fda904753def3d123cd25fdbaf479c88

This model is a fine-tuned version of google-t5/t5-3b on the Helsinki-NLP/opus_books [en-pt] dataset. It achieves the following results on the evaluation set:

• Loss: 0.8807
• Data Size: 1.0
• Epoch Runtime: 36.5586
• Bleu: 19.0050

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	1.6888	0	1.9636	8.1270
No log	1	35	1.4034	0.0078	2.8702	8.9262
No log	2	70	1.2250	0.0156	6.1203	12.4007
No log	3	105	1.1722	0.0312	11.3888	10.3772
No log	4	140	1.1197	0.0625	19.1300	10.5168
No log	5	175	1.0644	0.125	23.5111	11.2868
No log	6	210	0.9824	0.25	28.5935	13.1734
No log	7	245	0.8901	0.5	24.8551	16.2580
0.2182	8.0	280	0.8069	1.0	33.4287	16.1355
0.8651	9.0	315	0.7824	1.0	34.4039	16.7790
0.6654	10.0	350	0.7796	1.0	37.7468	18.0529
0.6654	11.0	385	0.7771	1.0	40.3196	17.7458
0.5116	12.0	420	0.7826	1.0	30.4486	18.1556
0.4119	13.0	455	0.8035	1.0	31.8901	18.4966
0.4119	14.0	490	0.8527	1.0	35.8862	17.9012
0.331	15.0	525	0.8807	1.0	36.5586	19.0050

Framework versions

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