gen

This is a mini-project to connect a discord bot to a language model so that it might generate interesting responses from a prompt.

Dependencies

I'm using the following:

• discord.py v1.3.3
• transformers v2.6.0
• Python v3.8.2

Of course these have their own dependencies, I'm not your package manager.

How to use

With any luck you only need to do the following:

1. Install the dependencies
2. Clone the repo
3. Create a config.json that has at least a single key token with your bot's token.
4. python main.py

Optionally you can pass the path to a JSON file as the first command-line argument to main.py.

The bot is currently configured to listen for messages that start with "$transform" and passes everything that follows in that message to the language model to generate follow-up text. The generated text is sent as a new message.

Optional

• Add a link to your GitHub repo as the github key in config.json to hook up the $github command (else it'll point back to my repo 😝 )

Disclaimer

This is for fun, it's not intended to be anything. There are probably many bots like this, but this one is mine.