|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- precision |
|
- recall |
|
- f1 |
|
- accuracy |
|
model-index: |
|
- name: pos_final_mono_de |
|
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. --> |
|
|
|
# pos_final_mono_de |
|
|
|
This model is a fine-tuned version of [dbmdz/bert-base-german-cased](https://huggingface.co/dbmdz/bert-base-german-cased) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1567 |
|
- Precision: 0.9771 |
|
- Recall: 0.9791 |
|
- F1: 0.9781 |
|
- Accuracy: 0.9810 |
|
|
|
## 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: 5e-05 |
|
- train_batch_size: 256 |
|
- eval_batch_size: 256 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 1024 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 40.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
| No log | 0.99 | 128 | 0.2357 | 0.9443 | 0.9413 | 0.9428 | 0.9475 | |
|
| No log | 1.99 | 256 | 0.0513 | 0.9843 | 0.9842 | 0.9842 | 0.9853 | |
|
| No log | 2.99 | 384 | 0.0406 | 0.9868 | 0.9866 | 0.9867 | 0.9875 | |
|
| 0.6822 | 3.99 | 512 | 0.0365 | 0.9877 | 0.9877 | 0.9877 | 0.9885 | |
|
| 0.6822 | 4.99 | 640 | 0.0352 | 0.9881 | 0.9882 | 0.9882 | 0.9890 | |
|
| 0.6822 | 5.99 | 768 | 0.0345 | 0.9887 | 0.9887 | 0.9887 | 0.9895 | |
|
| 0.6822 | 6.99 | 896 | 0.0353 | 0.9888 | 0.9888 | 0.9888 | 0.9896 | |
|
| 0.024 | 7.99 | 1024 | 0.0371 | 0.9886 | 0.9888 | 0.9887 | 0.9895 | |
|
| 0.024 | 8.99 | 1152 | 0.0387 | 0.9888 | 0.9888 | 0.9888 | 0.9896 | |
|
| 0.024 | 9.99 | 1280 | 0.0402 | 0.9890 | 0.9889 | 0.9890 | 0.9898 | |
|
| 0.024 | 10.99 | 1408 | 0.0429 | 0.9889 | 0.9890 | 0.9889 | 0.9897 | |
|
| 0.0128 | 11.99 | 1536 | 0.0454 | 0.9889 | 0.9889 | 0.9889 | 0.9896 | |
|
| 0.0128 | 12.99 | 1664 | 0.0461 | 0.9889 | 0.9889 | 0.9889 | 0.9897 | |
|
| 0.0128 | 13.99 | 1792 | 0.0477 | 0.9892 | 0.9891 | 0.9891 | 0.9899 | |
|
| 0.0128 | 14.99 | 1920 | 0.0507 | 0.9890 | 0.9891 | 0.9890 | 0.9898 | |
|
| 0.0069 | 15.99 | 2048 | 0.0514 | 0.9893 | 0.9893 | 0.9893 | 0.9901 | |
|
| 0.0069 | 16.99 | 2176 | 0.0530 | 0.9892 | 0.9892 | 0.9892 | 0.9899 | |
|
| 0.0069 | 17.99 | 2304 | 0.0552 | 0.9890 | 0.9891 | 0.9891 | 0.9898 | |
|
| 0.0069 | 18.99 | 2432 | 0.0567 | 0.9891 | 0.9892 | 0.9892 | 0.9898 | |
|
| 0.0037 | 19.99 | 2560 | 0.0577 | 0.9892 | 0.9893 | 0.9892 | 0.9900 | |
|
| 0.0037 | 20.99 | 2688 | 0.0592 | 0.9892 | 0.9893 | 0.9893 | 0.9899 | |
|
| 0.0037 | 21.99 | 2816 | 0.0606 | 0.9893 | 0.9893 | 0.9893 | 0.9900 | |
|
| 0.0037 | 22.99 | 2944 | 0.0628 | 0.9893 | 0.9893 | 0.9893 | 0.9900 | |
|
| 0.0023 | 23.99 | 3072 | 0.0629 | 0.9892 | 0.9891 | 0.9891 | 0.9899 | |
|
| 0.0023 | 24.99 | 3200 | 0.0625 | 0.9892 | 0.9893 | 0.9893 | 0.9900 | |
|
| 0.0023 | 25.99 | 3328 | 0.0636 | 0.9893 | 0.9893 | 0.9893 | 0.9900 | |
|
| 0.0023 | 26.99 | 3456 | 0.0650 | 0.9894 | 0.9894 | 0.9894 | 0.9901 | |
|
| 0.0017 | 27.99 | 3584 | 0.0644 | 0.9894 | 0.9894 | 0.9894 | 0.9901 | |
|
| 0.0017 | 28.99 | 3712 | 0.0656 | 0.9895 | 0.9895 | 0.9895 | 0.9901 | |
|
| 0.0017 | 29.99 | 3840 | 0.0668 | 0.9895 | 0.9895 | 0.9895 | 0.9902 | |
|
| 0.0017 | 30.99 | 3968 | 0.0666 | 0.9895 | 0.9894 | 0.9894 | 0.9901 | |
|
| 0.0011 | 31.99 | 4096 | 0.0678 | 0.9894 | 0.9894 | 0.9894 | 0.9900 | |
|
| 0.0011 | 32.99 | 4224 | 0.0685 | 0.9896 | 0.9896 | 0.9896 | 0.9902 | |
|
| 0.0011 | 33.99 | 4352 | 0.0692 | 0.9894 | 0.9894 | 0.9894 | 0.9901 | |
|
| 0.0011 | 34.99 | 4480 | 0.0698 | 0.9895 | 0.9895 | 0.9895 | 0.9902 | |
|
| 0.0009 | 35.99 | 4608 | 0.0698 | 0.9894 | 0.9894 | 0.9894 | 0.9901 | |
|
| 0.0009 | 36.99 | 4736 | 0.0695 | 0.9895 | 0.9895 | 0.9895 | 0.9902 | |
|
| 0.0009 | 37.99 | 4864 | 0.0696 | 0.9894 | 0.9895 | 0.9894 | 0.9902 | |
|
| 0.0009 | 38.99 | 4992 | 0.0699 | 0.9895 | 0.9895 | 0.9895 | 0.9902 | |
|
| 0.0007 | 39.99 | 5120 | 0.0697 | 0.9894 | 0.9894 | 0.9894 | 0.9901 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.25.1 |
|
- Pytorch 1.12.0 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.13.2 |
|
|
