高性能小模型排行榜*记最佳得分(模型名称需带有nlpcc标识，模型参数低于1/9的bert-base参数量)

排行	模型	研究机构	测评时间	Score	认证	CLUEWSC 2020	CSL	CLUENER	CMRC2018
1	nlpcc-tiny-NEZHA-8L	Huawei Cloud & Noah's Ark lab	20-08-25	78.897	待认证	80.345	82.933	78.058	74.250
2	nlpcc-tiny-NEZHA-6L	Huawei Cloud & Noah's Ark lab	20-09-08	77.587	待认证	80.690	81.200	76.459	72.000
3	nlpcc-tiny-bert-6L	Xiaomi AI Lab	20-06-28	76.860	待认证	80.690	78.567	79.082	69.100
4	nlpcc-bert-base-test	mi-ailab-nlp-app	20-03-31	76.047	待认证	74.828	81.267	77.441	70.650
5	nlpcc-bert-tiny	TencentPretrain & TI-ONE	20-05-20	73.507	待认证	73.448	74.167	75.613	70.800
6	nlpcc-tiny-bert	Xiaomi AI Lab	20-05-27	73.114	待认证	67.931	78.567	76.858	69.100
7	nlpcc-tiny-NEZHA-4L	Huawei Cloud & Noah's Ark lab	20-05-09	72.514	待认证	66.207	79.767	75.583	68.500
8	nlpcc-xiaomi-tiny-6L	mi-ailab-nlp-app	20-05-03	72.045	待认证	67.931	79.033	75.166	66.050
9	nlpcc-bert-small	Tencent TEG	20-05-01	69.704	待认证	63.448	76.767	74.851	63.750
10	nlpcc-RoBERTa-tiny-clue	NUDT	20-06-29	69.298	待认证	64.828	78.700	67.414	66.250
11	nlpcc-tiny-bert-6L	mi-ailab-nlp-app	20-04-25	68.577	待认证	62.414	76.733	73.262	61.900
12	nlpcc-BERT-4T	Huawei-ICS-NLP	20-05-20	67.885	待认证	63.793	75.500	72.996	59.250
13	nlpcc-roberta_tiny	mi-ailab-nlp-app	20-04-07	67.421	待认证	62.414	72.733	72.287	62.250
14	nlpcc-tiny	mi-ailab-nlp-app	20-04-02	67.237	待认证	62.759	72.433	71.507	62.250
15	nlpcc-roberta-tiny-clue-baseline1	test-baseline1	20-03-31	66.580	待认证	60.690	72.033	71.745	61.850
16	nlpcc-RoBERTa-tiny-clue	dlnu-5.20	20-05-20	63.423	待认证	62.414	66.000	66.429	58.850
17	nlpcc-RoBERTa-tiny-clue	dlnu-5.19	20-05-19	63.337	待认证	62.069	66.000	66.429	58.850
18	nlpcc-RoBERTa-tiny-clue	dlnu-xiaomi	20-05-15	62.043	待认证	62.414	66.000	60.908	58.850
19	nlpcc-RoBERTa-tiny-clue	民大F4	20-05-19	62.043	待认证	62.414	66.000	60.908	58.850
20	nlpcc-roberta-tiny-clue	中国平安财产险科技中心	20-05-11	61.191	待认证	60.690	68.567	60.805	54.700

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

	C	E	N
C	0.0	0.0	0.0
E	0.0	0.0	0.0
N	0.0	0.0	0.0

ALBERT(Ensemble)

GitHub/模型网址：

提交日期：9月17日

分数：9月17日

更多详情：

型号说明

阿尔伯特模型集合

参数说明

单任务微调。我们从MNLI为RTE、STS和MRPC优化的模型开始

总参数:-1

共享参数:-1

诊断信息

诊断主混淆矩阵

	C	N	E
C	182	36	40
N	81	189	116
E	17	69	374

C = 对立

N = 不包含

E = 包含

类别相关马修分数