3e98143c881a06f5c6cc58f04165b1cc

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

• Loss: 0.4411
• Data Size: 1.0
• Epoch Runtime: 19.4509
• Accuracy: 0.8961
• F1 Macro: 0.7260

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	1.1413	0	1.9749	0.0645	0.0418
No log	1	619	0.6892	0.0078	2.3202	0.7672	0.2894
No log	2	1238	0.6640	0.0156	2.2242	0.7672	0.2894
0.016	3	1857	0.5470	0.0312	2.6022	0.7812	0.3493
0.016	4	2476	0.4127	0.0625	3.1766	0.8350	0.5344
0.3294	5	3095	0.3322	0.125	4.2737	0.8711	0.6776
0.0271	6	3714	0.2982	0.25	6.3551	0.8951	0.6520
0.2801	7	4333	0.2793	0.5	10.8508	0.9016	0.7423
0.2492	8.0	4952	0.2708	1.0	19.6870	0.9097	0.6521
0.1836	9.0	5571	0.2912	1.0	19.5440	0.9046	0.7183
0.1498	10.0	6190	0.3440	1.0	19.4393	0.9042	0.7602
0.1134	11.0	6809	0.3975	1.0	19.2728	0.9085	0.7369
0.0761	12.0	7428	0.4411	1.0	19.4509	0.8961	0.7260

Framework versions

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