9417b5cbee076c45cbdb02eab59f22d8

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

• Loss: 0.2087
• Data Size: 1.0
• Epoch Runtime: 8.8965
• Accuracy: 0.9729
• F1 Macro: 0.9679

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	1.8773	0	0.7921	0.1333	0.0392
No log	1	170	1.7371	0.0078	1.0690	0.2708	0.1290
No log	2	340	1.6297	0.0156	1.1771	0.2979	0.1181
No log	3	510	1.5630	0.0312	1.5610	0.4146	0.2226
No log	4	680	1.4098	0.0625	1.8137	0.4062	0.2603
0.085	5	850	0.6977	0.125	2.4598	0.8708	0.7276
0.085	6	1020	0.2034	0.25	3.3882	0.9521	0.9274
0.2278	7	1190	0.2597	0.5	5.2150	0.9187	0.9152
0.1507	8.0	1360	0.1433	1.0	9.4180	0.9729	0.9739
0.0932	9.0	1530	0.1189	1.0	9.3945	0.9667	0.9568
0.0726	10.0	1700	0.1368	1.0	9.6073	0.9708	0.9670
0.0301	11.0	1870	0.1928	1.0	9.6883	0.9708	0.9738
0.0227	12.0	2040	0.1731	1.0	8.8580	0.9729	0.9685
0.0157	13.0	2210	0.2087	1.0	8.8965	0.9729	0.9679

Framework versions

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