HueyNemud's picture
update model card README.md
e78d099
|
raw
history blame
3.58 kB
metadata
tags:
  - generated_from_trainer
model-index:
  - name: icdar23-entrydetector_plaintext_breaks_indents_left_ref_right_ref
    results: []

icdar23-entrydetector_plaintext_breaks_indents_left_ref_right_ref

This model is a fine-tuned version of HueyNemud/das22-10-camembert_pretrained on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0033
  • Ebegin: {'precision': 0.9935849056603774, 'recall': 0.9902218879277924, 'f1': 0.9919005462422301, 'number': 2659}
  • Eend: {'precision': 0.996606334841629, 'recall': 0.9876681614349776, 'f1': 0.9921171171171171, 'number': 2676}
  • Overall Precision: 0.9951
  • Overall Recall: 0.9889
  • Overall F1: 0.9920
  • Overall Accuracy: 0.9989

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: 0.0001
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 7500

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 0.07 300 0.0262 0.9758 0.9906 0.9832 0.9973
0.1599 0.14 600 0.0116 0.9866 0.9919 0.9892 0.9982
0.1599 0.21 900 0.0111 0.9907 0.9856 0.9882 0.9980
0.0162 0.29 1200 0.0096 0.9813 0.9966 0.9889 0.9981
0.0099 0.36 1500 0.0060 0.9820 0.9955 0.9887 0.9982
0.0099 0.43 1800 0.0046 0.9925 0.9934 0.9929 0.9988
0.0074 0.5 2100 0.0057 0.9961 0.9880 0.9920 0.9987
0.0074 0.57 2400 0.0039 0.9911 0.9953 0.9932 0.9988
0.0072 0.64 2700 0.0075 0.9842 0.9949 0.9895 0.9982
0.0061 0.72 3000 0.0040 0.9906 0.9963 0.9934 0.9989
0.0061 0.79 3300 0.0034 0.9955 0.9936 0.9946 0.9991
0.005 0.86 3600 0.0034 0.9933 0.9946 0.9939 0.9990
0.005 0.93 3900 0.0047 0.9847 0.9976 0.9911 0.9985
0.0041 1.0 4200 0.0031 0.9972 0.9936 0.9954 0.9992
0.0031 1.07 4500 0.0030 0.9967 0.9945 0.9956 0.9992
0.0031 1.14 4800 0.0032 0.9966 0.9938 0.9952 0.9992
0.003 1.22 5100 0.0029 0.9960 0.9939 0.9949 0.9991
0.003 1.29 5400 0.0030 0.9935 0.9947 0.9941 0.9990
0.0023 1.36 5700 0.0028 0.9973 0.9933 0.9953 0.9992
0.0027 1.43 6000 0.0029 0.9968 0.9936 0.9952 0.9992

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.13.1+cu116
  • Datasets 2.9.0
  • Tokenizers 0.13.2