llama.cpp CUDA 13.2 gibberish for IQ3_S / IQ3_XXS, IQ2_M - `SOLUTION use CUDA < 13.0 or Unsloth Studio`

April 4, 2026 ยท #4849
View on GitHub
Python Difficulty: Medium

Labels

fixed bug

Sign in required

Authenticate to use favourites & bookmarks

5