c86ce1877086071c562c73a058a18346

This model is a fine-tuned version of google/mt5-large on the Helsinki-NLP/opus_books [fr-it] dataset. It achieves the following results on the evaluation set:

• Loss: 1.8298
• Data Size: 1.0
• Epoch Runtime: 151.8425
• Bleu: 7.2635

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	24.6284	0	12.1172	0.0072
No log	1	367	28.2720	0.0078	14.2842	0.0071
No log	2	734	28.6837	0.0156	15.9569	0.0081
No log	3	1101	22.3060	0.0312	19.5309	0.0102
No log	4	1468	15.8043	0.0625	25.7076	0.0192
1.1179	5	1835	8.5722	0.125	34.9587	0.0300
7.2271	6	2202	2.8501	0.25	50.8760	0.3819
2.8449	7	2569	2.1381	0.5	85.7075	4.8749
2.4527	8.0	2936	2.0038	1.0	155.4511	5.5158
2.2795	9.0	3303	1.9413	1.0	152.9994	5.7755
2.1449	10.0	3670	1.9087	1.0	152.7194	6.1533
2.0455	11.0	4037	1.8749	1.0	151.8434	6.5866
1.9907	12.0	4404	1.8610	1.0	154.1404	6.8924
1.9354	13.0	4771	1.8465	1.0	152.7489	6.9367
1.8288	14.0	5138	1.8354	1.0	152.6904	7.1516
1.7899	15.0	5505	1.8206	1.0	153.4511	7.1190
1.7398	16.0	5872	1.8251	1.0	154.0484	7.1627
1.6879	17.0	6239	1.8251	1.0	153.4658	7.2111
1.6392	18.0	6606	1.8287	1.0	151.2117	7.1964
1.6077	19.0	6973	1.8298	1.0	151.8425	7.2635

Framework versions

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