|
--- |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: icdar23-entrydetector_plaintext_breaks_indents_left_diff_right_ref |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# icdar23-entrydetector_plaintext_breaks_indents_left_diff_right_ref |
|
|
|
This model is a fine-tuned version of [HueyNemud/das22-10-camembert_pretrained](https://huggingface.co/HueyNemud/das22-10-camembert_pretrained) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0045 |
|
- Ebegin: {'precision': 0.9977203647416414, 'recall': 0.9875893192929672, 'f1': 0.9926289926289926, 'number': 2659} |
|
- Eend: {'precision': 0.9962221382697394, 'recall': 0.9854260089686099, 'f1': 0.9907946646627841, 'number': 2676} |
|
- Overall Precision: 0.9970 |
|
- Overall Recall: 0.9865 |
|
- Overall F1: 0.9917 |
|
- Overall Accuracy: 0.9986 |
|
|
|
## 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 | 0.0294 | 0.9739 | 0.9925 | 0.9831 | 0.9972 | |
|
| 0.1819 | 0.14 | 600 | 0.0152 | 0.9911 | 0.9804 | 0.9857 | 0.9978 | |
|
| 0.1819 | 0.21 | 900 | 0.0067 | 0.9871 | 0.9959 | 0.9915 | 0.9986 | |
|
| 0.0165 | 0.29 | 1200 | 0.0077 | 0.9871 | 0.9961 | 0.9916 | 0.9986 | |
|
| 0.0103 | 0.36 | 1500 | 0.0065 | 0.9872 | 0.9962 | 0.9917 | 0.9986 | |
|
| 0.0103 | 0.43 | 1800 | 0.0053 | 0.9903 | 0.9952 | 0.9927 | 0.9988 | |
|
| 0.0087 | 0.5 | 2100 | 0.0068 | 0.9974 | 0.9886 | 0.9930 | 0.9988 | |
|
| 0.0087 | 0.57 | 2400 | 0.0073 | 0.9951 | 0.9877 | 0.9914 | 0.9985 | |
|
| 0.0078 | 0.64 | 2700 | 0.0045 | 0.9899 | 0.9946 | 0.9923 | 0.9987 | |
|
| 0.0054 | 0.72 | 3000 | 0.0043 | 0.9978 | 0.9905 | 0.9941 | 0.9990 | |
|
| 0.0054 | 0.79 | 3300 | 0.0042 | 0.9976 | 0.9899 | 0.9938 | 0.9989 | |
|
| 0.0047 | 0.86 | 3600 | 0.0042 | 0.9955 | 0.9925 | 0.9940 | 0.9990 | |
|
| 0.0047 | 0.93 | 3900 | 0.0048 | 0.9865 | 0.9974 | 0.9920 | 0.9986 | |
|
| 0.0044 | 1.0 | 4200 | 0.0034 | 0.9979 | 0.9919 | 0.9949 | 0.9991 | |
|
| 0.0026 | 1.07 | 4500 | 0.0041 | 0.9954 | 0.9944 | 0.9949 | 0.9991 | |
|
| 0.0026 | 1.14 | 4800 | 0.0036 | 0.9979 | 0.9922 | 0.9950 | 0.9992 | |
|
| 0.0029 | 1.22 | 5100 | 0.0037 | 0.9956 | 0.9931 | 0.9944 | 0.9991 | |
|
| 0.0029 | 1.29 | 5400 | 0.0050 | 0.9899 | 0.9956 | 0.9927 | 0.9988 | |
|
| 0.0029 | 1.36 | 5700 | 0.0034 | 0.9975 | 0.9935 | 0.9955 | 0.9993 | |
|
| 0.0028 | 1.43 | 6000 | 0.0036 | 0.9970 | 0.9937 | 0.9954 | 0.9992 | |
|
| 0.0028 | 1.5 | 6300 | 0.0038 | 0.9932 | 0.9951 | 0.9942 | 0.9990 | |
|
| 0.0027 | 1.57 | 6600 | 0.0034 | 0.9969 | 0.9933 | 0.9951 | 0.9992 | |
|
| 0.0027 | 1.65 | 6900 | 0.0034 | 0.9974 | 0.9929 | 0.9952 | 0.9992 | |
|
| 0.0027 | 1.72 | 7200 | 0.0036 | 0.9970 | 0.9934 | 0.9952 | 0.9992 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.26.1 |
|
- Pytorch 1.13.1+cu116 |
|
- Datasets 2.9.0 |
|
- Tokenizers 0.13.2 |
|
|
