20230829231514
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: 0.5962
- Accuracy: 0.6827
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: 0.007
- 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 | 1.8158 | 0.5865 |
No log | 2.0 | 70 | 0.5893 | 0.625 |
No log | 3.0 | 105 | 0.8945 | 0.5962 |
No log | 4.0 | 140 | 0.5866 | 0.625 |
No log | 5.0 | 175 | 0.9890 | 0.3846 |
No log | 6.0 | 210 | 0.8076 | 0.5192 |
No log | 7.0 | 245 | 0.6353 | 0.5288 |
No log | 8.0 | 280 | 2.2871 | 0.3846 |
No log | 9.0 | 315 | 0.7403 | 0.6346 |
No log | 10.0 | 350 | 1.4011 | 0.4038 |
No log | 11.0 | 385 | 1.1139 | 0.4038 |
No log | 12.0 | 420 | 0.9394 | 0.6058 |
No log | 13.0 | 455 | 0.6693 | 0.5865 |
No log | 14.0 | 490 | 1.1625 | 0.4231 |
1.0588 | 15.0 | 525 | 0.6894 | 0.6346 |
1.0588 | 16.0 | 560 | 0.6938 | 0.3942 |
1.0588 | 17.0 | 595 | 0.6737 | 0.5 |
1.0588 | 18.0 | 630 | 0.7273 | 0.625 |
1.0588 | 19.0 | 665 | 0.6071 | 0.5385 |
1.0588 | 20.0 | 700 | 1.0395 | 0.5192 |
1.0588 | 21.0 | 735 | 0.6420 | 0.6058 |
1.0588 | 22.0 | 770 | 0.7194 | 0.6154 |
1.0588 | 23.0 | 805 | 1.3367 | 0.3942 |
1.0588 | 24.0 | 840 | 0.9467 | 0.4231 |
1.0588 | 25.0 | 875 | 0.6453 | 0.6058 |
1.0588 | 26.0 | 910 | 0.6247 | 0.6346 |
1.0588 | 27.0 | 945 | 0.6118 | 0.5577 |
1.0588 | 28.0 | 980 | 0.7381 | 0.4423 |
0.8818 | 29.0 | 1015 | 0.5847 | 0.6346 |
0.8818 | 30.0 | 1050 | 0.7924 | 0.3654 |
0.8818 | 31.0 | 1085 | 0.7978 | 0.4231 |
0.8818 | 32.0 | 1120 | 1.1682 | 0.3654 |
0.8818 | 33.0 | 1155 | 1.1758 | 0.6346 |
0.8818 | 34.0 | 1190 | 0.6784 | 0.6442 |
0.8818 | 35.0 | 1225 | 0.6660 | 0.4135 |
0.8818 | 36.0 | 1260 | 1.1904 | 0.3654 |
0.8818 | 37.0 | 1295 | 0.5965 | 0.6731 |
0.8818 | 38.0 | 1330 | 0.6026 | 0.6442 |
0.8818 | 39.0 | 1365 | 0.6658 | 0.6346 |
0.8818 | 40.0 | 1400 | 0.7463 | 0.3846 |
0.8818 | 41.0 | 1435 | 1.2989 | 0.3654 |
0.8818 | 42.0 | 1470 | 0.9206 | 0.3654 |
0.8069 | 43.0 | 1505 | 0.6119 | 0.6346 |
0.8069 | 44.0 | 1540 | 0.7291 | 0.4038 |
0.8069 | 45.0 | 1575 | 0.9749 | 0.3654 |
0.8069 | 46.0 | 1610 | 0.6391 | 0.4808 |
0.8069 | 47.0 | 1645 | 0.5934 | 0.6442 |
0.8069 | 48.0 | 1680 | 0.6020 | 0.6346 |
0.8069 | 49.0 | 1715 | 0.6096 | 0.6346 |
0.8069 | 50.0 | 1750 | 0.7630 | 0.3654 |
0.8069 | 51.0 | 1785 | 0.8983 | 0.3654 |
0.8069 | 52.0 | 1820 | 0.6252 | 0.5481 |
0.8069 | 53.0 | 1855 | 0.9840 | 0.3654 |
0.8069 | 54.0 | 1890 | 0.7640 | 0.3846 |
0.8069 | 55.0 | 1925 | 0.6074 | 0.6346 |
0.8069 | 56.0 | 1960 | 0.5978 | 0.6346 |
0.8069 | 57.0 | 1995 | 0.7187 | 0.375 |
0.7258 | 58.0 | 2030 | 0.6309 | 0.4423 |
0.7258 | 59.0 | 2065 | 0.6101 | 0.6442 |
0.7258 | 60.0 | 2100 | 0.6555 | 0.6346 |
0.7258 | 61.0 | 2135 | 0.6048 | 0.6346 |
0.7258 | 62.0 | 2170 | 0.6749 | 0.4038 |
0.7258 | 63.0 | 2205 | 0.6003 | 0.6538 |
0.7258 | 64.0 | 2240 | 0.6711 | 0.6346 |
0.7258 | 65.0 | 2275 | 0.5839 | 0.6346 |
0.7258 | 66.0 | 2310 | 0.5848 | 0.6346 |
0.7258 | 67.0 | 2345 | 0.6198 | 0.6346 |
0.7258 | 68.0 | 2380 | 0.6282 | 0.4904 |
0.7258 | 69.0 | 2415 | 0.5936 | 0.6346 |
0.7258 | 70.0 | 2450 | 0.5954 | 0.6346 |
0.7258 | 71.0 | 2485 | 0.5858 | 0.6346 |
0.6781 | 72.0 | 2520 | 0.6104 | 0.5769 |
0.6781 | 73.0 | 2555 | 0.6286 | 0.5192 |
0.6781 | 74.0 | 2590 | 0.6538 | 0.4231 |
0.6781 | 75.0 | 2625 | 0.6025 | 0.625 |
0.6781 | 76.0 | 2660 | 0.5940 | 0.6635 |
0.6781 | 77.0 | 2695 | 0.7307 | 0.3846 |
0.6781 | 78.0 | 2730 | 0.6168 | 0.5673 |
0.6781 | 79.0 | 2765 | 0.5995 | 0.6635 |
0.6781 | 80.0 | 2800 | 0.5962 | 0.6827 |
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.