Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

ComplexAbs and Resize antialias errors #2368

Open
ZachZhao96 opened this issue Nov 20, 2024 · 0 comments
Open

ComplexAbs and Resize antialias errors #2368

ZachZhao96 opened this issue Nov 20, 2024 · 0 comments
Labels
bug An unexpected problem or unintended behavior

Comments

@ZachZhao96
Copy link

ZachZhao96 commented Nov 20, 2024

Hi, I tried the latest tf2onnx and used tf2onnx.convert.from_keras(model,opset=19) to convert my keras h5 model. Here is my partial code for preprocessing:
image

however, I got two error messages shown below:
image

The first error says that ComplexAbs doesn't expect the same dimesion input as tf.abs used. This is not an issue for opset -13, so i hope this is just a bug.

then, the second one says antialias for resize is not supported. I wonder if the antialias supported in Resize Op is the same as the tf.image.resize(..., antialias = True).
image

Here is the resize bilinear with antialias example in ONNX documentation.
image

Thanks for your time.

@ZachZhao96 ZachZhao96 added the bug An unexpected problem or unintended behavior label Nov 20, 2024
# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
bug An unexpected problem or unintended behavior
Projects
None yet
Development

No branches or pull requests

1 participant