- Install TensorFlow 1.14 (not 2)
- Download https://github.com/tensorflow/models/ repository
- Add models/, models/research/, and models/research/slim/ to the
PYTHONPATH
variable. - Install
pycocotools
from this clone (works with Windows): https://github.com/philferriere/cocoapi - Follow installation guide from https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md (Except the cocoapi part)
- Download https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10 for inference script in webcam and TFRecord data generator script
- Mainly, try to follow this local training guide and fix the bug one by one by providing the requirements: https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/running_locally.md
There are 2 main concerns I have when following unofficial guide:
- I don't want to deviate much from the official TensorFlow object detection API documents.
- I don't want to meet with the difficulty, tedious, and buggy guide.