You already have the Sixfab AI HAT+ kit on hand, and it ships ready to assemble. Before you start, confirm the host platform you'll mount it on and the few items the kit does not include are in front of you.

What you'll need

What's in the box

Open the kit and confirm all six items are present before you start. The AI HAT+ board itself, every fastener, the PCIe cable, and the passive cooler ship in the same box, so nothing else is needed from Sixfab.

The 16-pin PCIe cable ships pre-attached to the AI HAT+ at the factory, so assembly starts on the Raspberry Pi 5 side. The kit's contents and a "Not in the box" reminder are in Step 1. Tighten every fastener finger-tight only.

Place the four M2.5 × 16 mm female-female spacers over the Pi 5's four mounting holes and fasten them from the underside of the Pi 5 with four of the M2.5 × 5 mm plastic screws.

Press the included 16 mm stacking header (2×20 female) straight down onto the Pi 5's 40-pin GPIO header until it is fully seated on the Pi 5. The header passes the 3.3 V and 5 V rails and the I²C ID lines up to the AI HAT+.

The header is intentionally not tall enough for its plastic body to reach the AI HAT+, and that gap is by design: it lets air move between the two boards instead of sealing the space off. When you mount the board in the next step, the AI HAT+ rests on the four spacers, not on the header. The header only carries the electrical connection. So press it firmly all the way down onto the Pi 5 and leave the clearance above it as it is.

The 16-pin FFC cable is already locked into the AI HAT+ at the factory; its free end is labeled Raspberry Pi 5. Hold the AI HAT+ upright above the Pi 5 so the cable reaches the Pi 5's PCIe connector, and connect that end before you lower the board.

1. Seat the cable in the Pi 5. Lift the latch on the Pi 5's PCIe connector, slide the free Raspberry Pi 5 end of the cable in with its label oriented as printed on the cable, seat it fully, and close the latch.

2. Lower the AI HAT+ onto the stack. Align the AI HAT+'s 40-pin female header with the stacking header on the Pi 5, then press the board straight down, firmly and evenly, until the headers are fully mated and the board rests on the four spacers. Let the FFC cable fold neatly into the gap between the two boards as the board comes down.

Fasten the AI HAT+ to the four spacers with the remaining four M2.5 × 5 mm plastic screws, driven from the top of the board. Eight screws now hold the stack: four securing the spacers to the Pi 5 from below, and four securing the AI HAT+ to the spacers from above. The hardware stack is complete. Finger-tight only.

The AI HAT+ kit includes a passive heatsink, pre-paired with a thermal pad, for the DEEPX NPU. Fitting it keeps the NPU cooler under sustained inference, so it is recommended for most workloads.

Peel the protective film off the thermal pad on the underside of the heatsink, then place the heatsink on the NPU exactly as shown below. Press down gently and evenly on the top of the heatsink so the thermal pad makes full contact across both surfaces.

Connect the official USB-C power supply to the Raspberry Pi 5 and boot it, then open a terminal. A single APT command installs the kernel driver, the DEEPX runtime, and the CLI tools (dxrt-cli, dxtop, run_model).

# 1. Refresh the package index
sudo apt update

# 2. Install the Sixfab APT repository
sudo apt install apt-repo-sixfab

# 3. Refresh again, then install the DEEPX runtime
sudo apt update && sudo apt install sixfab-dx

When the installer prompts Do you want to continue? [Y/n], type Y and press Enter. The driver compiles against your running kernel, so the first install can take 5–10 minutes depending on connection speed.

Sanity check: the runtime is reachable and the NPU enumerates

# 1. Confirm the runtime CLI is on PATH
dxrt-cli --version

# 2. Confirm the NPU enumerates over PCIe
lspci | grep -i deepx

The Raspberry Pi 5 negotiates the PCIe link at Gen 2 by default. To get the best inference throughput from the DEEPX NPU, switch the link to Gen 3 with two lines in the boot configuration file.

Open the boot configuration file:

sudo nano /boot/firmware/config.txt

Add the following lines at the end of the file:

dtparam=pciex1
dtparam=pciex1_gen=3

Save and exit (Ctrl+X, then Y, then Enter), then reboot:

sudo reboot

Before running inference, confirm that the kernel driver, the DEEPX runtime, and the firmware on the DEEPX NPU module are at the versions you expect. dxrt-cli -s prints a complete hardware and firmware status report in one shot.

# Full hardware and firmware status report
dxrt-cli -s

What each field in the report confirms

The sixfab-dx package ships with a pre-compiled YOLOv8 detection demo wired up to run_hello_world. A single command launches it on the NPU and opens a window with bounding boxes drawn in real time.

run_hello_world

A window opens showing the live feed with bounding boxes drawn around detected objects in real time.

Watch the NPU work in real time

Open a second terminal while the demo is running and launch dxtop. It prints per-core utilisation, voltage, clock, and temperature, much like htop for the NPU.

dxtop

Press q to quit.