010993abcb24623d12455dc96095e04b

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

• Loss: 1.5191
• Data Size: 1.0
• Epoch Runtime: 791.5257
• Bleu: 12.8342

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	3.7035	0	65.0916	10.1402
No log	1	3177	2.0743	0.0078	71.5583	18.4289
0.0341	2	6354	1.8378	0.0156	78.0708	20.4880
1.7782	3	9531	1.7200	0.0312	90.2172	20.9207
1.671	4	12708	1.6230	0.0625	112.4153	26.1966
1.5249	5	15885	1.5292	0.125	158.6532	19.3041
1.4223	6	19062	1.4315	0.25	249.1083	12.3889
1.2672	7	22239	1.3396	0.5	428.7141	13.9986
1.1322	8.0	25416	1.2630	1.0	795.9198	14.2121
0.9354	9.0	28593	1.2733	1.0	789.2542	13.3030
0.8045	10.0	31770	1.3261	1.0	785.1027	12.6968
0.6472	11.0	34947	1.4131	1.0	791.0912	12.9002
0.5071	12.0	38124	1.5191	1.0	791.5257	12.8342

Framework versions

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