Raspberry Pi is all set to ramp up its machine learning abilities, co-founder Eton Upton said at tinyML Summit 2021. Upton said the in-house chip-development team has been working towards improving the hardware’s machine learning skills. 

The Raspberry Pi 400 is a computer integrated into a single compact keyboard that runs on Raspberry Pi 4, released in 2019. The Raspberry Pi 4 came up with a significant upgrade in its hardware, architecture, and operating system. 

Pi 4 can run basic machine learning algorithms using its built-in camera input for image recognition. It can accomplish basic tasks such as recognising objects, observing movement or running basic inference tasks. It allows for better load and run-time algorithms since code can be compiled quickly; thanks to its faster CPU and RAM. Machine learning tasks also perform twice as much better on Pi 4 as it is vastly more powerful than its previous iterations.

Going by Upton’s tinyML summit presentation, most of their work appears to be focused on developing lightweight accelerators for ultra-low power machine learning applications. 

For this, Raspberry Pi will use either of the three current generations of ‘Pi Silicon’ boards or low-cost, high-performance boards, two of which – SparkFun’s MicroMod RP2040 and Arduino’s Nano RP2040 Connect – are from board partners.

The last one works on an ArduCam Pico4ML from ArduCam that incorporates a camera, microphone, screen, and machine learning into the Pico package. It is a standalone microcontroller that does not need a CPU.

He also said the future chips might have lightweight accelerators, possibly 4-8 multiply-accumulates (MACs) per clock cycle compared to less than one MACs per clock cycle of RP2040.

The post Raspberry Pi To Get Machine Learning Boost appeared first on Artificial Intelligence.

The Raspberry Pi Pico was released in January, and while this modest US$4 microcontroller didn’t make big waves in the PC community it was an important first – as the debut platform of the first ‘Pi Silicon’ CPU, the RP2040. Obviously at that price point this is a low power processor (40nm dual-core Arm Cortex-M0+ at 133MHz) with limited scope, but entirely capable of the microcontroller tasks for which it will be purposed. The Raspberry Pi Pico is reportedly selling very well with a quarter million units sold and three quarters of a million on backorder.

It is just a few weeks later and Raspberry Pi co-founder Eben Upton has already started to discuss what might be next for Pi Silicon – and the answer is a huge uplift in performance for machine learning tasks. Upton and his in-house ASIC development team reckon it is very worthwhile now to pursue the development of a lightweight accelerator to be used for ultra low power machine learning applications.

Above is the state of play comparing the current (bad) ML performance of RP2040 powered boards with rivals. In particular, you can see the Pico can complete a ‘detect a person’ task using ML processing in 2,200ms. That result is a bit slow compared to competitors. In the future, within 6 months or so according to Upton, the aim is for ML accelerators to make it into Pi Silicon to complete this detection work in 200ms or less.

The key forward-looking slide from Upton’s presentation is reproduced above and shows the competitive environment for an upcoming Pi Silicon based board. The final bullet point outlines how the second Pi Silicon processor might be made up. It will be designed as a ‘lightweight accelerator’ and feature four to eight multiply-accumulates (MACs) per clock cycle, according to this info.

The post Next Pi Silicon processor to feature machine learning acceleration appeared first on Artificial Intelligence.