<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
20230829213515
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: 0.5986
- Accuracy: 0.6538
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: 0.005
- 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 | 0.6658 | 0.625 |
No log | 2.0 | 70 | 0.6763 | 0.4808 |
No log | 3.0 | 105 | 0.6122 | 0.5769 |
No log | 4.0 | 140 | 0.7455 | 0.5673 |
No log | 5.0 | 175 | 0.7356 | 0.3654 |
No log | 6.0 | 210 | 0.7729 | 0.4423 |
No log | 7.0 | 245 | 0.7041 | 0.4327 |
No log | 8.0 | 280 | 1.1367 | 0.6346 |
No log | 9.0 | 315 | 1.1380 | 0.4135 |
No log | 10.0 | 350 | 0.6978 | 0.5865 |
No log | 11.0 | 385 | 0.9494 | 0.3942 |
No log | 12.0 | 420 | 0.6083 | 0.6346 |
No log | 13.0 | 455 | 0.6045 | 0.5962 |
No log | 14.0 | 490 | 1.3653 | 0.4038 |
0.8661 | 15.0 | 525 | 0.8699 | 0.6346 |
0.8661 | 16.0 | 560 | 0.6745 | 0.4231 |
0.8661 | 17.0 | 595 | 0.6906 | 0.5096 |
0.8661 | 18.0 | 630 | 0.6755 | 0.6154 |
0.8661 | 19.0 | 665 | 1.0554 | 0.375 |
0.8661 | 20.0 | 700 | 0.8385 | 0.4135 |
0.8661 | 21.0 | 735 | 0.6031 | 0.6346 |
0.8661 | 22.0 | 770 | 0.6460 | 0.4904 |
0.8661 | 23.0 | 805 | 1.0714 | 0.375 |
0.8661 | 24.0 | 840 | 0.7565 | 0.4135 |
0.8661 | 25.0 | 875 | 0.7257 | 0.5673 |
0.8661 | 26.0 | 910 | 0.6050 | 0.6538 |
0.8661 | 27.0 | 945 | 0.5938 | 0.6346 |
0.8661 | 28.0 | 980 | 0.6601 | 0.5769 |
0.7783 | 29.0 | 1015 | 0.5878 | 0.6346 |
0.7783 | 30.0 | 1050 | 0.7318 | 0.3558 |
0.7783 | 31.0 | 1085 | 0.5853 | 0.6346 |
0.7783 | 32.0 | 1120 | 0.8454 | 0.3558 |
0.7783 | 33.0 | 1155 | 0.7431 | 0.6346 |
0.7783 | 34.0 | 1190 | 0.5968 | 0.6346 |
0.7783 | 35.0 | 1225 | 0.6201 | 0.6346 |
0.7783 | 36.0 | 1260 | 0.6217 | 0.5481 |
0.7783 | 37.0 | 1295 | 0.6343 | 0.4808 |
0.7783 | 38.0 | 1330 | 0.6639 | 0.4519 |
0.7783 | 39.0 | 1365 | 0.7022 | 0.3846 |
0.7783 | 40.0 | 1400 | 0.6172 | 0.5192 |
0.7783 | 41.0 | 1435 | 1.0947 | 0.3654 |
0.7783 | 42.0 | 1470 | 0.6203 | 0.5481 |
0.7329 | 43.0 | 1505 | 0.5951 | 0.6346 |
0.7329 | 44.0 | 1540 | 0.6051 | 0.5673 |
0.7329 | 45.0 | 1575 | 0.8094 | 0.3654 |
0.7329 | 46.0 | 1610 | 0.6247 | 0.5288 |
0.7329 | 47.0 | 1645 | 0.5813 | 0.6538 |
0.7329 | 48.0 | 1680 | 0.5972 | 0.6346 |
0.7329 | 49.0 | 1715 | 0.6132 | 0.6346 |
0.7329 | 50.0 | 1750 | 0.6039 | 0.6538 |
0.7329 | 51.0 | 1785 | 0.7320 | 0.3846 |
0.7329 | 52.0 | 1820 | 0.5957 | 0.6346 |
0.7329 | 53.0 | 1855 | 0.6665 | 0.4231 |
0.7329 | 54.0 | 1890 | 0.7335 | 0.3846 |
0.7329 | 55.0 | 1925 | 0.6059 | 0.6346 |
0.7329 | 56.0 | 1960 | 0.5978 | 0.6346 |
0.7329 | 57.0 | 1995 | 0.6234 | 0.4808 |
0.688 | 58.0 | 2030 | 0.6427 | 0.4231 |
0.688 | 59.0 | 2065 | 0.6607 | 0.375 |
0.688 | 60.0 | 2100 | 0.6745 | 0.6346 |
0.688 | 61.0 | 2135 | 0.6068 | 0.6346 |
0.688 | 62.0 | 2170 | 0.6284 | 0.6346 |
0.688 | 63.0 | 2205 | 0.6015 | 0.6731 |
0.688 | 64.0 | 2240 | 0.6576 | 0.6346 |
0.688 | 65.0 | 2275 | 0.5950 | 0.6538 |
0.688 | 66.0 | 2310 | 0.5874 | 0.6346 |
0.688 | 67.0 | 2345 | 0.6258 | 0.4712 |
0.688 | 68.0 | 2380 | 0.5909 | 0.6538 |
0.688 | 69.0 | 2415 | 0.5862 | 0.6346 |
0.688 | 70.0 | 2450 | 0.5865 | 0.6346 |
0.688 | 71.0 | 2485 | 0.6265 | 0.4904 |
0.6632 | 72.0 | 2520 | 0.6135 | 0.5 |
0.6632 | 73.0 | 2555 | 0.5911 | 0.6346 |
0.6632 | 74.0 | 2590 | 0.6323 | 0.4615 |
0.6632 | 75.0 | 2625 | 0.6121 | 0.5673 |
0.6632 | 76.0 | 2660 | 0.6073 | 0.5577 |
0.6632 | 77.0 | 2695 | 0.6643 | 0.4231 |
0.6632 | 78.0 | 2730 | 0.6209 | 0.5288 |
0.6632 | 79.0 | 2765 | 0.6047 | 0.6346 |
0.6632 | 80.0 | 2800 | 0.5986 | 0.6538 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3