bf1462d89aff6f3487022728c0e619d3

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

• Loss: 1.2867
• Data Size: 1.0
• Epoch Runtime: 102.8343
• Bleu: 4.8831

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	3.8321	0	7.7511	0.2778
No log	1	419	3.1758	0.0078	10.2457	0.2173
No log	2	838	2.4281	0.0156	9.8304	0.1990
0.0781	3	1257	2.2940	0.0312	11.6407	0.4073
0.0781	4	1676	2.1725	0.0625	14.9945	0.4680
0.1468	5	2095	2.0765	0.125	19.9688	0.4470
0.2969	6	2514	1.9815	0.25	30.7923	0.4445
2.055	7	2933	1.8771	0.5	52.1986	1.2337
1.908	8.0	3352	1.7673	1.0	98.1835	1.7631
1.8323	9.0	3771	1.6963	1.0	104.2326	2.0477
1.7914	10.0	4190	1.6447	1.0	104.3759	2.1788
1.7083	11.0	4609	1.6005	1.0	97.1955	2.3249
1.6842	12.0	5028	1.5647	1.0	98.2808	2.5057
1.6198	13.0	5447	1.5387	1.0	98.4246	2.7196
1.6071	14.0	5866	1.5071	1.0	98.7859	2.8021
1.5506	15.0	6285	1.4813	1.0	103.0710	2.9360
1.5343	16.0	6704	1.4627	1.0	103.5573	2.9611
1.5054	17.0	7123	1.4434	1.0	100.6397	3.2404
1.4638	18.0	7542	1.4301	1.0	102.1796	3.3761
1.4393	19.0	7961	1.4137	1.0	101.3366	3.4453
1.4032	20.0	8380	1.4006	1.0	107.1915	3.5435
1.4009	21.0	8799	1.3863	1.0	102.6508	3.6507
1.3719	22.0	9218	1.3724	1.0	102.2086	3.6668
1.362	23.0	9637	1.3700	1.0	99.3133	3.8352
1.3195	24.0	10056	1.3549	1.0	103.5616	3.9424
1.3178	25.0	10475	1.3464	1.0	98.7277	3.9231
1.2825	26.0	10894	1.3382	1.0	108.8320	4.1158
1.2786	27.0	11313	1.3299	1.0	109.9823	4.0274
1.2738	28.0	11732	1.3250	1.0	107.1214	4.1404
1.256	29.0	12151	1.3199	1.0	101.2873	4.2051
1.2284	30.0	12570	1.3160	1.0	101.6196	4.2800
1.2079	31.0	12989	1.3049	1.0	99.4647	4.3756
1.2035	32.0	13408	1.3039	1.0	98.0781	4.3166
1.1767	33.0	13827	1.3071	1.0	97.6157	4.4239
1.1486	34.0	14246	1.2912	1.0	101.8700	4.4547
1.1646	35.0	14665	1.2933	1.0	102.8562	4.5101
1.1339	36.0	15084	1.2922	1.0	101.9525	4.5820
1.1166	37.0	15503	1.2862	1.0	105.2984	4.5991
1.1019	38.0	15922	1.2860	1.0	101.3865	4.5862
1.0878	39.0	16341	1.2851	1.0	98.2697	4.6133
1.0754	40.0	16760	1.2796	1.0	99.6733	4.6847
1.0683	41.0	17179	1.2873	1.0	99.9450	4.7070
1.0581	42.0	17598	1.2823	1.0	98.5332	4.7370
1.0576	43.0	18017	1.2793	1.0	98.4368	4.7392
1.032	44.0	18436	1.2821	1.0	105.1789	4.8431
1.0127	45.0	18855	1.2795	1.0	100.6899	4.8298
1.0096	46.0	19274	1.2804	1.0	104.6137	4.8207
0.9995	47.0	19693	1.2731	1.0	103.7977	4.7398
0.9828	48.0	20112	1.2887	1.0	108.5207	4.9181
0.9896	49.0	20531	1.2801	1.0	105.9333	4.7891
0.9738	50.0	20950	1.2867	1.0	102.8343	4.8831

Framework versions

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