df4197adbc18b8e2e711085ed3d505fa

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

• Loss: 0.7362
• Data Size: 1.0
• Epoch Runtime: 187.9114
• Bleu: 15.0047

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	3.0800	0	12.3362	0.9194
No log	1	447	2.1202	0.0078	14.4803	1.8458
0.0375	2	894	1.5253	0.0156	19.6325	3.6464
0.0384	3	1341	1.3652	0.0312	24.5179	6.7247
0.0581	4	1788	1.2629	0.0625	31.9134	4.9112
0.0983	5	2235	1.1519	0.125	43.5487	5.8782
1.25	6	2682	1.0470	0.25	61.5060	7.2867
1.0637	7	3129	0.9211	0.5	103.0064	9.1057
0.8718	8.0	3576	0.8047	1.0	186.3805	11.2624
0.7723	9.0	4023	0.7503	1.0	183.0542	12.2775
0.6801	10.0	4470	0.7202	1.0	184.2025	13.1528
0.6033	11.0	4917	0.7003	1.0	183.7012	13.6732
0.5501	12.0	5364	0.6939	1.0	186.6995	14.0794
0.4759	13.0	5811	0.6954	1.0	185.8033	14.4014
0.4388	14.0	6258	0.7130	1.0	185.1243	14.5703
0.3932	15.0	6705	0.7266	1.0	186.5666	14.7679
0.3538	16.0	7152	0.7362	1.0	187.9114	15.0047

Framework versions

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