bert-base-NER-finetuned-ner

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

• Loss: 0.2391
• Precision: 0.9245
• Recall: 0.9186
• F1: 0.9216
• Accuracy: 0.9168

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: 2e-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.37	100	0.5115	0.8204	0.8719	0.8454	0.8200
No log	0.75	200	0.3808	0.8684	0.8766	0.8725	0.8600
No log	1.12	300	0.3315	0.8900	0.8865	0.8882	0.8799
No log	1.49	400	0.3069	0.9036	0.8917	0.8976	0.8921
0.5306	1.87	500	0.2908	0.9066	0.8978	0.9022	0.8980
0.5306	2.24	600	0.2783	0.9114	0.9061	0.9087	0.9048
0.5306	2.61	700	0.2729	0.9151	0.9123	0.9137	0.9096
0.5306	2.99	800	0.2628	0.9157	0.9086	0.9121	0.9077
0.5306	3.36	900	0.2600	0.9207	0.9123	0.9165	0.9107
0.3037	3.73	1000	0.2539	0.9188	0.9134	0.9161	0.9110
0.3037	4.1	1100	0.2488	0.9229	0.9178	0.9203	0.9148
0.3037	4.48	1200	0.2449	0.9225	0.9170	0.9198	0.9146
0.3037	4.85	1300	0.2466	0.9230	0.9177	0.9203	0.9155
0.3037	5.22	1400	0.2415	0.9229	0.9188	0.9208	0.9161
0.2668	5.6	1500	0.2413	0.9237	0.9189	0.9213	0.9164
0.2668	5.97	1600	0.2391	0.9245	0.9186	0.9216	0.9168
0.2668	6.34	1700	0.2399	0.9245	0.9178	0.9211	0.9162
0.2668	6.72	1800	0.2369	0.9239	0.9181	0.9210	0.9164
0.2668	7.09	1900	0.2390	0.9239	0.9183	0.9211	0.9164

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2