c2b2bddde563193b24de391967f5f1e6

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

• Loss: 0.1672
• Data Size: 1.0
• Epoch Runtime: 27.5816
• Accuracy: 0.9729
• F1 Macro: 0.9676

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	6.1477	0	1.4406	0.1792	0.0507
No log	1	170	15.6145	0.0078	1.8442	0.1604	0.0743
No log	2	340	1.9834	0.0156	2.4700	0.1812	0.1333
No log	3	510	1.8505	0.0312	3.6845	0.25	0.1973
No log	4	680	1.4484	0.0625	4.6974	0.4146	0.3252
0.1137	5	850	0.2997	0.125	6.5294	0.9146	0.7694
0.1137	6	1020	0.1504	0.25	9.3258	0.9625	0.9368
0.2359	7	1190	0.1121	0.5	14.8208	0.9625	0.9479
0.1451	8.0	1360	0.1814	1.0	26.9786	0.9667	0.9694
0.0716	9.0	1530	0.1547	1.0	27.0222	0.9688	0.9653
0.054	10.0	1700	0.2016	1.0	27.4355	0.975	0.9775
0.0383	11.0	1870	0.1672	1.0	27.5816	0.9729	0.9676

Framework versions

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