Good point. It's much more of an issue when running dense models with tensor parallelism. In that case, I'd look for an MoE model instead.
Good point. It's much more of an issue when running dense models with tensor parallelism. In that case, I'd look for an MoE model instead.