8ca2e63766dbde8a7aa3097817fd2f3c

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

• Loss: 0.9198
• Data Size: 1.0
• Epoch Runtime: 178.2104
• Bleu: 14.4192

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	2.1967	0	13.1722	0.4158
No log	1	437	2.0449	0.0078	14.8412	0.9544
No log	2	874	1.9380	0.0156	15.9749	1.3087
No log	3	1311	1.8222	0.0312	20.7062	1.7295
No log	4	1748	1.7119	0.0625	26.9948	2.8520
1.81	5	2185	1.6065	0.125	38.1287	3.4831
1.6817	6	2622	1.4706	0.25	59.5105	4.3908
1.5213	7	3059	1.3347	0.5	100.7658	6.1842
1.3251	8.0	3496	1.1834	1.0	180.4133	8.4047
1.2087	9.0	3933	1.0970	1.0	182.8740	9.7600
1.091	10.0	4370	1.0375	1.0	177.3929	10.6758
1.0247	11.0	4807	1.0022	1.0	176.7033	11.4031
0.9658	12.0	5244	0.9670	1.0	177.8799	12.2733
0.9057	13.0	5681	0.9476	1.0	178.1959	12.5709
0.8377	14.0	6118	0.9361	1.0	176.9459	13.1417
0.8037	15.0	6555	0.9284	1.0	180.4766	13.0320
0.7443	16.0	6992	0.9108	1.0	179.0464	13.6398
0.7211	17.0	7429	0.9076	1.0	178.3307	13.3907
0.6807	18.0	7866	0.9184	1.0	181.1875	13.8668
0.6415	19.0	8303	0.9113	1.0	178.9721	13.9984
0.6186	20.0	8740	0.9104	1.0	180.3212	13.9662
0.592	21.0	9177	0.9198	1.0	178.2104	14.4192

Framework versions

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