Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Relay] Add TopPattern to nn.dropout #7685

Merged
merged 1 commit into from Mar 18, 2021

Conversation

comaniac
Copy link
Contributor

nn.dropout is usually simplified in inference, but not for training, so we need to make up this attribute to make sure the model with nn.dropout can pass all Relay passes.

cc @anijain2305

@masahi masahi merged commit 4976bb2 into apache:main Mar 18, 2021
@comaniac comaniac deleted the fuse_pattern_dropout branch March 18, 2021 05:34
trevor-m pushed a commit to trevor-m/tvm that referenced this pull request May 6, 2021
trevor-m pushed a commit to neo-ai/tvm that referenced this pull request May 11, 2021
@Lyken17
Copy link
Contributor

Lyken17 commented Nov 26, 2021

Was this related with https://discuss.tvm.apache.org/t/toppattern-has-not-been-registered-for-nn-dropout/11305?

@AndrewZhaoLuo mentioned that

Dropout isn’t really supported

But actually here is an registration.

@AndrewZhaoLuo
Copy link
Contributor

Hey Lyken, this isn't really registering compute, instead it's registering an attribute saying this operator should not be fused with anything if I understand it correctly. Don't know the full context though.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

5 participants