helsinki_new_ver5.1

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-ZH on the sarahwei/Taiwanese-Minnan-Example-Sentences dataset. It achieves the following results on the evaluation set:

• Loss: 0.1774
• Bleu: 31.4017
• Ter: 49.3755

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: 8e-06
• train_batch_size: 8
• eval_batch_size: 16
• 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
• lr_scheduler_warmup_steps: 1000
• training_steps: 14000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Ter
0.1528	0.5656	1000	0.1781	30.0868	50.0852
0.1622	1.1312	2000	0.1786	30.0483	49.7162
0.1481	1.6968	3000	0.1779	30.6107	49.3613
0.1369	2.2624	4000	0.1773	30.3317	49.4039
0.1434	2.8281	5000	0.1771	31.3488	48.7653
0.1283	3.3937	6000	0.1768	31.9413	48.7085
0.1295	3.9593	7000	0.1770	31.8531	49.2052
0.1256	4.5249	8000	0.1769	31.6029	49.2052
0.123	5.0905	9000	0.1774	31.4017	49.3755

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1