8000 Commits · fluckydog232/ScaleLLM · GitHub

More Web Proxy on the site http://driver.im/

Commits

Commits on Oct 25, 2024

Commits on Oct 23, 2024

misc: attention kernel refactoring (#339)
guocuimi
authored

Commits on Oct 18, 2024

ut: add more tests for different warp layout (#340)
guocuimi
authored

Commits on Oct 17, 2024

Upgrade pytorch to 2.5.0 (#347)
guocuimi
authored

Commits on Oct 15, 2024

Commits on Oct 13, 2024

upgrade pytorch to 2.4.1 (#341)
guocuimi
authored

Commits on Sep 11, 2024

Commits on Sep 10, 2024

Commits on Sep 9, 2024

ut: added unittests for flash infer kernels (#332)
guocuimi
authored

Commits on Sep 7, 2024

Commits on Sep 5, 2024

Commits on Sep 4, 2024

Commits on Sep 3, 2024

refactor: added static switch for marlin kernel dispatch (#325)
guocuimi
authored

Commits on Aug 30, 2024

fix: put item into asyncio.Queue in a thread-safe way (#324)
guocuimi
authored

Commits on Aug 29, 2024

ci: allow build without requiring a physical gpu device (#321)
guocuimi
authored

Commits on Aug 28, 2024

Commits on Aug 27, 2024

test: added unittests for marlin kernels (#317)
guocuimi
authored

Commits on Aug 25, 2024

build: speed up compilation for marlin kernels (#316)
guocuimi
authored

Commits on Aug 24, 2024

feat: added awq marlin qlinear (#315)
guocuimi
authored

Commits on Aug 21, 2024

bump up version to 0.2.0
guocuimi
committed

Commits on Aug 20, 2024

kernel: port awq repack kernel (#314)
guocuimi
authored

Commits on Aug 19, 2024

Commits on Aug 17, 2024

refactor: remove the logic loading individual weight from shared partitions (#311)
guocuimi
authored

0