Ultra_net : A FPGA-based Object Detection for the DAC-SDC 2020

This is a repository for FPGA-based neural network inference. The design won first place in the 57th IEEE/ACM Design Automation Conference System Design Contest (DAC-SDC).Designed by:

BJUT_runner Group, Beijing University of Technology

Kang ZHAN, Junnan GUO, Bingyan SONG, Wenbo ZHANG, Zhenshan BAO

The final rankings are published at https://dac-sdc-2020.groups.et.byu.net/doku.php?id=results

Repository Organization

• train: Contains the training script.
• model: Contains the pre-trained weights, model script and test script.
• quantization: Contains the python script which process the model generation header files to be used in Vivado HLS.
• hls: Contains the Vivado HLS implementation of Ultra_net.
• vivado: Contains the Vivado Block Design files.
• deploy: A Jupyter notebook showing how to use the FPGA based neural network to perform object detection on ultra96-PYNQ.

train

• cd train/yolov3/
• python3 train.py --multi-scale --img-size 320 --multi-scale --batch-size 32

quantization

• cd quantization/
• python3 torch_export.py
• python3 ultranet_param_gen.py