f9b04855a15c4e1ee578a846cbe843de

This model is a fine-tuned version of google-bert/bert-base-uncased on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 0.2729
• Data Size: 1.0
• Epoch Runtime: 9.3364
• Accuracy: 0.9521
• F1 Macro: 0.9127

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	1.9881	0	0.8013	0.1333	0.0393
No log	1	170	1.7450	0.0078	1.0081	0.2417	0.1701
No log	2	340	1.7545	0.0156	1.0634	0.1375	0.0464
No log	3	510	1.5215	0.0312	1.4727	0.5563	0.3592
No log	4	680	1.2185	0.0625	1.8107	0.4688	0.3429
0.0762	5	850	0.4806	0.125	2.4072	0.9104	0.7648
0.0762	6	1020	0.2103	0.25	3.4776	0.9437	0.7916
0.1962	7	1190	0.1975	0.5	5.4365	0.9563	0.9433
0.1472	8.0	1360	0.1064	1.0	9.6251	0.9771	0.9804
0.0997	9.0	1530	0.1224	1.0	9.4786	0.9667	0.9496
0.0532	10.0	1700	0.1818	1.0	9.9117	0.9542	0.9467
0.0288	11.0	1870	0.2221	1.0	10.0296	0.9646	0.9550
0.0109	12.0	2040	0.2729	1.0	9.3364	0.9521	0.9127

Framework versions

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