Sign in
Rare find
marlin.
FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.
github.com/IST-DASLab/marlin
by
@IST-DASLab
Problem
Audience
Keep exploring
Updates
Founder
0
Product
0
No founder updates yet. Check back soon.
🧪
staging
