opus-mt-en-bkm-Final-60

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

• Loss: 1.5584
• Bleu: 9.354
• Gen Len: 40.6029

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: 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
3.4876	1.0	601	2.2031	2.3601	50.4173
2.2475	2.0	1202	1.9329	4.8587	40.8697
2.0162	3.0	1803	1.7959	5.9413	39.0495
1.8926	4.0	2404	1.7144	6.9033	40.8797
1.71	5.0	3005	1.6537	7.7651	40.4224
1.6501	6.0	3606	1.6161	8.441	41.3464
1.6053	7.0	4207	1.5869	8.812	40.534
1.554	8.0	4808	1.5725	9.2092	40.4813
1.5409	9.0	5409	1.5608	9.3966	40.9083
1.5049	10.0	6010	1.5584	9.354	40.6029

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2