kyle/imladris
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

straylight: disable laguna, default to ornith

Browse Source
This commit is contained in:
kyle
2026-09-19 21:23:42 -07:00
parent 72261df517
commit 0543864f38
2 changed files with 9 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
hw/straylight/default.nix
+8 -5
View File
@@ -42,14 +42,17 @@ let
; DFlash GGUF is on disk (laguna-s-2.1-DFlash-BF16.gguf). nixpkgs llama.cpp 0.4.0
; rejects it at load ("expected 76, got 69" tensors); Poolside's file needs their
; laguna-branch loader. Leave unwired so Laguna still starts.
[laguna-s-2.1]
model = ${llamaModelsDir}/laguna-s-2.1-UD-Q4_K_S/Laguna-S-2.1-UD-Q4_K_S-00001-of-00003.gguf
; DISABLED 2026-09-19: resident beside Ornith 35B it left ~15 GB free, and a third
; model load then hit the 104 GiB TTM cap and OOM-killed the desktop session.
; Weights stay on disk; uncomment to re-enable.
; [laguna-s-2.1]
; model = ${llamaModelsDir}/laguna-s-2.1-UD-Q4_K_S/Laguna-S-2.1-UD-Q4_K_S-00001-of-00003.gguf
; Two slots sharing one unified 131072-token KV pool: a second request (subagent,
; another client) runs alongside instead of queueing, and either one may still use
; the full 131072 as long as their combined context fits.
parallel = 2
kv-unified = true
ctx-size = 131072
; parallel = 2
; kv-unified = true
; ctx-size = 131072
; Previous default model (a Qwen 3.5 derivative), kept for comparison runs. It was
; fetched by the old -hf flag into the HF cache; dedup hides the cache's own entry.