c27ea2b104c6d80756e0f294eacb9ff2

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

• Loss: 2.4926
• Data Size: 1.0
• Epoch Runtime: 106.1337
• Bleu: 6.2550

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	5.5511	0	9.2797	0.7034
No log	1	419	2.8508	0.0078	10.4928	3.9162
No log	2	838	2.6648	0.0156	11.6209	4.3573
0.0799	3	1257	2.5399	0.0312	14.6387	5.1101
0.0799	4	1676	2.4502	0.0625	17.9705	5.0032
0.1542	5	2095	2.3558	0.125	23.9706	5.5299
0.3054	6	2514	2.2597	0.25	35.2747	5.9908
2.003	7	2933	2.1603	0.5	59.9474	6.4589
1.7352	8.0	3352	2.0958	1.0	109.3495	7.3348
1.3838	9.0	3771	2.1351	1.0	106.3738	6.2668
1.0913	10.0	4190	2.2452	1.0	106.7843	10.9967
0.7823	11.0	4609	2.3457	1.0	106.3178	6.7105
0.5932	12.0	5028	2.4926	1.0	106.1337	6.2550

Framework versions

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