ICP
Deep learning power with Movidius accelerator
The Mustang V100 series from ICP Germany is a family of PCI Express cards with the Intel Movidius Myriad X MA2485 chip.
The Vision Processing Unit (VPU) accelerates deep learning inference at the edge.
© ICP GermanyThis Vision Processing Unit (VPU) accelerates deep learning inference at the edge - i.e. where quick decisions need to be made without uploading to the cloud. The card from ICP is available with four or eight VPU units in various designs: as a classic PCI Express card, in Mini PCIe and in M.2 format.
The multichannel capability allows a different DL topology to be assigned to each individual VPU in order to carry out simultaneous calculations. AlexNet, GoogleNet, Yolo Tiny, SSD300, ResNet, SqueezeNet or MobileNet are only a small part of the supported topologies. Compatibility with the Open Visual Inference Neural Network Optimization (OpenVINO) toolkit from Intel is provided.











