487dbc38ff2531112e3e2ce4f1de88aa

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

• Loss: 0.4323
• Data Size: 1.0
• Epoch Runtime: 69.5302
• Accuracy: 0.8962
• F1 Macro: 0.8965

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	9.2425	0	5.7339	0.0471	0.0117
No log	1	499	5.0235	0.0078	6.5192	0.0504	0.0168
0.0606	2	998	3.2964	0.0156	6.9029	0.0476	0.0193
0.0646	3	1497	2.9868	0.0312	8.0333	0.0620	0.0396
0.105	4	1996	2.6214	0.0625	9.9802	0.2218	0.1599
1.7195	5	2495	1.1128	0.125	13.9014	0.6530	0.6263
0.825	6	2994	0.7211	0.25	21.7331	0.7719	0.7681
0.5923	7	3493	0.5553	0.5	37.1279	0.8133	0.8161
0.4126	8.0	3992	0.3951	1.0	68.5513	0.8654	0.8660
0.2936	9.0	4491	0.4038	1.0	68.8398	0.8745	0.8722
0.1864	10.0	4990	0.4338	1.0	68.3976	0.8795	0.8758
0.1468	11.0	5489	0.4087	1.0	68.3846	0.8851	0.8851
0.1407	12.0	5988	0.4323	1.0	69.5302	0.8962	0.8965

Framework versions

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