|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- precision |
|
- recall |
|
- f1 |
|
- accuracy |
|
model-index: |
|
- name: xlm-roberta-base-conll2003 |
|
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. --> |
|
|
|
# xlm-roberta-base-conll2003 |
|
|
|
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the [conll2003-persian](https://huggingface.co/datasets/Amir13/conll2003-persian |
|
) dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1579 |
|
- Precision: 0.8794 |
|
- Recall: 0.8745 |
|
- F1: 0.8769 |
|
- Accuracy: 0.9758 |
|
|
|
## 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: 15 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
| No log | 1.0 | 430 | 0.1374 | 0.8043 | 0.7966 | 0.8004 | 0.9613 | |
|
| 0.2862 | 2.0 | 860 | 0.1093 | 0.8384 | 0.8482 | 0.8433 | 0.9695 | |
|
| 0.1043 | 3.0 | 1290 | 0.1121 | 0.8448 | 0.8556 | 0.8502 | 0.9708 | |
|
| 0.0689 | 4.0 | 1720 | 0.1094 | 0.8635 | 0.8650 | 0.8643 | 0.9737 | |
|
| 0.0473 | 5.0 | 2150 | 0.1225 | 0.8665 | 0.8625 | 0.8645 | 0.9736 | |
|
| 0.0342 | 6.0 | 2580 | 0.1186 | 0.8722 | 0.8730 | 0.8726 | 0.9745 | |
|
| 0.0245 | 7.0 | 3010 | 0.1292 | 0.8802 | 0.8717 | 0.8759 | 0.9755 | |
|
| 0.0245 | 8.0 | 3440 | 0.1309 | 0.8832 | 0.8689 | 0.8760 | 0.9749 | |
|
| 0.0177 | 9.0 | 3870 | 0.1388 | 0.8712 | 0.8717 | 0.8715 | 0.9743 | |
|
| 0.0135 | 10.0 | 4300 | 0.1466 | 0.8699 | 0.8728 | 0.8714 | 0.9752 | |
|
| 0.0103 | 11.0 | 4730 | 0.1486 | 0.8716 | 0.8747 | 0.8731 | 0.9756 | |
|
| 0.0081 | 12.0 | 5160 | 0.1521 | 0.8789 | 0.8736 | 0.8762 | 0.9759 | |
|
| 0.007 | 13.0 | 5590 | 0.1546 | 0.8804 | 0.8734 | 0.8769 | 0.9756 | |
|
| 0.0053 | 14.0 | 6020 | 0.1552 | 0.8750 | 0.8732 | 0.8741 | 0.9756 | |
|
| 0.0053 | 15.0 | 6450 | 0.1579 | 0.8794 | 0.8745 | 0.8769 | 0.9758 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.27.0.dev0 |
|
- Pytorch 1.13.1+cu116 |
|
- Datasets 2.8.0 |
|
- Tokenizers 0.13.2 |
|
|
|
### Citation |
|
If you used the datasets and models in this repository, please cite it. |
|
|
|
```bibtex |
|
@misc{https://doi.org/10.48550/arxiv.2302.09611, |
|
doi = {10.48550/ARXIV.2302.09611}, |
|
url = {https://arxiv.org/abs/2302.09611}, |
|
author = {Sartipi, Amir and Fatemi, Afsaneh}, |
|
keywords = {Computation and Language (cs.CL), Artificial Intelligence (cs.AI), FOS: Computer and information sciences, FOS: Computer and information sciences}, |
|
title = {Exploring the Potential of Machine Translation for Generating Named Entity Datasets: A Case Study between Persian and English}, |
|
publisher = {arXiv}, |
|
year = {2023}, |
|
copyright = {arXiv.org perpetual, non-exclusive license} |
|
} |
|
``` |
|
|
