martiansideofthemoon / style-transfer-paraphrase

Official code and data repository for our EMNLP 2020 long paper "Reformulating Unsupervised Style Transfer as Paraphrase Generation" (https://arxiv.org/abs/2010.05700).

Home Page:http://style.cs.umass.edu

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Classifier training has zero LR after 50% training

martiansideofthemoon opened this issue · comments

https://github.com/martiansideofthemoon/style-transfer-paraphrase/blob/master/style_paraphrase/style_classify/examples/train_cds.sh#L54

This is a mismatch from the template script, which would assign an update frequency of 8. Should not lead to a major change in performance though. More details in an email sent to kalpesh@cs.umass.edu (thread date Nov 16, 2020, email date Dec 16, 2020). Originally reported by @philno