3d6a29693a69f8cf09826201d445a79d

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

• Loss: 1.4244
• Data Size: 1.0
• Epoch Runtime: 229.3908
• Bleu: 12.3129

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	5.3762	0	16.1029	2.0943
No log	1	437	3.7703	0.0078	17.9230	8.6365
No log	2	874	2.9843	0.0156	24.6454	14.2601
No log	3	1311	2.4537	0.0312	32.1008	18.7244
No log	4	1748	2.0045	0.0625	39.7906	21.4055
2.2761	5	2185	1.6558	0.125	54.5317	8.9286
1.9204	6	2622	1.5167	0.25	77.2758	9.8104
1.7279	7	3059	1.4234	0.5	126.4218	10.6479
1.5085	8.0	3496	1.3485	1.0	232.0429	11.2266
1.359	9.0	3933	1.3306	1.0	227.8528	11.7087
1.1925	10.0	4370	1.3270	1.0	230.0269	11.9800
1.0712	11.0	4807	1.3286	1.0	235.3832	12.0977
0.9557	12.0	5244	1.3387	1.0	233.7459	12.2954
0.8333	13.0	5681	1.3848	1.0	226.9231	12.2241
0.7383	14.0	6118	1.4244	1.0	229.3908	12.3129

Framework versions

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