DeepBird: The smartest bird in the nest!

DeepBird is a completely automated Recurrent Neural Network and Twitter bot! My goal is to make it easier to create autonomous and intelligent Twitter bots. Mostly recycled code from Word-RNN and Char-RNN

To Do List:

• Make DeepBird easier to use.
• Write Documentation.
• Make this more modular.
• Add Synergy Mode (Tweets and evaluates over CPU, endless Training over GPU.)

Requirements:

• python 3.6.x or newer.
• Tensorflow 1.1.0rc0 (GPU version recommended).
• Tweepy.
• Word-RNN (You can also use Char-RNN).
• Twitter account with an app attached to it.

Optional, but recommended:

• Google sheet with twitter archiver enabled. (To record all tweets made by the account.)

Basic Usage

• Create a New profile! python3 DeepBirdWizard.py

• Start Training python3 DeepBird.py --profile_dir PROFILE_DIR --train 1

• Start Running python3 DeepBird.py --profile_dir PROFILE_DIR --rly_post BOOL --debug BOOL

Currently Deployed

@KaleBobRoss is a simple Word-RNN trained on the transcripts from Bob Ross: The Joy Of Painting. @KaleBobRoss even has a spreadsheet archive.