ca5e7b77ca12657bed3d8759d088eddb

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

• Loss: 3.0078
• Data Size: 1.0
• Epoch Runtime: 184.2464
• Bleu: 6.5777

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	8.4635	0	15.5562	0.3667
No log	1	721	5.0839	0.0078	16.9998	2.4507
No log	2	1442	4.6093	0.0156	19.0640	3.2587
0.0935	3	2163	3.4956	0.0312	23.0000	3.1308
0.2617	4	2884	2.8635	0.0625	28.8507	3.7224
2.8273	5	3605	2.6575	0.125	40.0532	6.2440
2.5513	6	4326	2.5166	0.25	60.2866	6.6559
10.0686	7	5047	6.9851	0.5	101.7979	0.0052
2.2053	8.0	5768	2.8342	1.0	186.0315	4.6538
1.8546	9.0	6489	2.3698	1.0	184.6548	7.4195
1.6162	10.0	7210	2.4484	1.0	184.2103	6.6773
1.3109	11.0	7931	2.5860	1.0	184.1730	7.5129
1.0946	12.0	8652	2.8058	1.0	184.8414	6.3458
0.8617	13.0	9373	3.0078	1.0	184.2464	6.5777

Framework versions

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