Skip to content

Latest commit

 

History

History
20 lines (12 loc) · 688 Bytes

README.md

File metadata and controls

20 lines (12 loc) · 688 Bytes

Exploring Multi-Category Object Detection with Bounding Boxes

How to use

Launch Jupter lab, open the Pytorch Lightning or FastAI based notebook and run!

Alternative (Run as Python Standalone Script)

Prerequisite

  • Python 3.8 and pip installed
  • Ensure IceVision is installed i.e. pip install icevision
  • Which should mean nbdev, jupyter, pytorch-lightning, fastai, pytorch, torchvision, fastcore, numpy ... are all installed

Optional Generate script

  • On command line, run nbdev_build_libs

Run e.g. FastAI version of Subcoco script

  • On command line, run python run_subcoco_effdet_icevision_fastai.py or python run_subcoco_frcnn_lightning.py