List your company
ResearchFormerly WizWand

Text Classification on TNEWS CLUE (test)

59.47AccuracyChineseBERT
9.893222.764135.63548.5059Jun 30, 2021Jul 11, 2021Jul 23, 2021Aug 4, 2021Aug 16, 2021Aug 28, 2021Sep 9, 2021
Updated 8mo ago

Evaluation results

MethodLinks
ChineseBERTModel Scale=Large, Pre-train Data=Standard2021.0659.47
ChineseBERTModel Scale=Base, Pre-train Data=Standard2021.0658.95
RoBERTaModel Scale=Large, Pre-train Data=Extended2021.0658.61
ERNIEModel Scale=Base, Pre-train Data=Standard2021.0658.33
RoBERTaModel Scale=Base, Pre-train Data=Extended2021.0656.94
BERTModel Scale=Base, Pre-train Data=Extended2021.0656.86
BERTModel Scale=Base, Pre-train Data=Standard2021.0656.58
Unified PPTTuning Mode=Prompt tuning, Parameters=410K2021.0950.6
FTTuning Mode=Full-model tuning, Parameters=11B2021.0943.2
PTTuning Mode=Prompt tuning, Parameters=410K2021.0941.2
PT (MC)Tuning Mode=Prompt tuning, Format=Multiple-choice classification format without prompt pre-training2021.0911.8