Closed tkoskela closed 1 year ago
By default the template type is always complex, which is not supported by TensorFlow. In practice we can expect to work with real data and it should be safe to drop the imaginary parts of the complex values before passing them into TF
This will be fixed on sopt side https://github.com/astro-informatics/sopt/pull/358
By default the template type is always complex, which is not supported by TensorFlow. In practice we can expect to work with real data and it should be safe to drop the imaginary parts of the complex values before passing them into TF