[Torch, QNN] Add missing upcast to uint8 avg_pool conversion #5089
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
anijain2305
approved these changes
Mar 18, 2020
|
Thanks @masahi This is merged |
trevor-m
pushed a commit
to trevor-m/tvm
that referenced
this pull request
Apr 16, 2020
…5089) * add missing upcast to avgpool * add avg pool test
zhiics
pushed a commit
to neo-ai/tvm
that referenced
this pull request
Apr 17, 2020
…5089) * add missing upcast to avgpool * add avg pool test
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Outdated suggestions cannot be applied.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.
Previously we were missing an upcast to int32 for avg_pool2d, fixed now. It seems avg_pool is used in inception v3, but I don't know why our accuracy is not affected by the missing upcast.
Beside the accuracy issue, without upcast it also causes a mysterious typing problem for inception v3 when compiling for AVX512. For AVX2 and other, there have been no issue.
Even with the upcast, the raw output from the new test case, quantize -> avgpool -> dequantize, has some gap with torch output. This depends on how Torch implements uint8 avg pool (probably with fp32 piggy back).
cc @anijain2305