Conversation
added 13 commits
October 6, 2025 20:37
Jammy2211
requested review from
CKrawczyk,
Sketos,
qiuhan96,
rhayes777 and
samlange04
added 7 commits
October 13, 2025 20:20
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
FFT and Real-Space Convolution Refactor
This PR refactors Kernel2D to support both FFT- and real-space convolution paths using JAX FFTs for accelerated computation.
FFT Convolution (default)
Uses jax.numpy.fft.rfft2 / irfft2 for efficient 2D convolution.
Requires precomputed padding shapes (fft_shape, full_shape, mask_shape) to ensure linear convolution (avoiding wrap-around).
Raises clear errors if shapes are inconsistent, with diagnostics on expected vs actual shapes.
Much faster for large kernels (>~5×5).
Real-Space Convolution (fallback)
Uses jax.scipy.signal.convolve with explicit masking.
Slower but avoids padding, useful for tests and small kernels.
Available for both images and mapping matrices.
Mapping Matrices
Convolution of mapping matrices now supported via both FFT and real space.
Blurring matrices are handled consistently in both modes.
Other improvements
Clearer warnings when blurring contributions are omitted.
Improved error messages and logging with full shape diagnostics.
Comprehensive docstrings explaining FFT padding, trade-offs, and usage.
This unifies convolution logic under JAX, making operations GPU/TPU compatible and more performant while preserving correctness for masked images.