Example - for building an inference engine via autoresearch type loop the model gets stuck on quantizing the projection matrices and focussed on that for almost 4-6 hours trying to extract every bit…