v1.5.2 Fix required numpy version

What's Changed

• Update required numpy by @mmuckley in #103

Full Changelog: v1.5.1...v1.5.2

v1.5.1 Update package versions + bug fixes

This release primarily fixes the release tag so the docs and PyPI upload all work. In addition, it updates all packages in torchkbnufft to their latest versions

What's Changed

• Remove some relative imports by @mmuckley in #102

Full Changelog: v1.5.0...v1.5.1

v1.5.0

What's Changed

• Fixes for release action by @mmuckley in #80
• Minor update to the bib for cite by @chaithyagr in #83
• Fix Numpy complex type by @Andrew-Dupuis in #94
• conftest.py: allow running tests with numpy 2.0 by @gabuzi in #100
• Update packages to latest versions by @mmuckley in #101

New Contributors

• @Andrew-Dupuis made their first contribution in #94
• @gabuzi made their first contribution in #100

Full Changelog: v1.4.0...v.1.5.0

Support PyTorch 1.13.0 + new build system

This release supports the latest version of PyTorch (1.13.0) and adds a new build and versioning system. A list of some work on the current release:

• Changed default normalization for Toeplitz to "ortho" (PR #43)
• From @ckolbPTB, fixed size of weights in calc_toeplitz_kernel (PR #47)
• From @ckolbPTB, allowed use of single smap for batches for Toeplitz NUFFT (PR #48)
• Updates for internal type system (PR #77)
• Updates for package dependencies and build system (PR #78, #79)

Fixes for Toeplitz NUFFT scaling

• Fixes Toeplitz NUFFT scaling issue (PR #39)
• Updates for requirements to current PyTorch (PR #40)

Documentation Formatting Update

Fixes formatting for a reference on the main documentation page.

Documentation Update

This pure-documentation release fixes an issue with how documentation headers were rendered on the ReadTheDocs website. See PR #29.

Update for PyTorch 1.8

This updates torchkbnufft for PyTorch version 1.8. It uses a new version of index_add that operates natively on complex tensors. The update also fixes a performance regression that arose due to thread management as identified in Issue #25.

Most changes came from PR #27, which has the below list of modifications:

• Update requirements.txt and dev-requirements.txt to latest packages.
• Remove calc_split_sizes - we can now use tensor_split.
• Removed some calls to tensor attributes - these can be expensive.
• Removal of kwarg usage for some torch.jit.script functions - these can behave strangely with scripted functions.
• Removal of index_put for accumulation. We now only use index_add.

Batched NUFFT

This adds support for a new batched NUFFT, which is substantially faster than using a Python for loop over the batch dimension when applying a NUFFT with many small k-space trajectories. It also updates the documentation and includes a new page for performance tips. See PR #24 and Issue #24 for details and testing.

Bug fixes and documentation updates

• Fixes an incredibly weird bug in the autograd for forward NUFFTs caused by an unsqueeze operation (PR #23).
• Remove references to batch dimension in notebooks (PR #20).
• Remove unnecessary adjoint NUFFT objects in docs (PR #19).
• Add a test for CPU/GPU forward matching (PR #17).