dda0f36ebebba3f91ececf49fde6706a

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

• Loss: 0.9009
• Data Size: 1.0
• Epoch Runtime: 111.4918
• Bleu: 11.5587

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	3.3957	0	8.6733	0.4226
No log	1	447	2.9355	0.0078	10.7657	0.4165
0.0488	2	894	2.3432	0.0156	10.3163	0.4884
0.0573	3	1341	2.2212	0.0312	12.3303	0.8807
0.0905	4	1788	2.0793	0.0625	15.5077	0.8417
0.1563	5	2235	1.9609	0.125	21.8305	1.0791
2.0802	6	2682	1.8493	0.25	34.7633	2.0794
1.9014	7	3129	1.7160	0.5	59.9261	2.9115
1.7279	8.0	3576	1.5688	1.0	111.8558	3.9321
1.6258	9.0	4023	1.4644	1.0	106.2814	4.6714
1.544	10.0	4470	1.3874	1.0	112.3671	5.3961
1.4653	11.0	4917	1.3200	1.0	107.5847	5.8752
1.4083	12.0	5364	1.2686	1.0	109.5193	6.4212
1.342	13.0	5811	1.2232	1.0	116.3281	6.8981
1.2931	14.0	6258	1.1928	1.0	121.9684	7.1075
1.263	15.0	6705	1.1596	1.0	120.9998	7.5103
1.2135	16.0	7152	1.1288	1.0	105.8727	7.8518
1.1871	17.0	7599	1.1031	1.0	105.8427	7.9929
1.1637	18.0	8046	1.0857	1.0	111.1980	8.2713
1.1241	19.0	8493	1.0636	1.0	106.4765	8.5642
1.0953	20.0	8940	1.0499	1.0	107.8300	8.7800
1.0958	21.0	9387	1.0354	1.0	105.1607	8.8858
1.0353	22.0	9834	1.0202	1.0	106.5439	9.0909
1.0098	23.0	10281	1.0077	1.0	101.3376	9.2164
0.9925	24.0	10728	0.9981	1.0	108.5114	9.4437
0.9704	25.0	11175	0.9906	1.0	104.2813	9.6297
0.9426	26.0	11622	0.9761	1.0	106.7828	9.8129
0.9492	27.0	12069	0.9694	1.0	107.0763	9.9101
0.9029	28.0	12516	0.9657	1.0	105.2038	10.1075
0.8811	29.0	12963	0.9548	1.0	106.0641	10.0841
0.8678	30.0	13410	0.9499	1.0	110.9597	10.2664
0.8721	31.0	13857	0.9407	1.0	105.4958	10.3908
0.8381	32.0	14304	0.9407	1.0	102.9635	10.4487
0.8413	33.0	14751	0.9335	1.0	105.2890	10.4928
0.8143	34.0	15198	0.9322	1.0	104.6580	10.6494
0.7941	35.0	15645	0.9236	1.0	111.0841	10.7106
0.7828	36.0	16092	0.9175	1.0	100.8563	10.7893
0.7575	37.0	16539	0.9191	1.0	104.8849	10.9428
0.7695	38.0	16986	0.9181	1.0	106.6565	11.0279
0.744	39.0	17433	0.9089	1.0	109.3794	11.1432
0.7226	40.0	17880	0.9080	1.0	109.9170	11.1099
0.7161	41.0	18327	0.9049	1.0	103.7046	11.2632
0.698	42.0	18774	0.9077	1.0	102.1417	11.2277
0.701	43.0	19221	0.9150	1.0	106.4206	11.2808
0.6753	44.0	19668	0.9004	1.0	105.7527	11.3886
0.6702	45.0	20115	0.9061	1.0	108.0678	11.4461
0.6565	46.0	20562	0.9050	1.0	105.7159	11.5096
0.6572	47.0	21009	0.9049	1.0	106.0853	11.5223
0.6343	48.0	21456	0.9009	1.0	111.4918	11.5587

Framework versions

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