llama.cpp
metal : add Q8_0 support
#2763
Merged

Commits
  • metal : add dequantize_q8_0 kernel
    ggerganov committed 2 years ago
  • metal : add mul_mat_q8_0_f32 kernel
    ggerganov committed 2 years ago
  • metal : add Q8_0 mul_mm kernel
    ggerganov committed 2 years ago
Loading