8d53e2a5b344b1d72aee24bf6fc1a937

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

• Loss: 0.0343
• Data Size: 1.0
• Epoch Runtime: 35.3202
• Accuracy: 0.9946
• F1 Macro: 0.9943

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
No log	0	0	0.6705	0	3.0614	0.6209	0.4127
No log	1	650	0.0419	0.0078	3.5020	0.9936	0.9933
No log	2	1300	0.0222	0.0156	3.7032	0.9958	0.9955
No log	3	1950	0.0204	0.0312	4.4154	0.9963	0.9961
No log	4	2600	0.0315	0.0625	5.3453	0.9952	0.9949
0.0018	5	3250	0.0049	0.125	7.3281	0.9986	0.9986
0.0103	6	3900	0.0099	0.25	11.4372	0.9985	0.9984
0.0321	7	4550	0.0144	0.5	19.7372	0.9983	0.9982
0.015	8.0	5200	0.0291	1.0	35.6272	0.9940	0.9937
0.0088	9.0	5850	0.0343	1.0	35.3202	0.9946	0.9943

Framework versions

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