File size: 3,969 Bytes
cf40f1d 3ab593a cf40f1d 3ab593a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 |
---
license: mit
tags:
- generated_from_trainer
datasets:
- conll2003
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: xlm-roberta-base-conll2003-en
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: conll2003
type: conll2003
config: conll2003
split: validation
args: conll2003
metrics:
- name: Precision
type: precision
value: 0.9478680879413725
- name: Recall
type: recall
value: 0.9588879528222409
- name: F1
type: f1
value: 0.9533461763966831
- name: Accuracy
type: accuracy
value: 0.9917972098823162
---
<!-- 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-en
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the [conll2003](https://huggingface.co/datasets/conll2003) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0534
- Precision: 0.9479
- Recall: 0.9589
- F1: 0.9533
- Accuracy: 0.9918
## 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 | 439 | 0.0535 | 0.9131 | 0.9238 | 0.9184 | 0.9865 |
| 0.1663 | 2.0 | 878 | 0.0461 | 0.9305 | 0.9390 | 0.9348 | 0.9887 |
| 0.0404 | 3.0 | 1317 | 0.0366 | 0.9431 | 0.9501 | 0.9466 | 0.9910 |
| 0.0252 | 4.0 | 1756 | 0.0381 | 0.9395 | 0.9516 | 0.9455 | 0.9908 |
| 0.0172 | 5.0 | 2195 | 0.0398 | 0.9409 | 0.9523 | 0.9466 | 0.9911 |
| 0.0119 | 6.0 | 2634 | 0.0429 | 0.9389 | 0.9560 | 0.9474 | 0.9910 |
| 0.0091 | 7.0 | 3073 | 0.0463 | 0.9451 | 0.9548 | 0.9500 | 0.9913 |
| 0.0063 | 8.0 | 3512 | 0.0446 | 0.9478 | 0.9575 | 0.9526 | 0.9919 |
| 0.0063 | 9.0 | 3951 | 0.0513 | 0.9424 | 0.9569 | 0.9496 | 0.9911 |
| 0.0049 | 10.0 | 4390 | 0.0494 | 0.9470 | 0.9545 | 0.9507 | 0.9915 |
| 0.0036 | 11.0 | 4829 | 0.0506 | 0.9477 | 0.9553 | 0.9515 | 0.9917 |
| 0.0029 | 12.0 | 5268 | 0.0518 | 0.9472 | 0.9586 | 0.9529 | 0.9919 |
| 0.0026 | 13.0 | 5707 | 0.0530 | 0.9451 | 0.9567 | 0.9508 | 0.9916 |
| 0.0021 | 14.0 | 6146 | 0.0526 | 0.9468 | 0.9567 | 0.9517 | 0.9917 |
| 0.0016 | 15.0 | 6585 | 0.0534 | 0.9479 | 0.9589 | 0.9533 | 0.9918 |
### Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2
### Citation
```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}
}
``` |