Mistral
1 update on Mistral.
Apple runs an 816B model across four Mac Studios at 36 tok/s with MLX
Apple Machine Learning Research reports over 36 tok/s for an 816B-parameter MoE model split across four Mac Studios over Thunderbolt 5 RDMA.
1 update on Mistral.
Apple Machine Learning Research reports over 36 tok/s for an 816B-parameter MoE model split across four Mac Studios over Thunderbolt 5 RDMA.