Buseak's picture
update model card README.md
d067bf6
metadata
license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: spellcorrector_2510_v15_canine-s
    results: []

spellcorrector_2510_v15_canine-s

This model is a fine-tuned version of google/canine-s on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1599
  • Precision: 0.9768
  • Recall: 0.9820
  • F1: 0.9794
  • Accuracy: 0.9786

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: 2e-05
  • train_batch_size: 4
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.1921 1.0 1951 0.1677 0.9417 0.9774 0.9592 0.9650
0.1627 2.0 3902 0.1436 0.9500 0.9779 0.9637 0.9674
0.1395 3.0 5853 0.1266 0.9545 0.9788 0.9665 0.9697
0.1266 4.0 7804 0.1172 0.9661 0.9698 0.9680 0.9702
0.1105 5.0 9755 0.1064 0.9669 0.9766 0.9717 0.9731
0.1011 6.0 11706 0.1011 0.9705 0.9757 0.9731 0.9745
0.0933 7.0 13657 0.0987 0.9718 0.9766 0.9742 0.9752
0.0851 8.0 15608 0.0973 0.9715 0.9787 0.9751 0.9755
0.0758 9.0 17559 0.0998 0.9734 0.9765 0.9750 0.9756
0.069 10.0 19510 0.0993 0.9732 0.9810 0.9771 0.9764
0.0635 11.0 21461 0.1055 0.9739 0.9808 0.9773 0.9766
0.0576 12.0 23412 0.1072 0.9751 0.9794 0.9772 0.9765
0.0493 13.0 25363 0.1078 0.9754 0.9807 0.9780 0.9776
0.0469 14.0 27314 0.1145 0.9757 0.9815 0.9786 0.9777
0.0409 15.0 29265 0.1174 0.9758 0.9806 0.9782 0.9764
0.0373 16.0 31216 0.1218 0.9763 0.9801 0.9782 0.9769
0.0338 17.0 33167 0.1239 0.9768 0.9805 0.9787 0.9773
0.0326 18.0 35118 0.1312 0.9770 0.9787 0.9779 0.9773
0.029 19.0 37069 0.1320 0.9764 0.9809 0.9786 0.9773
0.0245 20.0 39020 0.1376 0.9767 0.9802 0.9784 0.9777
0.0231 21.0 40971 0.1382 0.9763 0.9814 0.9788 0.9776
0.0212 22.0 42922 0.1473 0.9762 0.9826 0.9794 0.9780
0.0201 23.0 44873 0.1485 0.9762 0.9816 0.9789 0.9778
0.0187 24.0 46824 0.1494 0.9763 0.9818 0.9790 0.9775
0.0166 25.0 48775 0.1502 0.9769 0.9813 0.9791 0.9781
0.0163 26.0 50726 0.1560 0.9769 0.9813 0.9791 0.9785
0.0149 27.0 52677 0.1556 0.9764 0.9824 0.9794 0.9784
0.0143 28.0 54628 0.1587 0.9767 0.9818 0.9792 0.9784
0.0126 29.0 56579 0.1589 0.9766 0.9821 0.9793 0.9784
0.013 30.0 58530 0.1599 0.9768 0.9820 0.9794 0.9786

Framework versions

  • Transformers 4.28.0
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.6
  • Tokenizers 0.13.3