This is a thin wrapper around Tensorflow Object Detection API for easy installation and use. The original installation procedure contains multiple manual steps that make dependency management difficult. This repository creates a pip package that automate the installation so that you can install the API with a single pip install.
pip install tf-object-detection[tf]
Or for tensorflow with GPU support,
pip install tf-object-detection[tf-gpu]
All the scripts from tensorflow object detection APIs work out-of-box. You can find an example usages from the API's model_main.py.
import object_detection
- setup.py: The python packaging script.
PROTOC
Environment variable should be set so thatsetup.py
can compile as part of thesetup.py
execution.
# Download and setup all required dependencies required for tensorflow object
# detection libarary to work correctly.
$ python setup.py install
- tf_object_detection: A git submodule pointing to the version of tensorflow object detection this thin wrapper is for.