opus-mt-ar-en-finetuned-ar-to-en-13

This model is a fine-tuned version of Helsinki-NLP/opus-mt-ar-en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.7932
• Bleu: 0.2658
• Meteor: 0.5618
• Gen Len: 35.0292

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 20
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Gen Len
No log	1.0	360	2.0417	0.176	0.507	46.7987
2.2112	2.0	720	1.9312	0.2055	0.5315	43.4091
1.7576	3.0	1080	1.8670	0.2372	0.5443	38.0487
1.7576	4.0	1440	1.8335	0.2341	0.5487	38.7208
1.4873	5.0	1800	1.8100	0.2363	0.5477	38.0292
1.2916	6.0	2160	1.7920	0.2577	0.5521	36.0065
1.1372	7.0	2520	1.7834	0.2618	0.5559	35.3506
1.1372	8.0	2880	1.7881	0.2515	0.5589	36.9903
1.0072	9.0	3240	1.7868	0.265	0.5614	35.6981
0.9407	10.0	3600	1.7922	0.2704	0.5642	35.3701
0.9407	11.0	3960	1.7932	0.2658	0.5618	35.0292

Framework versions

• Transformers 4.50.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1