0becee0a296ac7ba1544f08f79fe24eb

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

• Loss: 1.7168
• Data Size: 1.0
• Epoch Runtime: 362.9564
• Bleu: 13.6791

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	6.4165	0	30.5876	0.5729
No log	1	1407	3.0088	0.0078	33.1366	3.1890
No log	2	2814	2.4049	0.0156	36.1601	5.5787
0.0699	3	4221	2.1396	0.0312	42.9476	7.2070
2.1567	4	5628	1.9503	0.0625	53.9853	8.6811
1.9505	5	7035	1.7966	0.125	74.7479	9.6196
1.7299	6	8442	1.6698	0.25	114.7956	11.0295
1.6018	7	9849	1.5431	0.5	196.3446	12.2364
1.3821	8.0	11256	1.4519	1.0	359.3257	13.1925
1.1669	9.0	12663	1.4362	1.0	358.1526	13.6128
1.0546	10.0	14070	1.4689	1.0	360.4168	13.7347
0.8565	11.0	15477	1.5358	1.0	378.2215	13.6197
0.7167	12.0	16884	1.5849	1.0	362.5242	13.6894
0.6186	13.0	18291	1.7168	1.0	362.9564	13.6791

Framework versions

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