This repository has been archived by the owner on Jan 15, 2024. It is now read-only.

Mixed precision training for language model #847

Open

eric-haibin-lin opened this issue Jul 25, 2019 · 1 comment

Labels

bug

Member

eric-haibin-lin commented Jul 25, 2019 •

edited

Loading

automatic mixed precision training does not work for the language model training script. The error log is attached amp.log.

The code to reproduce the issue is here: https://github.com/dmlc/gluon-nlp/compare/master...eric-haibin-lin:lm?expand=1

Uncommenting line 173 should reproduce the error.

eric-haibin-lin added the bug label

Contributor

ptrendx commented Aug 6, 2019

MXNet PR fixing this error: apache/mxnet#15740

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.