So, I read a cool article about a new text2image model released by OpenAI.
I decided to implement a simple webapp that would allow me to talk to a personal AIrtist.
* Also took the opportunity to test-drive GitHub's Copilot AI Code completion... Spoiler Alert: It didn't take long to impress me.
On the data engineering side of things, I want to store my prompt, the model's suggestions, and the piece I actually select.
- Rather than host the model locally, incorporate a serverless function which would call the model that is hosted on its own GPU-optimized container
- Develop pipeline to collect more data for active learning: User Prompt, Model Options, and final User Selection(s)