I've only read the abstract but they don't mention quantizing the weights or otherwise trying to shrink the model in any way. They're claiming to be able to ef

I've only read the abstract but they don't mention quantizing the weights or otherwise trying to shrink the model in any way. They're claiming to be able to efficiently run larger models without loading the entire thing into GPU memory. If they're using the same weights, the same architecture and just using tensor parallel operations to perform the forward pass that would imply no loss in quality. I'm sure there are trade-offs but they're not clear by just looking at the abstract.