bert-base-chinese-finetuned-ner

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

• Loss: 0.1016
• Precision: 0.9146
• Recall: 0.9414
• F1: 0.9278
• Accuracy: 0.9751

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-05
• train_batch_size: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	2	0.9181	0.1271	0.1255	0.1263	0.7170
No log	2.0	4	0.8048	0.1919	0.2385	0.2127	0.7669
No log	3.0	6	0.7079	0.2422	0.3264	0.2781	0.7980
No log	4.0	8	0.6201	0.3505	0.4854	0.4070	0.8338
No log	5.0	10	0.5462	0.3898	0.4812	0.4307	0.8611
No log	6.0	12	0.4851	0.4749	0.5941	0.5279	0.8802
No log	7.0	14	0.4338	0.5213	0.6151	0.5643	0.8936
No log	8.0	16	0.3843	0.5663	0.6611	0.6100	0.9076
No log	9.0	18	0.3451	0.6255	0.6987	0.6601	0.9214
No log	10.0	20	0.3058	0.6719	0.7197	0.6949	0.9293
No log	11.0	22	0.2783	0.6808	0.7406	0.7094	0.9344
No log	12.0	24	0.2497	0.7050	0.7699	0.7360	0.9427
No log	13.0	26	0.2235	0.7519	0.8117	0.7807	0.9506
No log	14.0	28	0.2031	0.7713	0.8326	0.8008	0.9552
No log	15.0	30	0.1861	0.7915	0.8577	0.8233	0.9593
No log	16.0	32	0.1726	0.8031	0.8703	0.8353	0.9613
No log	17.0	34	0.1619	0.8320	0.8912	0.8606	0.9641
No log	18.0	36	0.1521	0.8571	0.9038	0.8798	0.9674
No log	19.0	38	0.1420	0.8710	0.9038	0.8871	0.9695
No log	20.0	40	0.1352	0.8795	0.9163	0.8975	0.9700
No log	21.0	42	0.1281	0.8755	0.9121	0.8934	0.9712
No log	22.0	44	0.1209	0.8916	0.9289	0.9098	0.9728
No log	23.0	46	0.1155	0.8924	0.9372	0.9143	0.9733
No log	24.0	48	0.1115	0.904	0.9456	0.9243	0.9746
No log	25.0	50	0.1087	0.9116	0.9498	0.9303	0.9746
No log	26.0	52	0.1068	0.9146	0.9414	0.9278	0.9740
No log	27.0	54	0.1054	0.9146	0.9414	0.9278	0.9743
No log	28.0	56	0.1036	0.9146	0.9414	0.9278	0.9743
No log	29.0	58	0.1022	0.9146	0.9414	0.9278	0.9746
No log	30.0	60	0.1016	0.9146	0.9414	0.9278	0.9751

Framework versions

• Transformers 4.16.2
• Pytorch 1.10.0+cu111
• Datasets 1.18.3
• Tokenizers 0.11.0