af062ae77d27ffbd654d8da5454b105a

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

• Loss: 2.3050
• Data Size: 1.0
• Epoch Runtime: 57.7549
• Bleu: 8.2352

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	5.9983	0	3.7660	1.8370
No log	1	86	5.4381	0.0078	4.4692	2.7273
No log	2	172	4.6831	0.0156	11.9946	5.1253
No log	3	258	4.2062	0.0312	21.0001	7.5684
No log	4	344	3.5904	0.0625	29.8068	10.2243
0.2804	5	430	3.1909	0.125	23.1855	12.6905
0.9359	6	516	2.7979	0.25	27.6699	13.6143
1.1113	7	602	2.3778	0.5	44.9102	6.3871
1.5253	8.0	688	2.1438	1.0	58.9131	7.4955
2.195	9.0	774	2.1024	1.0	57.6172	7.6946
1.9853	10.0	860	2.0805	1.0	55.6679	8.1310
1.7598	11.0	946	2.0803	1.0	61.9601	8.1887
1.5109	12.0	1032	2.1305	1.0	56.6188	8.1766
1.3717	13.0	1118	2.1599	1.0	55.6418	8.2854
1.2224	14.0	1204	2.2332	1.0	54.3970	8.2995
1.0726	15.0	1290	2.3050	1.0	57.7549	8.2352

Framework versions

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