|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- news_commentary |
|
metrics: |
|
- bleu |
|
model-index: |
|
- name: opus-mt-ar-en-finetuned-ar-to-en |
|
results: |
|
- task: |
|
name: Sequence-to-sequence Language Modeling |
|
type: text2text-generation |
|
dataset: |
|
name: news_commentary |
|
type: news_commentary |
|
args: ar-en |
|
metrics: |
|
- name: Bleu |
|
type: bleu |
|
value: 32.8872 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# opus-mt-ar-en-finetuned-ar-to-en |
|
|
|
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-ar-en](https://huggingface.co/Helsinki-NLP/opus-mt-ar-en) on the news_commentary dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6933 |
|
- Bleu: 32.8872 |
|
- Gen Len: 56.084 |
|
|
|
## 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: 3e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 5 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |
|
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:| |
|
| No log | 1.0 | 188 | 0.7407 | 30.7259 | 56.296 | |
|
| No log | 2.0 | 376 | 0.6927 | 32.2038 | 58.602 | |
|
| 0.8066 | 3.0 | 564 | 0.6898 | 33.1091 | 57.72 | |
|
| 0.8066 | 4.0 | 752 | 0.6925 | 33.0842 | 56.574 | |
|
| 0.8066 | 5.0 | 940 | 0.6933 | 32.8872 | 56.084 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.20.0 |
|
- Pytorch 1.11.0+cu113 |
|
- Datasets 2.3.2 |
|
- Tokenizers 0.12.1 |
|
|
