Although the Raspberry Pi Pico comes with the RP2040 chip that lacks the performance to implement machine learning inference for its applications. However, we saw a person detection use case through ArduCAM and TensorFlow lite interface. But, the processing performance of the use case was on the slower side. Additionally, a recent Eben Upton presentation also unveiled that due to low power requirements the board compensates the processing efficiency. Hence, it offers low-performance for edge inference and machine learning use cases.
Eben Upton’s teaser on improvement in machine learning and the future scope of “Pi Silicon” revealed potential growth and development in edge inference applications. The demand for RP2040 boards has given rise to the market necessity for more boards. This demand can only be fulfilled if more boards with RP2040 chip are available in the market and company “partners such as Adafruit, Pimoroni, Adafruit and Sparkfun are start releasing their own hardware, many with features not found on the Pico.”
The RP2040 SoC enables the maximum performance for machine learning inference at the lowest power, this is due to its energy-efficient dual Arm Cortex-M0+ cores working at a comparatively higher frequency of 133 MHz. Hence, there are a few third-party Raspberry Pi RP2040 boards dedicated to ML applications. Some of these are:
Arduino Nano RP2040 Connect Board for Machine Learning Inference
Arduino Nano RP2040 Connect board is one such board featuring an STMicro MEMS sensor with 9-axis IMU and microphone. These are for the data collection on which the edge inference modeling can be done. Additionally, it includes 16MB external SPI flash, a u-blox NINA WiFi & Bluetooth module for flexible connectivity. “This can allow the users to develop connected products leveraging the hardware powered by Raspberry silicon. A solid radio module with efficient performance, and the Arduino Create IoT Cloud.”
SparkFun’s MicroMod RP2040 Processor
SparkFun also came up with a MicroMod RP2040 processor similar to its other processor cards which are compatible with different SparkFun carrier boards. Hence, the MicroMod RP2040 card can act as a dynamic add-on for various carrier boards depending on the applications of the users. Specifically, the ML carrier board would be a great fit for the RP2040 card due to its ML functionalities. For more information visit the detailed article on SparkFun’s processor cards and carrier boards.
Arducam Pico4ML: The new ArduinoML TensorFlow Lite board
Additionally, Arducam has specifically named its RP2040 board as Arducam Pico4ML. The board comes with all Tensorflow Lite Micro use-cases on a single platform. Arducam says “as the RP2040 SoC is based on a high-clocked dual Cortex-M0+, it’s also a remarkably good platform for endpoint AI, or more specifically TinyML.” The new ArduinoML TensorFlow Lite board comes with the following functionalities:
- Wake word detection
- Magic wand
- Person detection
- On-device LCD display
- Other sensor-based analysis
However, as discussed earlier the person detection example using Arducam and TensorFlow lite with Raspberry Pi Pico was quite slow. So, it is still a dilemma if the ML use cases will perform efficiently on the Arducam Pico4ML. The board is not yet released, so you can join the waitlist and follow the Twitter updates of the board. The links for the same are available on the product page.
Looking at the few drawbacks of the Raspberry Pi Pico in terms of machine learning, Eben Upton’s presentation has also unveiled the Raspberry Pi’s plan to work on its current RP2040 chip to enhance its edge computing capacity for machine learning. The organization aims to build lightweight on-device AI accelerators for low power machine learning inference and edge computing.
Source: Tom’s Hardware
Saumitra Jagdale is a Backend Developer, Freelance Technical Author, Global AI Ambassador (SwissCognitive), Open-source Contributor in Python projects, Leader of Tensorflow Community India and Passionate AI/ML Enthusiast