opus-mt-en-ar-evaluated-en-to-ar-4000instances-opus-leaningRate2e-05-batchSize8-11-action-1

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-ar on the opus100 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1717
• Bleu: 26.8232
• Meteor: 0.172
• Gen Len: 12.1288

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 11

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Gen Len
0.7364	0.25	100	0.1731	27.2753	0.1729	12.0887
0.2175	0.5	200	0.1731	27.2055	0.1722	11.5675
0.2193	0.75	300	0.1722	27.3277	0.1798	12.1325
0.2321	1.0	400	0.1750	27.5152	0.1762	11.925
0.1915	1.25	500	0.1690	27.5043	0.1751	11.9038
0.1794	1.5	600	0.1719	26.8607	0.1713	11.8138
0.1741	1.75	700	0.1725	26.974	0.1724	11.8462
0.1732	2.0	800	0.1717	26.8232	0.172	12.1288

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0
• Datasets 2.1.0
• Tokenizers 0.12.1