5ff575dd601e90aa158c480a0ed476ba

This model is a fine-tuned version of google-bert/bert-base-german-dbmdz-cased on the fancyzhx/dbpedia_14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0890
• Data Size: 1.0
• Epoch Runtime: 888.4984
• Accuracy: 0.9832
• F1 Macro: 0.9832
• Rouge1: 0.9832
• Rouge2: 0.0
• Rougel: 0.9832
• Rougelsum: 0.9832

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	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	2.6930	0	29.6646	0.0731	0.0251	0.0730	0.0	0.0731	0.0731
0.2964	1	17500	0.1652	0.0078	35.5760	0.9573	0.9573	0.9572	0.0	0.9572	0.9572
0.0995	2	35000	0.1447	0.0156	42.4478	0.9644	0.9646	0.9644	0.0	0.9645	0.9644
0.1136	3	52500	0.1179	0.0312	55.5914	0.9750	0.9749	0.9750	0.0	0.9750	0.9750
0.0993	4	70000	0.0985	0.0625	80.2408	0.9773	0.9773	0.9773	0.0	0.9773	0.9773
0.0667	5	87500	0.0692	0.125	128.7757	0.9846	0.9846	0.9846	0.0	0.9846	0.9846
0.0901	6	105000	0.0775	0.25	232.4449	0.9833	0.9834	0.9833	0.0	0.9833	0.9833
0.0005	7	122500	0.0705	0.5	438.0950	0.9849	0.9849	0.9849	0.0	0.9849	0.9849
0.0736	8.0	140000	0.0848	1.0	891.7947	0.9837	0.9837	0.9837	0.0	0.9837	0.9837
0.0843	9.0	157500	0.0890	1.0	888.4984	0.9832	0.9832	0.9832	0.0	0.9832	0.9832

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1