模型:

microsoft/DialogRPT-width

英文

演示

请尝试这个 ➤➤➤ Colab Notebook Demo (click me!)

Context Response width score
I love NLP! Can anyone recommend a nice review paper? 0.701
I love NLP! Me too! 0.029

宽度得分预测了回复的可能性有多大。

DialogRPT-width

对话排序预训练转换器

对话回复有多大可能获得赞同?和/或回复??

这是 DialogRPT 学会预测的。这是一组对话回复排序模型,由 Microsoft Research NLP Group 提出,使用了一亿多条人类反馈数据进行训练。它可以用于通过重新排序生成的响应候选项来改进现有的对话生成模型(例如 DialoGPT )。

快速链接:

我们考虑了以下任务,并提供了相应的预训练模型。

Task Description Pretrained model
Human feedback given a context and its two human responses, predict...
updown ... which gets more upvotes? 1238321
width ... which gets more direct replies? this model
depth ... which gets longer follow-up thread? 1239321
Human-like (human vs fake) given a context and one human response, distinguish it with...
human_vs_rand ... a random human response 12310321
human_vs_machine ... a machine generated response 12311321

联系方式:

请在 our repo 上创建一个问题

引用:

@inproceedings{gao2020dialogrpt,
    title={Dialogue Response RankingTraining with Large-Scale Human Feedback Data},
    author={Xiang Gao and Yizhe Zhang and Michel Galley and Chris Brockett and Bill Dolan},
    year={2020},
    booktitle={EMNLP}
}