b4bd21083cef9cc6fbc3677cffa57d82

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

• Loss: 4.3176
• Data Size: 1.0
• Epoch Runtime: 26.7447
• Bleu: 3.8473

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	10.7150	0	2.8688	0.0997
No log	1	89	8.8775	0.0078	3.7153	0.1756
No log	2	178	7.7197	0.0156	4.8349	0.3158
No log	3	267	6.9539	0.0312	6.9943	0.2144
No log	4	356	6.3301	0.0625	8.8442	0.3134
No log	5	445	5.7009	0.125	10.7525	0.7706
0.422	6	534	5.0720	0.25	12.7942	1.1431
1.7204	7	623	4.5197	0.5	17.2471	1.6827
4.0173	8.0	712	3.9938	1.0	27.9134	2.7140
3.2802	9.0	801	3.7948	1.0	28.9793	3.1322
2.7743	10.0	890	3.7258	1.0	26.2730	3.1813
2.2574	11.0	979	3.7807	1.0	26.4106	3.4789
1.8053	12.0	1068	3.9539	1.0	27.5233	3.6337
1.4524	13.0	1157	4.1589	1.0	26.4951	3.5387
1.1577	14.0	1246	4.3176	1.0	26.7447	3.8473

Framework versions

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