ahuston-0/nix-dotfiles
1
0
Fork 0
Code Issues Pull Requests 1 Actions 1 Packages Projects Releases Wiki Activity

ollama
All checks were successful
Check flake.lock / Check health of `flake.lock` (pull_request) Successful in 1m14s
Details
Check Nix flake / Perform Nix flake checks (pull_request) Successful in 4m15s
Details

Browse Source
This commit is contained in:
ahuston-0
2026-03-23 23:14:05 -04:00
parent 67e4dc15e7
commit 3557b88d7c
1 changed files with 6 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
systems/palatine-hill/ollama.nix
View File

@@ -12,19 +12,18 @@ in
package = pkgs.ollama; package = pkgs.ollama;
syncModels = true; syncModels = true;
loadModels = [ loadModels = [
"deepseek-r1:latest"
"deepseek-r1:1.5b" "deepseek-r1:1.5b"
"deepseek-r1:32b" "deepseek-r1:32b"
"deepseek-r1:70b"
"qwen3" "qwen3"
#"qwen3-coder-next" #"qwen3-coder-next"
"qwen3-coder" "qwen3-coder-next"
"lennyerik/zeta" "lennyerik/zeta"
"qwen2.5-coder:1.5b-base"
"qwen2.5-coder:14b" "qwen2.5-coder:14b"
"qwen2.5-coder:32b"
"nomic-embed-text:latest" "nomic-embed-text:latest"
"llama4:scout" "llama4:scout"
"mistral:7b" "mistral:7b"
"deepseek-v3.2:cloud"
"minimax-m2.7:cloud" "minimax-m2.7:cloud"
"lfm2:24b" "lfm2:24b"
"glm-4.7-flash" "glm-4.7-flash"
@@ -34,9 +33,9 @@ in
FLASH_ATTENTION = "1"; FLASH_ATTENTION = "1";
OLLAMA_KV_CACHE_TYPE = "q8_0"; OLLAMA_KV_CACHE_TYPE = "q8_0";
# Ollama memory configuration # Ollama memory configuration
OLLAMA_MAX_LOADED_MODELS = "2"; OLLAMA_MAX_LOADED_MODELS = "3";
OLLAMA_MAX_QUEUE = "4"; OLLAMA_MAX_QUEUE = "512";
OLLAMA_NUM_PARALLEL = "2"; OLLAMA_NUM_PARALLEL = "3";
# ROCm memory optimization # ROCm memory optimization
#HIP_VISIBLE_DEVICES = "0"; #HIP_VISIBLE_DEVICES = "0";