cb950fd819ea576ea12b161bd8b5769f

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 [de-pt] dataset. It achieves the following results on the evaluation set:

• Loss: 3.1192
• Data Size: 1.0
• Epoch Runtime: 3.3357
• Bleu: 6.4752

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	8.1198	0	1.3876	0.0371
No log	1	27	6.3264	0.0078	1.0782	0.2695
No log	2	54	5.4045	0.0156	1.4951	1.4121
No log	3	81	5.1551	0.0312	1.9245	1.7348
No log	4	108	4.8512	0.0625	2.8902	2.1320
No log	5	135	4.4644	0.125	2.6907	2.7545
No log	6	162	4.0748	0.25	2.9858	3.3529
No log	7	189	3.5504	0.5	3.2476	3.9064
0.8535	8.0	216	3.1393	1.0	4.6170	4.8284
0.8535	9.0	243	3.0497	1.0	4.0670	5.3345
2.5255	10.0	270	2.9773	1.0	4.3793	5.4799
2.5255	11.0	297	2.9997	1.0	3.4594	6.1256
1.6947	12.0	324	3.0297	1.0	3.1656	6.2854
1.1819	13.0	351	3.0842	1.0	3.4267	6.4015
1.1819	14.0	378	3.1192	1.0	3.3357	6.4752

Framework versions

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