This repo is built for the challenge in video https://www.youtube.com/watch?v=cAICT4Al5Ow by siraj.
The use of this is to identify if the given image has Cat or Dog in it.
Accuracy of the present model is 99% because of inception model used.
This script uses pre built model inceptionV3 and retrain tensor flow with new categories of images for prediciton. New data is present in data/train folder. Any number of new objects/categories can be trained by this script by just placing objects in this(train) folder.
##Requirements
- Tensorflow
- python
##Usage
- Change path(retrain_dir) to tensorflow retrain directory in script "retrain-tensorflow.sh"
- Run the script "retrain-tensorflow.sh"
- Run the label_image.py python script with argument as path to image.