<!-- 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. -->
20230829194716
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: 0.6036
- 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 | 0.6826 | 0.4615 |
No log | 2.0 | 70 | 1.4088 | 0.4135 |
No log | 3.0 | 105 | 0.6262 | 0.5385 |
No log | 4.0 | 140 | 0.6811 | 0.5 |
No log | 5.0 | 175 | 0.6179 | 0.5481 |
No log | 6.0 | 210 | 0.6164 | 0.5673 |
No log | 7.0 | 245 | 1.0753 | 0.3654 |
No log | 8.0 | 280 | 0.6432 | 0.6346 |
No log | 9.0 | 315 | 0.7495 | 0.4038 |
No log | 10.0 | 350 | 0.8799 | 0.4135 |
No log | 11.0 | 385 | 0.7895 | 0.375 |
No log | 12.0 | 420 | 1.5836 | 0.4135 |
No log | 13.0 | 455 | 0.6627 | 0.4231 |
No log | 14.0 | 490 | 0.6839 | 0.4904 |
0.789 | 15.0 | 525 | 0.6172 | 0.6346 |
0.789 | 16.0 | 560 | 0.7635 | 0.6058 |
0.789 | 17.0 | 595 | 0.6028 | 0.5769 |
0.789 | 18.0 | 630 | 0.8397 | 0.4135 |
0.789 | 19.0 | 665 | 0.8332 | 0.4038 |
0.789 | 20.0 | 700 | 0.5984 | 0.6346 |
0.789 | 21.0 | 735 | 0.8444 | 0.4231 |
0.789 | 22.0 | 770 | 0.6347 | 0.5481 |
0.789 | 23.0 | 805 | 0.8138 | 0.3846 |
0.789 | 24.0 | 840 | 0.6123 | 0.625 |
0.789 | 25.0 | 875 | 0.6067 | 0.6154 |
0.789 | 26.0 | 910 | 0.6349 | 0.4519 |
0.789 | 27.0 | 945 | 0.6530 | 0.6058 |
0.789 | 28.0 | 980 | 0.6060 | 0.6346 |
0.739 | 29.0 | 1015 | 0.6988 | 0.6154 |
0.739 | 30.0 | 1050 | 0.6460 | 0.5096 |
0.739 | 31.0 | 1085 | 0.6370 | 0.4712 |
0.739 | 32.0 | 1120 | 1.0492 | 0.4135 |
0.739 | 33.0 | 1155 | 0.7931 | 0.625 |
0.739 | 34.0 | 1190 | 0.5922 | 0.625 |
0.739 | 35.0 | 1225 | 0.6288 | 0.5385 |
0.739 | 36.0 | 1260 | 0.6562 | 0.3846 |
0.739 | 37.0 | 1295 | 0.6163 | 0.6346 |
0.739 | 38.0 | 1330 | 0.5908 | 0.6346 |
0.739 | 39.0 | 1365 | 0.7644 | 0.3942 |
0.739 | 40.0 | 1400 | 0.6363 | 0.4904 |
0.739 | 41.0 | 1435 | 0.6546 | 0.5288 |
0.739 | 42.0 | 1470 | 0.7025 | 0.4038 |
0.6941 | 43.0 | 1505 | 0.5878 | 0.6346 |
0.6941 | 44.0 | 1540 | 0.5859 | 0.6346 |
0.6941 | 45.0 | 1575 | 0.6765 | 0.4712 |
0.6941 | 46.0 | 1610 | 0.6047 | 0.625 |
0.6941 | 47.0 | 1645 | 0.6015 | 0.625 |
0.6941 | 48.0 | 1680 | 0.6275 | 0.5385 |
0.6941 | 49.0 | 1715 | 0.6394 | 0.6346 |
0.6941 | 50.0 | 1750 | 0.5939 | 0.6154 |
0.6941 | 51.0 | 1785 | 0.6507 | 0.4327 |
0.6941 | 52.0 | 1820 | 0.6356 | 0.625 |
0.6941 | 53.0 | 1855 | 0.6994 | 0.4135 |
0.6941 | 54.0 | 1890 | 0.6811 | 0.4135 |
0.6941 | 55.0 | 1925 | 0.6032 | 0.6346 |
0.6941 | 56.0 | 1960 | 0.6382 | 0.6346 |
0.6941 | 57.0 | 1995 | 0.5909 | 0.6346 |
0.6702 | 58.0 | 2030 | 0.5930 | 0.6346 |
0.6702 | 59.0 | 2065 | 0.6644 | 0.4038 |
0.6702 | 60.0 | 2100 | 0.6210 | 0.6346 |
0.6702 | 61.0 | 2135 | 0.5941 | 0.6346 |
0.6702 | 62.0 | 2170 | 0.5983 | 0.6346 |
0.6702 | 63.0 | 2205 | 0.6180 | 0.5385 |
0.6702 | 64.0 | 2240 | 0.5944 | 0.6346 |
0.6702 | 65.0 | 2275 | 0.6127 | 0.625 |
0.6702 | 66.0 | 2310 | 0.6350 | 0.4327 |
0.6702 | 67.0 | 2345 | 0.6102 | 0.5962 |
0.6702 | 68.0 | 2380 | 0.6076 | 0.6731 |
0.6702 | 69.0 | 2415 | 0.5925 | 0.6346 |
0.6702 | 70.0 | 2450 | 0.6078 | 0.6346 |
0.6702 | 71.0 | 2485 | 0.5985 | 0.6442 |
0.6475 | 72.0 | 2520 | 0.6140 | 0.6058 |
0.6475 | 73.0 | 2555 | 0.6000 | 0.6346 |
0.6475 | 74.0 | 2590 | 0.6181 | 0.5288 |
0.6475 | 75.0 | 2625 | 0.6062 | 0.6442 |
0.6475 | 76.0 | 2660 | 0.6169 | 0.5769 |
0.6475 | 77.0 | 2695 | 0.6235 | 0.5288 |
0.6475 | 78.0 | 2730 | 0.6160 | 0.5769 |
0.6475 | 79.0 | 2765 | 0.6108 | 0.6442 |
0.6475 | 80.0 | 2800 | 0.6036 | 0.6346 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3