738dc1200d147ffa176ee7ca0a39ea37

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

• Loss: 1.9426
• Data Size: 1.0
• Epoch Runtime: 37.5210
• Bleu: 10.6772

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	5.9683	0	1.6631	1.5941
No log	1	27	6.3306	0.0078	2.1335	1.1276
No log	2	54	6.7093	0.0156	6.4058	0.9344
No log	3	81	6.7105	0.0312	14.9482	0.8790
No log	4	108	6.5118	0.0625	20.4633	0.8710
No log	5	135	5.5037	0.125	27.4676	1.8894
No log	6	162	3.7515	0.25	23.8936	9.0043
No log	7	189	2.8397	0.5	24.0969	16.0224
1.0992	8.0	216	2.2077	1.0	36.7295	21.2278
1.0992	9.0	243	2.0107	1.0	37.3103	8.8538
2.5144	10.0	270	1.9231	1.0	28.3653	9.2239
2.5144	11.0	297	1.8696	1.0	32.5803	12.2707
1.8982	12.0	324	1.8491	1.0	35.2127	10.4182
1.5756	13.0	351	1.8521	1.0	40.4682	9.9941
1.5756	14.0	378	1.8920	1.0	29.3580	10.1127
1.2949	15.0	405	1.8897	1.0	34.5300	10.9814
1.2949	16.0	432	1.9426	1.0	37.5210	10.6772

Framework versions

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