a9c4e2985f2d1af80b1565892288806e

This model is a fine-tuned version of openai-community/gpt2-medium on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 0.4070
• Data Size: 1.0
• Epoch Runtime: 181.3512
• Accuracy: 0.8999
• F1 Macro: 0.8992

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	11.7047	0	13.5620	0.0496	0.0208
No log	1	499	4.2730	0.0078	15.1230	0.0408	0.0136
0.0664	2	998	3.1953	0.0156	16.8286	0.0612	0.0278
0.0593	3	1497	2.9059	0.0312	19.7461	0.0995	0.0652
0.1019	4	1996	1.2435	0.0625	25.1790	0.6371	0.6290
0.9594	5	2495	0.6716	0.125	35.9062	0.7853	0.7855
0.5914	6	2994	0.5146	0.25	56.8589	0.8339	0.8296
0.4185	7	3493	0.4210	0.5	98.4198	0.8654	0.8650
0.273	8.0	3992	0.3380	1.0	181.6549	0.8846	0.8846
0.22	9.0	4491	0.3767	1.0	181.1275	0.8868	0.8886
0.1379	10.0	4990	0.3871	1.0	181.3614	0.8959	0.8941
0.0981	11.0	5489	0.4224	1.0	181.1304	0.8972	0.8963
0.0823	12.0	5988	0.4070	1.0	181.3512	0.8999	0.8992

Framework versions

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