getnamo / Llama-Unreal

Llama.cpp plugin for Unreal Engine 5
MIT License
41 stars 5 forks source link

Track multiple and swap easily between contexts #5

Open getnamo opened 8 months ago

getnamo commented 8 months ago

NB: context meaning full prompt history

getnamo commented 7 months ago

Likely need some form of batched inference with different contexts: https://github.com/ggerganov/llama.cpp/tree/master/examples/batched-bench

NB: