bert-base-cased-NER-favsbot

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

• Loss: 0.1680
• Precision: 0.8462
• Recall: 0.88
• F1: 0.8627
• Accuracy: 0.9444

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: 1.5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	7	1.8761	0.0	0.0	0.0	0.5833
No log	2.0	14	1.3530	0.0	0.0	0.0	0.5972
No log	3.0	21	1.0400	1.0	0.12	0.2143	0.6389
No log	4.0	28	0.7987	0.7895	0.6	0.6818	0.8194
No log	5.0	35	0.6055	0.85	0.68	0.7556	0.875
No log	6.0	42	0.4749	0.8696	0.8	0.8333	0.9167
No log	7.0	49	0.3838	0.84	0.84	0.8400	0.9444
No log	8.0	56	0.3084	0.88	0.88	0.88	0.9583
No log	9.0	63	0.2643	0.88	0.88	0.88	0.9583
No log	10.0	70	0.2360	0.8462	0.88	0.8627	0.9444
No log	11.0	77	0.2168	0.8462	0.88	0.8627	0.9444
No log	12.0	84	0.2031	0.8462	0.88	0.8627	0.9444
No log	13.0	91	0.1937	0.88	0.88	0.88	0.9583
No log	14.0	98	0.1853	0.8462	0.88	0.8627	0.9444
No log	15.0	105	0.1791	0.8462	0.88	0.8627	0.9444
No log	16.0	112	0.1757	0.8462	0.88	0.8627	0.9444
No log	17.0	119	0.1718	0.8462	0.88	0.8627	0.9444
No log	18.0	126	0.1698	0.8148	0.88	0.8462	0.9444
No log	19.0	133	0.1686	0.8148	0.88	0.8462	0.9444
No log	20.0	140	0.1680	0.8462	0.88	0.8627	0.9444

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.1
• Datasets 2.4.0
• Tokenizers 0.12.1