---
license: cc-by-nc-sa-4.0
tags:
- generated_from_trainer
datasets:
- opus_infopankki
metrics:
- bleu
model-index:
- name: mt5-small-parsinlu-opus-translation_fa_en-finetuned-fa-to-en
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    dataset:
      name: opus_infopankki
      type: opus_infopankki
      args: en-fa
    metrics:
    - name: Bleu
      type: bleu
      value: 9.5106
---

<!-- 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. -->

# mt5-small-parsinlu-opus-translation_fa_en-finetuned-fa-to-en

This model is a fine-tuned version of [persiannlp/mt5-small-parsinlu-opus-translation_fa_en](https://huggingface.co/persiannlp/mt5-small-parsinlu-opus-translation_fa_en) on the opus_infopankki dataset.
It achieves the following results on the evaluation set:
- Loss: 2.5449
- Bleu: 9.5106
- Gen Len: 13.6434

## 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-06
- 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: 30
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Bleu   | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|
| No log        | 1.0   | 151  | 3.1656          | 7.194  | 14.1885 |
| No log        | 2.0   | 302  | 3.0419          | 7.7031 | 14.1005 |
| No log        | 3.0   | 453  | 2.9549          | 8.1502 | 13.9834 |
| 3.5336        | 4.0   | 604  | 2.8857          | 8.4488 | 13.9251 |
| 3.5336        | 5.0   | 755  | 2.8297          | 8.6606 | 13.786  |
| 3.5336        | 6.0   | 906  | 2.7808          | 8.8217 | 13.7983 |
| 3.2511        | 7.0   | 1057 | 2.7386          | 8.9221 | 13.7518 |
| 3.2511        | 8.0   | 1208 | 2.7006          | 9.1988 | 13.7159 |
| 3.2511        | 9.0   | 1359 | 2.6678          | 9.2751 | 13.676  |
| 3.1055        | 10.0  | 1510 | 2.6387          | 9.4142 | 13.6648 |
| 3.1055        | 11.0  | 1661 | 2.6154          | 9.5726 | 13.6841 |
| 3.1055        | 12.0  | 1812 | 2.5945          | 9.6571 | 13.6546 |
| 3.1055        | 13.0  | 1963 | 2.5813          | 9.8303 | 13.6571 |
| 3.0199        | 14.0  | 2114 | 2.5709          | 9.6726 | 13.5855 |
| 3.0199        | 15.0  | 2265 | 2.5619          | 9.632  | 13.6125 |
| 3.0199        | 16.0  | 2416 | 2.5563          | 9.5773 | 13.6256 |
| 2.9862        | 17.0  | 2567 | 2.5538          | 9.5425 | 13.6366 |
| 2.9862        | 18.0  | 2718 | 2.5515          | 9.5359 | 13.6326 |
| 2.9862        | 19.0  | 2869 | 2.5495          | 9.5544 | 13.642  |
| 2.9859        | 20.0  | 3020 | 2.5478          | 9.5183 | 13.6374 |
| 2.9859        | 21.0  | 3171 | 2.5466          | 9.5387 | 13.632  |
| 2.9859        | 22.0  | 3322 | 2.5458          | 9.5183 | 13.6355 |
| 2.9859        | 23.0  | 3473 | 2.5451          | 9.5019 | 13.6376 |
| 2.9731        | 24.0  | 3624 | 2.5449          | 9.5004 | 13.6405 |
| 2.9731        | 25.0  | 3775 | 2.5449          | 9.5106 | 13.6434 |
| 2.9731        | 26.0  | 3926 | 2.5449          | 9.5106 | 13.6434 |
| 2.9671        | 27.0  | 4077 | 2.5449          | 9.5106 | 13.6434 |
| 2.9671        | 28.0  | 4228 | 2.5449          | 9.5106 | 13.6434 |
| 2.9671        | 29.0  | 4379 | 2.5449          | 9.5106 | 13.6434 |
| 2.97          | 30.0  | 4530 | 2.5449          | 9.5106 | 13.6434 |


### Framework versions

- Transformers 4.19.2
- Pytorch 1.7.1+cu110
- Datasets 2.2.2
- Tokenizers 0.12.1