IST-DASLab/marlin

FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

by IST Austria Distributed Algorithms and Systems Lab Python Apache-2.0 licence 96 forks Created Jan 2024 View on GitHub

Stars
1,174
16% of IST-DASLab's stars
Rank
#57,577
of 847,120 repos, top 7%
In Python
#10,385
of 172,217 repos, top 7%

Among 847,120 repos

Compare with

1k–1.6k stars: 23,533 repos, including marlin

Stars along the bottom, and repos in each band on a log scale

Quick stats

Next rank
+1 star to reach #57,539
Next milestone
+6,058 stars to reach the top 1%
Popular
One of 66,327 repos with 1k+ stars
Yearly average
417 stars a year since Jan 2024

Stars counted Recounted daily Next count around