RH-AI-STEM

Pre-requisites on an empty MacOS

• Install brew:
  • /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
• Install pip, Cython, wget:
  • brew install pip
  • brew install wget
  • brew install python3
• Install virtualenv, Python3:
  • $ python3 -m venv ./venv
  • $ source ./venv/bin/activate
• Install Cython:
  • pip install Cython

Setting up environments

• Clone this directory to somewhere you like
• Run the command:
  • source env.sh

Traning Corgi models

• Traning:
  • $ cd models/research/object_detection
  • $ python model_main.py --logtostderr --model_dir=corgi_training/ --pipeline_config_path=corgi_training/corgi.config
• Export inference graph:
  • TODO: Looking for highest number of trained model in models/research/object_detection/corgi_training
  • $ python export_inference_graph.py --input_type image_tensor --pipeline_config_path corgi_training/corgi.config --trained_checkpoint_prefix corgi_training/model.ckpt-<highest_number> --output_directory inference_graph_corgi

Run testing

• Copy testing script into tensorflow models:
  • $ cd ../../..
  • $ cp corgi_detection.py models/research
• Run:
  • $ python models/research/corgi_detection.py

Notes:

• Change the backend of matplotlib in corgi_detection.py to a relevant one of running machine in order to using matplotlib for image rendering