[microNPU] Set output tolerance of codegen and network tests to 0 #10675
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
lhutton1
force-pushed
the
reference-kernel
branch
2 times, most recently
from
cd65869 to
4797434
Compare
Contributor
|
Can you check if we need these two output tolerances? |
Contributor
Author
|
Thanks for pointing these out @NicolaLancellotti! The identity tolerance cannot be changed since the output can be slightly different compared to the numpy reference due to quantization (cc @ekalda). However, we should be able to remove the tolerance from max. I'll update after #10762 to save an additional rebase :) |
After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added. Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM. Change-Id: Idcf408b84c67456d16da5d61f4865eb8b6042554
lhutton1
force-pushed
the
reference-kernel
branch
from
4797434 to
563e6f0
Compare
Contributor
Author
|
After #10889 this should now pass CI |
Change-Id: I174cc48a47c75590f76e0ee2a9634547b2442e87
Member
|
Nice to see zero tolerance @lhutton1 / @NicolaLancellotti 😸 |
pfk-beta
pushed a commit
to pfk-beta/tvm
that referenced
this pull request
Apr 11, 2022
…ache#10675) After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added. Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM.
mehrdadh
pushed a commit
to mehrdadh/tvm
that referenced
this pull request
Apr 11, 2022
…ache#10675) After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added. Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM.
Lucien0
pushed a commit
to Lucien0/tvm
that referenced
this pull request
Apr 19, 2022
…ache#10675) After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added. Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM.
altanh
pushed a commit
to altanh/tvm
that referenced
this pull request
Apr 28, 2022
…ache#10675) After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added. Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
After the recent upgrade of Tensorflow to 2.6, we are now able to use reference kernels in order to verify the output. Thus, removing the tolerances previously added.
Additionally, the network tests have been altered to use TFLite as a reference, rather than TVM.
cc @manupa-arm @ekalda @NicolaLancellotti @dchauhan-arm