eetfm_automation
July 8, 2018 ยท View on GitHub
eetfm_automation: Export and Evaluate TensorFlow Model Automation based on TensorFlow Object Detection API.
Note: Currently only faster_rcnn based models supported (e.g mask_rcnn)
How to use this repository
- Place standard TF Object_detection API model directories inside the
base_modelsfolder, containing at least the following files:model.ckpt.data-00000-of-00001 model.ckpt.index model.ckpt.meta pipeline.config - Create a copy of
config.sample.ymlnamedconfig.ymland modify the params to your needs - run
python eetfm_automation.py - the new models are exported to the
export_modelsfolder - evaluation results are saved to a new
evalfolder inside each model