bert-finetuned-ner-cfv

This model is a fine-tuned version of bert-base-cased on the conll2002 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1851
• Precision: 0.8077
• Recall: 0.8212
• F1: 0.8144
• Accuracy: 0.9741

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: 4e-05
• train_batch_size: 24
• eval_batch_size: 24
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 17

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	347	0.1278	0.7284	0.7475	0.7378	0.9646
0.1176	2.0	694	0.1212	0.7509	0.7806	0.7654	0.9681
0.0453	3.0	1041	0.1156	0.8062	0.8116	0.8089	0.9730
0.0453	4.0	1388	0.1270	0.8081	0.8031	0.8056	0.9720
0.0233	5.0	1735	0.1298	0.8145	0.8231	0.8187	0.9746
0.0145	6.0	2082	0.1431	0.7950	0.8091	0.8020	0.9728
0.0145	7.0	2429	0.1501	0.8103	0.8166	0.8135	0.9734
0.009	8.0	2776	0.1553	0.8118	0.8157	0.8138	0.9738
0.0061	9.0	3123	0.1572	0.7891	0.8084	0.7986	0.9720
0.0061	10.0	3470	0.1589	0.8142	0.8196	0.8169	0.9739
0.005	11.0	3817	0.1671	0.8092	0.8148	0.8120	0.9733
0.0032	12.0	4164	0.1716	0.8066	0.8139	0.8102	0.9733
0.0031	13.0	4511	0.1767	0.8025	0.8169	0.8096	0.9731
0.0031	14.0	4858	0.1756	0.8096	0.8217	0.8156	0.9741
0.0023	15.0	5205	0.1845	0.8109	0.8157	0.8133	0.9739
0.0018	16.0	5552	0.1850	0.8090	0.8203	0.8146	0.9739
0.0018	17.0	5899	0.1851	0.8077	0.8212	0.8144	0.9741

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1