titaniumtown/server-config
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

llm: use Q4_0 quants (faster)

Browse Source
This commit is contained in:
Simon Gardling 2025-03-31 18:33:24 -04:00
parent c31635bdd7
commit 2d47c441fe
Signed by: titaniumtown
GPG Key ID: 9AB28AC10ECE533D
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
services/llama-cpp.nix
View File

@ -26,8 +26,8 @@ in
enable = true;
model = builtins.toString (
pkgs.fetchurl {
url = "https://huggingface.co/bartowski/google_gemma-3-12b-it-GGUF/resolve/main/google_gemma-3-12b-it-IQ4_XS.gguf";
sha256 = "aa7b7ae0b17931c379ede82da59b01f246046925aeb752af1ab4285a3b0d69db";
url = "https://huggingface.co/bartowski/google_gemma-3-12b-it-GGUF/resolve/main/google_gemma-3-12b-it-Q4_0.gguf";
sha256 = "9a7b70be8727da9fb28523b35946dd42d4fe0f622cce03daa44fccff0775516d";
}
);
port = service_configs.ports.llama_cpp;