generative ai

Posted on by Jean-Luc Aufranc (CNXSoft) - 9 Comments on $199+ M5Stack AI Pyramid Computing Box Linux AI mini PC is based on Axera AX8850 SoC

$199+ M5Stack AI Pyramid Computing Box Linux AI mini PC is based on Axera AX8850 SoC

Axera AX8850 Linux Mini PC 4GB RAM Transparent case

M5Stack has just launched the “AI Pyramid Computer Box” AI mini PC powered by an Axera AX8850 octa-core Cortex-A55 SoC with a 24 TOPS NPU and featuring an unusual design since it’s housed in a pyramidal enclosure. The mini PC features 4GB or 8GB of RAM, 32GB eMMC flash, a microSD card slot, two HDMI 2.0 ports (see specs for details), a 4-mic array, a built-in speaker, two Gigabit Ethernet ports, four USB 3.2 ports, and two USB-C ports for data and power. It serves as a desktop-grade Edge AI computing platform for systems such as AI vision gateways, smart security systems, and local smart photo albums. M5Stack AI Pyramid Computing Box specifications: SoC – Axera AX8850 CPU – Octa-core Cortex‑A55 processor at 1.7 GHz NPU – 24 TOPS @ INT8 VPU Video Encoder – 8K @ 30 fps H.264/H.265 encoding, supports scaling / cropping Video Decoder – 8K @ 60 fps H.264/H.265 decoding, supports 16 channels 1080p parallel decoding, […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 2 Comments on Fusion HAT+ Review – Adding AI voice and servo/motor control to Raspberry Pi for robotics, Smart Home, or education

Fusion HAT+ Review – Adding AI voice and servo/motor control to Raspberry Pi for robotics, Smart Home, or education

Fusion HAT+ review LLM servo control camera

SunFounder has sent me a review sample of the Fusion HAT+ Raspberry Pi expansion board designed for motor and servo control using audio interactions with its built-in microphone and speaker, as well as LLM models. It can be used as an AI-enabled robot controller, a smart home hub, a voice assistant, or an interactive learning platform. In this review, after an unboxing and going through the installation of the Fusion HAT+ on a Raspberry Pi 5 2GB, I’ll mainly focus on the voice interaction part using text-to-speech (TTS), speech-to-text (STT), and local and cloud-based LLMs and VLMs, and also quickly test servo control to wave a flag using voice commands. SunFounder Fusion HAT+ unboxing I received the sample in the retail package reading “SunFounder Fusion HAT+ for Raspberry Pi” and detailing the key features, namely rechargeable battery, 12x PWM, onboard speaker and microphone, 4x 12-bit ADC, safe shutdown, 4x DC […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 2 Comments on M5Stack AI-8850 LLM Accelerator M.2 Kit offers an alternative to Raspberry Pi AI HAT+ 2

M5Stack AI-8850 LLM Accelerator M.2 Kit offers an alternative to Raspberry Pi AI HAT+ 2

M5Stack AI-8850 LLM accelerator M.2 kit

M5Stack has launched the “AI-88502 LLM Accelerator M.2 Kit 8GB Version” based on its LLM-8850 M.2 card with a 24 TOPS Axera AX8850 SoC, and offering an alternative to the Raspberry Pi AI HAT+ 2, supporting both LLM and AI vision workloads. The kit is comprised of the M.2 card and a Raspberry Pi-HAT 8850 board with USB PD power input for the card and Raspberry Pi 5, a 16-pin PCIe connector and 40-pin GPIO header for connection to the SBC, as well as accessories. M5Stack AI-8850 LLM accelerator M.2 kit specifications: M5Stack LLM‑8850 M.2 card SoC – Axera AX8850 CPU – Octa-core Cortex‑A55 processor at 1.7 GHz NPU – 24 TOPS @ INT8 VPU Video Encoder – 8K @ 30 fps H.264/H.265 encoding, supports scaling / cropping Video Decoder – 8K @ 60 fps H.264/H.265 decoding, supports 16 channels 1080p parallel decoding, supports scaling / cropping Memory (two options) 8GB 64‑bit LPDDR4x @ 4266 Mbps 4GB 64-bit LPDDR4x, 4266 Mbps (not […]

Posted on by Jean-Luc Aufranc (CNXSoft) - No Comments on MaixCAM2 modular 4K AI camera is based on Axera AX630 SoC with 3.2 TOPS NPU (Crowdfunding)

MaixCAM2 modular 4K AI camera is based on Axera AX630 SoC with 3.2 TOPS NPU (Crowdfunding)

MaixCAM2

Designed for makers and researchers, Sipeed’s MaixCAM2 is a modular AI camera built around an Axera AX630 AI SoC with 3.2 TOPS (INT8) to 12.8 TOPS (INT4) of AI performance, yet delivering 10 to 20x higher performance on OpenMV-N6 and Raspberry Pi 5, and about the same performance (140 FPS) as a 33 TOPS Jetson Orin Nano 4GB Super in YOLO11n. The camera comes with 1GB or 4GB RAM, a 32GB eMMC flash, and a microSD card slot for storage, a 2.4-inch touchscreen display, and an 8MP / 4Kp30 camera sensor with LED flash, as well as two microphones and a speaker, making it suitable for local LLM and VLM applications.  It also features two PMOD connectors for expansion, such as a thermal imager and a Time of Flight (ToF) ranging sensor. MaxiCAM2 specifications: SoC – Axera Tech AX630 CPU Dual-core Arm Cortex-A53 @ 1.2 GHz running Ubuntu Linux 32-bit […]

Posted on by Debashis Das - No Comments on Lanner EAI-I351 – An NVIDIA Jetson Thor Edge AI computer with 100GbE QSFP28 port and 8x GMSL2 camera

Lanner EAI-I351 – An NVIDIA Jetson Thor Edge AI computer with 100GbE QSFP28 port and 8x GMSL2 camera

Lanner EAI I351 Jetson Thor Edge AI computer with QSFP28 port

Lanner EAI-I351 is a rugged edge AI computer built around the NVIDIA Jetson Thor platform, featuring a 100GbE QSFP28 port for high-bandwidth networking and 8x GMSL2 camera inputs for low-latency vision processing. The system supports up to 128 GB of 256-bit LPDDR5X memory, features up to four 25GbE lanes via a QSFP28 port, a 5GbE RJ45 port, HDMI 2.0, four USB 3.2 Gen1 ports, two RS232/422/485 serial interfaces with optional CANBus, digital I/Os, and audio interfaces. Three expansion sockets are offered with an M.2 NVMe slot, an M.2 E-Key for Wi-Fi/Bluetooth, and an M.2 B-Key for 4G LTE/5G cellular with dual Nano-SIM support. Designed for industrial deployments, the embedded computer can operate in a –25°C to 70°C temperature range and takes 24V/48V DC power input. It targets robotics, smart manufacturing, and autonomous systems such as collaborative robots, AMRs, and machine-vision-based factory automation. Lanner EAI-I351 specifications: Supported system-on-module – NVIDIA Jetson […]

Posted on by Debashis Das - 2 Comments on Hiwonder WonderLLM – An ESP32-S3 smart chat module with 2-inch touch display, 2MP camera, and dedicated voice chip

Hiwonder WonderLLM – An ESP32-S3 smart chat module with 2-inch touch display, 2MP camera, and dedicated voice chip

WonderLLM AI module

Hiwonder has introduced the WonderLLM, an ESP32-S3-based smart chat module that combines a 2MP camera, a 2.0-inch touch display, a speaker, and a microphone array to support both offline computer vision tasks and cloud-based Large Language Models (LLMs) via the XiaoZhi AI platform. The device ships with a dedicated voice chip (CI1302) that enables always-on wake-word detection, and a 4-pin I2C interface allows it to act as a smart vision/voice sensor for external controllers like Arduino, STM32, or other ESP32 boards. Typical applications include small robots, STEM education kits, interactive assistants, and vision-enabled projects where this device handles perception and interaction, and another MCU manages motion and control logic. Hiwonder WonderLLM specifications: Wireless Module – ESP32-S3-WROOM-1 SoC – Espressif Systems ESP32-S3 CPU – Dual-core Tensilica LX7 up to 240 MHz with vector extension for AI/ML workloads RAM – 512KB SRAM Storage – TBD Wireless – WiFi 4 and Bluetooth LE 5 […]

Posted on by Jean-Luc Aufranc (CNXSoft) - 6 Comments on Raspberry Pi AI HAT+ 2 review – A 40 TOPS AI accelerator tested with Computer Vision, LLM, and VLM workloads

Raspberry Pi AI HAT+ 2 review – A 40 TOPS AI accelerator tested with Computer Vision, LLM, and VLM workloads

Raspberry Pi AI HAT+ 2 review VLM chat test

Raspberry Pi sent me a sample of their AI HAT+ 2 generative AI accelerator based on Hailo-10H for review. The 40 TOPS AI accelerator is advertised as being suitable for LLMs (Large Language Models) and VLM (Vision Language Models), while delivering about the same performance as the first-generation AI HAT+ (Hailo-8) for AI vision/computer vision models. After going through an unboxing, I’ll assemble the AI HAT+ 2 to a Raspberry Pi 5 with 2GB of RAM fitted with a Raspberry Pi Camera Module 3, before quickly checking whether AI vision models work as expected, and spending more time on testing LLM and VLM samples. Raspberry Pi AI HAT+ 2 unboxing My sample had a somewhat long and rough trip from the UK to Thailand, and the package did not look that good when DHL delivered it. But luckily, nothing was damaged, and I got the AI HAT+ 2 with a […]

Posted on by Debashis Das - 5 Comments on Tablet-like, ESP32-P4-based 7, 8, and 10.1-inch HMI displays integrate Wi-Fi 6 connectivity, 5MP camera

Tablet-like, ESP32-P4-based 7, 8, and 10.1-inch HMI displays integrate Wi-Fi 6 connectivity, 5MP camera

ESP32 P4 HMI Display

Waveshare has recently released the ESP32-P4-WIFI6-Touch-LCD, a family of tablet-like, fully enclosed HMI display development boards built around the ESP32-P4 SoC. The company offers 7-inch, 8-inch, or 10.1-inch configurations, all designed for industrial HMI, smart home terminals, and edge AI applications. Since the ESP32-P4 does not include built-in wireless connectivity, Waveshare has integrated an ESP32-C6-MINI module for WiFi 6 and Bluetooth 5 (LE) support. The board also leverages the ESP32-P4’s peripherals for MIPI CSI/DSI interfaces, a 5MP camera, and various I/Os, including USB 2.0 OTG and USB-to-UART Type-C ports, an SDIO 3.0 microSD card slot, dual microphones with echo cancellation, a speaker driven by an onboard audio codec, GPIO expansion headers, and optional battery support. Waveshare ESP32-P4-WIFI6-Touch-LCD specifications: Wireless module – ESP32-P4-Core SoC – Espressif Systems ESP32-P4NRW32 CPU Dual-core 32-bit RISC-V HP (High-performance) CPU @ up to 400 MHz with AI instructions extension and single-precision FPU Single-RISC-V LP (Low-power) MCU core […]

Exit mobile version