bert-base-uncased-grammar-ner-generic

This model is a fine-tuned version of bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1085
• Accuracy: 0.9887
• F1 Macro: 0.9103
• F1 Micro: 0.9103
• Precision Macro: 0.9448
• Precision Micro: 0.9448
• Recall Macro: 0.8782
• Recall Micro: 0.8782

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: 24
• eval_batch_size: 24
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Micro	Precision Macro	Precision Micro	Recall Macro	Recall Micro
0.0121	1.0	93	0.1199	0.9821	0.8430	0.8430	0.9373	0.9373	0.7660	0.7660
0.0124	2.0	186	0.0898	0.9832	0.8448	0.8448	0.8435	0.8435	0.8462	0.8462
0.0176	3.0	279	0.0690	0.9871	0.8797	0.8797	0.9051	0.9051	0.8558	0.8558
0.0182	4.0	372	0.0805	0.9858	0.8793	0.8793	0.9078	0.9078	0.8526	0.8526
0.01	5.0	465	0.0816	0.9824	0.8465	0.8465	0.8198	0.8198	0.875	0.875
0.0075	6.0	558	0.0887	0.9866	0.8867	0.8867	0.8954	0.8954	0.8782	0.8782
0.0038	7.0	651	0.1157	0.9839	0.8701	0.8701	0.8502	0.8502	0.8910	0.8910
0.0031	8.0	744	0.1000	0.9877	0.9034	0.9034	0.9231	0.9231	0.8846	0.8846
0.0017	9.0	837	0.1097	0.9877	0.9010	0.9010	0.9286	0.9286	0.875	0.875
0.0013	10.0	930	0.1044	0.9894	0.9161	0.9161	0.9613	0.9613	0.875	0.875
0.001	11.0	1023	0.1069	0.9889	0.9118	0.9118	0.9481	0.9481	0.8782	0.8782
0.0003	12.0	1116	0.1085	0.9887	0.9103	0.9103	0.9448	0.9448	0.8782	0.8782

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3