About

Number recognition with MNIST on Raspberry Pi Pico + TensorFlow Lite for Microcontrollers

Device

• Raspberry Pi Pico
• LCDディスプレイ 2.8"240x320 SPI TFT LCDタッチパネルシリアルポートモジュール+ PCB ILI9341 5V / 3.3V
  • LCD: Ili9341 (SPI)
  • Touch Panel: TSC2046 (SPI)

Wiring

• pj_mnist/wiring.txt

Build

git clone https://github.com/iwatake2222/pico-mnist.git
cd pico-mnist
git submodule update --init
cd pico-sdk && git submodule update --init && cd ..
mkdir build && cd build

# For Windows Visual Studio 2019 (Developer Command Prompt for VS 2019)
# cmake .. -G "NMake Makefiles" -DCMAKE_BUILD_TYPE=Debug -DPICO_DEOPTIMIZED_DEBUG=on
cmake .. -G "NMake Makefiles"
nmake

# For Windows MSYS2 (Run the following commands on MSYS2)
# cmake .. -G "MSYS Makefiles" -DCMAKE_BUILD_TYPE=Debug -DPICO_DEOPTIMIZED_DEBUG=on
cmake .. -G "MSYS Makefiles" 
make

Model

• Google Colaboratory script for model creation in Keras and model conversion with quantization
  • pj_mnist/conv_mnist/tflmicro_mnist.ipynb
• Note : It looks uint8 quantization is not supported. You need to use int8 quantization, or use FP32 model.
  • https://github.com/raspberrypi/pico-tflmicro/blob/fbfff01ab9adc9b2e83b8eee6e001de13cf3b982/src/tensorflow/lite/micro/kernels/quantize.cpp#L65

Acknowledgements

• pico-sdk

  • https://github.com/raspberrypi/pico-sdk
  • Copyright 2020 (c) 2020 Raspberry Pi (Trading) Ltd.

• pico-examples

  • https://github.com/raspberrypi/pico-examples
  • Copyright 2020 (c) 2020 Raspberry Pi (Trading) Ltd.

• pico-tflmicro

  • https://github.com/raspberrypi/pico-tflmicro
  • Copyright 2019 The TensorFlow Authors. All Rights Reserved.