air55555 opened this issue 2 years ago · 0 comments
Would be nice to make HCube inferencw faster using parallel model deploy into 2+ GPUs. Any options ?