Complex Neural Networks

Xiangteng NPU Chip (1) – Fully Domestic Autonomous Defined Neural Network Chip for Strong Real-Time, High-Performance, and High-Concurrency Application Scenarios

by
Xiangteng NPU Chip (1) - Fully Domestic Autonomous Defined Neural Network Chip for Strong Real-Time, High-Performance, and High-Concurrency Application Scenarios

The Xiangteng NPU chip HKN201 is a fully domestic autonomous defined neural network chip, aimed at strong real-time, high-performance, and high-concurrency application scenarios.1. Chip Specifications Overview of Chip Specifications Chip Specification List AI Processor 4-core 2048MAC NPU core Clock frequency 1GHz AI Computing Power 16TOPS INT8 8TFLOPS FP16 Software Specifications Supports PyTorch, Caffe, Darknet deep … Read more

Xiangteng NPU Chip (2) – AI Chip for Inference Applications

by
Xiangteng NPU Chip (2) - AI Chip for Inference Applications

The Xiangteng NPU chip is an AI chip designed for inference.When applying artificial intelligence, we encounter two concepts: Training and Inference, which are two stages in the implementation of AI. Let’s first understand these two issues. What is the difference between training and inference? What key points should we focus on to distinguish between AI … Read more

Custom Solutions! RK3588 Image Processing Boards with Various Interfaces to Suit Your Needs

by
Custom Solutions! RK3588 Image Processing Boards with Various Interfaces to Suit Your Needs

The RK3588, as Rockchip’s flagship chip, has garnered attention from many industries due to its industrial-grade computing power. However, not every RK3588 image processing board can be used directly, as it requires deep customization based on the camera interface and application scenario. Chengdu Huishi possesses the capability for rapid integration and customization. With years of … Read more

Smart Recognition of Handwritten Digits Using ESP32 4.3-Inch Screen

by
Smart Recognition of Handwritten Digits Using ESP32 4.3-Inch Screen

What functions can be achieved by combining the ESP32 screen module with an LED light board? The project presented today only requires a swipe on the 4.3-inch touchscreen, and the neural network can instantly understand your handwriting and display the number you wrote on an 8×8 LED light board. Isn’t it amazing? Let’s see how … Read more

Understanding the Core of Multi-Axis 3D Printing: Open Source Non-Planar Support-Free Slicing Algorithm

by
Understanding the Core of Multi-Axis 3D Printing: Open Source Non-Planar Support-Free Slicing Algorithm

AM Yidao Introduction:Since the inception of 3D printing technology, it has been plagued by a stubborn problem—support structures.These annoying supports not only waste materials but also increase post-processing workload, and even affect print quality and mechanical strength.However, a groundbreaking study from the University of Manchester is completely changing this situation!The neural network slicing algorithm proposed … Read more

Running Ollama and DeepSeek-R1 on Raspberry Pi 4

by
Running Ollama and DeepSeek-R1 on Raspberry Pi 4

My Hardware List Raspberry Pi 4 Board with 4GB RAM 64GB Memory Card Mobile Phone Charger Raspberry Pi 4 Specifications Processor: Broadcom BCM2711, quad-core Cortex-A72 (ARM v8) 64-bit SoC @ 1.5GHz Memory: 4GB LPDDR4 Connectivity: 2.4 GHz and 5.0 GHz IEEE 802.11b/g/n/ac wireless LAN, Bluetooth 5.0, BLE Gigabit Ethernet 2 × USB 3.0 ports 2 … Read more

Google TPU v2 Training Chip

by
Google TPU v2 Training Chip

In 2017, Google updated their TPU series. Google referred to this generation of TPU as a “domain-specific supercomputer for training neural networks.” It is evident that compared to TPU v1, which focused on inference scenarios, TPU v2 shifted its design focus towards training-related scenarios. Looking back at history, around 2017, groundbreaking work in deep learning … Read more

Why Implement GPT-2 in Pure C Language? Karpathy Responds to Online Criticism

by
Why Implement GPT-2 in Pure C Language? Karpathy Responds to Online Criticism

Machine Heart ReportEditor: Xiao Zhou Karpathy: for fun. A few days ago, Andrej Karpathy, former head of Tesla Autopilot and OpenAI scientist, released a project called “llm.c” that implements GPT-2 training in just 1000 lines of code on CPU/fp32. llm.c aims to simplify large model (LM) training — using pure C language / CUDA, without … Read more

Building Brain-like Visual Sensors with Atomic LEGO

by
Building Brain-like Visual Sensors with Atomic LEGO

A public academic platform initiated by overseas scholars Sharing information, integrating resources Exchanging academic ideas, occasionally discussing other topics Over 80% of the information processed by the human brain is obtained through the eyes. The human eye not only detects and synchronously processes information but also operates at an extremely low overall power consumption (far … Read more

How to Decide on the Next DSP Processor Solution to Meet Evolving SoC Designs

by
How to Decide on the Next DSP Processor Solution to Meet Evolving SoC Designs

Assuming you are a system architect making decisions regarding the digital signal processor (DSP) in an SoC, one major consideration is how to expedite the design process of the SoC, leading to faster tape-out and production. Another consideration is how to create, debug, and modify software-defined features post-tape-out to mitigate risks. Today’s in-depth article discusses … Read more