8000 GitHub - TinyTalks/CILLMs: CILLMs
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

TinyTalks/CILLMs

Repository files navigation

CILLMs —— Chinese Internet Large Language Models

Evaluated Datase

Sentiment Detection

数据集 解释 来源 链接
SMP2020-EWECT SMP2020微博情绪分类评测 中国中文信息学会社会媒体处理专业委员会(CIPS-SMP) https://smp2020ewect.github.io
SMP2019_ECISA SMP2019中文隐式情感分析评测 中国中文信息学会社会媒体处理专委会 http://www.cips-smp.org/smp_data/5
DDmkTCECT 弹幕情感标注语料 TinkTalks https://github.com/TinyTalks/DDmkTCCorpus

NER

数据集 解释 来源 链接
CLUENER2020 清华大学开源的文本分类数据集THUCNEWS,进行筛选过滤、实体标注生成的 清华大学 https://github.com/CLUEbenchmark/CLUENER2020
DDmkTCNER 弹幕命名实体识别标注语料 TinkTalks https://github.com/TinyTalks/DDmkTCCorpus

Chinese Internet Large Language Models

BERT

RoBERTa

chinese_danmaku_roberta

  • This model is a fine-tuned version of uer/chinese_roberta_L-8_H-512 on an Danmaku Corpus(2000k raw data) dataset.
Mask Accuracy Link
0.7780 https://huggingface.co/WUJUNCHAO/chinese_danmaku_roberta
Dataset Accuracy Precision Recall F1
DDmkTCECT 0.89 0.86 0.88 0.87

T5

About

CILLMs

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  
0