search

CLUEbenchmark / CLUE

中文语言理解测评基准 Chinese Language Understanding Evaluation Benchmark: datasets, baselines, pre-trained models, corpus and leaderboard
http://www.CLUEbenchmarks.com
4k stars 540 forks source link

bert-base在wsc任务的性能与readme里的好像有些差异 #138

Open wangwenqian21 opened 2 years ago

wangwenqian21 commented 2 years ago

用bert-base-chinese在wsc任务上跑出来有74.1，和readme里给的62相差较大，请问是数据更新了吗？