HackCMU2020

VN Prompts: A Twitter bot that posts visual novel story descriptions!

dataset.py

To add a new dataset, add a new subfolder in the data directory and subclass the Dataset base class to implement the get_lines method.

Before a dataset can be used for training, the Dataset.process method must be called with the source material to create a raw_lines.txt.

train.py

To train a model, combine samples of the dataset as desired in make_data, then run the main script.

run_model.py

A command-line tool for generating model text, optionally based on an input.

tweet.py

Automatically tweet generated descriptions.

Dataset

vndb: https://dl.vndb.org/dump/

Share this project:

Updates