opus-mt-en-bkm-Final-80

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: 1.4265
• Bleu: 10.3853
• Gen Len: 49.8443

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
3.4583	1.0	781	2.0553	3.6067	54.2885
1.9985	2.0	1562	1.8020	5.5427	48.141
1.8553	3.0	2343	1.6664	6.9654	48.4025
1.6865	4.0	3124	1.5875	7.949	47.8417
1.6207	5.0	3905	1.5302	8.6008	48.1286
1.5265	6.0	4686	1.4857	9.3569	49.3219
1.4961	7.0	5467	1.4598	9.7246	50.4949
1.4409	8.0	6248	1.4413	10.0815	49.4964
1.4101	9.0	7029	1.4288	10.3204	49.9725
1.3954	10.0	7810	1.4265	10.3853	49.8443

Framework versions

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