You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
Currently the mask generated by dropout op uses the same dtype and shape as the input. 1 bit per coordinate should be sufficient. Since dropout is heavily used after activations, this will save a lot of memory for training.
Hey, this is the MXNet Label Bot.
Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it.
Here are my recommended labels: Feature
Currently the mask generated by dropout op uses the same dtype and shape as the input. 1 bit per coordinate should be sufficient. Since dropout is heavily used after activations, this will save a lot of memory for training.
@ptrendx @DickJC123 @PatricZhao @TaoLv
The text was updated successfully, but these errors were encountered: