<!-- 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. -->
20230829194638
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: 0.6050
- 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 | 0.5812 | 0.6346 |
No log | 2.0 | 70 | 0.6027 | 0.5481 |
No log | 3.0 | 105 | 0.5878 | 0.6346 |
No log | 4.0 | 140 | 0.7205 | 0.4135 |
No log | 5.0 | 175 | 0.7094 | 0.4327 |
No log | 6.0 | 210 | 0.7357 | 0.4519 |
No log | 7.0 | 245 | 0.8854 | 0.3846 |
No log | 8.0 | 280 | 0.5948 | 0.6346 |
No log | 9.0 | 315 | 0.8744 | 0.4038 |
No log | 10.0 | 350 | 1.1156 | 0.375 |
No log | 11.0 | 385 | 0.7347 | 0.4038 |
No log | 12.0 | 420 | 0.7712 | 0.6346 |
No log | 13.0 | 455 | 0.9681 | 0.3942 |
No log | 14.0 | 490 | 0.7136 | 0.6346 |
0.7475 | 15.0 | 525 | 0.6871 | 0.6346 |
0.7475 | 16.0 | 560 | 0.6093 | 0.625 |
0.7475 | 17.0 | 595 | 0.6090 | 0.6058 |
0.7475 | 18.0 | 630 | 0.7972 | 0.3654 |
0.7475 | 19.0 | 665 | 1.0616 | 0.3654 |
0.7475 | 20.0 | 700 | 0.6116 | 0.5385 |
0.7475 | 21.0 | 735 | 0.6122 | 0.6346 |
0.7475 | 22.0 | 770 | 0.5964 | 0.6346 |
0.7475 | 23.0 | 805 | 0.9926 | 0.375 |
0.7475 | 24.0 | 840 | 0.6146 | 0.6346 |
0.7475 | 25.0 | 875 | 0.6174 | 0.5577 |
0.7475 | 26.0 | 910 | 0.7911 | 0.4135 |
0.7475 | 27.0 | 945 | 0.6156 | 0.5577 |
0.7475 | 28.0 | 980 | 0.6016 | 0.625 |
0.7595 | 29.0 | 1015 | 0.5903 | 0.6346 |
0.7595 | 30.0 | 1050 | 0.7155 | 0.4904 |
0.7595 | 31.0 | 1085 | 0.6956 | 0.5 |
0.7595 | 32.0 | 1120 | 0.8902 | 0.3846 |
0.7595 | 33.0 | 1155 | 0.6929 | 0.6346 |
0.7595 | 34.0 | 1190 | 0.5932 | 0.625 |
0.7595 | 35.0 | 1225 | 0.9652 | 0.3654 |
0.7595 | 36.0 | 1260 | 0.6075 | 0.6442 |
0.7595 | 37.0 | 1295 | 0.6217 | 0.6346 |
0.7595 | 38.0 | 1330 | 0.5904 | 0.625 |
0.7595 | 39.0 | 1365 | 0.7185 | 0.3942 |
0.7595 | 40.0 | 1400 | 0.6530 | 0.4135 |
0.7595 | 41.0 | 1435 | 0.6031 | 0.5962 |
0.7595 | 42.0 | 1470 | 0.7759 | 0.3846 |
0.7018 | 43.0 | 1505 | 0.5980 | 0.625 |
0.7018 | 44.0 | 1540 | 0.5870 | 0.6346 |
0.7018 | 45.0 | 1575 | 0.6563 | 0.5 |
0.7018 | 46.0 | 1610 | 0.6183 | 0.6058 |
0.7018 | 47.0 | 1645 | 0.5909 | 0.6346 |
0.7018 | 48.0 | 1680 | 0.6099 | 0.625 |
0.7018 | 49.0 | 1715 | 0.6232 | 0.6346 |
0.7018 | 50.0 | 1750 | 0.6034 | 0.6346 |
0.7018 | 51.0 | 1785 | 0.6536 | 0.4231 |
0.7018 | 52.0 | 1820 | 0.6477 | 0.625 |
0.7018 | 53.0 | 1855 | 0.6981 | 0.3846 |
0.7018 | 54.0 | 1890 | 0.6281 | 0.4519 |
0.7018 | 55.0 | 1925 | 0.6001 | 0.6346 |
0.7018 | 56.0 | 1960 | 0.6461 | 0.6346 |
0.7018 | 57.0 | 1995 | 0.5900 | 0.6346 |
0.6645 | 58.0 | 2030 | 0.5908 | 0.6346 |
0.6645 | 59.0 | 2065 | 0.6601 | 0.3942 |
0.6645 | 60.0 | 2100 | 0.6277 | 0.625 |
0.6645 | 61.0 | 2135 | 0.6025 | 0.6154 |
0.6645 | 62.0 | 2170 | 0.5973 | 0.6346 |
0.6645 | 63.0 | 2205 | 0.6140 | 0.5769 |
0.6645 | 64.0 | 2240 | 0.5938 | 0.6346 |
0.6645 | 65.0 | 2275 | 0.6088 | 0.6731 |
0.6645 | 66.0 | 2310 | 0.6162 | 0.5769 |
0.6645 | 67.0 | 2345 | 0.6078 | 0.625 |
0.6645 | 68.0 | 2380 | 0.6127 | 0.6154 |
0.6645 | 69.0 | 2415 | 0.5926 | 0.6346 |
0.6645 | 70.0 | 2450 | 0.6059 | 0.6346 |
0.6645 | 71.0 | 2485 | 0.6044 | 0.6154 |
0.6456 | 72.0 | 2520 | 0.6098 | 0.6538 |
0.6456 | 73.0 | 2555 | 0.5981 | 0.6346 |
0.6456 | 74.0 | 2590 | 0.6173 | 0.4904 |
0.6456 | 75.0 | 2625 | 0.6066 | 0.6538 |
0.6456 | 76.0 | 2660 | 0.6192 | 0.5 |
0.6456 | 77.0 | 2695 | 0.6191 | 0.5481 |
0.6456 | 78.0 | 2730 | 0.6183 | 0.5865 |
0.6456 | 79.0 | 2765 | 0.6142 | 0.5673 |
0.6456 | 80.0 | 2800 | 0.6050 | 0.6346 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3