9a3e32507a87ae986e2a61de836a8bee

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

• Loss: 1.1425
• Data Size: 1.0
• Epoch Runtime: 966.9192
• Bleu: 14.7168

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	2.3636	0	69.4507	3.3313
No log	1	2336	2.0102	0.0078	83.9350	7.5377
0.0324	2	4672	1.9195	0.0156	93.3999	8.4411
0.0468	3	7008	1.8391	0.0312	103.4198	9.7646
1.9541	4	9344	1.7461	0.0625	128.2039	10.8839
1.8482	5	11680	1.6449	0.125	196.8387	11.4094
1.7016	6	14016	1.5278	0.25	310.5450	12.7442
1.5894	7	16352	1.4131	0.5	552.2568	13.9636
1.4512	8.0	18688	1.3037	1.0	1116.3571	14.6789
1.3209	9.0	21024	1.2410	1.0	1105.0303	14.7549
1.2325	10.0	23360	1.2049	1.0	1102.1752	15.2807
1.1882	11.0	25696	1.1791	1.0	1026.3898	15.7608
1.1145	12.0	28032	1.1635	1.0	1034.3636	15.6938
1.0567	13.0	30368	1.1489	1.0	1005.4051	15.0973
1.0315	14.0	32704	1.1389	1.0	1024.3694	15.7429
0.9644	15.0	35040	1.1368	1.0	1007.0582	15.5204
0.9753	16.0	37376	1.1333	1.0	1035.7918	15.1784
0.8878	17.0	39712	1.1289	1.0	1014.8950	15.2695
0.8874	18.0	42048	1.1360	1.0	992.9356	15.2767
0.8534	19.0	44384	1.1413	1.0	978.4691	15.2869
0.8061	20.0	46720	1.1390	1.0	989.3805	15.2742
0.7735	21.0	49056	1.1425	1.0	966.9192	14.7168

Framework versions

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