gewuek

vitis_ai_custom_platform_flow. This project is trying to create a base vitis platform to run with DPU

flower_classification_vai_tf_numpy_array. The project is a simple example about how to use TensorFlow to train a ConNet model from labeled dataset and then use Vitis AI tools to deploy the model into Xilinx FPGA(ZCU102 board). To make it easier to deploy with Vitis AI. I just use numpy array as input data and OpenCV function to open images. And there is backup for DNNDK3.1 tag

vitis_ai_custom_platform_v1.2_tmp. For v1.2 test

flower_classification_vai_tf_dataset. The project show how to quick train a model and deploy it on Xilinx FPGA using TensorFlow and Vitis AI. It uses tf.data.Dataset API to handle input data, so that the input function need to be rewritten. And there is a backup for DNNDK 3.1 tag

imagenet_vai_tf_train2deploy. This tutorial descibe about how to train a custom network from imagenet dataset and deploy it on ZCU102 via Vitis AI