84af8aa70e34fba0bfd0a2787d37e01c

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 [fr-sv] dataset. It achieves the following results on the evaluation set:

• Loss: 2.3855
• Data Size: 1.0
• Epoch Runtime: 6.2713
• Bleu: 2.9322

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	7.7749	0	0.9722	0.0272
No log	1	75	6.6330	0.0078	1.6036	0.0813
No log	2	150	5.3214	0.0156	1.7175	0.2307
No log	3	225	4.4848	0.0312	2.1804	0.2994
No log	4	300	3.8709	0.0625	2.3024	0.5750
No log	5	375	3.3103	0.125	2.6586	0.9877
No log	6	450	2.8784	0.25	3.2501	1.3959
No log	7	525	2.5247	0.5	4.6764	2.0154
2.1751	8.0	600	2.2620	1.0	7.6483	2.5002
1.805	9.0	675	2.2159	1.0	5.9042	2.8034
1.5341	10.0	750	2.2306	1.0	6.1104	2.8231
1.2859	11.0	825	2.2669	1.0	6.1395	2.9006
1.0959	12.0	900	2.3122	1.0	6.4357	2.9696
0.899	13.0	975	2.3855	1.0	6.2713	2.9322

Framework versions

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