FPGA AI Suite Custom Models - 15 Minutes FPGA AI Suite enables inference IP generation for Altera® FPGAs. This training will discuss using a model that has layers not supported by the FPGA AI Suite. It will cover a quick example of a model unsupported by removing a layer from the end of a model to test performance metrics. It also looks at an in-depth example of taking a model with an unsupported layer, modifying, retraining, and generating IP to deploy to the FPGA. Course Objective: At course completion you will be able to: Understand what a Custom Model is! How to make the model supported fully by FPGA through: · Modify layers · Tuning hyperparameters · Retraining How to use FPGA AI Suite to: · Generate the OpenVINO™ Model Optimizer Intermediate Representation (IR) files · Compile the model · Use the Architecture Optimizer · How to take the new model and generate Quartus® Prime Software IP · How to import your IP into Platform Designer and use it in your design Skills Required: Basic Knowledge of: Python* Machine Learning (ML) frameworks such as TensorFlow* and PyTorch* FPGA design flow Working knowledge of Quartus Prime software Some knowledge of Altera® OpenVINO toolkit If the audio for the course does not start automatically, press pause and then play on the course player. The transcript of the course audio is available in the Notes or closed captioning (CC) feature of the player. If you need assistance with this course, please email fpgatraining@altera.com . Reference Course Code: FPGA_OFPGAAISUITECSTMMDL. FPGA_OFPGAAISUITECSTMMDL. <p>FPGA AI Suite Custom Models</p>