Cuda programming

- From Stanford CS231N(2017) lectures, "TensorFlow is a safe bet for most projects. Not perfect but has huge community, wide usage. Maybe pair with high-level wrapper (Keras, Sonnet, etc).I think PyTorch is best for research. However still new, there can be rough patches. Use TensorFlow for one graph over many machines.Consider Caffe, Caffe2, or TensorFlow for production deployment Consider Tens..

- 초보자가 접근하기에 홍콩과기대 김성훈 교수님 자료가 좋음- https://hunkim.github.io/ml/ 참고- 유투브 강좌를 처음부터 보면 개념이 잡힘. 그다음은 유명한 cs231n강좌http://cs231n.stanford.edu/