This is the code for the "How to Make an Image Classifier" - Intro to Deep Learning #6 by Siraj Raval on Youtube
This is the code for this video on Youtube by Siraj Raval as part of the Intro to Deep Learning Udacity nanodegree. The code uses Keras & Tensorflow to train a convolutional neural network on a labeled dataset of cats and dogs. Then, it will be able to classify novel cats and dogs pretty well.
- keras (https://keras.io/#installation)
- tensorflow (https://www.tensorflow.org/versions/r0.10/get_started/os_setup)
Install Jupyter notebook from here.
jupyter notebook in terminal when you are in the main directory, then a browser window will pop up. You can run the code in the browser easily.
Coding Challenge - Due Date: Thursday, February 23rd 2017 at 12 PM PST
The challenge for this video is to build an image classifer for 2 different animals. They can be any 2 animals. You can use Keras & Tensorflow. The demo notebook attached will act as a good guide, as well as the links i've provided in the video description.
Bonus points if you do one or both of the following.
Use only Tensorflow
You use transfer learning to improve your model.
Credits go to a Wizard named Rajiv Shah! I've merely created a wrapper to get people started.