bert-finetuned-ner

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

• Loss: 0.1220
• Precision: 0.7806
• Recall: 0.8738
• F1: 0.8246
• Accuracy: 0.9617

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.5575	1.0	2844	0.2163	0.5293	0.7189	0.6097	0.9211
0.2173	2.0	5688	0.1625	0.6841	0.8002	0.7376	0.9421
0.165	3.0	8532	0.1599	0.6889	0.8593	0.7647	0.9423
0.1365	4.0	11376	0.1468	0.75	0.8655	0.8036	0.9535
0.1181	5.0	14220	0.1281	0.7686	0.8736	0.8178	0.9580
0.1031	6.0	17064	0.1220	0.7806	0.8738	0.8246	0.9617
0.0928	7.0	19908	0.1232	0.8008	0.8725	0.8351	0.9625
0.0852	8.0	22752	0.1301	0.8042	0.8804	0.8406	0.9634
0.0774	9.0	25596	0.1228	0.8128	0.8863	0.8480	0.9646
0.0713	10.0	28440	0.1302	0.8040	0.8956	0.8473	0.9630
0.0662	11.0	31284	0.1338	0.8215	0.8944	0.8564	0.9646
0.0615	12.0	34128	0.1385	0.8162	0.9016	0.8568	0.9647
0.0579	13.0	36972	0.1442	0.8111	0.9048	0.8554	0.9644
0.0548	14.0	39816	0.1447	0.8125	0.9003	0.8542	0.9645
0.0512	15.0	42660	0.1371	0.8360	0.8969	0.8654	0.9673
0.049	16.0	45504	0.1414	0.8288	0.9028	0.8642	0.9668
0.0466	17.0	48348	0.1386	0.8347	0.9016	0.8669	0.9670
0.0455	18.0	51192	0.1438	0.8432	0.8995	0.8705	0.9678
0.0437	19.0	54036	0.1463	0.8363	0.9014	0.8676	0.9669
0.0427	20.0	56880	0.1488	0.8358	0.9050	0.8690	0.9667

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0