A dog and cat classifier created using convolutional neural networks using the library keras in python.The model has two convolution layer and uses adam as optimizer and accuracy as metric. Max pooling is used for pooling. The model achieved a test set accuracy of 85%.
Building this projects requires Python3 to run the script Code/kerasModel.py . The model is trained and saved in the file Code/trained_model.h5 .
The following python packages need to be installed for the script kerasModel.py to work:
- Keras
- PIL
- numpy
- webrowser
Set the directory name in the variable 'image_predict' and run the code under the comment 'Train the model' The result will open in a separate browser window.