Adjust ollama settings

This commit is contained in:
Chris (wolcen) Thompson 2026-02-03 16:55:59 -05:00
parent cccadd6eb6
commit 414fb71a84

View file

@ -433,6 +433,12 @@ in
environmentVariables = { environmentVariables = {
HCC_AMDGPU_TARGET = "gfx1031"; # used to be necessary, but doesn't seem to anymore HCC_AMDGPU_TARGET = "gfx1031"; # used to be necessary, but doesn't seem to anymore
OLLAMA_LOAD_TIMEOUT = "2m"; # Reduce load timeout (from 5 min)...if it's that big, forget it. OLLAMA_LOAD_TIMEOUT = "2m"; # Reduce load timeout (from 5 min)...if it's that big, forget it.
OLLAMA_CONTEXT_LENGTH = "65536";
OLLAMA_KEEP_ALIVE = "10m";
OLLAMA_MAX_LOADED_MODELS = "1";
OLLAMA_KV_CACHE_TYPE = "Q4_K_0";
OLLAMA_DEBUG = "1";
OLLAMA_FLASH_ATTENTION = "true";
# OK, so origins is more about setting up CORS than firewalling things. (and requires e.g. http[s]://IP|Host/* etc) # OK, so origins is more about setting up CORS than firewalling things. (and requires e.g. http[s]://IP|Host/* etc)
#OLLAMA_ORIGINS = "10.40.4.2"; # Only accept connections from locally. #OLLAMA_ORIGINS = "10.40.4.2"; # Only accept connections from locally.