Make LCPAN also pass the activator to DepthwiseSeparable #9280
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
Without this, nn.Hardswish() will be selected in ConvBNLayer, which causes the opset version to be bumped to 14, in turn breaking a ESP-DL onnx quantization, which only supports 13. I would also suspect that this is more correct also in other cases as one would not be likely to not want to use a specific activation function everywhere, would one supply it. Also, it aligns with how the act-parameter is handled in other cases.
|
Thanks for your contribution! |
|
|
simoberny
approved these changes
Feb 7, 2025
There was a problem hiding this comment.
These changes effectively allow passing the configured activation function. Without them, setting act=relu6 in LCPAN still results in a hardswish activation function in the exported ONNX.
I successfully exported an ONNX network with opset 11 for ESP-DL (which does not support hardswish activation and opset 14) using these changes.
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Outdated suggestions cannot be applied.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.
Without this, nn.Hardswish() will be selected in ConvBNLayer, which causes the opset version to be bumped to 14, in turn, and in my case, breaks ESP-DL ONNX quantization, which only supports 13.
I would also suspect that this is more correct also in other cases as one would not be likely to not want to use a specific activation function everywhere, would one supply it.
Also, it aligns with how the act-parameter is handled in other cases.
(accidentally first did the PR against a release branch)