f9d7db2b7ddca1778c918d3f359208f5

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

• Loss: 0.8352
• Data Size: 1.0
• Epoch Runtime: 197.4582
• Bleu: 13.1874

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	3.0883	0	14.1751	0.4827
No log	1	447	2.3676	0.0078	16.1659	1.0953
0.0413	2	894	1.9266	0.0156	17.9630	1.6517
0.0469	3	1341	1.7889	0.0312	21.8603	1.7657
0.0739	4	1788	1.6813	0.0625	27.1166	2.2834
0.1276	5	2235	1.5740	0.125	39.6788	3.2473
1.6784	6	2682	1.4593	0.25	63.4252	4.4493
1.5029	7	3129	1.3210	0.5	105.9890	5.5874
1.2876	8.0	3576	1.1613	1.0	193.8574	7.4531
1.1839	9.0	4023	1.0697	1.0	193.1934	8.5501
1.0911	10.0	4470	1.0156	1.0	197.1107	9.3000
1.0205	11.0	4917	0.9645	1.0	190.9796	9.7673
0.9745	12.0	5364	0.9334	1.0	189.2375	10.1733
0.9031	13.0	5811	0.9056	1.0	194.9996	10.8554
0.8574	14.0	6258	0.8889	1.0	187.1327	11.1684
0.825	15.0	6705	0.8677	1.0	197.2455	11.4799
0.7736	16.0	7152	0.8537	1.0	200.4439	11.8776
0.7444	17.0	7599	0.8436	1.0	196.0900	11.9945
0.7215	18.0	8046	0.8334	1.0	197.2098	12.1491
0.6925	19.0	8493	0.8372	1.0	206.5470	12.3215
0.6551	20.0	8940	0.8323	1.0	197.6972	12.6482
0.6515	21.0	9387	0.8252	1.0	195.9774	12.6647
0.5968	22.0	9834	0.8207	1.0	203.3854	12.7550
0.5777	23.0	10281	0.8197	1.0	201.8217	12.9205
0.5442	24.0	10728	0.8301	1.0	199.0957	12.8124
0.5252	25.0	11175	0.8256	1.0	198.2719	13.0060
0.5003	26.0	11622	0.8358	1.0	200.2850	13.0495
0.5018	27.0	12069	0.8352	1.0	197.4582	13.1874

Framework versions

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