happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
llama.cpp/examples/duo/README.md

202 B
Raw Blame History

duo

Minimal example. What's not implemented, but can be implemented separately in pieces:

  • tree-based speculation
  • correct sampling
  • support more than 2 instances
  • just one instance speculates