HK
Heykuki News
Top
New
Best
Ask
Show
Jobs
Toggle theme
Top
New
Best
Ask
Show
Jobs
Request
1.
▲
A new CUDA kernel for quantized LLMs achieves up to 2.6x latency improvements
github.com/HanGuo97
1 comment
2 years ago
radichoml
2 points
2.
▲
Show HN: Tokenusage – Rust CLI that tracks Claude Code/Codex tokens 214x faster
github.com/hanbu97
3 comments
4 months ago
hanbu97
1 points