tatoeba-vi-tok

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

• Loss: 0.6910
• Bleu: 46.4356

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: 64
• eval_batch_size: 64
• 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: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
1.2609	1.0	1167	1.0286	33.8552
0.98	2.0	2334	0.8700	39.5255
0.8347	3.0	3501	0.8036	41.4872
0.7753	4.0	4668	0.7661	43.3025
0.7272	5.0	5835	0.7389	44.2770
0.6876	6.0	7002	0.7254	45.1330
0.659	7.0	8169	0.7153	45.4847
0.6383	8.0	9336	0.7067	45.4353
0.6122	9.0	10503	0.7015	45.7612
0.5998	10.0	11670	0.6979	46.0799
0.5836	11.0	12837	0.6960	46.1173
0.5732	12.0	14004	0.6928	46.2538
0.5607	13.0	15171	0.6919	46.3801
0.553	14.0	16338	0.6911	46.3629
0.5518	15.0	17505	0.6910	46.4356

Framework versions

• Transformers 4.52.4
• Pytorch 2.7.1+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1