I want to build an LLM that generates a basic course on drawing. Every video is 30seconds to 1 min that depicts basics of drawing by basic examples.
How can I achive it? Where should I start from ?
LLM’s work by having a huge training set of written text, from which the model learns to make predictions about sequences of letters.
For what you’re discussing, you’d need a huge training set of instructions for how to create a drawing. Not just a training set of finished drawings, but a training set that details the sequence of actions in creating a drawing.
I am not aware that any such training set exists.
Thanks TMosh. Do you think if I split a bunch of videos into a smaller chunks , enrich it with a data and train my model this way ? I don’t think for such kind of model I need lots of examples ? I can surface recipes from popular hubs and try to make it done this way
You’re certainly welcome to give it a try. Please report back your results, I’d like to see how it works.