Here's the prototype for LCM and physical objects demo created for the prompt "LCM" during the AIxUIdailies coding challenge. You can see the demo video here.
See all of the prototypes for the challenge here.
- Do the fal authentication: https://www.fal.ai/docs/authentication/key-based
- Install the dependencies:
pip install -r requirements.txt
-
run
listen.py
to get the voice to text for prompts -
then run
see.py
to take regular snapshots from the cam -
then run
run_lcm_fal.py
to loop taking the prompt and the image and running LCM, saving the image link to a txt file -
use something like Photobooth to see what's on the webcam
-
Run server.py to see the image
There's a ton of latency, so the next thing to do here is rewrite it in the browser, which fal is great for.