llama.cpp
Move to backend sampling for MTP draft path
#23287
Merged

Commits
  • Move to backend sampling for MTP draft path
    gaugarg-nv committed 123 days ago
  • Allow sampler chains to be partially offloaded to backend
    gaugarg-nv committed 123 days ago
  • Add --spec-draft-backend-sampling argument. Enabled by default.
    gaugarg-nv committed 123 days ago
Loading