Shape inference for NMS #2269
Merged
Shape inference for NMS #2269
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
stevenlix
approved these changes
Aug 29, 2019
|
@ebarsoum @gramalingam - @stevenlix said he wanted this logic. So could you kindly take a look when you have a couple of minutes please ? |
gramalingam
approved these changes
Sep 12, 2019
|
@gramalingam - thanks for reviewing. I resolved conflict with master. Please let me know if you have any concerns merging. |
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Outdated suggestions cannot be applied.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.
Implement shape inference for NMS op and add corresponding test.
CC: @linkerzhang @gramalingam @stevenlix
One open question I had was since NMS has multiple optional scalar inputs that do not determine output shape, do we still need to validate their shape (if they are present) or let the runtime deal with faulty inputs ? There also seems to be a debate as to whether a "scalar" should strictly be a one-element tensor with 0 dims or whether it can be a one-element tensor with shape [1]. Strictly checking for something here might cause ShapeInference exceptions to be thrown for any existing working model and hence I did not validate the optional input shapes as they do not affect output shape anyway.