Skip to content

Fix 4-bit quantization for weight matrices not divisible by blocksize #142

Fix 4-bit quantization for weight matrices not divisible by blocksize

Fix 4-bit quantization for weight matrices not divisible by blocksize #142

Annotations

5 errors

CUDA (linux-x64, T4, 12.8.1)  /  test

cancelled Mar 3, 2026 in 8m 40s