- Set up the CUDA Toolkit and CUDNN versions that support TensorFlow-GPU version 2.7.0. You can find the installation instructions and tested build configuration at https://www.tensorflow.org/install/source#gpu.
- Open main.pynb
- Install the requirement:
!pip install -r requirements.txt
- Run:
from pipeline_func import *
- Modify the input_video with the desired input video path, and change out_folder with the desired output folder name.
print("Start") # To see the progress (if running in VSCode)
input_video = "input/1.mp4"
out_folder = "video_output"
# function definition located in pipeline_func.py file (line 498)
road_damage_detection(input_video, out_folder)
- Wait