|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
datasets: papluca/language-identification |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: xlm-roberta-base-finetuned-language-identification |
|
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-finetuned-language-detection-new |
|
|
|
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the [Language Identification dataset](https://huggingface.co/datasets/papluca/language-identification). |
|
|
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0436 |
|
- Accuracy: 0.9959 |
|
|
|
## Model description |
|
|
|
The model used in this task is XLM-RoBERTa, a transformer model with a classification head on top. |
|
|
|
## Intended uses & limitations |
|
|
|
It identifies the language a document is written in and it supports 20 different langauges: |
|
|
|
Arabic (ar), Bulgarian (bg), German (de), Modern greek (el), English (en), Spanish (es), French (fr), Hindi (hi), Italian (it), Japanese (ja), Dutch (nl), Polish (pl), Portuguese (pt), Russian (ru), Swahili (sw), Thai (th), Turkish (tr), Urdu (ur), Vietnamese (vi), Chinese (zh) |
|
|
|
|
|
## Training and evaluation data |
|
|
|
The model is fine-tuned on the [Language Identification dataset](https://huggingface.co/datasets/papluca/language-identification), a corpus consists of text from 20 different languages. The dataset is split with 7000 sentences for training, 1000 for validating, and 1000 for testing. The accuracy on the test set is 99.5%. |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 2e-05 |
|
- train_batch_size: 2 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 2 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:-----:|:-----:|:---------------:|:--------:| |
|
| 0.0493 | 1.0 | 35000 | 0.0407 | 0.9955 | |
|
| 0.018 | 2.0 | 70000 | 0.0436 | 0.9959 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.19.2 |
|
- Pytorch 1.11.0+cu113 |
|
- Datasets 2.2.2 |
|
- Tokenizers 0.12.1 |
|
|
