ef0fc60307b5ae3907cd626160675744

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

• Loss: 2.2386
• Data Size: 1.0
• Epoch Runtime: 45.3162
• Bleu: 3.2749

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	7.4356	0	3.0217	0.0092
No log	1	70	5.7502	0.0078	3.8717	0.0183
No log	2	140	4.5681	0.0156	11.2349	0.0596
No log	3	210	4.1004	0.0312	16.1280	0.1179
No log	4	280	3.9024	0.0625	22.6958	0.1069
No log	5	350	3.3192	0.125	29.0845	0.1753
No log	6	420	2.6971	0.25	29.2950	0.3854
0.6003	7	490	2.4368	0.5	36.5773	1.8668
2.7265	8.0	560	2.2457	1.0	59.0803	2.1418
2.3962	9.0	630	2.1955	1.0	43.3304	2.4797
2.1264	10.0	700	2.1957	1.0	47.0717	2.3341
1.9469	11.0	770	2.1982	1.0	45.4346	2.7379
1.8626	12.0	840	2.2279	1.0	52.1217	3.0321
1.6317	13.0	910	2.2386	1.0	45.3162	3.2749

Framework versions

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