1 repository
Deferred calculation of derivatives until a backward pass is explicitly triggered.
Distinct from Lazy Evaluation Patterns: Specializes lazy evaluation patterns specifically for the timing of gradient calculations in autodiff
Explore 1 awesome GitHub repository matching software engineering & architecture · Gradient Evaluation. Refine with filters or upvote what's useful.
Autograd is an automatic differentiation library and numerical gradient engine for Python. Its primary purpose is to compute the gradients of mathematical functions to enable numerical optimization and the training of mathematical models. The library automates the calculation of derivatives to simplify the implementation of optimization algorithms. This supports activities such as machine learning research, gradient-based learning, and the optimization of numerical models.
Defers the numerical calculation of derivatives until an explicit backward call is triggered on a scalar.