bert_bilstm_mega_crf-ner-weibo

This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2341
• Precision: 0.6657
• Recall: 0.7075
• F1: 0.6860
• Accuracy: 0.9683

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: 5e-05
• train_batch_size: 128
• eval_batch_size: 128
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.7329	1.0	11	0.4907	0.0	0.0	0.0	0.9274
0.4493	2.0	22	0.3486	0.0	0.0	0.0	0.9274
0.3203	3.0	33	0.2384	0.2941	0.0629	0.1036	0.9354
0.2259	4.0	44	0.1618	0.5219	0.4874	0.5041	0.9586
0.1617	5.0	55	0.1318	0.5476	0.5975	0.5714	0.9642
0.1171	6.0	66	0.1202	0.5718	0.6509	0.6088	0.9676
0.0956	7.0	77	0.1210	0.6022	0.6855	0.6412	0.9692
0.0666	8.0	88	0.1208	0.5951	0.6887	0.6385	0.9690
0.0567	9.0	99	0.1205	0.5963	0.7107	0.6485	0.9687
0.0433	10.0	110	0.1219	0.6230	0.7170	0.6667	0.9699
0.0333	11.0	121	0.1365	0.6375	0.6635	0.6502	0.9687
0.0309	12.0	132	0.1421	0.6011	0.6918	0.6433	0.9672
0.0239	13.0	143	0.1460	0.6398	0.6981	0.6677	0.9687
0.0235	14.0	154	0.1539	0.6518	0.6887	0.6697	0.9687
0.0188	15.0	165	0.1604	0.6656	0.6824	0.6739	0.9694
0.0193	16.0	176	0.1625	0.6471	0.6918	0.6687	0.9687
0.0155	17.0	187	0.1758	0.6770	0.6855	0.6813	0.9683
0.0148	18.0	198	0.1714	0.6506	0.6792	0.6646	0.9688
0.014	19.0	209	0.1626	0.6391	0.7296	0.6814	0.9674
0.0116	20.0	220	0.1718	0.6459	0.7170	0.6796	0.9687
0.0111	21.0	231	0.1840	0.6718	0.6824	0.6771	0.9694
0.0097	22.0	242	0.1807	0.6479	0.6887	0.6677	0.9677
0.0098	23.0	253	0.1787	0.6391	0.7296	0.6814	0.9664
0.0089	24.0	264	0.1877	0.6518	0.6887	0.6697	0.9688
0.0077	25.0	275	0.1896	0.6519	0.6950	0.6728	0.9693
0.008	26.0	286	0.1915	0.6608	0.7107	0.6848	0.9690
0.0079	27.0	297	0.2008	0.6606	0.6792	0.6698	0.9687
0.0072	28.0	308	0.1961	0.6486	0.7138	0.6796	0.9681
0.0067	29.0	319	0.2040	0.6617	0.7013	0.6809	0.9691
0.0063	30.0	330	0.2028	0.6725	0.7296	0.6998	0.9688
0.0056	31.0	341	0.2053	0.6716	0.7201	0.6950	0.9689
0.0073	32.0	352	0.2088	0.6465	0.6730	0.6595	0.9674
0.0061	33.0	363	0.1936	0.6138	0.7296	0.6667	0.9673
0.0057	34.0	374	0.2061	0.6596	0.6824	0.6708	0.9683
0.0062	35.0	385	0.2077	0.6627	0.7044	0.6829	0.9680
0.0046	36.0	396	0.2133	0.6738	0.6950	0.6842	0.9689
0.0062	37.0	407	0.2029	0.6696	0.7201	0.6939	0.9680
0.0058	38.0	418	0.2039	0.6707	0.7044	0.6871	0.9678
0.0047	39.0	429	0.2055	0.6667	0.7233	0.6938	0.9685
0.0049	40.0	440	0.2105	0.6757	0.7075	0.6912	0.9692
0.0048	41.0	451	0.2052	0.6667	0.7107	0.6880	0.9683
0.0049	42.0	462	0.2081	0.6590	0.7170	0.6867	0.9687
0.0063	43.0	473	0.2011	0.6552	0.7170	0.6847	0.9683
0.0046	44.0	484	0.1994	0.6477	0.7170	0.6806	0.9676
0.0047	45.0	495	0.2122	0.6790	0.6918	0.6854	0.9693
0.0048	46.0	506	0.2082	0.6609	0.7233	0.6907	0.9687
0.0042	47.0	517	0.2140	0.6769	0.6918	0.6843	0.9695
0.0054	48.0	528	0.2054	0.6514	0.7170	0.6826	0.9681
0.0037	49.0	539	0.2070	0.6686	0.7107	0.6890	0.9689
0.0045	50.0	550	0.2093	0.6514	0.7170	0.6826	0.9686
0.004	51.0	561	0.2163	0.6787	0.7107	0.6943	0.9698
0.0038	52.0	572	0.2173	0.6706	0.7107	0.6901	0.9694
0.0042	53.0	583	0.2156	0.6745	0.7233	0.6980	0.9694
0.0039	54.0	594	0.2190	0.6727	0.6981	0.6852	0.9689
0.0037	55.0	605	0.2213	0.6767	0.7044	0.6903	0.9687
0.0043	56.0	616	0.2247	0.6829	0.7044	0.6935	0.9690
0.0034	57.0	627	0.2291	0.6789	0.6981	0.6884	0.9689
0.0046	58.0	638	0.2258	0.6737	0.7075	0.6902	0.9686
0.0033	59.0	649	0.2254	0.6736	0.7138	0.6931	0.9689
0.0036	60.0	660	0.2255	0.6758	0.7013	0.6883	0.9690
0.0038	61.0	671	0.2200	0.6580	0.7138	0.6848	0.9682
0.0036	62.0	682	0.2210	0.6657	0.7075	0.6860	0.9687
0.0039	63.0	693	0.2237	0.6647	0.7107	0.6869	0.9682
0.0039	64.0	704	0.2295	0.6727	0.6981	0.6852	0.9688
0.0032	65.0	715	0.2271	0.6707	0.7044	0.6871	0.9687
0.0038	66.0	726	0.2290	0.6677	0.7013	0.6840	0.9687
0.0033	67.0	737	0.2260	0.6617	0.7013	0.6809	0.9682
0.0038	68.0	748	0.2250	0.6676	0.7138	0.6900	0.9686
0.0037	69.0	759	0.2254	0.6618	0.7075	0.6839	0.9684
0.0039	70.0	770	0.2281	0.6687	0.6981	0.6831	0.9687
0.0036	71.0	781	0.2317	0.6687	0.6981	0.6831	0.9687
0.0034	72.0	792	0.2272	0.6609	0.7170	0.6878	0.9686
0.0036	73.0	803	0.2278	0.6756	0.7138	0.6942	0.9687
0.0035	74.0	814	0.2287	0.6677	0.7075	0.6870	0.9683
0.0034	75.0	825	0.2283	0.6686	0.7107	0.6890	0.9681
0.0032	76.0	836	0.2331	0.6657	0.7075	0.6860	0.9672
0.0041	77.0	847	0.2357	0.6598	0.7075	0.6829	0.9675
0.0033	78.0	858	0.2352	0.6706	0.7170	0.6930	0.9676
0.0039	79.0	869	0.2363	0.6696	0.7075	0.6881	0.9689
0.0036	80.0	880	0.2367	0.6627	0.6918	0.6769	0.9685
0.0032	81.0	891	0.2369	0.6607	0.6981	0.6789	0.9683
0.0036	82.0	902	0.2331	0.6696	0.7201	0.6939	0.9687
0.0036	83.0	913	0.2286	0.6599	0.7138	0.6858	0.9682
0.0034	84.0	924	0.2276	0.6637	0.7138	0.6879	0.9687
0.0035	85.0	935	0.2286	0.6647	0.7107	0.6869	0.9687
0.0031	86.0	946	0.2296	0.6667	0.7044	0.6850	0.9689
0.0036	87.0	957	0.2296	0.6677	0.7075	0.6870	0.9687
0.0033	88.0	968	0.2299	0.6706	0.7170	0.6930	0.9688
0.0033	89.0	979	0.2301	0.6618	0.7138	0.6868	0.9683
0.0034	90.0	990	0.2320	0.6766	0.7170	0.6962	0.9687
0.0031	91.0	1001	0.2309	0.6766	0.7170	0.6962	0.9686
0.0033	92.0	1012	0.2315	0.6736	0.7138	0.6931	0.9685
0.0037	93.0	1023	0.2333	0.6696	0.7075	0.6881	0.9684
0.0031	94.0	1034	0.2342	0.6696	0.7075	0.6881	0.9684
0.0029	95.0	1045	0.2351	0.6687	0.7044	0.6861	0.9683
0.004	96.0	1056	0.2347	0.6667	0.7044	0.6850	0.9683
0.0032	97.0	1067	0.2346	0.6667	0.7044	0.6850	0.9683
0.0033	98.0	1078	0.2343	0.6667	0.7044	0.6850	0.9683
0.0032	99.0	1089	0.2341	0.6647	0.7044	0.6840	0.9682
0.0034	100.0	1100	0.2341	0.6657	0.7075	0.6860	0.9683

Framework versions

• Transformers 4.46.2
• Pytorch 2.4.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3