CNN_Cat_vs_Dog

Convolutional Neural Network to distinguish between cats and dogs given only an image.
Uses Tensorflow library for the CNN model.

Requirements

• Kaggle Dataset - Dogs vs. Cats Redux: Kernels Edition - only train.zip
• Tensorflow / Tensorflow-gpu
  • CUDA Toolkit 8.0 (optional - to use with the gpu version)
  • cuDNN 7 (optional - to use with the gpu version)
• TFlearn
• OpenCV
• Numpy
• TQDM (optional - no added functionality or better performance - only for pretty progress bars)

How To Use

#Set TRAIN_DIR in CNN_catdog_train.py:17 to the directory where the downloaded images are.
TRAIN_DIR = ''

#Run CNN_catdog_train.py to train/load the model. Input an image path (to test) when asked.
$ python CNN_catdog_train.py