[ACL] Prevent offloading of per-channel quantized operators #14484
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
Collaborator
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment. Generated by tvm-bot |
NicolaLancellotti
changed the title
NicolaLancellotti
force-pushed
the
acl/prevent-per-channel-quantizatio-offload
branch
from
a16b15d to
21454d3
Compare
Currently, only per-layer quantization is supported in the Arm Compute Library runtime. However, there is no check that prevents the offloading of per-channel quantized operators, as a consequence, tvm fails during inference if such operators are found.
NicolaLancellotti
force-pushed
the
acl/prevent-per-channel-quantizatio-offload
branch
from
21454d3 to
d5fe3be
Compare
lhutton1
approved these changes
Apr 5, 2023
Contributor
lhutton1
left a comment
lhutton1
left a comment
There was a problem hiding this comment.
Thanks @NicolaLancellotti, LGTM!
| test_qnn_dense() | ||
| test_codegen_dense() | ||
| test_codegen_qnn_dense() | ||
| test_codegen_qnn_dense_per_channel_quantization() |
Contributor
There was a problem hiding this comment.
No need to add now, we should probably convert to the:
if __name__ == "__main__":
tvm.testing.main()
format when we get a another chance
Contributor
|
Thanks @NicolaLancellotti! |
NicolaLancellotti
restored the
acl/prevent-per-channel-quantizatio-offload
branch
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Outdated suggestions cannot be applied.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.
Currently, only per-layer quantization is supported in the Arm Compute Library runtime. However, there is no check that prevents the offloading of per-channel quantized operators, as a consequence, tvm fails during inference if such operators are found.
This patch adds a check that prevents offloading per-channel quantized operators.