Skip to content

Using tensorflow for training custom images of Corgi

Notifications You must be signed in to change notification settings

vietstacker/RH-AI-STEM

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Mar 11, 2020
a8896e2 · Mar 11, 2020

History

31 Commits
Jan 7, 2020
Jan 7, 2020
Jan 7, 2020
Jan 7, 2020
Jan 22, 2020
Jan 12, 2020
Jan 7, 2020
Mar 11, 2020
Jan 10, 2020
Jan 14, 2020
Jan 7, 2020
Jan 7, 2020
Jan 7, 2020
Jan 7, 2020

Repository files navigation

RH-AI-STEM

Pre-requisites on an empty MacOS

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

About

Using tensorflow for training custom images of Corgi

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published