picoLLM is a cross-platform, on-device LLM inference engine

picoLLM Raspberry Pi 5

Large Language Models (LLMs) can run locally on mini PCs or single board computers like the Raspberry Pi 5 but with limited performance due to high memory usage and bandwidth requirements. That’s why Picovoice has developed the picoLLM Inference Engine cross-platform SDK optimized for running compressed large language models on systems running Linux (x86_64), macOS (arm64, x86_64), and Windows (x86_64), Raspberry Pi OS on Pi 5 and 4, Android and iOS mobile operating systems, as well as web browsers such as Chrome, Safari, Edge, and Firefox. Alireza Kenarsari, Picovoice CEO, told CNX Software that “picoLLM is a joint effort of Picovoice deep learning researchers who developed the X-bit quantization algorithm and engineers who built the cross-platform LLM inference engine to bring any LLM to any device and control back to enterprises”. The company says picoLLM delivers better accuracy than GPTQ when using Llama-3.8B MMLU (Massive Multitask Language Understanding) as a […]

Easily build a robot car with the Car Base Board for the STM32F411 “Black Pill” board

Car Base Board Mounted PCB chassis

The Car Base Board from Applying Microcontroller Solutions is a modular platform for building robot car projects powered by the WeAct Studio Black Pill development board. The Black Pill board is an upgrade to the “Blue Pill 2” board and features the STM32F411CEU6 microcontroller running at 100MHz with 512 KB of flash memory, 128 KB SRAM, and a USB Type-C port for power and programming. The Care Base Board printed circuit board is a base controller that takes hardware expansions such as wireless modules, servos, and sensors to monitor and control a robot car. The onboard headers provide a straightforward way to wire these connections and help prevent a tangled mess (rat’s nest) of wires. The PCB’s design makes it easy to use widely-available, “generic” devices and boards in development and to power all of them with batteries. It also allows the developer to select their favorite wireless communication device. […]

Sipeed MaixCAM is a RISC-V AI camera devkit with up to 5MP camera, 2.3-inch color touchscreen display, GPIOs

Sipeed MaixCAM

Sipeed MaixCAM is an AI camera based on SOPHGO SG2002 RISC-V (and Arm, and 8051) SoC with a 1 TOPS NPU that takes up to 5MP camera modules and comes with a 2.3-inch color touchscreen display. The development kit also comes with WiFi 6 and BLE 5.4 connectivity, optional Ethernet, audio input and output ports, a USB Type-C port, and two 14-pin GPIO headers for expansion that makes it suitable for a range of computer vision, Smart audio, and AIoT applications. Sipeed MaixCAM specifications: SoC – SOPHGO SG2002 CPU 1 GHz RISC-V C906 processor or Arm Cortex-A53 core (selectable at boot) running Linux 700 MHz RISC-V C906 core running an RTOS 25 to 300 MHz low-power 8051 processor NPU – 1 TOPS @ INT8 with support for models such as Mobilenetv2, YOLOv5, YOLOv8, etc… Video Codec – H.264, H.265, MJPEG hardware encoding and decoding up to 2K @ 30fps Memory […]

BitNetMCU project enables Machine Learning on CH32V003 RISC-V MCU

Neural networks on the CH32V003

Neural networks and other machine learning processes are often associated with powerful processors and GPUs. However, as we’ve seen on the page, AI is also moving to the very edge, and the BitNetMCU open-source project further showcases that it is possible to run low-bit quantized neural networks on low-end RISC-V microcontrollers such as the inexpensive CH32V003. As a reminder, the CH32V003 is based on the QingKe 32-bit RISC-V2A processor, which supports two levels of interrupt nesting. It is a compact, low-power, general-purpose 48MHz microcontroller that has 2KB SRAM with 16KB flash. The chip comes in a TSSOP20, QFN20, SOP16, or SOP8 package. To run machine learning on the CH32V003 microcontroller, the BitNetMCU project does Quantization Aware Training (QAT) and fine-tunes the inference code and model structure, which makes it possible to surpass 99% test accuracy on a 16×16 MNIST dataset without using any multiplication instructions. This performance is impressive, considering […]

The Lark Weather Station works with Arduino, ESP32, micro:bit, Raspberry Pi, and other boards

Lark Weather Station Arduino ESP32 Raspberry Pi

The Lark Weather Station measures wind speed, wind direction, temperature, humidity, and air pressure through a range of sensors and connects to popular development boards such as Arduino UNO, ESP32, BBC micro:bit, Raspberry Pi, or DFRobot Unihiker through I2C or UART. We’ve seen several projects for Internet-connection weather stations that retrieve weather data from the web and display the results locally, but the Lark Weather Station allows the users to get atmospheric data right in his/her current location thanks to its built-in anemometer, wind vane, and built-in sensors, as well as expansion interfaces for additional sensors. Lark Weather Station specifications: Storage – 16MB flash good to store about 160 days of data (when data is recorded once per minute) Sensors Compass Anemometer Wind Speed: 0.5~12m/s Cover to protect the anemometer during storage/transport Wind vane and wind direction shaft to report the wind direction (eight directions) Temperature Range –20~60℃ ±0.2℃ Humidity […]

Raspberry Pi RP2040-powered 0.99″ rounded display is housed in CNC metal case

Waveshare RP2040 LCD 0.99 B Rounded Display

The Waveshare RP2040-LCD-0.99-B rounded display is the latest display module by Waveshare. This board is built around the Raspberry Pi RP2040 MCU and hosts a 128×115 pixels 65K color IPS LCD, along with a QMI8658C IMU, all inside a CNC metal case with an acrylic dull-polish bottom plate. The board also hosts a USB Type-C connector, an LDO, a 2MB NOR-Flash, a SH1.0 6PIN connector (adapting 4x GPIO pins), and a BOOT button. Waveshare RP2040-LCD-0.99-B rounded display module specification: Microcontroller – Raspberry Pi RP2040 dual-core Arm Cortex M0+ processor up to 133 MHz with 264KB SRAM Storage – 2MB on-board flash memory Display – 0.99-inch 128×115 pixel 65K color IPS LCD USB – 1x USB 1.1 Type-C port (host/device) for power and programming using drag-and-drop via USB mass storage Sensor – QMI8658C IMU Expansion – 4x GPIO via SH1.0 6PIN connector Power Management – Low-power sleep and dormant modes Dimension – […]

WeAct STM32G4 is a tiny board based on STMicro STM32G4 mixed-signal microcontroller

WeAct STM32G4 mixed-signal board

WeAct STM32G4 is a tiny development board based on a 170 MHz STMicro STM32G4 Arm Cortex-M4F mixed-signal microcontroller with DSP instructions and suitable for applications such as motor control, building automation, lighting, digital power meters, and more. Two versions of the board are offered one with an STM32G474CEU6 “Hi-resolution line” microcontroller equipped with 128KB RAM and 512KB flash, and the other with the lower-end STM32G431CBU6 “Access Line” MCU with just 32KB RAM and 128KB flash. The board also comes with a USB-C port for power and programming, three buttons, and two 24-pin headers. WeAct STM32G4 specifications: Microcontroller (one or the other) STMicro STM32G431CBU6 – Arm Cortex-M4F MCU @ 170 MHz with DSP instructions, 32KB RAM, 128KB flash, and math accelerator STMicro STM32G474CEU6 – Arm Cortex-M4F MCU @ 170 MHz with DSP instructions, 128KB RAM, 512KB flash, and math accelerator; high-resolution timer and complex waveform builder plus event handler (HRTIM) for […]

Waveshare RP2040-BLE is a Raspberry Pi RP2040 module with Bluetooth LE connectivity

RP2040-BLE

If you could not care less about WiFi, the Waveshare RP2040-BLE module combines a Raspberry Pi RP2040 microcontroller with an unnamed Bluetooth 5.1 dual-mode chip, and supports a USB-C add-on board useful during development. There are several Raspberry Pi RP2040 boards with a WiFi and Bluetooth module including the Raspberry Pi Pico W, but if you are only interested in using Bluetooth, the RP2040-BLE board from Waveshare may be more power-efficient and its design may be more suitable for integration into products. Waveshare RP2040-BLE specifications: MCU – Raspberry Pi RP2040 dual-core Arm Cortex M0+ microcontroller up to 133 MHz with 264KB of SRAM Storage – 2MB of onboard Flash memory. Connectivity Bluetooth 5.1 dual-mode function (BLE and classic) Controlled by serial AT commands Ceramic antenna Expansion 24x through and castellated holes with 14x multi-function GPIO pins. 2x SPI, 2x I2C, 2x UART, 3x 12-bit ADC, 14 x PWM FPC connector […]