de1dffef4207f6fec938d98242b37c7a

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

• Loss: 1.6406
• Data Size: 1.0
• Epoch Runtime: 794.3154
• Bleu: 15.1494

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	6.0084	0	65.8213	1.2509
No log	1	3177	3.2378	0.0078	72.3778	9.6568
0.0542	2	6354	2.2685	0.0156	77.8192	11.8766
9.9744	3	9531	7.5505	0.0312	89.6887	0.0016
2.1183	4	12708	1.8953	0.0625	113.7882	19.2824
1.7258	5	15885	1.6635	0.125	159.3793	21.9236
1.531	6	19062	1.4965	0.25	247.8980	17.6415
1.3825	7	22239	1.4083	0.5	428.8253	14.0973
1.2144	8.0	25416	1.3038	1.0	790.6624	15.7468
1.0109	9.0	28593	1.3003	1.0	794.5531	16.7103
0.8794	10.0	31770	1.3401	1.0	795.7410	17.3701
0.7221	11.0	34947	1.4224	1.0	791.2622	14.7897
0.5718	12.0	38124	1.5171	1.0	797.5035	15.8861
0.4526	13.0	41301	1.6406	1.0	794.3154	15.1494

Framework versions

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