22d8b0d285b5e283cf9ac09aee743c2b

This model is a fine-tuned version of google-bert/bert-base-german-cased on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.4046
• Data Size: 1.0
• Epoch Runtime: 34.2407
• Accuracy: 0.8931
• F1 Macro: 0.7158

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	0.9905	0	2.8818	0.7634	0.2931
No log	1	619	0.6721	0.0078	3.3581	0.7672	0.2894
No log	2	1238	0.5441	0.0156	3.6335	0.7490	0.4541
0.0136	3	1857	0.4316	0.0312	4.4059	0.8446	0.5394
0.0136	4	2476	0.3533	0.0625	5.2466	0.8809	0.6511
0.322	5	3095	0.3296	0.125	7.4018	0.8908	0.6806
0.0275	6	3714	0.2970	0.25	11.1527	0.9030	0.6408
0.2861	7	4333	0.3038	0.5	18.8954	0.8927	0.7329
0.2733	8.0	4952	0.2745	1.0	35.0378	0.9056	0.6491
0.2176	9.0	5571	0.3002	1.0	33.5553	0.8985	0.7635
0.2226	10.0	6190	0.3072	1.0	33.9894	0.9046	0.7620
0.1827	11.0	6809	0.3475	1.0	33.2486	0.8981	0.7344
0.1291	12.0	7428	0.4046	1.0	34.2407	0.8931	0.7158

Framework versions

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