34c403562c40d3d0e8ca2963ea1949cc

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

• Loss: 1.8597
• Data Size: 1.0
• Epoch Runtime: 205.6935
• Bleu: 11.1760

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	5.6370	0	14.0258	1.9325
No log	1	390	4.1417	0.0078	15.7780	6.8556
No log	2	780	3.6315	0.0156	22.8919	9.6346
No log	3	1170	3.0688	0.0312	31.5033	13.1035
No log	4	1560	2.5825	0.0625	35.4414	15.7837
0.2014	5	1950	2.2403	0.125	51.9248	7.6039
0.3869	6	2340	2.0152	0.25	70.3411	8.5262
2.2406	7	2730	1.9107	0.5	114.2732	9.3904
2.023	8.0	3120	1.8053	1.0	208.0181	10.1580
1.7625	9.0	3510	1.7579	1.0	203.5810	10.6035
1.5747	10.0	3900	1.7482	1.0	201.0889	10.8837
1.4453	11.0	4290	1.7580	1.0	204.3553	11.1109
1.2926	12.0	4680	1.7783	1.0	200.3883	11.1654
1.1655	13.0	5070	1.8155	1.0	203.1205	11.1456
1.0504	14.0	5460	1.8597	1.0	205.6935	11.1760

Framework versions

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