add initial framework desktop config

active/device_framework_desktop/framework_desktop.md

@@ -10,45 +10,64 @@
 3. Update grub `grub2-mkconfig -o /boot/grub2/grub.cfg`
 4. Reboot
 
-## Docker
+## References
+
+<https://docs.podman.io/en/latest/markdown/podman-systemd.unit.5.html>
+
+## Notes
+
+### Update quadlets
 
 ```bash
-dnf config-manager addrepo \
+scp -r active/device_framework_desktop/quadlets/* deskwork-ai:quadlets/
---from-repofile https://download.docker.com/linux/fedora/docker-ce.repo
+podman quadlet install --replace quadlets/*
+```
 
-dnf install \
+### Volume Locations
-docker-ce \
+
-docker-ce-cli \
+`~/.local/share/containers/storage/volumes/`
-containerd.io \
+
-docker-buildx-plugin \
+## User
-docker-compose-plugin
+
+```bash
+# Create your local ai user. This will be the user you launch podman processes from.
+useradd -m ai
+loginctl enable-linger ai
+su -l ai
+mkdir -p ~/.config/containers/systemd/
 ```
 
 ## Llama.cpp
 
 ```bash
-# Build the llama.cpp docker image
+# Build the llama.cpp container image
 git clone https://github.com/ggml-org/llama.cpp.git
 export BUILD_TAG=$(date +"%Y-%m-%d-%H-%M-%S")
-docker build -t llama-cpp-vulkan:${BUILD_TAG} -f .devops/vulkan.Dockerfile .
+podman build -t llama-cpp-vulkan:${BUILD_TAG} -f .devops/vulkan.Dockerfile .
 
 # Run llama server with gpt-oss-120b
-docker run \
+podman run \
 -d \
+--replace \
 --restart always \
 --name=llama-server \
---network=host \
+-p 8000:8000 \
 --device=/dev/kfd \
 --device=/dev/dri \
 -v llama-server-cache:/root/.cache \
-llama-cpp-vulkan:${BUILD_TAG} \
+localhost/llama-cpp-vulkan:2026-01-12-10-13-30 \
--hf ggml-org/gpt-oss-120b-GGUF  --ctx-size 0 --jinja -ub 2048 -b 2048 \
+-hf ggml-org/gpt-oss-120b-GGUF  --ctx-size 32000 --jinja -ub 2048 -b 2048 \
 --port 8000 --host 0.0.0.0 -n -1 --n-gpu-layers 999
 
+# To enable autostart, you'll need to create a quadlet
+# Quadlets are documented in podman manual pages
+# Search for "EXAMPLES" when you run the below command
+# Put your quadlet at ~/.config/containers/systemd/
+man "podman-systemd.unit(5)"
+
 # Run llama server with devstral-small-2 24b
-docker run \
+podman run \
 -d \
---restart always \
 --name=llama-server-devstral \
 --network=host \
 --device=/dev/kfd \
@@ -66,10 +85,18 @@ firewall-cmd --reload
 
 ## Ollama
 
+```bash
+# Install CLI
+curl -fsSL https://ollama.com/download/ollama-linux-amd64.tgz | tar xz -C ~/.local
+
+# Add export OLLAMA_HOST=127.0.0.1
+vim ~/.bashrc.d/ollama.sh
+```
+
 ```bash
 # Run ollama
 # Will be available on port 11434
-docker run \
+podman run \
 -d \
 --restart always \
 --device /dev/kfd \
@@ -81,7 +108,7 @@ docker run \
 docker.io/ollama/ollama:0.13.5
 
 # Run an image
-docker exec -it ollama ollama run gpt-oss:20b
+podman exec -it ollama ollama run gpt-oss:20b
 
 # Firewall
 firewall-cmd --add-port=11434/tcp --permanent
@@ -107,9 +134,9 @@ vim /etc/anything-llm/.env
 
 # Server will be accessible on port 3001
 # Connect llama.cpp as a generic OpenAI LLM provider and use host
-# http://172.17.0.1:8000/v1
+# http://172.17.0.1:3001/v1
 # Chat model name doesn't matter.
-docker run \
+podman run \
 -d \
 --restart always \
 --network host \
@@ -129,9 +156,9 @@ firewall-cmd --reload
 
 ```bash
 # z-turbo
-docker run --rm \
+podman run --rm \
--v ./models:/models \
+-v /home/ai/stable-diffusion.cpp/models:/models:z \
--v ./build:/output \
+-v /home/ai/stable-diffusion.cpp/output:/output:z \
 --device /dev/kfd \
 --device /dev/dri \
 ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
@@ -140,17 +167,37 @@ ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
 --llm /models/z_turbo/qwen_3_4b.safetensors \
 --cfg-scale 1.0 \
 -v \
---diffusion-fa \
 -H 1024 \
 -W 512 \
--o /output/output.png \
 --seed -1 \
+-o /output/output.png \
 -p "Framework Laptop 13"
 
+# Flux
+podman run --rm \
+-v /srv/stable-diffusion.cpp/models:/models:z \
+-v ./output:/output:z \
+--device /dev/kfd \
+--device /dev/dri \
+ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
+--diffusion-model  /models/flux/flux1-dev-q4_k.gguf \
+--vae /models/flux/ae.safetensors \
+--clip_l /models/flux/clip_l.safetensors \
+--t5xxl /models/flux/t5xxl_fp16.safetensors \
+--cfg-scale 1.0 \
+--sampling-method euler \
+-v \
+-H 512 \
+-W 512 \
+--seed -1 \
+--steps 20 \
+-o /output/output.png \
+-p "An Everquest video game poster but with ribeye steaks for heads with the words 'EverSteak'"
+
 # Flux2
-docker run --rm \
+podman run --rm \
--v ./models:/models \
+-v /home/ai/stable-diffusion.cpp/models:/models:z \
--v ./build:/output \
+-v /home/ai/stable-diffusion.cpp/output:/output:z \
 --device /dev/kfd \
 --device /dev/dri \
 ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
@@ -160,7 +207,90 @@ ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
 --cfg-scale 1.0 \
 --sampling-method euler \
 -v \
---diffusion-fa \
+-H 512 \
+-W 1024 \
+--seed -1 \
+--steps 10 \
 -o /output/output.png \
 -p "A picture of sign that says 'framework'"
+
+# Qwen
+podman run --rm \
+-v /home/ai/stable-diffusion.cpp/models:/models:z \
+-v /home/ai/stable-diffusion.cpp/output:/output:z \
+--device /dev/kfd \
+--device /dev/dri \
+ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
+--diffusion-model /models/qwen_image/Qwen_Image-Q4_K_M.gguf \
+--vae /models/qwen_image/qwen_image_vae.safetensors \
+--llm /models/qwen_image/Qwen2.5-VL-7B-Instruct.Q4_K_M.gguf \
+--cfg-scale 2.5 \
+--sampling-method euler \
+-v \
+--offload-to-cpu \
+-H 512 -W 512 \
+--flow-shift 3 \
+--seed -1 \
+-o /output/output.png \
+-p 'Everquest DND mash up poster that says "ever dungeons and dragons"'
+
+# SD3
+podman run --rm \
+-v /home/ai/stable-diffusion.cpp/models:/models:z \
+-v /home/ai/stable-diffusion.cpp/output:/output:z \
+--device /dev/kfd \
+--device /dev/dri \
+ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
+-m  /models/sd3/sd3.5_large.safetensors \
+--clip_l /models/sd3/clip_l.safetensors \
+--clip_g /models/sd3/clip_g.safetensors \
+--t5xxl /models/sd3/t5xxl_fp16.safetensors \
+-H 512 -W 512 \
+--cfg-scale 4.5 \
+--sampling-method euler \
+-v \
+--seed -1 \
+-o /output/output.png \
+-p 'Everquest DND mash up poster that says "ever dungeons and dragons"'
+```
+
+### Stable Diffusion CPP Server
+
+Uses OpenAI Compatible Endpoints
+
+```bash
+# z-turbo server
+podman run \
+-d \
+--name stable-diffusion-cpp-server \
+-v /srv/stable-diffusion.cpp/models:/models \
+-v /srv/stable-diffusion.cpp/build:/output \
+--device /dev/kfd \
+--device /dev/dri \
+--entrypoint "/sd-server" \
+--network host \
+ghcr.io/leejet/stable-diffusion.cpp:master-vulkan \
+--diffusion-model /models/z_turbo/z_image_turbo_bf16.safetensors \
+--vae /models/z_turbo/ae.safetensors  \
+--llm /models/z_turbo/qwen_3_4b.safetensors \
+--cfg-scale 1.0 \
+-v \
+--diffusion-fa \
+-H 1024 \
+-W 512 \
+--seed -1 \
+-l 0.0.0.0
+```
+
+## Openai API Web UI
+
+```bash
+# Will be available on port 8080
+podman run \
+-d \
+--network host \
+-v open-webui:/app/backend/data \
+--name open-webui \
+--restart always \
+ghcr.io/open-webui/open-webui:main
 ```

active/device_framework_desktop/quadlets/ai.pod

@@ -0,0 +1,9 @@
+[Pod]
+# llama.cpp
+PublishPort=8000:8000/tcp
+# open-webui
+PublishPort=8080:8080/tcp
+# anything-llm
+PublishPort=3001:3001/tcp
+# ollama
+PublishPort=11434:11434/tcp

active/device_framework_desktop/quadlets/anything-llm.container

@@ -0,0 +1,21 @@
+[Unit]
+Description=An Anything LLM Frontend for Local AI Services
+
+[Container]
+Pod=ai.pod
+Image=docker.io/mintplexlabs/anythingllm
+Volume=anythingllm:/app/server/storage
+Volume=/home/ai/anything-llm/.env:/app/server/.env:z
+Environment=STORAGE_DIR=/app/server/storage
+AddCapability=SYS_ADMIN
+User=1000
+Group=1000
+
+[Service]
+Restart=always
+# Extend Timeout to allow time to pull the image
+TimeoutStartSec=900
+
+[Install]
+# Start by default on boot
+WantedBy=multi-user.target default.target

active/device_framework_desktop/quadlets/llama-server.container

@@ -0,0 +1,28 @@
+[Unit]
+Description=A Llama CPP Server Running GPT OSS 120b
+
+[Container]
+Pod=ai.pod
+Image=localhost/llama-cpp-vulkan:2026-01-12-10-13-30
+Volume=llama-server-cache:/root/.cache
+AddDevice=/dev/kfd
+AddDevice=/dev/dri
+
+Exec=-hf ggml-org/gpt-oss-120b-GGUF \
+--ctx-size 32000 \
+--jinja \
+-ub 2048 \
+-b 2048 \
+--port 8000 \
+--host 0.0.0.0 \
+-n -1 \
+--n-gpu-layers 999
+
+[Service]
+Restart=always
+# Extend Timeout to allow time to pull the image
+TimeoutStartSec=900
+
+[Install]
+# Start by default on boot
+WantedBy=multi-user.target default.target

active/device_framework_desktop/quadlets/ollama-server.container

@@ -0,0 +1,19 @@
+[Unit]
+Description=An Ollama Server
+
+[Container]
+Pod=ai.pod
+Image=docker.io/ollama/ollama:0.13.5
+Volume=ollama:/root/.ollama
+AddDevice=/dev/kfd
+AddDevice=/dev/dri
+Environment=OLLAMA_VULKAN=1
+
+[Service]
+Restart=always
+# Extend Timeout to allow time to pull the image
+TimeoutStartSec=900
+
+[Install]
+# Start by default on boot
+WantedBy=multi-user.target default.target

active/device_framework_desktop/quadlets/open-webui.container

@@ -0,0 +1,16 @@
+[Unit]
+Description=An Open Webui Frontend for Local AI Services
+
+[Container]
+Pod=ai.pod
+Image=ghcr.io/open-webui/open-webui:main
+Volume=open-webui-data:/app/backend/data
+
+[Service]
+Restart=always
+# Extend Timeout to allow time to pull the image
+TimeoutStartSec=900
+
+[Install]
+# Start by default on boot
+WantedBy=multi-user.target default.target