4053ccba4766f3012b943e0f2a6ba1cc

This model is a fine-tuned version of google/mt5-base on the Helsinki-NLP/opus_books [en-no] dataset. It achieves the following results on the evaluation set:

• Loss: 2.1552
• Data Size: 1.0
• Epoch Runtime: 21.0298
• Bleu: 7.9534

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	16.3903	0	2.4919	0.0012
No log	1	87	16.2652	0.0078	3.0402	0.0020
No log	2	174	15.7682	0.0156	3.1024	0.0020
No log	3	261	15.1807	0.0312	3.5727	0.0012
No log	4	348	16.0009	0.0625	4.7808	0.0012
0.6847	5	435	13.4265	0.125	5.6736	0.0012
3.4724	6	522	9.1619	0.25	7.9955	0.0045
4.0559	7	609	7.2011	0.5	12.7481	0.0064
5.0514	8.0	696	5.6457	1.0	21.7061	0.0198
5.5651	9.0	783	3.4591	1.0	20.4395	1.1073
3.9865	10.0	870	2.7854	1.0	21.3194	2.8684
3.436	11.0	957	2.5479	1.0	21.7505	4.3926
3.2065	12.0	1044	2.4041	1.0	22.9681	5.1570
2.9505	13.0	1131	2.3554	1.0	21.5941	5.3145
2.8146	14.0	1218	2.3420	1.0	20.0549	5.6208
2.6455	15.0	1305	2.2882	1.0	19.9195	5.8671
2.5617	16.0	1392	2.2256	1.0	20.2517	6.4782
2.4692	17.0	1479	2.2113	1.0	20.2933	6.7027
2.3672	18.0	1566	2.1917	1.0	21.1114	6.9196
2.3172	19.0	1653	2.1784	1.0	20.8452	6.9665
2.2092	20.0	1740	2.1631	1.0	21.1128	6.9706
2.1428	21.0	1827	2.1567	1.0	20.8195	7.2431
2.109	22.0	1914	2.1541	1.0	21.3080	7.2208
2.0272	23.0	2001	2.1516	1.0	22.4732	7.4292
1.9716	24.0	2088	2.1522	1.0	23.0559	7.4031
1.9407	25.0	2175	2.1517	1.0	21.1709	7.6519
1.8952	26.0	2262	2.1393	1.0	21.6689	7.7828
1.8579	27.0	2349	2.1565	1.0	22.4174	7.7715
1.7953	28.0	2436	2.1438	1.0	22.9117	7.7857
1.7689	29.0	2523	2.1526	1.0	24.4328	7.9907
1.7437	30.0	2610	2.1552	1.0	21.0298	7.9534

Framework versions

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