Skip to content

v0.1.9

Compare
Choose a tag to compare
@ySingularity ySingularity released this 01 Apr 03:42
· 1133 commits to main since this release

feat

  • support awq
  • mv attention mask when use FMHA
  • support sparse&robert embedding, support calc similarity

refactor

  • use asyncio.future to avoid resource exclusivity
  • mv asyncio lock to asyncmodel

fix

  • tmp fix filelock version
  • moe model size
  • add headers for image downloading
  • update whl version
  • cutlass interface

docs

  • update pipeline usage