088a54e139902c61c842f2b04fb533d1

This model is a fine-tuned version of facebook/mbart-large-50 on the Helsinki-NLP/opus_books [de-pt] dataset. It achieves the following results on the evaluation set:

• Loss: 6.7054
• Data Size: 1.0
• Epoch Runtime: 15.2267
• Bleu: 0.0115

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	8.2307	0	1.1976	0.5322
No log	1	27	6.2402	0.0078	1.8401	2.7954
No log	2	54	6.7653	0.0156	2.3078	1.8057
No log	3	81	5.5517	0.0312	3.1797	1.8714
No log	4	108	5.1352	0.0625	5.0648	2.5835
No log	5	135	4.7036	0.125	6.6002	3.9773
No log	6	162	4.2887	0.25	8.1320	5.1322
No log	7	189	3.4370	0.5	9.3524	6.4658
0.8516	8.0	216	2.7302	1.0	12.5395	6.8434
0.8516	9.0	243	3.5286	1.0	12.5885	5.4000
3.9383	10.0	270	12.7432	1.0	13.3199	0.0
3.9383	11.0	297	7.9491	1.0	14.3823	0.0271
12.3579	12.0	324	6.7054	1.0	15.2267	0.0115

Framework versions

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