gpt2moe_het2_1000mb
This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.7218
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 37035
- training_steps: 370358
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 0 | 0 | 11.0404 |
| 8.6089 | 0.0540 | 2000 | 7.9850 |
| 7.5891 | 0.1080 | 4000 | 6.9885 |
| 7.021 | 0.1620 | 6000 | 6.4072 |
| 6.601 | 0.2160 | 8000 | 5.9924 |
| 6.3052 | 0.2700 | 10000 | 5.6787 |
| 6.0684 | 0.3240 | 12000 | 5.4355 |
| 5.8674 | 0.3780 | 14000 | 5.2383 |
| 5.7119 | 0.4320 | 16000 | 5.0759 |
| 5.5771 | 0.4860 | 18000 | 4.9257 |
| 5.4478 | 0.5400 | 20000 | 4.8131 |
| 5.3629 | 0.5940 | 22000 | 4.7186 |
| 5.2861 | 0.6480 | 24000 | 4.6508 |
| 5.2247 | 0.7020 | 26000 | 4.5947 |
| 5.1728 | 0.7560 | 28000 | 4.5430 |
| 5.1195 | 0.8100 | 30000 | 4.5001 |
| 5.0843 | 0.8640 | 32000 | 4.4586 |
| 5.0484 | 0.9180 | 34000 | 4.4238 |
| 5.0174 | 0.9720 | 36000 | 4.3903 |
| 4.9677 | 1.0260 | 38000 | 4.3583 |
| 4.9396 | 1.0800 | 40000 | 4.3316 |
| 4.9187 | 1.1340 | 42000 | 4.3021 |
| 4.8834 | 1.1880 | 44000 | 4.2751 |
| 4.8697 | 1.2420 | 46000 | 4.2509 |
| 4.8421 | 1.2960 | 48000 | 4.2305 |
| 4.831 | 1.3500 | 50000 | 4.2102 |
| 4.812 | 1.4040 | 52000 | 4.1905 |
| 4.789 | 1.4580 | 54000 | 4.1759 |
| 4.7778 | 1.5120 | 56000 | 4.1602 |
| 4.765 | 1.5660 | 58000 | 4.1463 |
| 4.7513 | 1.6200 | 60000 | 4.1310 |
| 4.7314 | 1.6740 | 62000 | 4.1172 |
| 4.7206 | 1.7280 | 64000 | 4.1064 |
| 4.7216 | 1.7820 | 66000 | 4.0936 |
| 4.709 | 1.8361 | 68000 | 4.0826 |
| 4.6926 | 1.8901 | 70000 | 4.0741 |
| 4.6861 | 1.9441 | 72000 | 4.0628 |
| 4.6841 | 1.9981 | 74000 | 4.0539 |
| 4.6384 | 2.0521 | 76000 | 4.0461 |
| 4.635 | 2.1061 | 78000 | 4.0370 |
| 4.6324 | 2.1601 | 80000 | 4.0301 |
| 4.6261 | 2.2141 | 82000 | 4.0234 |
| 4.6182 | 2.2681 | 84000 | 4.0163 |
| 4.6178 | 2.3221 | 86000 | 4.0089 |
| 4.6005 | 2.3761 | 88000 | 4.0028 |
| 4.6002 | 2.4301 | 90000 | 3.9960 |
| 4.6008 | 2.4841 | 92000 | 3.9891 |
| 4.5939 | 2.5381 | 94000 | 3.9844 |
| 4.5945 | 2.5921 | 96000 | 3.9790 |
| 4.5836 | 2.6461 | 98000 | 3.9717 |
| 4.5798 | 2.7001 | 100000 | 3.9657 |
| 4.5736 | 2.7541 | 102000 | 3.9619 |
| 4.5673 | 2.8081 | 104000 | 3.9566 |
| 4.568 | 2.8621 | 106000 | 3.9517 |
| 4.5528 | 2.9161 | 108000 | 3.9466 |
| 4.5535 | 2.9701 | 110000 | 3.9410 |
| 4.5124 | 3.0241 | 112000 | 3.9394 |
| 4.5236 | 3.0781 | 114000 | 3.9346 |
| 4.5183 | 3.1321 | 116000 | 3.9305 |
| 4.5173 | 3.1861 | 118000 | 3.9269 |
| 4.5177 | 3.2401 | 120000 | 3.9213 |
| 4.5127 | 3.2941 | 122000 | 3.9193 |
| 4.5143 | 3.3481 | 124000 | 3.9153 |
| 4.5073 | 3.4021 | 126000 | 3.9115 |
| 4.5079 | 3.4561 | 128000 | 3.9072 |
| 4.5062 | 3.5101 | 130000 | 3.9031 |
| 4.5012 | 3.5641 | 132000 | 3.9004 |
| 4.5043 | 3.6181 | 134000 | 3.8961 |
| 4.496 | 3.6721 | 136000 | 3.8935 |
| 4.4957 | 3.7261 | 138000 | 3.8898 |
| 4.4946 | 3.7801 | 140000 | 3.8871 |
| 4.4902 | 3.8341 | 142000 | 3.8854 |
| 4.4888 | 3.8881 | 144000 | 3.8803 |
| 4.4893 | 3.9421 | 146000 | 3.8768 |
| 4.4828 | 3.9961 | 148000 | 3.8741 |
| 4.4497 | 4.0501 | 150000 | 3.8737 |
| 4.454 | 4.1041 | 152000 | 3.8716 |
| 4.4552 | 4.1581 | 154000 | 3.8688 |
| 4.452 | 4.2121 | 156000 | 3.8662 |
| 4.4561 | 4.2661 | 158000 | 3.8633 |
| 4.4511 | 4.3201 | 160000 | 3.8612 |
| 4.4481 | 4.3741 | 162000 | 3.8574 |
| 4.4442 | 4.4281 | 164000 | 3.8554 |
| 4.449 | 4.4821 | 166000 | 3.8528 |
| 4.4401 | 4.5361 | 168000 | 3.8508 |
| 4.4439 | 4.5901 | 170000 | 3.8482 |
| 4.4422 | 4.6441 | 172000 | 3.8460 |
| 4.4414 | 4.6981 | 174000 | 3.8429 |
| 4.4374 | 4.7521 | 176000 | 3.8406 |
| 4.4391 | 4.8061 | 178000 | 3.8383 |
| 4.4355 | 4.8601 | 180000 | 3.8360 |
| 4.4375 | 4.9141 | 182000 | 3.8344 |
| 4.4319 | 4.9681 | 184000 | 3.8311 |
| 4.4008 | 5.0221 | 186000 | 3.8310 |
| 4.3976 | 5.0761 | 188000 | 3.8296 |
| 4.4069 | 5.1301 | 190000 | 3.8282 |
| 4.4045 | 5.1841 | 192000 | 3.8265 |
| 4.4073 | 5.2381 | 194000 | 3.8240 |
| 4.4018 | 5.2921 | 196000 | 3.8221 |
| 4.4043 | 5.3461 | 198000 | 3.8203 |
| 4.4059 | 5.4002 | 200000 | 3.8173 |
| 4.4053 | 5.4542 | 202000 | 3.8157 |
| 4.4035 | 5.5082 | 204000 | 3.8145 |
| 4.4002 | 5.5622 | 206000 | 3.8124 |
| 4.3997 | 5.6162 | 208000 | 3.8113 |
| 4.3893 | 5.6702 | 210000 | 3.8085 |
| 4.3951 | 5.7242 | 212000 | 3.8063 |
| 4.4003 | 5.7782 | 214000 | 3.8042 |
| 4.4008 | 5.8322 | 216000 | 3.8029 |
| 4.3973 | 5.8862 | 218000 | 3.8011 |
| 4.3942 | 5.9402 | 220000 | 3.7991 |
| 4.3916 | 5.9942 | 222000 | 3.7972 |
| 4.3605 | 6.0482 | 224000 | 3.7986 |
| 4.3661 | 6.1022 | 226000 | 3.7968 |
| 4.3624 | 6.1562 | 228000 | 3.7953 |
| 4.3667 | 6.2102 | 230000 | 3.7943 |
| 4.371 | 6.2642 | 232000 | 3.7933 |
| 4.3705 | 6.3182 | 234000 | 3.7902 |
| 4.3704 | 6.3722 | 236000 | 3.7885 |
| 4.3714 | 6.4262 | 238000 | 3.7880 |
| 4.3646 | 6.4802 | 240000 | 3.7869 |
| 4.3686 | 6.5342 | 242000 | 3.7839 |
| 4.3632 | 6.5882 | 244000 | 3.7828 |
| 4.3679 | 6.6422 | 246000 | 3.7814 |
| 4.3678 | 6.6962 | 248000 | 3.7798 |
| 4.3646 | 6.7502 | 250000 | 3.7778 |
| 4.3628 | 6.8042 | 252000 | 3.7761 |
| 4.3636 | 6.8582 | 254000 | 3.7749 |
| 4.3608 | 6.9122 | 256000 | 3.7730 |
| 4.3595 | 6.9662 | 258000 | 3.7717 |
| 4.3304 | 7.0202 | 260000 | 3.7723 |
| 4.3374 | 7.0742 | 262000 | 3.7719 |
| 4.3326 | 7.1282 | 264000 | 3.7703 |
| 4.3401 | 7.1822 | 266000 | 3.7696 |
| 4.3437 | 7.2362 | 268000 | 3.7676 |
| 4.3332 | 7.2902 | 270000 | 3.7667 |
| 4.3424 | 7.3442 | 272000 | 3.7651 |
| 4.3331 | 7.3982 | 274000 | 3.7639 |
| 4.3385 | 7.4522 | 276000 | 3.7621 |
| 4.3303 | 7.5062 | 278000 | 3.7605 |
| 4.3409 | 7.5602 | 280000 | 3.7602 |
| 4.3344 | 7.6142 | 282000 | 3.7580 |
| 4.3371 | 7.6682 | 284000 | 3.7569 |
| 4.338 | 7.7222 | 286000 | 3.7557 |
| 4.3347 | 7.7762 | 288000 | 3.7544 |
| 4.335 | 7.8302 | 290000 | 3.7530 |
| 4.3308 | 7.8842 | 292000 | 3.7526 |
| 4.3293 | 7.9382 | 294000 | 3.7510 |
| 4.3318 | 7.9922 | 296000 | 3.7499 |
| 4.3089 | 8.0462 | 298000 | 3.7501 |
| 4.308 | 8.1002 | 300000 | 3.7489 |
| 4.3088 | 8.1542 | 302000 | 3.7485 |
| 4.3076 | 8.2082 | 304000 | 3.7470 |
| 4.3023 | 8.2622 | 306000 | 3.7467 |
| 4.3057 | 8.3162 | 308000 | 3.7447 |
| 4.3107 | 8.3702 | 310000 | 3.7444 |
| 4.3086 | 8.4242 | 312000 | 3.7432 |
| 4.3074 | 8.4782 | 314000 | 3.7421 |
| 4.3182 | 8.5322 | 316000 | 3.7409 |
| 4.3085 | 8.5862 | 318000 | 3.7398 |
| 4.3085 | 8.6402 | 320000 | 3.7390 |
| 4.3086 | 8.6942 | 322000 | 3.7377 |
| 4.3069 | 8.7482 | 324000 | 3.7367 |
| 4.309 | 8.8022 | 326000 | 3.7352 |
| 4.3075 | 8.8562 | 328000 | 3.7350 |
| 4.3071 | 8.9102 | 330000 | 3.7329 |
| 4.3074 | 8.9643 | 332000 | 3.7330 |
| 4.2814 | 9.0183 | 334000 | 3.7325 |
| 4.2861 | 9.0723 | 336000 | 3.7323 |
| 4.2886 | 9.1263 | 338000 | 3.7316 |
| 4.2859 | 9.1803 | 340000 | 3.7312 |
| 4.2897 | 9.2343 | 342000 | 3.7299 |
| 4.287 | 9.2883 | 344000 | 3.7296 |
| 4.2855 | 9.3423 | 346000 | 3.7286 |
| 4.2817 | 9.3963 | 348000 | 3.7278 |
| 4.2816 | 9.4503 | 350000 | 3.7271 |
| 4.2857 | 9.5043 | 352000 | 3.7265 |
| 4.2845 | 9.5583 | 354000 | 3.7255 |
| 4.279 | 9.6123 | 356000 | 3.7250 |
| 4.2798 | 9.6663 | 358000 | 3.7243 |
| 4.2833 | 9.7203 | 360000 | 3.7236 |
| 4.2837 | 9.7743 | 362000 | 3.7233 |
| 4.2842 | 9.8283 | 364000 | 3.7224 |
| 4.2922 | 9.8823 | 366000 | 3.7222 |
| 4.2801 | 9.9363 | 368000 | 3.7220 |
| 4.2817 | 9.9903 | 370000 | 3.7218 |
Framework versions
- Transformers 4.57.1
- Pytorch 2.9.1+cu128
- Datasets 3.6.0
- Tokenizers 0.22.1
- Downloads last month
- -