[Frontend][Keras] Make keras reshape less restrictive #7446
Conversation
This file contains 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
trevor-m
force-pushed
the
keras-reshape
branch
from
d365069
to
5d3ae19
Compare
|
@siju-samuel @jwfromm @mbrookhart Could you please review? |
Lokiiiiii
pushed a commit
to Lokiiiiii/tvm
that referenced
this pull request
Mar 2, 2021
trevor-m
pushed a commit
to neo-ai/tvm
that referenced
this pull request
Mar 2, 2021
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Outdated suggestions cannot be applied.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.
The previous implementation of reshape in the Keras frontend attempted to map the target shape, which is specified in NHWC format, to NCHW format. However, this approach had many limitations of what reshapes can be supported and introduced some complicated logic.
Instead, I am proposing to just transpose input to NHWC, apply the unmodified reshape, then transpose back to NCHW. This is simpler and can support more reshapes which are valid in keras.
I added some test cases for reshapes which were not supported in the previous implementation, but are now possible.