jojajovai-opus-mt-tc-bible-big-spa-grn

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

• Loss: 1.8078
• Bleu: 26.7899
• Chrf: 51.656
• Combined: 39.2229
• Gen Len: 36.4961

Model description

More information needed

Intended uses & limitations

Translation Spanish-Guarani...

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
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Bleu Gen Len	Chrf	Chrf Gen Len	Combined
No log	1.0	316	2.0189	24.4896	37.1848	48.0852	37.1848	36.2874
1.9749	2.0	632	1.8937	25.536	36.1965	49.6548	36.1965	37.5954
1.9749	3.0	948	1.8614	26.0179	36.5	50.2907	36.5	38.1543

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1