Simple Tinder algorithm able to swipe left and right based on the recommendations of a pre-trained deep neural network (Machine Learning).
Python
Switch branches/tags
Nothing to show
Clone or download
Latest commit 5eb9776 Dec 19, 2016

README.md

Deep Learning Tinder (Machine Learning)

With DIGITS, Caffe and Tinder token auto fetching

Welcome on the Tinder Bot app whose recommandations are based on Deep Learning for image recognition. Don't forget to star or fork this repo if you appreciate the bot!

How does it work?

Everything is explained here: http://philipperemy.github.io/tinder-deep-learning/

How to run the bot

git clone https://github.com/philipperemy/Deep-Learning-Tinder.git
cd Deep-Learning-Tinder
mv credentials.json.example credentials.json
vim credentials.json # edit and fill the variables (explained in the next section: Configuration)
python main.py

If the configuration file is correct, you will see in the logs: Successfully connected to Tinder servers.

Configuration of credentials.json

  • FB_ID : The id of your facebook. Your profile is available at https://www.facebook.com/FB_ID where FB_ID is your id.
  • FB_EMAIL_ADDRESS : Your Facebook email address.
  • FB_PASSWORD : Your Facebook password.
  • API_HOST: URL of your NVIDIA Digits server. More information available here : https://github.com/NVIDIA/DIGITS
  • MODEL_ID: ID of your trained model. It appears in the URL when you click on your model in the DIGITS interface, like this: API_HOST/models/MODEL_ID

FB_EMAIL_ADDRESS and FB_PASSWORD are used to retrieve your FB_AUTH_TOKEN, useful to request the Tinder token.

Your configuration should look like this:

{
  "FB_ID": "tim.cook",
  "FB_EMAIL_ADDRESS": "tim.cook@apple.com",
  "FB_PASSWORD": "i_love_apple",
  "API_HOST": "http://localhost:5000/",
  "MODEL_ID": "20160619-000820-19f6"
}

Hope you will have some fun !