3798b2451aed3cbf6dd09863bbcc2b53

This model is a fine-tuned version of google/mt5-base on the Helsinki-NLP/opus_books [it-ru] dataset. It achieves the following results on the evaluation set:

• Loss: 1.3212
• Data Size: 1.0
• Epoch Runtime: 102.9030
• Bleu: 11.4523

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	18.7101	0	8.1375	0.0070
No log	1	447	15.5222	0.0078	8.8450	0.0112
0.2642	2	894	15.0198	0.0156	10.1062	0.0104
0.346	3	1341	15.2283	0.0312	12.6545	0.0085
0.5442	4	1788	8.9476	0.0625	15.6935	0.0130
0.8016	5	2235	6.7790	0.125	21.4150	0.0207
6.8299	6	2682	3.6551	0.25	32.0205	0.3783
3.2328	7	3129	2.0647	0.5	54.0593	4.1573
2.3429	8.0	3576	1.6954	1.0	97.2096	6.3831
2.1504	9.0	4023	1.5869	1.0	95.6565	7.4095
1.9874	10.0	4470	1.5282	1.0	96.8454	7.9612
1.88	11.0	4917	1.4898	1.0	96.9167	8.5106
1.8172	12.0	5364	1.4523	1.0	99.1246	8.8419
1.691	13.0	5811	1.4316	1.0	97.7661	9.2087
1.6377	14.0	6258	1.4068	1.0	96.0381	9.4061
1.6041	15.0	6705	1.3921	1.0	99.0179	9.9002
1.5197	16.0	7152	1.3805	1.0	100.8643	9.9666
1.4845	17.0	7599	1.3618	1.0	100.6985	10.2203
1.4617	18.0	8046	1.3472	1.0	102.0336	10.3916
1.4036	19.0	8493	1.3414	1.0	103.3909	10.5841
1.3769	20.0	8940	1.3384	1.0	104.4361	10.7687
1.3606	21.0	9387	1.3333	1.0	103.3733	10.7833
1.2803	22.0	9834	1.3284	1.0	103.2838	10.7905
1.2383	23.0	10281	1.3167	1.0	103.3366	10.9550
1.2123	24.0	10728	1.3226	1.0	104.1142	10.9717
1.1761	25.0	11175	1.3208	1.0	103.8856	11.0127
1.1347	26.0	11622	1.3156	1.0	103.1935	11.1601
1.1619	27.0	12069	1.3206	1.0	101.5619	11.2488
1.0916	28.0	12516	1.3234	1.0	103.9717	11.1902
1.0614	29.0	12963	1.3304	1.0	103.1075	11.1793
1.0344	30.0	13410	1.3212	1.0	102.9030	11.4523

Framework versions

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