About this document

Elements of Differentiable Programming by nigel989 is a document available to read on EtoBox.

Differentiable programming allows numeric computer programs to be differentiated via automatic differentiation, enabling gradient-based optimization of parameters via techniques like gradient descent. There are two main approaches: static graph-based frameworks like TensorFlow that allow for optimization but limit program types, and dynamic graph frameworks like PyTorch that are more interactive but have overhead. Differentiable programming has applications in scientific computing, deep learning, robotics,

Author

nigel989

Language

EN