fba67bacd106771d859bb46312bd4a0c

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

• Loss: 0.8261
• Data Size: 1.0
• Epoch Runtime: 212.9900
• Bleu: 12.9615

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	2.3764	0	13.4695	0.4921
No log	1	434	1.9339	0.0078	14.7386	1.5038
No log	2	868	1.4971	0.0156	23.1625	3.9921
No log	3	1302	1.3828	0.0312	26.1764	3.6147
No log	4	1736	1.2868	0.0625	35.2394	4.6082
0.06	5	2170	1.2015	0.125	50.2450	5.5186
1.246	6	2604	1.1012	0.25	74.2602	6.7794
1.1094	7	3038	0.9999	0.5	117.9687	8.2004
0.9752	8.0	3472	0.8919	1.0	207.6366	9.8824
0.8354	9.0	3906	0.8372	1.0	207.8709	10.8329
0.7488	10.0	4340	0.8135	1.0	200.4027	11.6028
0.6618	11.0	4774	0.7902	1.0	203.8508	12.1933
0.5837	12.0	5208	0.7915	1.0	205.4988	12.3600
0.5313	13.0	5642	0.7980	1.0	210.1147	12.6810
0.4782	14.0	6076	0.8127	1.0	209.0812	12.9391
0.44	15.0	6510	0.8261	1.0	212.9900	12.9615

Framework versions

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