Fix input mutation in branched networks #38
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This fixes a bug in networks that have in-place operations (ReLU, ELU, LeakyReLU) right after branching.
Currently, a layer output that is used by several branches can be mutated by an activation function, since they all have
inplace=Truehardcoded.Standalone reproduction of the bug:
https://colab.research.google.com/drive/1VBRYou450YNgJdwiDZVRsCfJKUs4FiWh#scrollTo=mKvLPpKz29D
Change: I made the activation functions'
inplaceproperty depend on whether the input of the current node has a branching factor higher than one.Tests: I verified that this passes the tests and does not break conversion of models from
download_fixtures.sh.Example relevant network architecture: Openpilot's
supercombo.onnxmodel:Let me know if there is any way I can improve it!