224b06d8f617f248a9ebd5155a2a49c5

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

• Loss: 1.8325
• Data Size: 1.0
• Epoch Runtime: 53.4114
• Bleu: 1.8111

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	2.6619	0	2.8572	0.2666
No log	1	70	2.5426	0.0078	3.4201	0.3144
No log	2	140	2.3585	0.0156	7.8639	0.4791
No log	3	210	2.2649	0.0312	11.3587	0.5753
No log	4	280	2.1838	0.0625	17.3585	1.0647
No log	5	350	2.0830	0.125	24.8194	1.5429
No log	6	420	1.9863	0.25	30.6833	1.0717
0.3264	7	490	1.9024	0.5	33.1851	1.1965
1.8645	8.0	560	1.8183	1.0	54.4430	1.2549
1.7325	9.0	630	1.7821	1.0	53.1267	1.4422
1.554	10.0	700	1.7727	1.0	43.6281	1.5300
1.4343	11.0	770	1.7821	1.0	46.1418	1.6494
1.3619	12.0	840	1.7959	1.0	42.0194	1.7865
1.2361	13.0	910	1.8162	1.0	43.1086	1.8647
1.1369	14.0	980	1.8325	1.0	53.4114	1.8111

Framework versions

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