340e50a792c763605d83b343433815b1

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

• Loss: 0.4651
• Data Size: 1.0
• Epoch Runtime: 17.7907
• Accuracy: 0.8927
• F1 Macro: 0.7344

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	1.1561	0	1.9924	0.0615	0.0403
No log	1	619	0.6718	0.0078	2.3985	0.7672	0.2894
No log	2	1238	0.6038	0.0156	2.1801	0.7672	0.2894
0.0142	3	1857	0.3836	0.0312	2.5688	0.8720	0.5675
0.0142	4	2476	0.3169	0.0625	3.0954	0.8994	0.5998
0.3026	5	3095	0.3262	0.125	4.2372	0.8864	0.7068
0.0246	6	3714	0.2950	0.25	6.3210	0.9081	0.6093
0.2648	7	4333	0.2723	0.5	10.4750	0.9089	0.7519
0.2379	8.0	4952	0.2483	1.0	19.0608	0.9095	0.7620
0.165	9.0	5571	0.2720	1.0	19.2192	0.9107	0.7532
0.1418	10.0	6190	0.3401	1.0	19.4040	0.9095	0.7706
0.1131	11.0	6809	0.3716	1.0	18.1863	0.9079	0.7394
0.0584	12.0	7428	0.4651	1.0	17.7907	0.8927	0.7344

Framework versions

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