4 ms·We do GPU-specific training and inference speedups, at CentML.by pavelstoev 4y agoWe do GPU-specific training and inference speedups, at CentML.genewitch 4y agoGrata, well deserved.