a0c101c99b154177b388bc9384eb5156

This model is a fine-tuned version of facebook/mbart-large-cc25 on the Helsinki-NLP/opus_books [en-no] dataset. It achieves the following results on the evaluation set:

• Loss: 4.3388
• Data Size: 1.0
• Epoch Runtime: 26.1810
• Bleu: 7.0422

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	13.7917	0	2.8516	0.0560
No log	1	87	10.8231	0.0078	3.3023	0.2957
No log	2	174	8.7231	0.0156	4.6169	0.2979
No log	3	261	7.8121	0.0312	6.0099	0.2345
No log	4	348	7.1728	0.0625	8.2905	0.3959
0.3544	5	435	6.2155	0.125	10.3024	0.7289
1.533	6	522	5.4494	0.25	12.9780	1.2815
1.8628	7	609	4.8213	0.5	16.7492	2.4923
2.6213	8.0	696	12.8008	1.0	27.6169	0.0000
8.3875	9.0	783	5.4283	1.0	27.1007	2.0589
4.1986	10.0	870	3.8781	1.0	25.9280	3.1750
3.2376	11.0	957	3.5052	1.0	26.2485	4.5206
2.6636	12.0	1044	3.5104	1.0	27.8327	5.6326
2.0037	13.0	1131	3.4747	1.0	25.7378	7.2927
1.5914	14.0	1218	3.6475	1.0	26.0810	6.0995
1.1724	15.0	1305	3.8761	1.0	26.8978	7.2212
0.877	16.0	1392	4.0529	1.0	27.2078	6.4394
0.6739	17.0	1479	4.3388	1.0	26.1810	7.0422

Framework versions

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