e39e5b5d1d1a9fef750ca4ffc62fa637

This model is a fine-tuned version of Helsinki-NLP/opus-mt-tc-bible-big-deu_eng_fra_por_spa-mul on the Helsinki-NLP/opus_books [es-fi] dataset. It achieves the following results on the evaluation set:

• Loss: 2.4591
• Data Size: 1.0
• Epoch Runtime: 7.7919
• Bleu: 2.0670

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	8.2015	0	0.9712	0.0322
No log	1	83	6.6433	0.0078	1.3389	0.1368
No log	2	166	5.6313	0.0156	1.4539	0.2131
No log	3	249	4.7119	0.0312	1.6496	0.2628
0.1697	4	332	3.9187	0.0625	1.8090	0.5802
0.1697	5	415	3.3701	0.125	2.2723	0.9817
0.1697	6	498	2.9063	0.25	3.4089	1.1561
0.549	7	581	2.5450	0.5	4.9824	1.8603
2.2673	8.0	664	2.2748	1.0	8.3775	2.1352
1.9608	9.0	747	2.2037	1.0	6.4700	2.3766
1.5251	10.0	830	2.2133	1.0	6.8663	2.1368
1.2932	11.0	913	2.2946	1.0	7.1879	2.3002
1.1041	12.0	996	2.3549	1.0	7.5619	2.1326
0.8786	13.0	1079	2.4591	1.0	7.7919	2.0670

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1