df2ec487db69725dff7faeebdec684fd

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

• Loss: 2.7366
• Data Size: 1.0
• Epoch Runtime: 185.3542
• Bleu: 5.3451

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	6.8083	0	15.4169	0.3904
No log	1	721	3.2524	0.0078	17.0675	3.2322
No log	2	1442	3.0505	0.0156	20.1139	3.7278
0.0613	3	2163	2.9022	0.0312	22.9839	4.3023
0.2014	4	2884	2.7807	0.0625	28.3371	4.9489
2.7545	5	3605	2.6759	0.125	39.7855	5.6347
2.5588	6	4326	2.5606	0.25	59.6311	6.3126
2.4027	7	5047	2.4370	0.5	102.1237	7.7033
2.1059	8.0	5768	2.3380	1.0	184.7286	6.2981
1.769	9.0	6489	2.3692	1.0	185.0756	5.9444
1.5192	10.0	7210	2.4797	1.0	184.1943	5.6407
1.2383	11.0	7931	2.5675	1.0	183.9788	5.3725
1.0197	12.0	8652	2.7366	1.0	185.3542	5.3451

Framework versions

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