[mattbaconz]

2.4GB in browser is crazy. I’m curious whether quantization works here or if the model quality falls apart too much.