Skip to content

Releases: delveopers/Axon

v0.0.2

01 Aug 20:02
22d1a57

Choose a tag to compare

v0.0.2 Pre-release
Pre-release

Major Updates:

  • axon.linalg functions support
  • fixed previous cross-platform installation issues & several major bugs
  • CMake build support
  • documentation updates
  • fixed util functions & also added several new functions
  • added scalar support

What's Changed

  • codebase restructured, changed few file names by @shivendrra in #14
  • codebase restructured, changed few file names (#14) by @shivendrra in #16
  • added ('CMake') support for the library by @shivendrra in #17
  • i don't know what's happening by @shivendrra in #18
  • updated the directory structure & added new test cases by @shivendrra in #20
  • ('axon.linalg') module & functions implementation done! by @shivendrra in #21
  • fixed broadcasting, added comparison ops, & fixed some functions by @shivendrra in #23
  • done with axon.linalg, wrote most of the linalg functions by @shivendrra in #24
  • fixed multiple bugs, removed redundant codes, updated docs & tests by @shivendrra in #25

Full Changelog: 0.0.1...0.0.2

v0.0.1

18 Jun 13:08
da3d2ed

Choose a tag to compare

v0.0.1 Pre-release
Pre-release

Major Updates

  • first release / deployed pre-alpha version on pypi
  • implemented optimized matmul(), 5-10x faster than naive C matmul
  • added cuda backend too, not integrated in the main class yet
  • dtype casting like numpy & pytorch
  • all computation ops in float_32
  • python interface with util functions

What's Changed

  • done with basic core Array class & functions in c/c++ & python by @shivendrra in #6
  • added new utility functions along with few upgrades in array class by @shivendrra in #8
  • added ("dtype") casting related functionalities, implemented more ops & functions & added basic documentation by @shivendrra in #9
  • added matmul() ops & fixed the bug in transpose() & updated test cases by @shivendrra in #10
  • optimized the naive matmul using O(n) transpose & fixed errors with RNG by @shivendrra in #11
  • CUDA kernels borrowed from shivendrra/axgrad, not complied though by @shivendrra in #12
  • ready for deployment by @shivendrra in #13

New Contributors

Full Changelog: https://github.com/delveopers/Axon/commits/0.0.1