ad7cb7d81ea23144dc3870bc1ca6c73a

This model is a fine-tuned version of facebook/mbart-large-50-many-to-one-mmt on the Helsinki-NLP/opus_books [en-no] dataset. It achieves the following results on the evaluation set:

• Loss: 3.8958
• Data Size: 1.0
• Epoch Runtime: 25.5405
• Bleu: 12.2813

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	7.8297	0	2.4313	0.3365
No log	1	87	7.2703	0.0078	3.3232	0.3016
No log	2	174	6.7399	0.0156	4.4234	0.5142
No log	3	261	6.3615	0.0312	5.5580	0.5806
No log	4	348	5.8594	0.0625	7.2170	0.8034
0.2597	5	435	5.3109	0.125	9.2783	1.4105
1.2417	6	522	4.6317	0.25	11.9091	2.3443
1.5152	7	609	4.0807	0.5	15.8098	4.7346
2.1386	8.0	696	3.5947	1.0	27.3172	6.4501
2.8276	9.0	783	3.4030	1.0	26.5607	12.2341
2.2884	10.0	870	3.3712	1.0	25.2790	13.4284
1.803	11.0	957	3.4962	1.0	25.7365	14.3027
1.4159	12.0	1044	3.6205	1.0	27.2240	11.6812
1.0557	13.0	1131	3.7585	1.0	25.4835	14.7968
0.8476	14.0	1218	3.8958	1.0	25.5405	12.2813

Framework versions

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