99768a5ad6c8d997fe8d35bc4119cc42

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

• Loss: 1.4983
• Data Size: 1.0
• Epoch Runtime: 35.3967
• Bleu: 8.0107

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	2.7221	0	1.8492	1.9089
No log	1	27	2.4792	0.0078	2.2026	2.2985
No log	2	54	2.3030	0.0156	7.0519	3.0032
No log	3	81	2.1576	0.0312	12.0543	3.6427
No log	4	108	2.0204	0.0625	17.2036	4.3920
No log	5	135	1.8587	0.125	22.3992	4.2654
No log	6	162	1.6864	0.25	27.5897	4.9785
No log	7	189	1.5871	0.5	27.3922	5.2589
0.3884	8.0	216	1.4611	1.0	34.3287	6.8204
0.3884	9.0	243	1.4090	1.0	33.0873	7.0825
1.3865	10.0	270	1.3996	1.0	34.5781	6.5440
1.3865	11.0	297	1.4082	1.0	36.1248	6.6709
1.1107	12.0	324	1.3948	1.0	41.6256	7.0375
0.9038	13.0	351	1.4126	1.0	29.4171	7.2052
0.9038	14.0	378	1.4552	1.0	32.5662	7.6491
0.7444	15.0	405	1.4830	1.0	33.6072	7.9917
0.7444	16.0	432	1.4983	1.0	35.3967	8.0107

Framework versions

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