[Bug] Unable to load Gemma-4-26B-A4B-IT with Intel AutoRound quantization - yields UnboundLocalError: cannot access local variable 'GPTQMarlinMoEMethod' error before model loads.
April 8, 2026 · #22370
Python
Difficulty: Medium
Parent Repository
sgl-project/sglang
Python repository
25,687 5,300