Microchip edge AI solution "VectorBlox"

・A convolutional neural network-based edge AI/machine learning/inference accelerator.
・This solution has extremely low power consumption per processing performance (GOPS) and is ideal for image recognition processing.
・VecotrBlox operates on Microchip's mid-range FPGA "PolarFire FPGA & PolarFire SoC", which is ideal for industrial cameras and IoT edge devices.

VectorBlox Accelerator Software Development Kit (SDK)

・Convert neural network model to VectorBlox model.

- Supports a wide range of frameworks.
-TensorFlow, Caffe, MxNet, PyTorch, etc.

・CNN can be verified with a simulator before hardware implementation.

pretrained model

・We provide multiple tutorials using pre-trained models published on Github, etc.
・Face recognition and object recognition can be easily evaluated and verified using simulators and development kits.

PolarFire FPGA + VectorBlox Reference Design

・A reference design that works with the PolarFire Video Kit has been released.
・The reference design includes an FPGA hardware design and a Mi-V soft CPU (RISC-V)
Contains software code.
・By using this reference design, even those without FPGA design experience can quickly check and evaluate the operation of CNN.

PolarFire Video and Imaging Kit & PolarFire SoC Video Kit

Edge AI (VectorBlox) can be operated with these kits.

* Please refer to the following URL for the lineup of development kits and dedicated daughter cards.
https://www.microchip.com/en-us/products/fpgas-and-plds/boards-and-kits

Inquiry / Quotation

If you have any questions about this product or would like a quote, please contact us using the form below.