mbart-large-50-English_Spanish_Translation
This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0290
- Bleu: 41.4437
- Rouge:
- Rouge1: 0.6751402780531002
- Rouge2: 0.49769602014143044
- RougeL: 0.6371513427059108
- RougeLsum: 0.6376403149816605
- Meteor: 0.6479226630466496
Model description
This project translates Spanish text inputs into English.
Here is the link to the script I created to train this model: https://github.com/DunnBC22/NLP_Projects/blob/main/Machine%20Translation/NLP%20Translation%20Project-EN:ES.ipynb
Intended uses & limitations
This model is intended to demonstrate my ability to solve a complex problem using technology.
Training and evaluation data
Dataset Source: https://www.kaggle.com/datasets/hgultekin/paralel-translation-corpus-in-22-languages
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: 2
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Rouge1 | Rouge2 | RougeL | RougeLsum | Meteor |
|---|---|---|---|---|---|---|---|---|---|
| 1.5608 | 1.0 | 900 | 1.0899 | 39.9184 | 0.6645 | 0.4846 | 0.6254 | 0.6259 | 0.6376 |
| 0.9734 | 2.0 | 1800 | 1.0290 | 41.4436 | 0.6751 | 0.4977 | 0.6371 | 0.6376 | 0.6479 |
Framework versions
- Transformers 4.22.2
- Pytorch 1.12.1
- Datasets 2.5.2
- Tokenizers 0.12.1
