pos_tagger_3112_v3 / README.md
Buseak's picture
End of training
3b4e8b0
metadata
license: mit
base_model: dbmdz/bert-base-turkish-cased
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: pos_tagger_3112_v3
    results: []

pos_tagger_3112_v3

This model is a fine-tuned version of dbmdz/bert-base-turkish-cased on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7728
  • Precision: 0.8922
  • Recall: 0.8955
  • F1: 0.8938
  • Accuracy: 0.9244

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: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 1.0 244 0.3040 0.8905 0.8924 0.8915 0.9215
No log 2.0 488 0.2915 0.8981 0.9006 0.8994 0.9279
0.3896 3.0 732 0.3109 0.8933 0.8932 0.8933 0.9234
0.3896 4.0 976 0.3004 0.8954 0.8983 0.8969 0.9263
0.159 5.0 1220 0.3338 0.8929 0.8946 0.8937 0.9242
0.159 6.0 1464 0.3419 0.8914 0.8958 0.8936 0.9240
0.1038 7.0 1708 0.3840 0.8892 0.8930 0.8911 0.9223
0.1038 8.0 1952 0.3923 0.8857 0.8930 0.8894 0.9213
0.0629 9.0 2196 0.4441 0.8888 0.8914 0.8901 0.9213
0.0629 10.0 2440 0.4769 0.8886 0.8929 0.8908 0.9231
0.0357 11.0 2684 0.4846 0.8859 0.8913 0.8886 0.9199
0.0357 12.0 2928 0.5256 0.8877 0.8895 0.8886 0.9211
0.0212 13.0 3172 0.5554 0.8896 0.8900 0.8898 0.9219
0.0212 14.0 3416 0.5748 0.8870 0.8911 0.8890 0.9207
0.0143 15.0 3660 0.5988 0.8877 0.8916 0.8896 0.9220
0.0143 16.0 3904 0.6047 0.8874 0.8903 0.8888 0.9209
0.0098 17.0 4148 0.6161 0.8846 0.8914 0.8880 0.9199
0.0098 18.0 4392 0.6158 0.8883 0.8929 0.8906 0.9217
0.0072 19.0 4636 0.6216 0.8858 0.8928 0.8893 0.9209
0.0072 20.0 4880 0.6497 0.8892 0.8926 0.8909 0.9215
0.0058 21.0 5124 0.6698 0.8887 0.8919 0.8903 0.9216
0.0058 22.0 5368 0.6582 0.8858 0.8916 0.8887 0.9208
0.0046 23.0 5612 0.6915 0.8866 0.8925 0.8896 0.9212
0.0046 24.0 5856 0.6725 0.8898 0.8928 0.8913 0.9222
0.004 25.0 6100 0.6678 0.8912 0.8961 0.8936 0.9238
0.004 26.0 6344 0.6899 0.8891 0.8933 0.8912 0.9224
0.0034 27.0 6588 0.7082 0.8890 0.8922 0.8906 0.9215
0.0034 28.0 6832 0.7066 0.8903 0.8920 0.8911 0.9228
0.0026 29.0 7076 0.7243 0.8882 0.8938 0.8910 0.9228
0.0026 30.0 7320 0.7322 0.8891 0.8923 0.8907 0.9226
0.0023 31.0 7564 0.7292 0.8909 0.8930 0.8920 0.9230
0.0023 32.0 7808 0.7227 0.8922 0.8947 0.8934 0.9244
0.0027 33.0 8052 0.7231 0.8885 0.8922 0.8903 0.9222
0.0027 34.0 8296 0.7236 0.8907 0.8936 0.8922 0.9233
0.0019 35.0 8540 0.7313 0.8875 0.8895 0.8885 0.9214
0.0019 36.0 8784 0.7240 0.8902 0.8935 0.8919 0.9234
0.0017 37.0 9028 0.7364 0.8903 0.8939 0.8921 0.9233
0.0017 38.0 9272 0.7479 0.8896 0.8929 0.8913 0.9232
0.0013 39.0 9516 0.7511 0.8895 0.8937 0.8916 0.9230
0.0013 40.0 9760 0.7689 0.8896 0.8948 0.8922 0.9234
0.001 41.0 10004 0.7597 0.8909 0.8958 0.8933 0.9238
0.001 42.0 10248 0.7581 0.8897 0.8929 0.8913 0.9230
0.001 43.0 10492 0.7512 0.8919 0.8952 0.8935 0.9244
0.0012 44.0 10736 0.7622 0.8921 0.8957 0.8939 0.9244
0.0012 45.0 10980 0.7707 0.8907 0.8952 0.8930 0.9237
0.001 46.0 11224 0.7700 0.8922 0.8963 0.8942 0.9244
0.001 47.0 11468 0.7742 0.8895 0.8938 0.8916 0.9231
0.0009 48.0 11712 0.7753 0.8911 0.8945 0.8928 0.9239
0.0009 49.0 11956 0.7746 0.8909 0.8944 0.8927 0.9236
0.0008 50.0 12200 0.7728 0.8922 0.8955 0.8938 0.9244

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0