llama.cpp
llama : KV cache view API + better KV cache management
#4170
Merged

Commits
  • llama : keep track of used KV cells + better KV cache management
    ggerganov committed 2 years ago
  • llama : zero KV cache used upon clear
    ggerganov committed 2 years ago
  • llama : allow exporting a view of the KV cache (#4180)
    KerfuffleV2 committed 2 years ago
  • common : add -dkvc arg for enabling kv cache dumps
    ggerganov committed 2 years ago
Loading