In order to optimize your browsing experience we uses cookies. You agree to the usage of cookies when you browsing this site.
For more information regarding cookies and the processing of your personal data please read our Privacy Policy.
I understand
inner_banner_support
SUPPORT
Listen to Your Voices and Let Us Know Your Needs
Documentation Database
IP and Reference Design
Interface
GoAI
GoAI 2.0 Demo - Digit Classification
In this video, we introduce Gowin's GoAI 2.0 machine learning and inference solution, demonstrating its use for digit classification on GoAI 2.0 development boards. This setup accepts camera input, which can capture handwritten or machine-typed digits (0-9). The camera feeds the image into a buffer, from which it is processed by an ARM Cortex processor and then loaded into the GoAI 2.0 machine learning coprocessor. The coprocessor, preloaded with a trained model, classifies the digits and outputs results through GPIOs that can control LEDs and display text overlays on HDMI output. Watch to see edge-based processing in action, showcasing real-time classification directly on embedded devices. For more details, visit www.gowinsemi.com or contact info@gowinsemi.com.
GoAI 2.0 Demo- Car and Person classification with HDMI input
In this video, we demonstrate the capabilities of Gowin’s GoAI 2.0 development boards for object classification using HDMI input. The GoAI 2.0 platform supports a range of sensory inputs and performs real-time classification of objects such as cars and people. By leveraging HDMI input on the GW2A series boards, developers can easily stream video from a PC during the machine learning solution development phase, enabling efficient testing and debugging. Although the GW1 and SR4 development boards do not include HDMI input due to lower pin count, they still support camera-based classification for cars and people, ensuring versatile performance. Watch this demonstration to learn more about how GoAI 2.0 streamlines object classification at the edge.
擷取
GoAI2.0 Demo - Audio Phrase Detection
GoAI2.0 is an edge AI solution which is based on GOWIN FPGA, scalable from low cost, low power, low density devices to high performance devices. Audio Phrase detection is demonstrated here.
擷取
GoAI2.0 presentation
Join Grant Jennings, Director of International Marketing at GOWIN Semiconductor, as he delves into the innovative capabilities of GoAI 2.0, GOWIN’s edge-targeted machine learning inference solution. This comprehensive video explores the benefits of FPGA technology, particularly in parallel and pipelined computations, which make it ideal for real-time processing in AI edge applications. Grant explains how GoAI 2.0 seamlessly integrates software and hardware, leveraging TensorFlow and FPGA tools to deploy trained models for various sensory inputs like cameras, microphones, and accelerometers. Discover the architecture, key features, development boards, and use cases, such as person and audio detection, all designed for flexible and efficient edge AI solutions. For more information, visit our GoAI webpage or reach out via gowinsemi.com.
擷取
GoAI2.0 Demo - Person detection
GOWIN is excited to showcase a human presence detection demo using the GoAI 2.0 platform, designed for edge AI applications. This demonstration highlights how GoAI 2.0 integrates with TensorFlow for model development, utilizing quantization through TensorFlow Lite, and compiles firmware with the GoAI 2.0 SDK for efficient processing on FPGA accelerators and MCUs. Watch this video to see real-time human detection in action and learn how edge-based AI can offer powerful, localized intelligence without relying on cloud connectivity.
擷取
GoAI update 2
In this video, we highlight the advanced capabilities of GOWIN’s GoAI platform, focusing on human presence detection and LED indicator response based on detected images. The latest version of the development board now includes HDMI Rx functionality, external PS RAM for enhanced scratchpad memory, and additional components for improved performance. This upgrade allows HDMI data to be streamed through the board, making it easier to debug neural networks. With GoAI 2.0, PS RAM serves as scratchpad memory, while SPI flash is used for storing coefficients, enabling larger neural network models. Future benchmarks on popular networks like TinyYOLO and MobileNet will demonstrate market-ready solutions. Watch to learn more about the evolving capabilities of GOWIN's GoAI technology.
擷取
GoAI update
A single chip FPGA AI solution demonstrated here. HDMI interface is included.