78e45edfab2e76d0d369e52b61e2100a

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

• Loss: 0.4578
• Data Size: 1.0
• Epoch Runtime: 27.7194
• Bleu: 13.1403

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	7.4763	0	2.5179	0.0572
No log	1	434	2.5525	0.0078	3.9581	0.0049
No log	2	868	1.9816	0.0156	3.2139	0.0009
No log	3	1302	1.6718	0.0312	3.8519	0.0244
No log	4	1736	1.3498	0.0625	4.5576	0.3289
0.0714	5	2170	1.0614	0.125	6.0921	3.2350
0.9337	6	2604	0.7786	0.25	9.1924	5.3485
0.6619	7	3038	0.5972	0.5	15.4139	7.9549
0.5168	8.0	3472	0.4797	1.0	28.9889	10.9862
0.407	9.0	3906	0.4377	1.0	28.0562	11.5697
0.3527	10.0	4340	0.4187	1.0	27.0718	12.5934
0.3028	11.0	4774	0.4128	1.0	27.8363	12.7454
0.2623	12.0	5208	0.4204	1.0	27.2489	13.2087
0.2243	13.0	5642	0.4303	1.0	27.9318	13.0685
0.201	14.0	6076	0.4414	1.0	28.1184	13.3063
0.1808	15.0	6510	0.4578	1.0	27.7194	13.1403

Framework versions

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