I wonder why we still have techniques like diffusions and GANs, we won't reach very far with only a few techniques like this. The field needs new approaches.
It sounds like we need to stop abandoning technologies when the new hotness comes around. Revisiting old technologies with new hardware and data sets seems to be the better approach.
Comments
I wonder why we still have techniques like diffusions and GANs, we won't reach very far with only a few techniques like this. The field needs new approaches.
It sounds like we need to stop abandoning technologies when the new hotness comes around. Revisiting old technologies with new hardware and data sets seems to be the better approach.