opus-mt-en-es-finetuned-es-to-maq

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-es on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.7414
• Bleu: 6.663
• Gen Len: 94.437

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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	199	2.3505	2.7386	127.0327
No log	2.0	398	2.0862	4.4403	97.3489
2.643	3.0	597	1.9576	5.2104	98.7116
2.643	4.0	796	1.8831	5.4016	98.4962
2.643	5.0	995	1.8320	5.6026	96.1826
1.9678	6.0	1194	1.7944	6.374	95.1398
1.9678	7.0	1393	1.7726	6.514	94.83
1.8281	8.0	1592	1.7551	6.7802	95.194
1.8281	9.0	1791	1.7451	6.7625	94.2091
1.8281	10.0	1990	1.7414	6.663	94.437

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3