d4f391010aec32c8d2e0ab6b6898b51e

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

• Loss: 1.5131
• Data Size: 1.0
• Epoch Runtime: 188.7451
• Bleu: 3.4104

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	203.3179	0	13.8725	0.0048
No log	1	437	151.9397	0.0078	16.8310	0.0021
No log	2	874	91.9245	0.0156	17.5949	0.0049
No log	3	1311	41.3338	0.0312	21.7833	0.0037
No log	4	1748	14.5198	0.0625	28.3048	0.0124
20.006	5	2185	9.2048	0.125	40.2275	0.0077
15.1068	6	2622	7.0029	0.25	61.3442	0.0165
10.3681	7	3059	5.1010	0.5	104.5574	0.0282
7.4323	8.0	3496	3.9191	1.0	193.2738	0.1163
6.0454	9.0	3933	3.3345	1.0	192.0627	0.2352
5.1294	10.0	4370	3.0954	1.0	191.1681	0.4104
4.5298	11.0	4807	2.6860	1.0	191.6684	0.5311
4.0702	12.0	5244	2.5652	1.0	191.5956	0.5507
3.6954	13.0	5681	2.4747	1.0	191.6501	0.4395
3.3925	14.0	6118	2.3057	1.0	190.4953	1.0742
3.1576	15.0	6555	2.1997	1.0	190.1354	0.9730
2.9524	16.0	6992	2.0628	1.0	190.5867	1.0125
2.7976	17.0	7429	2.0462	1.0	190.6328	1.1332
2.659	18.0	7866	1.9582	1.0	189.9537	1.2786
2.5601	19.0	8303	1.9270	1.0	191.3812	1.1806
2.4688	20.0	8740	1.8754	1.0	191.0130	1.3126
2.3734	21.0	9177	1.8524	1.0	189.7151	1.7209
2.2997	22.0	9614	1.8289	1.0	189.3750	1.5993
2.2331	23.0	10051	1.8041	1.0	191.0508	1.6395
2.1879	24.0	10488	1.7909	1.0	190.0420	1.4962
2.1437	25.0	10925	1.7521	1.0	190.8916	1.6839
2.0761	26.0	11362	1.7351	1.0	191.1157	1.9509
2.0375	27.0	11799	1.7174	1.0	188.1737	1.7784
1.984	28.0	12236	1.7307	1.0	188.5870	2.0249
1.9444	29.0	12673	1.6861	1.0	189.2382	2.0246
1.9286	30.0	13110	1.6741	1.0	188.5746	2.0330
1.9027	31.0	13547	1.6614	1.0	189.0473	2.0770
1.8799	32.0	13984	1.6574	1.0	188.4332	2.1513
1.8246	33.0	14421	1.6412	1.0	188.1285	2.0970
1.8204	34.0	14858	1.6290	1.0	187.5204	2.4002
1.7772	35.0	15295	1.6164	1.0	187.6700	2.3430
1.7504	36.0	15732	1.6102	1.0	188.1867	2.3951
1.737	37.0	16169	1.5956	1.0	188.8280	2.5689
1.7137	38.0	16606	1.5900	1.0	189.6627	2.5392
1.6757	39.0	17043	1.5782	1.0	189.1879	2.5737
1.6486	40.0	17480	1.5642	1.0	189.7931	2.6443
1.6366	41.0	17917	1.5826	1.0	189.2834	2.8002
1.6132	42.0	18354	1.5468	1.0	191.9030	2.9299
1.5917	43.0	18791	1.5420	1.0	188.9345	2.7305
1.5572	44.0	19228	1.5362	1.0	190.2572	3.0314
1.5642	45.0	19665	1.5299	1.0	188.9003	3.0041
1.5197	46.0	20102	1.5411	1.0	190.9117	3.1187
1.5004	47.0	20539	1.5202	1.0	190.3094	3.1452
1.487	48.0	20976	1.5408	1.0	190.0726	3.2623
1.4657	49.0	21413	1.5182	1.0	188.5026	3.2668
1.4396	50.0	21850	1.5131	1.0	188.7451	3.4104

Framework versions

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