Buseak's picture
update model card README.md
d158c42
metadata
license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: spellcorrector_3110_v17
    results: []

spellcorrector_3110_v17

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: 0.0086
  • Precision: 0.9991
  • Recall: 0.9990
  • F1: 0.9990
  • Accuracy: 0.9977

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: 4
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.1377 1.0 1949 0.1114 0.9542 0.9820 0.9679 0.9757
0.1079 2.0 3898 0.0851 0.9680 0.9801 0.9740 0.9795
0.0904 3.0 5847 0.0717 0.9733 0.9842 0.9787 0.9823
0.0788 4.0 7796 0.0612 0.9773 0.9859 0.9816 0.9845
0.0709 5.0 9745 0.0548 0.9824 0.9843 0.9833 0.9858
0.0646 6.0 11694 0.0483 0.9847 0.9890 0.9868 0.9876
0.0579 7.0 13643 0.0426 0.9875 0.9889 0.9882 0.9889
0.0532 8.0 15592 0.0385 0.9897 0.9889 0.9893 0.9898
0.0477 9.0 17541 0.0320 0.9913 0.9932 0.9922 0.9916
0.044 10.0 19490 0.0268 0.9926 0.9952 0.9939 0.9929
0.0401 11.0 21439 0.0232 0.9937 0.9960 0.9949 0.9936
0.0366 12.0 23388 0.0200 0.9957 0.9961 0.9959 0.9944
0.0317 13.0 25337 0.0172 0.9968 0.9969 0.9968 0.9953
0.0294 14.0 27286 0.0146 0.9971 0.9979 0.9975 0.9959
0.0269 15.0 29235 0.0126 0.9979 0.9982 0.9981 0.9965
0.0248 16.0 31184 0.0119 0.9984 0.9982 0.9983 0.9968
0.0228 17.0 33133 0.0098 0.9987 0.9987 0.9987 0.9973
0.0203 18.0 35082 0.0091 0.9989 0.9987 0.9988 0.9975
0.0189 19.0 37031 0.0087 0.9990 0.9989 0.9990 0.9976
0.0198 20.0 38980 0.0086 0.9991 0.9990 0.9990 0.9977

Framework versions

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