This bot analyzes a user's tinder profile using two different machine learning methods:
- TextBlob uses NLP to get sentiment of text data (user.bio)
- CNN trained on tinder faces labeled = ['happy', 'neutral']
The model's architecture is a less deep and less perfected version of VGG-16's
Currently using predictions from a transfer learned version of InceptionV3 by google while I get the hyperparams of the model above down.
Coded with love <3, coffee and Lady Gaga
To run this bot, you need to make the folders
- all_faces/
- -happy_imgs/
- -neutral_imgs/
- db/
you will also need the requirements which I cba to write out
- Call 'connect.py' and use --help to guide you through it. This will make a creds.txt file for you
- auto_liker.py will like all users and add their data to the DB
- auto_mess.py is the message responder, sends new matches their stats
Currently you need to run auto_mess.py all the time, and it will get http errors - working on a fix
read notes.txt if you really care
Thank You Mary for tinderpro
- Pynder by charlie/wolf - git clone the repo
- Tensorflow - Google
- TextBlob