opus-mt-en-bkm-10e64

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: 1.4935
• Bleu: 7.795
• Gen Len: 60.8227

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: 64
• eval_batch_size: 64
• 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
No log	1.0	279	2.2834	1.3309	68.9546
3.1609	2.0	558	1.8907	4.0661	60.289
3.1609	3.0	837	1.7445	5.2814	59.5739
1.9255	4.0	1116	1.6564	6.0017	61.6317
1.9255	5.0	1395	1.5978	6.5742	60.6533
1.7347	6.0	1674	1.5572	7.0291	61.3661
1.7347	7.0	1953	1.5287	7.4536	60.8991
1.642	8.0	2232	1.5092	7.6945	60.6611
1.5888	9.0	2511	1.4971	7.7436	60.8317
1.5888	10.0	2790	1.4935	7.795	60.8227

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2