NPU Environment Setup
The NIO 12L is powered by the MediaTek MT8395 (Genio 1200), which integrates an APU (AI Processor Unit) and AI Accelerator (Aia). You can use the NeuroPilot software stack on Ubuntu to access hardware acceleration for AI inference.
Install Firmware and Runtime
First, install the APU firmware and Neuron runtime:
sudo apt install mediatek-apusys-firmware-genio1200
sudo apt install mediatek-libneuron mediatek-neuron-utils mediatek-libneuron-dev
sudo reboot
Install Neuron SDK
After rebooting, install the Neuron SDK:
snap install mtk-neuropilot --edge
The --edge channel publishes the latest packages, which may update more frequently than the stable channel. For stable production use, check the Snap Store for each channel's version and support status.
Verify Installation
vpu5_test
sudo vpu5_test -a ks -l 10
The result should show PASS.
ncc-tflite
ncc-tflite is a command-line tool for converting models from tflite format to dla format:
ncc-tflite --help
neuronrt
neuronrt can run dla models to verify model integrity:
neuronrt -v
Run Benchmark
sudo mkdir -p /usr/share/benchmark_dla
sudo cp /usr/share/neuropilot/benchmark_dla/* /usr/share/benchmark_dla/
sudo apt install python3-pip
sudo pip3 install numpy
sudo python3 /usr/share/benchmark_dla/benchmark.py --auto
Supported Model Formats
- TFLite (
.tflite) - ONNX (
.onnx) — requires loading via the Neuron API - Caffe (
.caffemodel)
For more details, refer to the official MediaTek documentation: