Skip to content

v2.0.3

Latest

Choose a tag to compare

@LoserCheems LoserCheems released this 07 Jun 02:48
· 45 commits to main since this release

What's Changed

  • Migrate all Triton kernels from make_block_ptr to make_tensor_descriptor by @LoserCheems in #307
  • [Feature] Fused FP8 Projection Kernels (forward + backward) by @LoserCheems in #309

Full Changelog: v2.0.2...v2.0.3

What's Changed

  • Migrate all Triton kernels from make_block_ptr to make_tensor_descriptor by @LoserCheems in #307
  • [Feature] Fused FP8 Projection Kernels (forward + backward) by @LoserCheems in #309

Full Changelog: v2.0.2...v2.0.3