Skip to content

Latest commit

 

History

History
17 lines (14 loc) · 796 Bytes

README.md

File metadata and controls

17 lines (14 loc) · 796 Bytes

Deploy to custom hardware

This section is about deploying the trained models of a hardware accelerator, specifically an FPGA from type Xilinx ZCU102.

The deployment is done using the following steps:

  1. Inference the model to be checked if the model is valid.
  2. Quantize the model using VitisAI.
  3. Compile the quantized model for a defined DPU.
  4. Finally, the target application for the FPGA is created.

To deploy a model TensorFlow model to FPGA , run the script deploy.sh

user@hostnmame:~$ source ./deploy.sh <dataset path> <model path> <target name>

dataset path: path to the datasets, which has been used to train the model.
model path: path to the saved model.
target name: the name used to save the output application for the FPGA.