5a851fe9daa6ecf99f3cc167b85d96bd

This model is a fine-tuned version of google/long-t5-tglobal-xl on the Helsinki-NLP/opus_books [fr-pt] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5514
• Data Size: 1.0
• Epoch Runtime: 32.8648
• Bleu: 5.3455

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	3.1616	0	2.0234	0.3357
No log	1	31	2.6356	0.0078	2.7049	0.7545
No log	2	62	2.2805	0.0156	6.3397	1.5095
No log	3	93	2.1553	0.0312	10.1454	2.0191
No log	4	124	2.0390	0.0625	14.2696	2.0133
No log	5	155	1.9438	0.125	20.4120	2.3003
No log	6	186	1.8920	0.25	22.5184	2.6820
0.4078	7	217	1.7802	0.5	24.7564	2.8725
0.4078	8.0	248	1.6672	1.0	36.3107	3.7166
1.4794	9.0	279	1.6093	1.0	37.5052	3.9596
1.8538	10.0	310	1.5589	1.0	28.9424	4.2740
1.8538	11.0	341	1.5322	1.0	30.1822	4.5934
1.6539	12.0	372	1.5296	1.0	33.4240	4.7288
1.4668	13.0	403	1.5187	1.0	35.0200	4.8495
1.4668	14.0	434	1.5087	1.0	29.9666	4.9276
1.3222	15.0	465	1.5174	1.0	32.5331	5.0111
1.3222	16.0	496	1.5145	1.0	36.1044	5.1490
1.1801	17.0	527	1.5198	1.0	29.5243	5.3126
1.0676	18.0	558	1.5514	1.0	32.8648	5.3455

Framework versions

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