cdbd9519200627023626eaf4f5a9c2ec

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

• Loss: 1.8382
• Data Size: 1.0
• Epoch Runtime: 296.1652
• Bleu: 10.3689

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	3.2834	0	32.3240	3.1665
No log	1	1286	2.8676	0.0078	31.2946	4.9302
0.061	2	2572	2.6478	0.0156	39.1063	6.0021
0.0781	3	3858	2.5222	0.0312	40.5774	6.1777
0.1162	4	5144	2.4145	0.0625	44.6000	5.9555
2.5374	5	6430	2.3278	0.125	71.0763	6.6924
2.4421	6	7716	2.2364	0.25	114.0874	6.9143
2.3398	7	9002	2.1406	0.5	203.1443	7.6527
2.2267	8.0	10288	2.0375	1.0	320.0082	8.2874
2.1245	9.0	11574	1.9762	1.0	327.2068	8.7267
2.0126	10.0	12860	1.9342	1.0	318.2429	9.0232
1.9288	11.0	14146	1.9076	1.0	313.7659	9.4612
1.9147	12.0	15432	1.8856	1.0	306.3946	9.5535
1.8118	13.0	16718	1.8670	1.0	301.9939	9.6642
1.7869	14.0	18004	1.8577	1.0	303.7155	9.7512
1.7684	15.0	19290	1.8451	1.0	311.2393	9.8912
1.7035	16.0	20576	1.8364	1.0	301.0892	10.0150
1.6895	17.0	21862	1.8324	1.0	312.8599	10.0120
1.6357	18.0	23148	1.8328	1.0	298.5094	10.0325
1.615	19.0	24434	1.8274	1.0	303.1685	10.1694
1.5417	20.0	25720	1.8263	1.0	301.3516	10.1965
1.5421	21.0	27006	1.8260	1.0	308.1764	10.1196
1.5266	22.0	28292	1.8275	1.0	300.0012	10.2606
1.4999	23.0	29578	1.8345	1.0	303.9673	10.3856
1.4494	24.0	30864	1.8391	1.0	291.5877	10.4315
1.4225	25.0	32150	1.8382	1.0	296.1652	10.3689

Framework versions

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