bert_bilstm_dst_crf-ner-weibo

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

• Loss: 0.2064
• Precision: 0.6286
• Recall: 0.7224
• F1: 0.6722
• Accuracy: 0.9691

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.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: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4101	1.0	22	0.3430	0.0	0.0	0.0	0.9330
0.2448	2.0	44	0.1469	0.5153	0.4756	0.4947	0.9626
0.138	3.0	66	0.1119	0.5918	0.7044	0.6432	0.9715
0.0899	4.0	88	0.1064	0.5565	0.6967	0.6187	0.9699
0.0616	5.0	110	0.1064	0.5978	0.6915	0.6412	0.9716
0.0553	6.0	132	0.1112	0.6078	0.6812	0.6424	0.9702
0.0396	7.0	154	0.1165	0.6366	0.7249	0.6779	0.9705
0.0343	8.0	176	0.1204	0.6208	0.7069	0.6611	0.9689
0.0274	9.0	198	0.1365	0.6191	0.7481	0.6775	0.9674
0.0291	10.0	220	0.1403	0.6288	0.6838	0.6552	0.9689
0.0199	11.0	242	0.1415	0.6330	0.7095	0.6691	0.9688
0.0204	12.0	264	0.1447	0.5979	0.7224	0.6542	0.9685
0.0162	13.0	286	0.1499	0.5822	0.7378	0.6508	0.9669
0.0163	14.0	308	0.1441	0.6138	0.7069	0.6571	0.9691
0.0156	15.0	330	0.1543	0.6157	0.7044	0.6571	0.9678
0.0107	16.0	352	0.1546	0.5957	0.7121	0.6487	0.9673
0.0134	17.0	374	0.1558	0.5860	0.7095	0.6419	0.9654
0.0103	18.0	396	0.1557	0.6030	0.7147	0.6541	0.9669
0.0087	19.0	418	0.1596	0.6031	0.6915	0.6443	0.9665
0.0094	20.0	440	0.1568	0.6105	0.6889	0.6473	0.9683
0.0106	21.0	462	0.1547	0.6561	0.6915	0.6733	0.9696
0.0088	22.0	484	0.1627	0.6483	0.6967	0.6716	0.9696
0.0077	23.0	506	0.1628	0.6059	0.7429	0.6674	0.9669
0.0076	24.0	528	0.1695	0.6174	0.6761	0.6454	0.9660
0.0081	25.0	550	0.1644	0.6387	0.7044	0.6699	0.9690
0.0066	26.0	572	0.1674	0.6225	0.7121	0.6643	0.9684
0.0067	27.0	594	0.1640	0.6281	0.7121	0.6675	0.9691
0.0065	28.0	616	0.1693	0.6091	0.7249	0.6620	0.9672
0.0063	29.0	638	0.1737	0.6299	0.7044	0.6650	0.9688
0.0141	30.0	660	0.1772	0.6205	0.7147	0.6643	0.9673
0.0064	31.0	682	0.1817	0.6233	0.7275	0.6714	0.9685
0.0082	32.0	704	0.1704	0.6392	0.6967	0.6667	0.9689
0.0051	33.0	726	0.1663	0.6236	0.7069	0.6627	0.9678
0.0041	34.0	748	0.1767	0.6278	0.7198	0.6707	0.9676
0.0053	35.0	770	0.1749	0.6529	0.6915	0.6717	0.9687
0.0066	36.0	792	0.1810	0.6382	0.7121	0.6731	0.9677
0.0044	37.0	814	0.1721	0.6351	0.7069	0.6691	0.9683
0.0043	38.0	836	0.1833	0.6283	0.7301	0.6754	0.9683
0.0047	39.0	858	0.1862	0.6176	0.7224	0.6659	0.9676
0.0038	40.0	880	0.1826	0.6106	0.7095	0.6564	0.9677
0.0045	41.0	902	0.1888	0.6069	0.7224	0.6596	0.9674
0.004	42.0	924	0.1862	0.6180	0.7069	0.6595	0.9682
0.0054	43.0	946	0.1903	0.6	0.7095	0.6502	0.9674
0.0052	44.0	968	0.1838	0.6379	0.7018	0.6683	0.9680
0.004	45.0	990	0.1850	0.6114	0.7198	0.6612	0.9676
0.0051	46.0	1012	0.1830	0.6412	0.7121	0.6748	0.9683
0.0045	47.0	1034	0.1939	0.6134	0.7301	0.6667	0.9683
0.0039	48.0	1056	0.1876	0.6559	0.6812	0.6683	0.9689
0.0041	49.0	1078	0.1904	0.6188	0.7095	0.6611	0.9675
0.0039	50.0	1100	0.1848	0.6242	0.7172	0.6675	0.9681
0.0043	51.0	1122	0.1823	0.6288	0.6967	0.6610	0.9685
0.0041	52.0	1144	0.1951	0.6137	0.7147	0.6603	0.9677
0.004	53.0	1166	0.1878	0.6026	0.7095	0.6517	0.9678
0.0047	54.0	1188	0.1843	0.6247	0.6889	0.6553	0.9687
0.0042	55.0	1210	0.1947	0.6132	0.7172	0.6611	0.9685
0.0039	56.0	1232	0.1902	0.6330	0.7095	0.6691	0.9690
0.0038	57.0	1254	0.1915	0.6339	0.7121	0.6707	0.9691
0.0035	58.0	1276	0.1887	0.6264	0.7198	0.6699	0.9686
0.0044	59.0	1298	0.1907	0.6247	0.7147	0.6667	0.9686
0.0026	60.0	1320	0.1927	0.6362	0.7147	0.6731	0.9687
0.004	61.0	1342	0.1904	0.6374	0.7095	0.6715	0.9689
0.0041	62.0	1364	0.1914	0.6222	0.7198	0.6675	0.9681
0.0037	63.0	1386	0.1878	0.6298	0.7172	0.6707	0.9684
0.0042	64.0	1408	0.1934	0.6074	0.7198	0.6588	0.9674
0.0047	65.0	1430	0.1992	0.6092	0.7172	0.6588	0.9676
0.0042	66.0	1452	0.1968	0.6186	0.7172	0.6643	0.9679
0.0038	67.0	1474	0.1970	0.6189	0.7224	0.6667	0.9683
0.0033	68.0	1496	0.1976	0.6173	0.7172	0.6635	0.9680
0.0037	69.0	1518	0.1983	0.6247	0.7147	0.6667	0.9684
0.0037	70.0	1540	0.1955	0.6247	0.7147	0.6667	0.9685
0.0038	71.0	1562	0.1970	0.6290	0.7147	0.6691	0.9682
0.0034	72.0	1584	0.2001	0.6242	0.7172	0.6675	0.9681
0.0039	73.0	1606	0.2023	0.6293	0.7069	0.6659	0.9676
0.0027	74.0	1628	0.2003	0.6381	0.7069	0.6707	0.9685
0.0037	75.0	1650	0.2009	0.6203	0.7224	0.6675	0.9683
0.0039	76.0	1672	0.2017	0.6275	0.7147	0.6683	0.9687
0.0035	77.0	1694	0.2016	0.6166	0.7275	0.6675	0.9688
0.0034	78.0	1716	0.2031	0.6108	0.7301	0.6651	0.9687
0.0028	79.0	1738	0.2029	0.6116	0.7326	0.6667	0.9682
0.003	80.0	1760	0.2036	0.6233	0.7275	0.6714	0.9683
0.0038	81.0	1782	0.2063	0.6303	0.7275	0.6754	0.9676
0.0042	82.0	1804	0.2040	0.6378	0.7198	0.6763	0.9685
0.0035	83.0	1826	0.2023	0.6149	0.7224	0.6643	0.9681
0.0033	84.0	1848	0.1991	0.6335	0.7198	0.6739	0.9685
0.0043	85.0	1870	0.2013	0.6306	0.7198	0.6723	0.9686
0.0036	86.0	1892	0.1988	0.6364	0.7018	0.6675	0.9694
0.0037	87.0	1914	0.2041	0.6217	0.7224	0.6683	0.9689
0.0031	88.0	1936	0.2043	0.6231	0.7224	0.6690	0.9689
0.0027	89.0	1958	0.2041	0.625	0.7198	0.6691	0.9688
0.0026	90.0	1980	0.2053	0.6284	0.7172	0.6699	0.9691
0.0031	91.0	2002	0.2049	0.6306	0.7198	0.6723	0.9690
0.003	92.0	2024	0.2056	0.6315	0.7224	0.6739	0.9687
0.0028	93.0	2046	0.2066	0.6149	0.7224	0.6643	0.9684
0.0031	94.0	2068	0.2075	0.6135	0.7224	0.6635	0.9684
0.0038	95.0	2090	0.2070	0.6198	0.7249	0.6682	0.9685
0.003	96.0	2112	0.2063	0.6253	0.7249	0.6714	0.9689
0.0028	97.0	2134	0.2062	0.6275	0.7275	0.6738	0.9692
0.0031	98.0	2156	0.2063	0.6272	0.7224	0.6714	0.9692
0.0026	99.0	2178	0.2062	0.6286	0.7224	0.6722	0.9691
0.002	100.0	2200	0.2064	0.6286	0.7224	0.6722	0.9691

Framework versions

• Transformers 4.46.1
• Pytorch 2.4.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.2