bb84e5956fa4843c4c1fa82337692137

This model is a fine-tuned version of google/umt5-xl on the Helsinki-NLP/opus_books [fr-it] dataset. It achieves the following results on the evaluation set:

• Loss: 2.0099
• Data Size: 1.0
• Epoch Runtime: 205.5781
• Bleu: 8.8433

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	5.2927	0	13.5941	2.2924
No log	1	367	4.3096	0.0078	15.5771	5.5344
No log	2	734	3.5125	0.0156	23.7748	10.6850
No log	3	1101	2.9758	0.0312	31.8804	14.5616
No log	4	1468	2.6384	0.0625	40.5064	15.9381
0.1899	5	1835	2.3644	0.125	52.6170	6.5791
2.7147	6	2202	2.1766	0.25	75.6631	7.1049
2.4397	7	2569	2.0553	0.5	117.8008	7.3811
2.1857	8.0	2936	1.9766	1.0	206.8776	8.0779
1.9886	9.0	3303	1.9395	1.0	206.6883	8.4148
1.761	10.0	3670	1.9255	1.0	205.8516	8.6401
1.6096	11.0	4037	1.9284	1.0	203.4900	8.7878
1.5012	12.0	4404	1.9467	1.0	205.5317	8.7877
1.4092	13.0	4771	1.9821	1.0	205.3684	8.8410
1.2512	14.0	5138	2.0099	1.0	205.5781	8.8433

Framework versions

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