opus-mt-ar-en-finetuned-src-to-trg-testing

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

• Loss: 3.3973
• Bleu: 0.1939
• Gen Len: 37.6364

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: 3
• mixed_precision_training: Apex, opt level O1

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	5	3.4353	0.1994	36.6364
No log	2.0	10	3.4015	0.1994	36.0909
No log	3.0	15	3.3973	0.1939	37.6364

Framework versions

• Transformers 4.14.1
• Pytorch 1.5.0
• Datasets 1.17.0
• Tokenizers 0.10.3