bert-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: 0.2292
• Precision: 0.6382
• Recall: 0.7121
• F1: 0.6731
• Accuracy: 0.9680

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: 32
• eval_batch_size: 32
• 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: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.262	1.0	43	0.1853	0.2802	0.2442	0.2610	0.9527
0.1455	2.0	86	0.1038	0.6031	0.7069	0.6509	0.9705
0.0958	3.0	129	0.0981	0.6633	0.6787	0.6709	0.9722
0.0756	4.0	172	0.1011	0.6189	0.7558	0.6806	0.9699
0.0389	5.0	215	0.1058	0.6627	0.7172	0.6889	0.9715
0.0339	6.0	258	0.1236	0.6205	0.7147	0.6643	0.9665
0.0259	7.0	301	0.1170	0.6376	0.7326	0.6818	0.9698
0.0182	8.0	344	0.1389	0.6110	0.7429	0.6705	0.9668
0.0184	9.0	387	0.1368	0.6063	0.7404	0.6667	0.9651
0.0128	10.0	430	0.1403	0.6283	0.7301	0.6754	0.9683
0.0122	11.0	473	0.1407	0.6275	0.7404	0.6792	0.9677
0.0147	12.0	516	0.1505	0.5967	0.7455	0.6629	0.9663
0.01	13.0	559	0.1406	0.6167	0.7404	0.6729	0.9675
0.0079	14.0	602	0.1527	0.6473	0.7172	0.6805	0.9692
0.0112	15.0	645	0.1549	0.6545	0.7352	0.6925	0.9681
0.0061	16.0	688	0.1585	0.6432	0.7275	0.6828	0.9691
0.0086	17.0	731	0.1598	0.6507	0.7326	0.6892	0.9683
0.0077	18.0	774	0.1677	0.6611	0.7172	0.6880	0.9685
0.0053	19.0	817	0.1674	0.6351	0.7249	0.6771	0.9687
0.0049	20.0	860	0.1777	0.6675	0.7121	0.6891	0.9687
0.0088	21.0	903	0.1579	0.6578	0.7018	0.6791	0.9676
0.0085	22.0	946	0.1729	0.6618	0.6941	0.6775	0.9675
0.0062	23.0	989	0.1788	0.6395	0.7249	0.6795	0.9685
0.0052	24.0	1032	0.1782	0.6458	0.7172	0.6797	0.9683
0.0084	25.0	1075	0.1803	0.6345	0.7275	0.6778	0.9670
0.006	26.0	1118	0.1972	0.6154	0.7198	0.6635	0.9651
0.0045	27.0	1161	0.1852	0.625	0.7198	0.6691	0.9674
0.0035	28.0	1204	0.1847	0.6412	0.7121	0.6748	0.9680
0.0045	29.0	1247	0.1823	0.6675	0.6915	0.6793	0.9687
0.0094	30.0	1290	0.1962	0.6362	0.7147	0.6731	0.9682
0.0036	31.0	1333	0.2092	0.6319	0.7018	0.6650	0.9667
0.0045	32.0	1376	0.1872	0.6242	0.7301	0.6730	0.9650
0.0051	33.0	1419	0.2008	0.6112	0.7275	0.6643	0.9649
0.0057	34.0	1462	0.2018	0.6088	0.7481	0.6713	0.9662
0.003	35.0	1505	0.1941	0.6539	0.7044	0.6782	0.9680
0.0074	36.0	1548	0.1978	0.6741	0.7018	0.6877	0.9683
0.0045	37.0	1591	0.1940	0.6563	0.7069	0.6807	0.9674
0.0031	38.0	1634	0.2075	0.6220	0.7275	0.6706	0.9674
0.0058	39.0	1677	0.1979	0.6429	0.7172	0.6780	0.9678
0.0029	40.0	1720	0.2002	0.6447	0.7044	0.6732	0.9689
0.0041	41.0	1763	0.1962	0.6222	0.7069	0.6619	0.9678
0.0028	42.0	1806	0.2035	0.6298	0.7172	0.6707	0.9672
0.0033	43.0	1849	0.2208	0.6144	0.7249	0.6651	0.9668
0.0024	44.0	1892	0.2208	0.6330	0.7095	0.6691	0.9668
0.0043	45.0	1935	0.2250	0.5872	0.7095	0.6426	0.9647
0.0043	46.0	1978	0.2151	0.6425	0.6838	0.6625	0.9676
0.0054	47.0	2021	0.2121	0.6692	0.6761	0.6726	0.9690
0.0048	48.0	2064	0.1978	0.6231	0.7224	0.6690	0.9671
0.0049	49.0	2107	0.1963	0.6453	0.7249	0.6828	0.9689
0.0043	50.0	2150	0.2090	0.6683	0.7095	0.6883	0.9691
0.0032	51.0	2193	0.2017	0.6317	0.7275	0.6762	0.9679
0.0046	52.0	2236	0.2036	0.6409	0.7249	0.6803	0.9694
0.0052	53.0	2279	0.2047	0.6210	0.7455	0.6776	0.9676
0.0027	54.0	2322	0.1953	0.6359	0.7095	0.6707	0.9688
0.0048	55.0	2365	0.1935	0.6555	0.7044	0.6791	0.9701
0.0037	56.0	2408	0.1975	0.6212	0.7378	0.6745	0.9688
0.0064	57.0	2451	0.2016	0.6337	0.7249	0.6763	0.9690
0.0039	58.0	2494	0.2087	0.6152	0.7275	0.6667	0.9669
0.0027	59.0	2537	0.2056	0.6388	0.7275	0.6803	0.9679
0.0028	60.0	2580	0.2067	0.6421	0.7378	0.6866	0.9687
0.0031	61.0	2623	0.1963	0.6300	0.7352	0.6785	0.9685
0.0042	62.0	2666	0.2048	0.6207	0.7404	0.6753	0.9670
0.0034	63.0	2709	0.2000	0.6332	0.7455	0.6848	0.9689
0.004	64.0	2752	0.1914	0.6484	0.7301	0.6868	0.9692
0.0038	65.0	2795	0.1983	0.6185	0.7378	0.6729	0.9685
0.0039	66.0	2838	0.2068	0.6214	0.7301	0.6714	0.9683
0.003	67.0	2881	0.2129	0.6236	0.7198	0.6683	0.9685
0.0036	68.0	2924	0.2118	0.6131	0.7455	0.6729	0.9676
0.0033	69.0	2967	0.1997	0.6513	0.7249	0.6861	0.9691
0.003	70.0	3010	0.2066	0.6217	0.7224	0.6683	0.9686
0.0042	71.0	3053	0.2064	0.6201	0.7301	0.6706	0.9682
0.0029	72.0	3096	0.2113	0.6196	0.7326	0.6714	0.9676
0.0021	73.0	3139	0.2051	0.6341	0.7172	0.6731	0.9685
0.0035	74.0	3182	0.2059	0.6353	0.7121	0.6715	0.9681
0.0042	75.0	3225	0.2085	0.6304	0.7147	0.6699	0.9678
0.0038	76.0	3268	0.2137	0.6284	0.7172	0.6699	0.9676
0.0023	77.0	3311	0.2134	0.6231	0.7224	0.6690	0.9682
0.003	78.0	3354	0.2149	0.6467	0.7198	0.6813	0.9689
0.0034	79.0	3397	0.2121	0.6406	0.7147	0.6756	0.9685
0.0034	80.0	3440	0.2146	0.6407	0.7198	0.6780	0.9685
0.0033	81.0	3483	0.2162	0.6430	0.7224	0.6804	0.9685
0.0031	82.0	3526	0.2233	0.6264	0.7198	0.6699	0.9678
0.0043	83.0	3569	0.2279	0.6355	0.7172	0.6739	0.9678
0.0032	84.0	3612	0.2247	0.6357	0.7224	0.6763	0.9682
0.0046	85.0	3655	0.2240	0.6495	0.7147	0.6805	0.9683
0.0047	86.0	3698	0.2262	0.6284	0.7172	0.6699	0.9684
0.0036	87.0	3741	0.2214	0.6435	0.7147	0.6772	0.9682
0.0034	88.0	3784	0.2199	0.6353	0.7121	0.6715	0.9685
0.0034	89.0	3827	0.2231	0.6414	0.7172	0.6772	0.9682
0.0024	90.0	3870	0.2239	0.6427	0.7121	0.6756	0.9683
0.0019	91.0	3913	0.2243	0.6397	0.7121	0.6740	0.9681
0.0032	92.0	3956	0.2264	0.6333	0.7147	0.6715	0.9680
0.0021	93.0	3999	0.2276	0.6304	0.7147	0.6699	0.9680
0.0029	94.0	4042	0.2277	0.6339	0.7121	0.6707	0.9680
0.0039	95.0	4085	0.2281	0.6353	0.7121	0.6715	0.9680
0.0021	96.0	4128	0.2289	0.6368	0.7121	0.6723	0.9681
0.0027	97.0	4171	0.2292	0.6382	0.7121	0.6731	0.9680
0.0028	98.0	4214	0.2289	0.6382	0.7121	0.6731	0.9682
0.0027	99.0	4257	0.2291	0.6382	0.7121	0.6731	0.9682
0.002	100.0	4300	0.2292	0.6382	0.7121	0.6731	0.9680

Framework versions

• Transformers 4.46.1
• Pytorch 1.13.1+cu116
• Datasets 3.1.0
• Tokenizers 0.20.1