CNN_IMAGE_CLASSIFICATION

A CNN classification model that achieves 85% accuracy on a small dataset of cats and dogs This model is really showing the power of a CNN classification model, in just 64 lines of code. This model is able too classify cats and dogs with a very high accuracy and in a limited training time. It is a simple 3 layers CNN written in keras, the focus is on many filters for finding diffrent patterns, and maxpooling to make it easier too compute. The dataset can be found on https://www.kaggle.com/c/dogs-vs-cats there is alot more pictures then the 2048 pictures i used for training, and with using more pictures the accuracy will defenitly rise. Tensorflow 1.8 has been used. The data folder is made up like the following folder called data, in that folder two folders called train and validation in both of these folders there are two more folders called cats and dogs. Where in the train there is 1024 pictures in the cat folder and 1024 pictures in the dog folder, in the validation folder it is 426 pictures in each. To run this model, simply install tensorflow via pip: pip install tensorflow This will get you the cpu version and the training time will not be too long.