fccfc9bdf1d4cd1c0a140ff7ecab1ee0

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

• Loss: 1.6262
• Data Size: 1.0
• Epoch Runtime: 60.3813
• Bleu: 6.5933

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	7.1683	0	4.9779	0.0449
No log	1	966	3.7093	0.0078	5.6907	1.2297
No log	2	1932	2.9808	0.0156	6.6172	2.2061
0.0765	3	2898	2.5623	0.0312	7.4455	3.3169
2.49	4	3864	2.2690	0.0625	8.9966	3.9312
2.2014	5	4830	2.0488	0.125	12.3951	4.5686
1.9877	6	5796	1.8650	0.25	19.2530	5.1543
1.7341	7	6762	1.7117	0.5	32.5290	5.8865
1.5436	8.0	7728	1.5788	1.0	61.3965	6.3768
1.3898	9.0	8694	1.5300	1.0	59.3007	6.5862
1.267	10.0	9660	1.5224	1.0	58.2819	6.8568
1.1135	11.0	10626	1.5239	1.0	58.7258	6.7260
1.02	12.0	11592	1.5512	1.0	59.0393	6.6845
0.9199	13.0	12558	1.5852	1.0	59.7723	6.7647
0.8566	14.0	13524	1.6262	1.0	60.3813	6.5933

Framework versions

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