83c702494a2c8fa78dd38c1b100bc0f5

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

• Loss: 0.9608
• Data Size: 1.0
• Epoch Runtime: 35.9424
• Bleu: 13.9213

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	2.3633	0	2.2425	2.9582
No log	1	35	2.0983	0.0078	2.9833	3.4714
No log	2	70	1.7876	0.0156	6.1930	5.8459
No log	3	105	1.5389	0.0312	10.4325	9.0450
No log	4	140	1.3976	0.0625	15.3639	8.0682
No log	5	175	1.3392	0.125	19.6033	7.6702
No log	6	210	1.2809	0.25	20.5096	7.9182
No log	7	245	1.1624	0.5	25.4581	8.7427
0.3629	8.0	280	1.0569	1.0	40.2946	10.7128
1.4816	9.0	315	1.0025	1.0	39.5400	11.2317
1.2493	10.0	350	0.9695	1.0	32.6111	11.9216
1.2493	11.0	385	0.9447	1.0	34.4315	12.1721
1.0496	12.0	420	0.9390	1.0	36.4801	12.7830
0.9033	13.0	455	0.9307	1.0	31.6877	13.0193
0.9033	14.0	490	0.9417	1.0	34.8782	13.3722
0.776	15.0	525	0.9330	1.0	39.7475	13.5117
0.6831	16.0	560	0.9448	1.0	31.6756	13.7563
0.6831	17.0	595	0.9608	1.0	35.9424	13.9213

Framework versions

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