3 ms·
Tensorflow or Pytorch. Plain Numpy for quick prototyping/testing. Sometimes have to write/modify Cuda kernels.
by p1esk 9y ago
Tensorflow or Pytorch. Plain Numpy for quick prototyping/testing. Sometimes have to write/modify Cuda kernels.
loading thread…