3a894b5c3df4433e0502b3246be974fc

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-nl] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5722
• Data Size: 1.0
• Epoch Runtime: 51.4148
• Bleu: 6.9552

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	7.3422	0	4.5202	0.0642
No log	1	806	3.8070	0.0078	5.1222	0.6618
No log	2	1612	3.0436	0.0156	5.4709	1.9150
No log	3	2418	2.6079	0.0312	6.4282	3.0784
0.0983	4	3224	2.2950	0.0625	7.9800	3.7184
2.223	5	4030	2.0417	0.125	10.8865	4.5234
1.9409	6	4836	1.8506	0.25	16.6736	5.2555
1.7363	7	5642	1.6873	0.5	27.7954	5.7061
1.5487	8.0	6448	1.5518	1.0	49.7051	6.5226
1.36	9.0	7254	1.5104	1.0	51.1680	6.8433
1.2423	10.0	8060	1.4871	1.0	53.1129	6.9580
1.1256	11.0	8866	1.5099	1.0	51.0642	6.9491
1.0185	12.0	9672	1.5126	1.0	51.0169	7.1589
0.9013	13.0	10478	1.5390	1.0	51.3129	6.9137
0.8255	14.0	11284	1.5722	1.0	51.4148	6.9552

Framework versions

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