4d23fa94893cd3634ba61f07ba0e7d07

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

• Loss: 0.4272
• Data Size: 1.0
• Epoch Runtime: 340.6608
• Accuracy: 0.9012
• F1 Macro: 0.9009

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	3.2826	0	24.7062	0.0524	0.0179
No log	1	499	3.0783	0.0078	27.6106	0.0696	0.0389
0.0314	2	998	2.6413	0.0156	31.5399	0.2009	0.1453
0.0519	3	1497	0.9160	0.0312	37.5638	0.6993	0.6878
0.0434	4	1996	0.7957	0.0625	47.6808	0.7427	0.7444
0.7257	5	2495	0.6020	0.125	67.4073	0.8112	0.8091
0.4597	6	2994	0.4780	0.25	107.6878	0.8531	0.8508
0.332	7	3493	0.3830	0.5	184.9563	0.8818	0.8809
0.2277	8.0	3992	0.3136	1.0	338.9666	0.8949	0.8954
0.1623	9.0	4491	0.3982	1.0	340.3727	0.8881	0.8888
0.1027	10.0	4990	0.3985	1.0	339.2017	0.9022	0.9002
0.1023	11.0	5489	0.3942	1.0	339.8123	0.9103	0.9101
0.0789	12.0	5988	0.4272	1.0	340.6608	0.9012	0.9009

Framework versions

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