dac4660cff700b7f8fdea3c93d043200

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-no] dataset. It achieves the following results on the evaluation set:

• Loss: 2.2548
• Data Size: 1.0
• Epoch Runtime: 8.1863
• Bleu: 4.3282

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	7.7712	0	1.0351	0.0166
No log	1	89	6.4004	0.0078	1.8103	0.0756
No log	2	178	4.9687	0.0156	1.7625	1.1373
No log	3	267	3.9881	0.0312	2.2696	0.7548
No log	4	356	3.3210	0.0625	3.0449	1.3163
No log	5	445	2.8553	0.125	3.6183	1.9417
0.2422	6	534	2.5131	0.25	4.5102	2.9639
0.8539	7	623	2.2323	0.5	5.7452	3.8006
1.9116	8.0	712	2.0048	1.0	8.3677	4.2909
1.5123	9.0	801	1.9542	1.0	7.6888	4.3819
1.2081	10.0	890	1.9957	1.0	7.6131	4.4627
0.9951	11.0	979	2.0654	1.0	7.5701	4.3196
0.8106	12.0	1068	2.1442	1.0	8.0270	4.3392
0.6811	13.0	1157	2.2548	1.0	8.1863	4.3282

Framework versions

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