llama.cpp
metal : fix kernel_norm
#3057
Merged
Overview
Commits
5
Changes
View On GitHub
Commits
  • metal : fix kernel_norm
    ggerganov committed 2 years ago
  • metal : put warning in kernel_norm to not combine the loops
    ggerganov committed 2 years ago
  • metal : restore original F16 mat-vec multiplication
    ggerganov committed 2 years ago
  • common : don't do warm-up with more than n_batch tokens (close #3058)
    ggerganov committed 2 years ago
  • metal : minor
    ggerganov committed 2 years ago
FAQ Terms Privacy Refunds Impressum
Loading