Switch text-to-image and automatic speech recognition (ASR) back to using the Hugging Face inference client; Zero GPU cannot accommodate the time it takes for those tasks

Switch to use GPU instead of inference client

Move environment variable querying code out of the Gradio UI-construction functions all the way to the root of the application, `app.py`

Move environment variable querying code out of the inference functions

Add documentation

Reorganize structure for even less code clutter; `app.py` is greatly slimmed down

Reorganize structure for less code clutter