En-Ts

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

Before training:

• Loss: 3.17
• Bleu: 14.513

After Training

• Loss: 1.3320
• Bleu: 36.7687

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: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
1.7082	1.0	5929	1.6902	32.1311
1.4606	2.0	11858	1.4996	34.1129
1.3182	3.0	17787	1.4107	35.7428
1.2543	4.0	23716	1.3631	36.2009
1.2116	5.0	29645	1.3389	36.5876
1.1723	6.0	35574	1.3320	36.7481

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1