20230829194700
This model is a fine-tuned version of bert-large-cased on the super_glue dataset. It achieves the following results on the evaluation set:
- Loss: 0.3338
- Accuracy: 0.6346
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: 0.003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 44
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 80.0
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
No log | 1.0 | 35 | 0.3465 | 0.5962 |
No log | 2.0 | 70 | 0.3324 | 0.5769 |
No log | 3.0 | 105 | 0.3771 | 0.6154 |
No log | 4.0 | 140 | 0.3892 | 0.5481 |
No log | 5.0 | 175 | 0.3777 | 0.5481 |
No log | 6.0 | 210 | 0.4086 | 0.4615 |
No log | 7.0 | 245 | 0.3702 | 0.4135 |
No log | 8.0 | 280 | 0.3553 | 0.4519 |
No log | 9.0 | 315 | 0.3547 | 0.4135 |
No log | 10.0 | 350 | 0.6069 | 0.3942 |
No log | 11.0 | 385 | 0.3542 | 0.4712 |
No log | 12.0 | 420 | 0.4133 | 0.625 |
No log | 13.0 | 455 | 0.4395 | 0.6346 |
No log | 14.0 | 490 | 0.3549 | 0.6346 |
0.45 | 15.0 | 525 | 0.3869 | 0.4231 |
0.45 | 16.0 | 560 | 0.3793 | 0.6346 |
0.45 | 17.0 | 595 | 0.3991 | 0.4135 |
0.45 | 18.0 | 630 | 0.3405 | 0.6442 |
0.45 | 19.0 | 665 | 0.3948 | 0.4038 |
0.45 | 20.0 | 700 | 0.3327 | 0.6442 |
0.45 | 21.0 | 735 | 0.3452 | 0.6346 |
0.45 | 22.0 | 770 | 0.3510 | 0.5962 |
0.45 | 23.0 | 805 | 0.3443 | 0.625 |
0.45 | 24.0 | 840 | 0.3563 | 0.6346 |
0.45 | 25.0 | 875 | 0.5409 | 0.3846 |
0.45 | 26.0 | 910 | 0.3971 | 0.4519 |
0.45 | 27.0 | 945 | 0.7386 | 0.4038 |
0.45 | 28.0 | 980 | 0.3423 | 0.6058 |
0.4313 | 29.0 | 1015 | 0.3482 | 0.5096 |
0.4313 | 30.0 | 1050 | 0.3383 | 0.5769 |
0.4313 | 31.0 | 1085 | 0.5153 | 0.4038 |
0.4313 | 32.0 | 1120 | 0.6008 | 0.3654 |
0.4313 | 33.0 | 1155 | 0.4639 | 0.6346 |
0.4313 | 34.0 | 1190 | 0.3641 | 0.6346 |
0.4313 | 35.0 | 1225 | 0.3407 | 0.5577 |
0.4313 | 36.0 | 1260 | 0.3406 | 0.5769 |
0.4313 | 37.0 | 1295 | 0.3353 | 0.6346 |
0.4313 | 38.0 | 1330 | 0.3465 | 0.6346 |
0.4313 | 39.0 | 1365 | 0.3408 | 0.6346 |
0.4313 | 40.0 | 1400 | 0.3325 | 0.625 |
0.4313 | 41.0 | 1435 | 0.3983 | 0.3942 |
0.4313 | 42.0 | 1470 | 0.3435 | 0.5577 |
0.3946 | 43.0 | 1505 | 0.3315 | 0.6346 |
0.3946 | 44.0 | 1540 | 0.3454 | 0.5577 |
0.3946 | 45.0 | 1575 | 0.3314 | 0.6346 |
0.3946 | 46.0 | 1610 | 0.3326 | 0.6346 |
0.3946 | 47.0 | 1645 | 0.3506 | 0.5385 |
0.3946 | 48.0 | 1680 | 0.3370 | 0.6154 |
0.3946 | 49.0 | 1715 | 0.3354 | 0.6346 |
0.3946 | 50.0 | 1750 | 0.3302 | 0.6442 |
0.3946 | 51.0 | 1785 | 0.3400 | 0.5865 |
0.3946 | 52.0 | 1820 | 0.3844 | 0.4423 |
0.3946 | 53.0 | 1855 | 0.3378 | 0.6058 |
0.3946 | 54.0 | 1890 | 0.3673 | 0.4327 |
0.3946 | 55.0 | 1925 | 0.3340 | 0.6346 |
0.3946 | 56.0 | 1960 | 0.3464 | 0.6346 |
0.3946 | 57.0 | 1995 | 0.3565 | 0.5192 |
0.375 | 58.0 | 2030 | 0.3356 | 0.6346 |
0.375 | 59.0 | 2065 | 0.4202 | 0.3942 |
0.375 | 60.0 | 2100 | 0.3495 | 0.6442 |
0.375 | 61.0 | 2135 | 0.3374 | 0.6346 |
0.375 | 62.0 | 2170 | 0.3323 | 0.6635 |
0.375 | 63.0 | 2205 | 0.3362 | 0.6731 |
0.375 | 64.0 | 2240 | 0.3767 | 0.6346 |
0.375 | 65.0 | 2275 | 0.3345 | 0.6346 |
0.375 | 66.0 | 2310 | 0.3451 | 0.6346 |
0.375 | 67.0 | 2345 | 0.3403 | 0.6058 |
0.375 | 68.0 | 2380 | 0.3347 | 0.6538 |
0.375 | 69.0 | 2415 | 0.3419 | 0.6346 |
0.375 | 70.0 | 2450 | 0.3479 | 0.6346 |
0.375 | 71.0 | 2485 | 0.3330 | 0.6346 |
0.3666 | 72.0 | 2520 | 0.3442 | 0.5192 |
0.3666 | 73.0 | 2555 | 0.3335 | 0.6346 |
0.3666 | 74.0 | 2590 | 0.3474 | 0.4615 |
0.3666 | 75.0 | 2625 | 0.3364 | 0.6538 |
0.3666 | 76.0 | 2660 | 0.3368 | 0.6538 |
0.3666 | 77.0 | 2695 | 0.3498 | 0.5192 |
0.3666 | 78.0 | 2730 | 0.3407 | 0.5577 |
0.3666 | 79.0 | 2765 | 0.3352 | 0.6346 |
0.3666 | 80.0 | 2800 | 0.3338 | 0.6346 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
- Downloads last month
- 2
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.