Edit model card

2_2e-3_1_0.1

This model is a fine-tuned version of bert-large-uncased on the super_glue dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5541
  • Accuracy: 0.7003

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: 0.002
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 11
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 60.0

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.8034 1.0 590 0.6537 0.6217
0.8338 2.0 1180 0.7014 0.6217
0.8142 3.0 1770 0.6716 0.5596
0.7701 4.0 2360 0.6599 0.6217
0.7412 5.0 2950 0.7053 0.6217
0.7414 6.0 3540 0.6539 0.6217
0.7411 7.0 4130 0.9828 0.3817
0.7237 8.0 4720 0.6571 0.6061
0.7339 9.0 5310 0.6448 0.6232
0.7005 10.0 5900 0.6632 0.6223
0.7171 11.0 6490 0.6442 0.6220
0.7084 12.0 7080 0.7522 0.4477
0.6985 13.0 7670 0.6253 0.6336
0.7044 14.0 8260 0.7021 0.6217
0.6752 15.0 8850 0.6321 0.6183
0.6817 16.0 9440 0.6388 0.6073
0.6715 17.0 10030 0.6276 0.6358
0.6591 18.0 10620 0.6297 0.6474
0.6681 19.0 11210 0.6139 0.6407
0.6595 20.0 11800 0.6048 0.6541
0.6463 21.0 12390 0.6135 0.6541
0.6391 22.0 12980 0.6181 0.6437
0.6407 23.0 13570 0.6047 0.6615
0.6226 24.0 14160 0.6077 0.6615
0.6271 25.0 14750 0.6129 0.6642
0.6288 26.0 15340 0.6329 0.6343
0.6254 27.0 15930 0.5903 0.6728
0.6085 28.0 16520 0.5946 0.6743
0.6107 29.0 17110 0.5848 0.6737
0.5917 30.0 17700 0.6179 0.6725
0.5997 31.0 18290 0.5991 0.6618
0.5877 32.0 18880 0.6386 0.6709
0.5894 33.0 19470 0.5830 0.6771
0.5804 34.0 20060 0.5765 0.6856
0.5751 35.0 20650 0.5944 0.6615
0.5825 36.0 21240 0.5702 0.6890
0.5824 37.0 21830 0.5807 0.6774
0.5671 38.0 22420 0.5671 0.6838
0.573 39.0 23010 0.5678 0.6862
0.5615 40.0 23600 0.5685 0.6893
0.5658 41.0 24190 0.5820 0.6792
0.5669 42.0 24780 0.5692 0.6902
0.5663 43.0 25370 0.5665 0.6881
0.5533 44.0 25960 0.5599 0.6920
0.5552 45.0 26550 0.5637 0.6905
0.5515 46.0 27140 0.5616 0.6893
0.5593 47.0 27730 0.5650 0.6887
0.5487 48.0 28320 0.5620 0.6948
0.5563 49.0 28910 0.5631 0.6911
0.5486 50.0 29500 0.5604 0.6972
0.5464 51.0 30090 0.5590 0.6939
0.5469 52.0 30680 0.5561 0.6969
0.5458 53.0 31270 0.5573 0.7
0.5425 54.0 31860 0.5558 0.6976
0.5412 55.0 32450 0.5552 0.6991
0.5434 56.0 33040 0.5564 0.6979
0.5363 57.0 33630 0.5536 0.6982
0.5404 58.0 34220 0.5556 0.6982
0.5378 59.0 34810 0.5542 0.6991
0.5431 60.0 35400 0.5541 0.7003

Framework versions

  • Transformers 4.30.0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.4
  • Tokenizers 0.13.3
Downloads last month
4
Inference API
This model can be loaded on Inference API (serverless).

Dataset used to train Onutoa/2_2e-3_1_0.1