metadata
tags:
- generated_from_trainer
model-index:
- name: icdar23-entrydetector_plaintext_breaks_indents_left_diff_right_ref
results: []
icdar23-entrydetector_plaintext_breaks_indents_left_diff_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: 0.0052
- Ebegin: {'precision': 0.9891263592050994, 'recall': 0.9921022940955246, 'f1': 0.9906120916259857, 'number': 2659}
- Eend: {'precision': 0.9947029890276201, 'recall': 0.9824364723467862, 'f1': 0.9885316788870088, 'number': 2676}
- Overall Precision: 0.9919
- Overall Recall: 0.9873
- Overall F1: 0.9896
- Overall Accuracy: 0.9984
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 | 0.0329 | 0.9706 | 0.9804 | 0.9755 | 0.9968 |
0.1902 | 0.14 | 600 | 0.0141 | 0.9815 | 0.9919 | 0.9867 | 0.9978 |
0.1902 | 0.21 | 900 | 0.0130 | 0.9853 | 0.9866 | 0.9860 | 0.9976 |
0.0162 | 0.29 | 1200 | 0.0110 | 0.9835 | 0.9932 | 0.9883 | 0.9981 |
0.0102 | 0.36 | 1500 | 0.0086 | 0.9856 | 0.9943 | 0.9899 | 0.9983 |
0.0102 | 0.43 | 1800 | 0.0052 | 0.9921 | 0.9909 | 0.9915 | 0.9987 |
0.0071 | 0.5 | 2100 | 0.0061 | 0.9915 | 0.9913 | 0.9914 | 0.9986 |
0.0071 | 0.57 | 2400 | 0.0053 | 0.9938 | 0.9915 | 0.9927 | 0.9988 |
0.0083 | 0.64 | 2700 | 0.0054 | 0.9905 | 0.9902 | 0.9904 | 0.9984 |
0.0058 | 0.72 | 3000 | 0.0060 | 0.9843 | 0.9953 | 0.9898 | 0.9983 |
0.0058 | 0.79 | 3300 | 0.0050 | 0.9919 | 0.9933 | 0.9926 | 0.9988 |
0.0067 | 0.86 | 3600 | 0.0062 | 0.9905 | 0.9935 | 0.9920 | 0.9987 |
0.0067 | 0.93 | 3900 | 0.0049 | 0.9883 | 0.9956 | 0.9919 | 0.9986 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2