 |
"model.layers.2": {
|
| 180 |
+
"accuracy": 0.8546795845031738,
|
| 181 |
+
"total_bits": 360997440,
|
| 182 |
"q_proj": {
|
| 183 |
"group_size": {
|
| 184 |
+
"2": 64
|
| 185 |
},
|
| 186 |
"bits": [
|
| 187 |
+
2
|
| 188 |
],
|
| 189 |
"bits_prop": [
|
| 190 |
1
|
|
|
|
| 193 |
},
|
| 194 |
"k_proj": {
|
| 195 |
"group_size": {
|
| 196 |
+
"2": 64
|
| 197 |
},
|
| 198 |
"bits": [
|
| 199 |
+
2
|
| 200 |
],
|
| 201 |
"bits_prop": [
|
| 202 |
1
|
|
|
|
| 217 |
},
|
| 218 |
"o_proj": {
|
| 219 |
"group_size": {
|
| 220 |
+
"2": 64
|
| 221 |
},
|
| 222 |
"bits": [
|
| 223 |
+
2
|
| 224 |
],
|
| 225 |
"bits_prop": [
|
| 226 |
1
|
|
|
|
| 253 |
},
|
| 254 |
"down_proj": {
|
| 255 |
"group_size": {
|
| 256 |
+
"2": 64
|
| 257 |
},
|
| 258 |
"bits": [
|
| 259 |
+
2
|
| 260 |
],
|
| 261 |
"bits_prop": [
|
| 262 |
1
|
|
|
|
| 265 |
}
|
| 266 |
},
|
| 267 |
"model.layers.3": {
|
| 268 |
+
"accuracy": 0.9084997177124023,
|
| 269 |
+
"total_bits": 449761824,
|
| 270 |
"q_proj": {
|
| 271 |
"group_size": {
|
| 272 |
+
"2": 64
|
| 273 |
},
|
| 274 |
"bits": [
|
| 275 |
+
2
|
| 276 |
],
|
| 277 |
"bits_prop": [
|
| 278 |
1
|
|
|
|
| 305 |
},
|
| 306 |
"o_proj": {
|
| 307 |
"group_size": {
|
| 308 |
+
"2": 64
|
| 309 |
},
|
| 310 |
"bits": [
|
| 311 |
+
2
|
| 312 |
],
|
| 313 |
"bits_prop": [
|
| 314 |
1
|
|
|
|
| 353 |
}
|
| 354 |
},
|
| 355 |
"model.layers.4": {
|
| 356 |
+
"accuracy": 0.8643641471862793,
|
| 357 |
+
"total_bits": 394026048,
|
| 358 |
"q_proj": {
|
| 359 |
"group_size": {
|
| 360 |
+
"2": 64
|
| 361 |
},
|
| 362 |
"bits": [
|
| 363 |
+
2
|
| 364 |
],
|
| 365 |
"bits_prop": [
|
| 366 |
1
|
|
|
|
| 369 |
},
|
| 370 |
"k_proj": {
|
| 371 |
"group_size": {
|
| 372 |
+
"2": 64
|
| 373 |
},
|
| 374 |
"bits": [
|
| 375 |
+
2
|
| 376 |
],
|
| 377 |
"bits_prop": [
|
| 378 |
1
|
|
|
|
| 429 |
},
|
| 430 |
"down_proj": {
|
| 431 |
"group_size": {
|
| 432 |
+
"2": 64
|
| 433 |
},
|
| 434 |
"bits": [
|
| 435 |
+
2
|
| 436 |
],
|
| 437 |
"bits_prop": [
|
| 438 |
1
|
|
|
|
| 441 |
}
|
| 442 |
},
|
| 443 |
"model.layers.5": {
|
| 444 |
+
"accuracy": 0.8657441139221191,
|
| 445 |
+
"total_bits": 365124672,
|
| 446 |
"q_proj": {
|
| 447 |
"group_size": {
|
| 448 |
+
"2": 64
|
| 449 |
},
|
| 450 |
"bits": [
|
| 451 |
+
2
|
| 452 |
],
|
| 453 |
"bits_prop": [
|
| 454 |
1
|
|
|
|
| 481 |
},
|
| 482 |
"o_proj": {
|
| 483 |
"group_size": {
|
| 484 |
+
"2": 64
|
| 485 |
},
|
| 486 |
"bits": [
|
| 487 |
+
2
|
| 488 |
],
|
| 489 |
"bits_prop": [
|
| 490 |
1
|
|
|
|
| 493 |
},
|
| 494 |
"up_proj": {
|
| 495 |
"group_size": {
|
| 496 |
+
"2": 64
|
| 497 |
},
|
| 498 |
"bits": [
|
| 499 |
+
2
|
| 500 |
],
|
| 501 |
"bits_prop": [
|
| 502 |
1
|
|
|
|
| 505 |
},
|
| 506 |
"gate_proj": {
|
| 507 |
"group_size": {
|
| 508 |
+
"2": 64
|
| 509 |
},
|
| 510 |
"bits": [
|
| 511 |
+
2
|
| 512 |
],
|
| 513 |
"bits_prop": [
|
| 514 |
1
|
|
|
|
| 517 |
},
|
| 518 |
"down_proj": {
|
| 519 |
"group_size": {
|
| 520 |
+
"2": 64
|
| 521 |
},
|
| 522 |
"bits": [
|
| 523 |
+
2
|
| 524 |
],
|
| 525 |
"bits_prop": [
|
| 526 |
1
|
|
|
|
| 529 |
}
|
| 530 |
},
|
| 531 |
"model.layers.6": {
|
| 532 |
+
"accuracy": 0.877474308013916,
|
| 533 |
+
"total_bits": 449761824,
|
| 534 |
"q_proj": {
|
| 535 |
"group_size": {
|
| 536 |
+
"2": 64
|
| 537 |
},
|
| 538 |
"bits": [
|
| 539 |
+
2
|
| 540 |
],
|
| 541 |
"bits_prop": [
|
| 542 |
1
|
|
|
|
| 545 |
},
|
| 546 |
"k_proj": {
|
| 547 |
"group_size": {
|
| 548 |
+
"2": 64
|
| 549 |
},
|
| 550 |
"bits": [
|
| 551 |
+
2
|
| 552 |
],
|
| 553 |
"bits_prop": [
|
| 554 |
1
|
|
|
|
| 569 |
},
|
| 570 |
"o_proj": {
|
| 571 |
"group_size": {
|
| 572 |
+
"2": 64
|
| 573 |
},
|
| 574 |
"bits": [
|
| 575 |
+
2
|
| 576 |
],
|
| 577 |
"bits_prop": [
|
| 578 |
1
|
|
|
|
| 581 |
},
|
| 582 |
"up_proj": {
|
| 583 |
"group_size": {
|
| 584 |
+
"2": 64
|
| 585 |
},
|
| 586 |
"bits": [
|
| 587 |
+
2
|
| 588 |
],
|
| 589 |
"bits_prop": [
|
| 590 |
1
|
|
|
|
| 593 |
},
|
| 594 |
"gate_proj": {
|
| 595 |
"group_size": {
|
| 596 |
+
"2": 64
|
| 597 |
},
|
| 598 |
"bits": [
|
| 599 |
+
2
|
| 600 |
],
|
| 601 |
"bits_prop": [
|
| 602 |
1
|
|
|
|
| 617 |
}
|
| 618 |
},
|
| 619 |
"model.layers.7": {
|
| 620 |
+
"accuracy": 0.8887453079223633,
|
| 621 |
+
"total_bits": 482790432,
|
| 622 |
"q_proj": {
|
| 623 |
"group_size": {
|
| 624 |
+
"2": 64
|
| 625 |
},
|
| 626 |
"bits": [
|
| 627 |
+
2
|
| 628 |
],
|
| 629 |
"bits_prop": [
|
| 630 |
1
|
|
|
|
| 633 |
},
|
| 634 |
"k_proj": {
|
| 635 |
"group_size": {
|
| 636 |
+
"2": 64
|
| 637 |
},
|
| 638 |
"bits": [
|
| 639 |
+
2
|
| 640 |
],
|
| 641 |
"bits_prop": [
|
| 642 |
1
|
|
|
|
| 669 |
},
|
| 670 |
"up_proj": {
|
| 671 |
"group_size": {
|
| 672 |
+
"2": 64
|
| 673 |
},
|
| 674 |
"bits": [
|
| 675 |
+
2
|
| 676 |
],
|
| 677 |
"bits_prop": [
|
| 678 |
1
|
|
|
|
| 681 |
},
|
| 682 |
"gate_proj": {
|
| 683 |
"group_size": {
|
| 684 |
+
"2": 64
|
| 685 |
},
|
| 686 |
"bits": [
|
| 687 |
+
2
|
| 688 |
],
|
| 689 |
"bits_prop": [
|
| 690 |
1
|
|
|
|
| 705 |
}
|
| 706 |
},
|
| 707 |
"model.layers.8": {
|
| 708 |
+
"accuracy": 0.9228010177612305,
|
| 709 |
+
"total_bits": 449761824,
|
| 710 |
"q_proj": {
|
| 711 |
"group_size": {
|
| 712 |
"2": 64
|
|
|
|
| 721 |
},
|
| 722 |
"k_proj": {
|
| 723 |
"group_size": {
|
| 724 |
+
"2": 64
|
| 725 |
},
|
| 726 |
"bits": [
|
| 727 |
+
2
|
| 728 |
],
|
| 729 |
"bits_prop": [
|
| 730 |
1
|
|
|
|
| 745 |
},
|
| 746 |
"o_proj": {
|
| 747 |
"group_size": {
|
| 748 |
+
"2": 64
|
| 749 |
},
|
| 750 |
"bits": [
|
| 751 |
+
2
|
| 752 |
],
|
| 753 |
"bits_prop": [
|
| 754 |
1
|
|
|
|
| 793 |
}
|
| 794 |
},
|
| 795 |
"model.layers.9": {
|
| 796 |
+
"accuracy": 0.9577234387397766,
|
| 797 |
+
"total_bits": 365124672,
|
| 798 |
"q_proj": {
|
| 799 |
"group_size": {
|
| 800 |
+
"2": 64
|
| 801 |
},
|
| 802 |
"bits": [
|
| 803 |
+
2
|
| 804 |
],
|
| 805 |
"bits_prop": [
|
| 806 |
1
|
|
|
|
| 809 |
},
|
| 810 |
"k_proj": {
|
| 811 |
"group_size": {
|
| 812 |
+
"4": 128
|
| 813 |
},
|
| 814 |
"bits": [
|
| 815 |
+
4
|
| 816 |
],
|
| 817 |
"bits_prop": [
|
| 818 |
1
|
|
|
|
| 881 |
}
|
| 882 |
},
|
| 883 |
"model.layers.10": {
|
| 884 |
+
"accuracy": 0.9458887577056885,
|
| 885 |
+
"total_bits": 360997440,
|
| 886 |
"q_proj": {
|
| 887 |
"group_size": {
|
| 888 |
"2": 64
|
|
|
|
| 921 |
},
|
| 922 |
"o_proj": {
|
| 923 |
"group_size": {
|
| 924 |
+
"2": 64
|
| 925 |
},
|
| 926 |
"bits": [
|
| 927 |
+
2
|
| 928 |
],
|
| 929 |
"bits_prop": [
|
| 930 |
1
|
|
|
|
| 969 |
}
|
| 970 |
},
|
| 971 |
"model.layers.11": {
|
| 972 |
+
"accuracy": 0.9322950839996338,
|
| 973 |
+
"total_bits": 360997440,
|
| 974 |
"q_proj": {
|
| 975 |
"group_size": {
|
| 976 |
+
"2": 64
|
| 977 |
},
|
| 978 |
"bits": [
|
| 979 |
+
2
|
| 980 |
],
|
| 981 |
"bits_prop": [
|
| 982 |
1
|
|
|
|
| 1057 |
}
|
| 1058 |
},
|
| 1059 |
"model.layers.12": {
|
| 1060 |
+
"accuracy": 0.9404451847076416,
|
| 1061 |
+
"total_bits": 365124672,
|
| 1062 |
"q_proj": {
|
| 1063 |
"group_size": {
|
| 1064 |
"2": 64
|
|
|
|
| 1073 |
},
|
| 1074 |
"k_proj": {
|
| 1075 |
"group_size": {
|
| 1076 |
+
"4": 128
|
| 1077 |
},
|
| 1078 |
"bits": [
|
| 1079 |
+
4
|
| 1080 |
],
|
| 1081 |
"bits_prop": [
|
| 1082 |
1
|
|
|
|
| 1097 |
},
|
| 1098 |
"o_proj": {
|
| 1099 |
"group_size": {
|
| 1100 |
+
"2": 64
|
| 1101 |
},
|
| 1102 |
"bits": [
|
| 1103 |
+
2
|
| 1104 |
],
|
| 1105 |
"bits_prop": [
|
| 1106 |
1
|
|
|
|
| 1145 |
}
|
| 1146 |
},
|
| 1147 |
"model.layers.13": {
|
| 1148 |
+
"accuracy": 0.9363645315170288,
|
| 1149 |
+
"total_bits": 365124672,
|
| 1150 |
"q_proj": {
|
| 1151 |
"group_size": {
|
| 1152 |
"2": 64
|
|
|
|
| 1161 |
},
|
| 1162 |
"k_proj": {
|
| 1163 |
"group_size": {
|
| 1164 |
+
"4": 128
|
| 1165 |
},
|
| 1166 |
"bits": [
|
| 1167 |
+
4
|
| 1168 |
],
|
| 1169 |
"bits_prop": [
|
| 1170 |
1
|
|
|
|
| 1185 |
},
|
| 1186 |
"o_proj": {
|
| 1187 |
"group_size": {
|
| 1188 |
+
"2": 64
|
| 1189 |
},
|
| 1190 |
"bits": [
|
| 1191 |
+
2
|
| 1192 |
],
|
| 1193 |
"bits_prop": [
|
| 1194 |
1
|
|
|
|
| 1233 |
}
|
| 1234 |
},
|
| 1235 |
"model.layers.14": {
|
| 1236 |
+
"accuracy": 0.9359749555587769,
|
| 1237 |
+
"total_bits": 365124672,
|
| 1238 |
"q_proj": {
|
| 1239 |
"group_size": {
|
| 1240 |
"2": 64
|
|
|
|
| 1249 |
},
|
| 1250 |
"k_proj": {
|
| 1251 |
"group_size": {
|
| 1252 |
+
"4": 128
|
| 1253 |
},
|
| 1254 |
"bits": [
|
| 1255 |
+
4
|
| 1256 |
],
|
| 1257 |
"bits_prop": [
|
| 1258 |
1
|
|
|
|
| 1273 |
},
|
| 1274 |
"o_proj": {
|
| 1275 |
"group_size": {
|
| 1276 |
+
"2": 64
|
| 1277 |
},
|
| 1278 |
"bits": [
|
| 1279 |
+
2
|
| 1280 |
],
|
| 1281 |
"bits_prop": [
|
| 1282 |
1
|
|
|
|
| 1321 |
}
|
| 1322 |
},
|
| 1323 |
"model.layers.15": {
|
| 1324 |
+
"accuracy": 0.9322938919067383,
|
| 1325 |
+
"total_bits": 365124672,
|
| 1326 |
"q_proj": {
|
| 1327 |
"group_size": {
|
| 1328 |
"2": 64
|
|
|
|
| 1361 |
},
|
| 1362 |
"o_proj": {
|
| 1363 |
"group_size": {
|
| 1364 |
+
"2": 64
|
| 1365 |
},
|
| 1366 |
"bits": [
|
| 1367 |
+
2
|
| 1368 |
],
|
| 1369 |
"bits_prop": [
|
| 1370 |
1
|
|
|
|
| 1397 |
},
|
| 1398 |
"down_proj": {
|
| 1399 |
"group_size": {
|
| 1400 |
+
"2": 64
|
| 1401 |
},
|
| 1402 |
"bits": [
|
| 1403 |
+
2
|
| 1404 |
],
|
| 1405 |
"bits_prop": [
|
| 1406 |
1
|
|
|
|
| 1409 |
}
|
| 1410 |
},
|
| 1411 |
"model.layers.16": {
|
| 1412 |
+
"accuracy": 0.939303994178772,
|
| 1413 |
+
"total_bits": 449761824,
|
| 1414 |
"q_proj": {
|
| 1415 |
"group_size": {
|
| 1416 |
"2": 64
|
|
|
|
| 1425 |
},
|
| 1426 |
"k_proj": {
|
| 1427 |
"group_size": {
|
| 1428 |
+
"2": 64
|
| 1429 |
},
|
| 1430 |
"bits": [
|
| 1431 |
+
2
|
| 1432 |
],
|
| 1433 |
"bits_prop": [
|
| 1434 |
1
|
|
|
|
| 1449 |
},
|
| 1450 |
"o_proj": {
|
| 1451 |
"group_size": {
|
| 1452 |
+
"2": 64
|
| 1453 |
},
|
| 1454 |
"bits": [
|
| 1455 |
+
2
|
| 1456 |
],
|
| 1457 |
"bits_prop": [
|
| 1458 |
1
|
|
|
|
| 1497 |
}
|
| 1498 |
},
|
| 1499 |
"model.layers.17": {
|
| 1500 |
+
"accuracy": 0.9451323747634888,
|
| 1501 |
+
"total_bits": 486917664,
|
| 1502 |
"q_proj": {
|
| 1503 |
"group_size": {
|
| 1504 |
+
"2": 64
|
| 1505 |
},
|
| 1506 |
"bits": [
|
| 1507 |
+
2
|
| 1508 |
],
|
| 1509 |
"bits_prop": [
|
| 1510 |
1
|
|
|
|
| 1513 |
},
|
| 1514 |
"k_proj": {
|
| 1515 |
"group_size": {
|
| 1516 |
+
"4": 128
|
| 1517 |
},
|
| 1518 |
"bits": [
|
| 1519 |
+
4
|
| 1520 |
],
|
| 1521 |
"bits_prop": [
|
| 1522 |
1
|
|
|
|
| 1585 |
}
|
| 1586 |
},
|
| 1587 |
"model.layers.18": {
|
| 1588 |
+
"accuracy": 0.9493275880813599,
|
| 1589 |
+
"total_bits": 519946272,
|
| 1590 |
"q_proj": {
|
| 1591 |
"group_size": {
|
| 1592 |
"4": 128
|
|
|
|
| 1601 |
},
|
| 1602 |
"k_proj": {
|
| 1603 |
"group_size": {
|
| 1604 |
+
"4": 128
|
| 1605 |
},
|
| 1606 |
"bits": [
|
| 1607 |
+
4
|
| 1608 |
],
|
| 1609 |
"bits_prop": [
|
| 1610 |
1
|
|
|
|
| 1673 |
}
|
| 1674 |
},
|
| 1675 |
"model.layers.19": {
|
| 1676 |
+
"accuracy": 0.9514966011047363,
|
| 1677 |
+
"total_bits": 571557408,
|
| 1678 |
"q_proj": {
|
| 1679 |
"group_size": {
|
| 1680 |
+
"2": 64
|
| 1681 |
},
|
| 1682 |
"bits": [
|
| 1683 |
+
2
|
| 1684 |
],
|
| 1685 |
"bits_prop": [
|
| 1686 |
1
|
|
|
|
| 1725 |
},
|
| 1726 |
"up_proj": {
|
| 1727 |
"group_size": {
|
| 1728 |
+
"4": 128
|
| 1729 |
},
|
| 1730 |
"bits": [
|
| 1731 |
+
4
|
| 1732 |
],
|
| 1733 |
"bits_prop": [
|
| 1734 |
1
|
|
|
|
| 1761 |
}
|
| 1762 |
},
|
| 1763 |
"model.layers.20": {
|
| 1764 |
+
"accuracy": 0.955375075340271,
|
| 1765 |
+
"total_bits": 604586016,
|
| 1766 |
"q_proj": {
|
| 1767 |
"group_size": {
|
| 1768 |
"4": 128
|
|
|
|
| 1777 |
},
|
| 1778 |
"k_proj": {
|
| 1779 |
"group_size": {
|
| 1780 |
+
"2": 64
|
| 1781 |
},
|
| 1782 |
"bits": [
|
| 1783 |
+
2
|
| 1784 |
],
|
| 1785 |
"bits_prop": [
|
| 1786 |
1
|
|
|
|
| 1813 |
},
|
| 1814 |
"up_proj": {
|
| 1815 |
"group_size": {
|
| 1816 |
+
"4": 128
|
| 1817 |
},
|
| 1818 |
"bits": [
|
| 1819 |
+
4
|
| 1820 |
],
|
| 1821 |
"bits_prop": [
|
| 1822 |
1
|
|
|
|
| 1849 |
}
|
| 1850 |
},
|
| 1851 |
"model.layers.21": {
|
| 1852 |
+
"accuracy": 0.9731628894805908,
|
| 1853 |
+
"total_bits": 664451616,
|
| 1854 |
"q_proj": {
|
| 1855 |
"group_size": {
|
| 1856 |
+
"2": 64
|
| 1857 |
},
|
| 1858 |
"bits": [
|
| 1859 |
+
2
|
| 1860 |
],
|
| 1861 |
"bits_prop": [
|
| 1862 |
1
|
|
|
|
| 1901 |
},
|
| 1902 |
"up_proj": {
|
| 1903 |
"group_size": {
|
| 1904 |
+
"4": 128
|
| 1905 |
},
|
| 1906 |
"bits": [
|
| 1907 |
+
4
|
| 1908 |
],
|
| 1909 |
"bits_prop": [
|
| 1910 |
1
|
|
|
|
| 1913 |
},
|
| 1914 |
"gate_proj": {
|
| 1915 |
"group_size": {
|
| 1916 |
+
"4": 128
|
| 1917 |
},
|
| 1918 |
"bits": [
|
| 1919 |
+
4
|
| 1920 |
],
|
| 1921 |
"bits_prop": [
|
| 1922 |
1
|
|
|
|
| 1937 |
}
|
| 1938 |
},
|
| 1939 |
"model.layers.22": {
|
| 1940 |
+
"accuracy": 0.9785275459289551,
|
| 1941 |
+
"total_bits": 697480224,
|
| 1942 |
"q_proj": {
|
| 1943 |
"group_size": {
|
| 1944 |
"4": 128
|
|
|
|
| 1953 |
},
|
| 1954 |
"k_proj": {
|
| 1955 |
"group_size": {
|
| 1956 |
+
"4": 128
|
| 1957 |
},
|
| 1958 |
"bits": [
|
| 1959 |
+
4
|
| 1960 |
],
|
| 1961 |
"bits_prop": [
|
| 1962 |
1
|
|
|
|
| 2001 |
},
|
| 2002 |
"gate_proj": {
|
| 2003 |
"group_size": {
|
| 2004 |
+
"4": 128
|
| 2005 |
},
|
| 2006 |
"bits": [
|
| 2007 |
+
4
|
| 2008 |
],
|
| 2009 |
"bits_prop": [
|
| 2010 |
1
|
|
|
|
| 2025 |
}
|
| 2026 |
},
|
| 2027 |
"model.layers.23": {
|
| 2028 |
+
"accuracy": 0.9788622856140137,
|
| 2029 |
+
"total_bits": 697480224,
|
| 2030 |
"q_proj": {
|
| 2031 |
"group_size": {
|
| 2032 |
"4": 128
|
|
|
|
| 2041 |
},
|
| 2042 |
"k_proj": {
|
| 2043 |
"group_size": {
|
| 2044 |
+
"4": 128
|
| 2045 |
},
|
| 2046 |
"bits": [
|
| 2047 |
+
4
|
| 2048 |
],
|
| 2049 |
"bits_prop": [
|
| 2050 |
1
|
|
|
|
| 2089 |
},
|
| 2090 |
"gate_proj": {
|
| 2091 |
"group_size": {
|
| 2092 |
+
"4": 128
|
| 2093 |
},
|
| 2094 |
"bits": [
|
| 2095 |
+
4
|
| 2096 |
],
|
| 2097 |
"bits_prop": [
|
| 2098 |
1
|
|
|
|
| 2113 |
}
|
| 2114 |
},
|
| 2115 |
"model.layers.24": {
|
| 2116 |
+
"accuracy": 0.9794007539749146,
|
| 2117 |
+
"total_bits": 697480224,
|
| 2118 |
"q_proj": {
|
| 2119 |
"group_size": {
|
| 2120 |
"4": 128
|
|
|
|
| 2129 |
},
|
| 2130 |
"k_proj": {
|
| 2131 |
"group_size": {
|
| 2132 |
+
"4": 128
|
| 2133 |
},
|
| 2134 |
"bits": [
|
| 2135 |
+
4
|
| 2136 |
],
|
| 2137 |
"bits_prop": [
|
| 2138 |
1
|
|
|
|
| 2177 |
},
|
| 2178 |
"gate_proj": {
|
| 2179 |
"group_size": {
|
| 2180 |
+
"4": 128
|
| 2181 |
},
|
| 2182 |
"bits": [
|
| 2183 |
+
4
|
| 2184 |
],
|
| 2185 |
"bits_prop": [
|
| 2186 |
1
|
|
|
|
| 2201 |
}
|
| 2202 |
},
|
| 2203 |
"model.layers.25": {
|
| 2204 |
+
"accuracy": 0.9806145429611206,
|
| 2205 |
+
"total_bits": 697480224,
|
| 2206 |
"q_proj": {
|
| 2207 |
"group_size": {
|
| 2208 |
"4": 128
|
|
|
|
| 2217 |
},
|
| 2218 |
"k_proj": {
|
| 2219 |
"group_size": {
|
| 2220 |
+
"4": 128
|
| 2221 |
},
|
| 2222 |
"bits": [
|
| 2223 |
+
4
|
| 2224 |
],
|
| 2225 |
"bits_prop": [
|
| 2226 |
1
|
|
|
|
| 2265 |
},
|
| 2266 |
"gate_proj": {
|
| 2267 |
"group_size": {
|
| 2268 |
+
"4": 128
|
| 2269 |
},
|
| 2270 |
"bits": [
|
| 2271 |
+
4
|
| 2272 |
],
|
| 2273 |
"bits_prop": [
|
| 2274 |
1
|
|
|
|
| 2289 |
}
|
| 2290 |
},
|
| 2291 |
"model.layers.26": {
|
| 2292 |
+
"accuracy": 0.9806764125823975,
|
| 2293 |
+
"total_bits": 697480224,
|
| 2294 |
"q_proj": {
|
| 2295 |
"group_size": {
|
| 2296 |
+
"4": 128
|
| 2297 |
},
|
| 2298 |
"bits": [
|
| 2299 |
+
4
|
| 2300 |
],
|
| 2301 |
"bits_prop": [
|
| 2302 |
1
|
|
|
|
| 2305 |
},
|
| 2306 |
"k_proj": {
|
| 2307 |
"group_size": {
|
| 2308 |
+
"4": 128
|
| 2309 |
},
|
| 2310 |
"bits": [
|
| 2311 |
+
4
|
| 2312 |
],
|
| 2313 |
"bits_prop": [
|
| 2314 |
1
|
|
|
|
| 2341 |
},
|
| 2342 |
"up_proj": {
|
| 2343 |
"group_size": {
|
| 2344 |
+
"4": 128
|
| 2345 |
},
|
| 2346 |
"bits": [
|
| 2347 |
+
4
|
| 2348 |
],
|
| 2349 |
"bits_prop": [
|
| 2350 |
1
|
|
|
|
| 2377 |
}
|
| 2378 |
},
|
| 2379 |
"model.layers.27": {
|
| 2380 |
+
"accuracy": 0.9815640449523926,
|
| 2381 |
+
"total_bits": 697480224,
|
| 2382 |
"q_proj": {
|
| 2383 |
"group_size": {
|
| 2384 |
+
"4": 128
|
| 2385 |
},
|
| 2386 |
"bits": [
|
| 2387 |
+
4
|
| 2388 |
],
|
| 2389 |
"bits_prop": [
|
| 2390 |
1
|
|
|
|
| 2393 |
},
|
| 2394 |
"k_proj": {
|
| 2395 |
"group_size": {
|
| 2396 |
+
"4": 128
|
| 2397 |
},
|
| 2398 |
"bits": [
|
| 2399 |
+
4
|
| 2400 |
],
|
| 2401 |
"bits_prop": [
|
| 2402 |
1
|
|
|
|
| 2441 |
},
|
| 2442 |
"gate_proj": {
|
| 2443 |
"group_size": {
|
| 2444 |
+
"4": 128
|
| 2445 |
},
|
| 2446 |
"bits": [
|
| 2447 |
+
4
|
| 2448 |
],
|
| 2449 |
"bits_prop": [
|
| 2450 |
1
|
|
|
|
| 2465 |
}
|
| 2466 |
},
|
| 2467 |
"model.layers.28": {
|
| 2468 |
+
"accuracy": 0.9820178747177124,
|
| 2469 |
+
"total_bits": 697480224,
|
| 2470 |
"q_proj": {
|
| 2471 |
"group_size": {
|
| 2472 |
"4": 128
|
|
|
|
| 2517 |
},
|
| 2518 |
"up_proj": {
|
| 2519 |
"group_size": {
|
| 2520 |
+
"4": 128
|
| 2521 |
},
|
| 2522 |
"bits": [
|
| 2523 |
+
4
|
| 2524 |
],
|
| 2525 |
"bits_prop": [
|
| 2526 |
1
|
|
|
|
| 2529 |
},
|
| 2530 |
"gate_proj": {
|
| 2531 |
"group_size": {
|
| 2532 |
+
"4": 128
|
| 2533 |
},
|
| 2534 |
"bits": [
|
| 2535 |
+
4
|
| 2536 |
],
|
| 2537 |
"bits_prop": [
|
| 2538 |
1
|
|
|
|
| 2553 |
}
|
| 2554 |
},
|
| 2555 |
"model.layers.29": {
|
| 2556 |
+
"accuracy": 0.9836413264274597,
|
| 2557 |
+
"total_bits": 697480224,
|
| 2558 |
"q_proj": {
|
| 2559 |
"group_size": {
|
| 2560 |
+
"4": 128
|
| 2561 |
},
|
| 2562 |
"bits": [
|
| 2563 |
+
4
|
| 2564 |
],
|
| 2565 |
"bits_prop": [
|
| 2566 |
1
|
|
|
|
| 2569 |
},
|
| 2570 |
"k_proj": {
|
| 2571 |
"group_size": {
|
| 2572 |
+
"4": 128
|
| 2573 |
},
|
| 2574 |
"bits": [
|
| 2575 |
+
4
|
| 2576 |
],
|
| 2577 |
"bits_prop": [
|
| 2578 |
1
|
|
|
|
| 2605 |
},
|
| 2606 |
"up_proj": {
|
| 2607 |
"group_size": {
|
| 2608 |
+
"4": 128
|
| 2609 |
},
|
| 2610 |
"bits": [
|
| 2611 |
+
4
|
| 2612 |
],
|
| 2613 |
"bits_prop": [
|
| 2614 |
1
|
|
|
|
| 2617 |
},
|
| 2618 |
"gate_proj": {
|
| 2619 |
"group_size": {
|
| 2620 |
+
"4": 128
|
| 2621 |
},
|
| 2622 |
"bits": [
|
| 2623 |
+
4
|
| 2624 |
],
|
| 2625 |
"bits_prop": [
|
| 2626 |
1
|
|
|
|
| 2641 |
}
|
| 2642 |
},
|
| 2643 |
"model.layers.30": {
|
| 2644 |
+
"accuracy": 0.9838729500770569,
|
| 2645 |
+
"total_bits": 697480224,
|
| 2646 |
"q_proj": {
|
| 2647 |
"group_size": {
|
| 2648 |
+
"4": 128
|
| 2649 |
},
|
| 2650 |
"bits": [
|
| 2651 |
+
4
|
| 2652 |
],
|
| 2653 |
"bits_prop": [
|
| 2654 |
1
|
|
|
|
| 2657 |
},
|
| 2658 |
"k_proj": {
|
| 2659 |
"group_size": {
|
| 2660 |
+
"4": 128
|
| 2661 |
},
|
| 2662 |
"bits": [
|
| 2663 |
+
4
|
| 2664 |
],
|
| 2665 |
"bits_prop": [
|
| 2666 |
1
|
|
|
|
| 2693 |
},
|
| 2694 |
"up_proj": {
|
| 2695 |
"group_size": {
|
| 2696 |
+
"4": 128
|
| 2697 |
},
|
| 2698 |
"bits": [
|
| 2699 |
+
4
|
| 2700 |
],
|
| 2701 |
"bits_prop": [
|
| 2702 |
1
|
|
|
|
| 2705 |
},
|
| 2706 |
"gate_proj": {
|
| 2707 |
"group_size": {
|
| 2708 |
+
"4": 128
|
| 2709 |
},
|
| 2710 |
"bits": [
|
| 2711 |
+
4
|
| 2712 |
],
|
| 2713 |
"bits_prop": [
|
| 2714 |
1
|
tokenizer.json
CHANGED
|
@@ -14,20 +14,20 @@
|
|
| 14 |
},
|
| 15 |
{
|
| 16 |
"id": 1,
|
| 17 |
-
"content": "
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
-
"normalized":
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
| 25 |
"id": 2,
|
| 26 |
-
"content": "
|
| 27 |
"single_word": false,
|
| 28 |
"lstrip": false,
|
| 29 |
"rstrip": false,
|
| 30 |
-
"normalized":
|
| 31 |
"special": true
|
| 32 |
},
|
| 33 |
{
|
|
@@ -56,24 +56,6 @@
|
|
| 56 |
"rstrip": false,
|
| 57 |
"normalized": false,
|
| 58 |
"special": true
|
| 59 |
-
},
|
| 60 |
-
{
|
| 61 |
-
"id": 64000,
|
| 62 |
-
"content": "<|startoftext|>",
|
| 63 |
-
"single_word": false,
|
| 64 |
-
"lstrip": false,
|
| 65 |
-
"rstrip": false,
|
| 66 |
-
"normalized": true,
|
| 67 |
-
"special": true
|
| 68 |
-
},
|
| 69 |
-
{
|
| 70 |
-
"id": 64001,
|
| 71 |
-
"content": "<|endoftext|>",
|
| 72 |
-
"single_word": false,
|
| 73 |
-
"lstrip": false,
|
| 74 |
-
"rstrip": false,
|
| 75 |
-
"normalized": true,
|
| 76 |
-
"special": true
|
| 77 |
}
|
| 78 |
],
|
| 79 |
"normalizer": {
|
|
@@ -153,8 +135,8 @@
|
|
| 153 |
"byte_fallback": true,
|
| 154 |
"vocab": {
|
| 155 |
"<unk>": 0,
|
| 156 |
-
"
|
| 157 |
-
"
|
| 158 |
"<|Human|>": 3,
|
| 159 |
"<|Assistant|>": 4,
|
| 160 |
"<|System|>": 5,
|
|
|
|
| 14 |
},
|
| 15 |
{
|
| 16 |
"id": 1,
|
| 17 |
+
"content": "<|startoftext|>",
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
+
"normalized": true,
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
| 25 |
"id": 2,
|
| 26 |
+
"content": "<|endoftext|>",
|
| 27 |
"single_word": false,
|
| 28 |
"lstrip": false,
|
| 29 |
"rstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
"special": true
|
| 32 |
},
|
| 33 |
{
|
|
|
|
| 56 |
"rstrip": false,
|
| 57 |
"normalized": false,
|
| 58 |
"special": true
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
}
|
| 60 |
],
|
| 61 |
"normalizer": {
|
|
|
|
| 135 |
"byte_fallback": true,
|
| 136 |
"vocab": {
|
| 137 |
"<unk>": 0,
|
| 138 |
+
"<|startoftext|>": 1,
|
| 139 |
+
"<|endoftext|>": 2,
|
| 140 |
"<|Human|>": 3,
|
| 141 |
"<|Assistant|>": 4,
|
| 142 |
"<|System|>": 5,
|
tokenizer_config.json
CHANGED
|
@@ -12,17 +12,17 @@
|
|
| 12 |
"special": true
|
| 13 |
},
|
| 14 |
"1": {
|
| 15 |
-
"content": "
|
| 16 |
"lstrip": false,
|
| 17 |
-
"normalized":
|
| 18 |
"rstrip": false,
|
| 19 |
"single_word": false,
|
| 20 |
"special": true
|
| 21 |
},
|
| 22 |
"2": {
|
| 23 |
-
"content": "
|
| 24 |
"lstrip": false,
|
| 25 |
-
"normalized":
|
| 26 |
"rstrip": false,
|
| 27 |
"single_word": false,
|
| 28 |
"special": true
|
|
@@ -50,22 +50,6 @@
|
|
| 50 |
"rstrip": false,
|
| 51 |
"single_word": false,
|
| 52 |
"special": true
|
| 53 |
-
},
|
| 54 |
-
"64000": {
|
| 55 |
-
"content": "<|startoftext|>",
|
| 56 |
-
"lstrip": false,
|
| 57 |
-
"normalized": true,
|
| 58 |
-
"rstrip": false,
|
| 59 |
-
"single_word": false,
|
| 60 |
-
"special": true
|
| 61 |
-
},
|
| 62 |
-
"64001": {
|
| 63 |
-
"content": "<|endoftext|>",
|
| 64 |
-
"lstrip": false,
|
| 65 |
-
"normalized": true,
|
| 66 |
-
"rstrip": false,
|
| 67 |
-
"single_word": false,
|
| 68 |
-
"special": true
|
| 69 |
}
|
| 70 |
},
|
| 71 |
"additional_special_tokens": [
|
|
|
|
| 12 |
"special": true
|
| 13 |
},
|
| 14 |
"1": {
|
| 15 |
+
"content": "<|startoftext|>",
|
| 16 |
"lstrip": false,
|
| 17 |
+
"normalized": true,
|
| 18 |
"rstrip": false,
|
| 19 |
"single_word": false,
|
| 20 |
"special": true
|
| 21 |
},
|
| 22 |
"2": {
|
| 23 |
+
"content": "<|endoftext|>",
|
| 24 |
"lstrip": false,
|
| 25 |
+
"normalized": true,
|
| 26 |
"rstrip": false,
|
| 27 |
"single_word": false,
|
| 28 |
"special": true
|
|
|
|
| 50 |
"rstrip": false,
|
| 51 |
"single_word": false,
|
| 52 |
"special": true
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
}
|
| 54 |
},
|
| 55 |
"additional_special_tokens": [
|