fd96cd898bba8dbe3bfa5a5b7b1a56c3

This model is a fine-tuned version of google/umt5-xl on the Helsinki-NLP/opus_books [en-no] dataset. It achieves the following results on the evaluation set:

• Loss: 1.8440
• Data Size: 1.0
• Epoch Runtime: 56.4380
• Bleu: 11.8935

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	5.7519	0	3.8324	2.6667
No log	1	87	5.3766	0.0078	4.5080	3.6279
No log	2	174	4.5978	0.0156	9.4975	6.6975
No log	3	261	3.9958	0.0312	14.2208	9.8415
No log	4	348	3.2172	0.0625	21.8229	15.4143
0.2201	5	435	2.8475	0.125	23.4467	17.9165
0.8228	6	522	2.3381	0.25	28.2647	20.7722
0.9515	7	609	2.0036	0.5	42.3805	9.1580
1.3167	8.0	696	1.7552	1.0	62.3441	10.7515
1.7823	9.0	783	1.7107	1.0	59.5657	11.3962
1.5709	10.0	870	1.6947	1.0	56.8749	11.6802
1.3463	11.0	957	1.7127	1.0	61.4083	11.8069
1.1895	12.0	1044	1.7487	1.0	54.1224	11.9382
1.0372	13.0	1131	1.7962	1.0	58.4049	11.7635
0.9109	14.0	1218	1.8440	1.0	56.4380	11.8935

Framework versions

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