more framework desktop local ai fixes

active/device_framework_desktop/framework_desktop.md

@@ -17,6 +17,7 @@
   - [stable-diffusion.cpp](#stable-diffusioncpp)
   - [open-webui](#open-webui)
   - [Install the whole thing with quadlets (TM)](#install-the-whole-thing-with-quadlets-tm)
+    - [Install the update script](#install-the-update-script)
 
 ## BIOS
 
@@ -91,6 +92,10 @@ hf auth login
 mkdir /home/ai/models/text/gpt-oss-120b
 hf download --local-dir /home/ai/models/text/gpt-oss-120b ggml-org/gpt-oss-120b-GGUF
 
+# gpt-oss-20b
+mkdir /home/ai/models/text/gpt-oss-20b
+hf download --local-dir /home/ai/models/text/gpt-oss-20b ggml-org/gpt-oss-20b-GGUF
+
 # devstral-2-123b
 mkdir /home/ai/models/text/devstral-2-123b
 hf download --local-dir /home/ai/models/text/devstral-2-123b unsloth/Devstral-2-123B-Instruct-2512-GGUF Q4_K_M/Devstral-2-123B-Instruct-2512-Q4_K_M-00001-of-00002.gguf
@@ -99,6 +104,14 @@ hf download --local-dir /home/ai/models/text/devstral-2-123b unsloth/Devstral-2-
 # devstral-small-2-24b
 mkdir /home/ai/models/text/devstral-small-2-24b
 hf download --local-dir /home/ai/models/text/devstral-small-2-24b unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF Devstral-Small-2-24B-Instruct-2512-Q4_K_M.gguf
+
+# ministral-3-14b
+mkdir /home/ai/models/text/ministral-3-14b
+hf download --local-dir /home/ai/models/text/ministral-3-14b ggml-org/Ministral-3-14B-Reasoning-2512-GGUF
+
+# nemotron-nano-30b
+mkdir /home/ai/models/text/nemotron-nano-30b
+hf download --local-dir /home/ai/models/text/nemotron-nano-30b ggml-org/Nemotron-Nano-3-30B-A3B-GGUF Nemotron-Nano-3-30B-A3B-Q4_K_M.gguf
 ```
 
 #### Image models
@@ -144,6 +157,10 @@ localhost/llama-cpp-vulkan:2026-01-19-18-00-02 \
 
 ## stable-diffusion.cpp
 
+Server: <https://github.com/leejet/stable-diffusion.cpp/tree/master/examples/server>
+
+CLI: <https://github.com/leejet/stable-diffusion.cpp/tree/master/examples/cli>
+
 ```bash
 git clone https://github.com/leejet/stable-diffusion.cpp.git
 cd stable-diffusion.cpp
@@ -165,12 +182,13 @@ localhost/stable-diffusion-cpp:latest \
 --llm /models/image/z-turbo/Qwen3-4B-Instruct-2507-Q4_K_M.gguf \
 --cfg-scale 1.0 \
 -v \
--H 512 \
+-H 1024 \
 -W 1024 \
 --seed -1 \
 --steps 8 \
+--vae-conv-direct \
 -o /output/output.png \
--p "A watercolor dragon with flowing ink lines, pastel palette, white paper background, soft brush strokes, high-resolution"
+-p "A photorealistic dragon"
 
 # Edit with flux kontext
 podman run --rm \
@@ -187,7 +205,7 @@ localhost/stable-diffusion-cpp:latest \
 --sampling-method euler \
 --seed -1 \
 --steps 20 \
--H 512 \
+-H 1024 \
 -W 1024 \
 -r /output/everquest_logo.png \
 -p "change 'EverQuest' to 'EverSteak'" \
@@ -197,6 +215,10 @@ localhost/stable-diffusion-cpp:latest \
 ## open-webui
 
 ```bash
+mkdir /home/ai/.env
+# Create a file called open-webui-env with `WEBUI_SECRET_KEY="some-random-key"
+scp active/device_framework_desktop/secrets/open-webui-env deskwork-ai:.env/
+
 # Will be available on port 8080
 podman run \
 -d \
@@ -215,3 +237,13 @@ ssh deskwork-ai
 systemctl --user daemon-reload
 systemctl --user restart ai-pod.service
 ```
+
+### Install the update script
+
+```bash
+# Copy update script and run it (assumes you have llama.cpp and stable-diffusion.cpp)
+scp active/device_framework_desktop/update-script.sh deskwork:
+ssh deskwork-ai
+chmod +x update-script.sh
+./update-script.sh
+```

active/device_framework_desktop/quadlets/open-webui.container

@@ -11,6 +11,9 @@ Image=ghcr.io/open-webui/open-webui:main
 # Nothing too complicated here. Open Webui will basically configure itself.
 Volume=open-webui-data:/app/backend/data
 
+# WEBUI_SECRET_KEY is required to prevent logout on Restart
+EnvironmentFile=/home/ai/.env/open-webui-env
+
 [Service]
 Restart=always
 # Extend Timeout to allow time to pull the image

active/device_framework_desktop/quadlets/stable-diffusion-edit-server.container

@@ -27,6 +27,7 @@ Exec=-l 0.0.0.0 \
     --t5xxl /models/image/flux-1-kontext/t5xxl_fp16.safetensors \
     --cfg-scale 1.0 \
     --sampling-method euler \
+    --vae-conv-direct \
     -v \
     --seed -1 \
     --steps 28

active/device_framework_desktop/quadlets/stable-diffusion-gen-server.container

@@ -27,6 +27,7 @@ Exec=-l 0.0.0.0 \
     -l 0.0.0.0 \
     --listen-port 1234 \
     --cfg-scale 1.0 \
+    --vae-conv-direct \
     -v \
     --seed -1 \
     --steps 8

active/device_framework_desktop/update-script.sh

@@ -0,0 +1,20 @@
+#!/bin/bash
+
+set -x
+
+export BUILD_TAG=$(date +"%Y-%m-%d-%H-%M-%S")
+
+echo "Updates stable-diffusion.cpp, llama.cpp, and open-webui"
+
+cd /home/ai/llama.cpp
+git pull
+podman build -t llama-cpp-vulkan:${BUILD_TAG} -t llama-cpp-vulkan:latest -f .devops/vulkan.Dockerfile .
+
+cd /home/ai/stable-diffusion.cpp
+git pull
+git submodule update --init --recursive
+podman build -f Dockerfile.vulkan -t stable-diffusion-cpp:${BUILD_TAG} -t stable-diffusion-cpp:latest .
+
+podman image pull ghcr.io/open-webui/open-webui:main
+
+systemctl --user restart ai-pod