just checking out the google docs file attached which lists the benchmarked models. Obviously, those architectures are not really getting any significant speedups with torchdynamo? or Is there any updated list of speedups? @jansel
just checking out the google docs file attached which lists the benchmarked models. Obviously, those architectures are not really getting any significant speedups with torchdynamo? or Is there any updated list of speedups? @jansel