Vision pytorch cpu

[jasg (jorge)]

how can I make the <> iteration run in parallel on all 32 cpus of my cluster