This workshop provides practical insight into AMD Vitis™ AI development for AMD Versal AI Edge and Versal AI Edge Gen 2 devices. The session outlines the core architecture of the Vitis AI stack, focusing on three primary components: the Neural Processing Unit (NPU) IP for concurrent model execution, the model compilation toolchain (Quantizer and Compiler) for INT8 and BF16 variants,
and the Vitis AI Runtime (VART) for C++ and Python application integration.
Attendees will experience the standard deployment workflow, which includes model framework based preparation on a x86 host (PyTorch, TensorFlow) and execution and postprocessing on base of the VART or ONNX applications within target platforms the NPUs are built into. The discussion will
close with touching on debugging and performance evaluation techniques to enable deeper dive for optimization.
11/17/2026 - 11/17/2026 Time Zone : (GMT+01:00) Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna Seats Remaining : 93 Venue : Online - PLC2 Address :
