This looks worth a try. Great test results, very good example output. No way to know if it’s cherry picked / overtuned without giving it a spin, but it will go on my list. Should fit on an M2 Max at full precision.
How do you figure out the required memory? The MoE aspect complicates it.
Each model added, no?
It does; in this case, though, a 25b f16 model will fit. The paper mentions an A100 80G is sufficient but a 40 is not; M2 Max has up to 192G. That said, MoEs are popular in lower memory devices because you can swap out the experts layers -- their expert layers are like 3-4b parameters, so if you are willing to have a sort of pause on generation where you load up the desired expert, you could do it in a lot less RAM. They pitch the main benefit here as faster generation, it's a lot less matmul to do per token generated.