7963be5d7c8e711497421ba8f5e52055

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

• Loss: 0.8853
• Data Size: 1.0
• Epoch Runtime: 203.0420
• Bleu: 15.8962

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	1.8069	0	13.8837	1.3891
No log	1	437	1.6123	0.0078	16.6482	2.6554
No log	2	874	1.4805	0.0156	21.4981	5.1919
No log	3	1311	1.3885	0.0312	29.1070	7.3363
No log	4	1748	1.2893	0.0625	37.0546	5.6148
1.3512	5	2185	1.1933	0.125	50.5094	6.6125
1.2509	6	2622	1.0797	0.25	68.3128	8.2764
1.0932	7	3059	0.9746	0.5	117.4071	9.7611
0.925	8.0	3496	0.8671	1.0	207.1769	12.2936
0.8115	9.0	3933	0.8198	1.0	204.0337	13.6432
0.6984	10.0	4370	0.7980	1.0	211.0713	14.3979
0.6237	11.0	4807	0.7907	1.0	206.7455	14.9244
0.5564	12.0	5244	0.7887	1.0	204.3353	15.1615
0.4977	13.0	5681	0.7936	1.0	199.5988	15.5572
0.4354	14.0	6118	0.8153	1.0	206.8478	15.4884
0.3921	15.0	6555	0.8439	1.0	204.7876	15.4676
0.3376	16.0	6992	0.8853	1.0	203.0420	15.8962

Framework versions

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