4 ms·

I would say keras or mxnet for speed and production. PyTorch for research. By this point there are hardly any cases when it’s worth it to descend to lower Tenso

by mlthoughts2018 8y ago

I would say keras or mxnet for speed and production. PyTorch for research. By this point there are hardly any cases when it’s worth it to descend to lower TensorFlow levels.