[cker] Fix min-max calculation in MaxPool2D #14133
Merged
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
This PR fixes min-max calculation in MaxPool2D kernel. ONE-DCO-1.0-Signed-off-by: seunghui youn <[email protected]>
zetwhite
commented
Sep 30, 2024
| out_mat.cwiseMin(params.float_activation_min).cwiseMax(params.float_activation_max); | ||
| out_mat = out_mat.cwiseMin(params.float_activation_max).cwiseMax(params.float_activation_min); |
There was a problem hiding this comment.
(note) This line is a main change in this PR
zetwhite
commented
Sep 30, 2024
Comment on lines
236
to
278
| // with min, max params | ||
| { | ||
| nnfw::cker::PoolParams op_param; | ||
| { | ||
| op_param.stride_height = 1; | ||
| op_param.stride_width = 1; | ||
| op_param.filter_height = 2; | ||
| op_param.filter_width = 2; | ||
| op_param.padding_values.height = 0; | ||
| op_param.padding_values.width = 0; | ||
| op_param.float_activation_min = 0.0; | ||
| op_param.float_activation_max = 6.0; | ||
| } | ||
| nnfw::cker::Shape in = {1, 3, 3, 1}; | ||
| nnfw::cker::Shape out = {1, 2, 2, 1}; | ||
|
|
||
| MaxPoolOpVerifier<float> verifier(op_param, in, out); | ||
|
|
||
| /** | ||
| * input(index) : output(arg-index): | ||
| * | ||
| * 1(0) 2(1) 3(2) | ||
| * 4(3) 5(4) 6(5) - (forward) -> 5(4) 6(5) | ||
| * 7(6) 8(7) 9(8) 6(7) 6(8) | ||
| */ | ||
|
|
||
| std::vector<float> input = {1, 2, 3, 4, 5, 6, 7, 8, 9}; | ||
| std::vector<float> expected_output = {5, 6, 6, 6}; | ||
| verifier.verifyForward(input, expected_output); | ||
|
|
||
| /** | ||
| * output_deriv: input_deriv: | ||
| * (randomly filled) | ||
| * | ||
| * 0.1 0.2 0 0 0 | ||
| * 0.3 0.4 - (backward) -> 0 0.1 0.2 | ||
| * 0 0.3 0.4 | ||
| */ | ||
|
|
||
| std::vector<float> output_deriv = {0.1, 0.2, 0.3, 0.4}; | ||
| std::vector<float> expected_input_deriv = {0, 0, 0, 0, 0.1, 0.2, 0, 0.3, 0.4}; | ||
| verifier.verifyBackward(output_deriv, expected_input_deriv); | ||
| } |
There was a problem hiding this comment.
(note) This is a newly added test - to check that min/max works as expected
ragmani
reviewed
Sep 30, 2024
Co-authored-by: Jang Jiseob <[email protected]>
ragmani
approved these changes
Oct 2, 2024
glistening
approved these changes
Oct 2, 2024
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.
This PR fixes min-max calculation in MaxPool2D kernel.
ONE-DCO-1.0-Signed-off-by: seunghui youn [email protected]