b41eb43d0bd8cc64c301f7cf431da554

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

• Loss: 2.3322
• Data Size: 1.0
• Epoch Runtime: 80.6379
• Bleu: 6.1766

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	7.9623	0	6.5503	0.0778
No log	1	1286	4.3364	0.0078	7.5239	2.4505
0.0897	2	2572	3.8425	0.0156	8.0332	2.7907
0.1029	3	3858	3.4386	0.0312	10.2682	3.1908
0.1467	4	5144	3.1281	0.0625	11.4994	3.8348
2.9633	5	6430	2.8771	0.125	16.6812	4.3154
2.7139	6	7716	2.6416	0.25	25.9702	4.9043
2.4351	7	9002	2.4623	0.5	43.7680	5.3610
2.2384	8.0	10288	2.2743	1.0	82.8114	5.8681
1.9918	9.0	11574	2.2136	1.0	80.5981	6.0774
1.7637	10.0	12860	2.1991	1.0	83.5776	6.2049
1.58	11.0	14146	2.2155	1.0	81.6651	6.2671
1.4801	12.0	15432	2.2387	1.0	81.3572	6.2061
1.2937	13.0	16718	2.2727	1.0	81.3294	6.2218
1.1756	14.0	18004	2.3322	1.0	80.6379	6.1766

Framework versions

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