b3dc16b7eaad78fae3a1be7d42047e94

This model is a fine-tuned version of openai-community/gpt2-medium on the contemmcm/hate-speech-and-offensive-language dataset. It achieves the following results on the evaluation set:

• Loss: 0.6391
• Data Size: 1.0
• Epoch Runtime: 87.5050
• Accuracy: 0.9067
• F1 Macro: 0.7569

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	4.3931	0	7.7849	0.7425	0.2892
No log	1	619	1.4848	0.0078	8.9086	0.6843	0.3121
No log	2	1238	0.7177	0.0156	9.4507	0.7608	0.3568
0.0296	3	1857	0.5054	0.0312	12.4153	0.8667	0.5503
0.0296	4	2476	0.3462	0.0625	14.6188	0.8963	0.5938
0.3402	5	3095	0.2804	0.125	20.3349	0.9026	0.7008
0.0256	6	3714	0.2641	0.25	30.6227	0.9083	0.6482
0.248	7	4333	0.2591	0.5	51.0064	0.9046	0.7481
0.2164	8.0	4952	0.2413	1.0	91.9164	0.9125	0.7483
0.1448	9.0	5571	0.2779	1.0	91.3241	0.9058	0.7711
0.0939	10.0	6190	0.3680	1.0	92.1140	0.9048	0.7296
0.0733	11.0	6809	0.5630	1.0	87.8674	0.8851	0.7387
0.0572	12.0	7428	0.6391	1.0	87.5050	0.9067	0.7569

Framework versions

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