Skip to content

Releases: haruishi43/equilib

Release Version 0.5.8

17 Jan 03:27
65ccb7c
Compare
Choose a tag to compare
  • clip_output argument for all functions. The default is clip_output=True, but this can be turned off in the case that you want to return a raw output
  • clipping is based on the input array/tensor instead of assuming that the range is 0.0 to 1.0.

Release Version 0.5.7

16 Aug 05:04
120d055
Compare
Choose a tag to compare
  • removed deprecated numpy functions
  • merged #16

Release Version 0.5.6

13 Oct 13:22
Compare
Choose a tag to compare

Reduced artifacts for cube2equi and equi2cube transform #12 .

Contributor: @Oletus

Release Version 0.5.5

09 Oct 03:35
Compare
Choose a tag to compare

Release Version 0.5.2

22 Nov 06:45
e1848e1
Compare
Choose a tag to compare

Fixed bugs.
Added naive BFoV script.

Release Version 0.5.1

07 Oct 12:00
a519513
Compare
Choose a tag to compare

Fixed bugs for base.py

Release Version 0.5.0: Stable Release

29 Aug 14:40
Compare
Choose a tag to compare

Skipped 0.4.0.

Major Changes:

  • Minor changes to arguments for functions and classes
  • Major changes to grid_sample function. Finally got around to compare the implemented grid_sample with some baselines such as cv2.remap and scipy's map_coordinates. Wrote a lot of test codes to make sure that this implementation comes close to them as much as possible. Also compared numpy and torch's grid_sample and made sure they were close too.
  • mypy compatible (sort of). This still needs work, but it's a start.
  • Revamp to test codes (pytest). Back then I only checked if it ran, but now I made test codes (and benchmark codes) to test most of the functions (+ alpha). In the future, I might add codecov.

Version 0.3.0

22 Jun 10:33
Compare
Choose a tag to compare
  • Fixed coordinate system (previously, README said it was z-axis up, but it was actually z-axis down)
  • Refactored code

0.2.1

22 Dec 03:08
4108732
Compare
Choose a tag to compare
  • edit readme
  • minor bugs

0.2.0

22 Dec 01:52
Compare
Choose a tag to compare
  • New APIs
  • Automatic detection of input types (numpy.ndarray, torch.Tensor, etc.)
  • Revised test codes