3 ms·How optimized are current neural networks to specific GPU architectures?by enchiridion 6y agoHow optimized are current neural networks to specific GPU architectures?