GitHub

Backends

DifferentiationInterface.jl is based on two concepts: operators and backends. This page is about the latter, check out that page to learn about the former.

List of backends

We support the following dense backend choices from ADTypes.jl:

We also support the sparse wrapper AutoSparse.

Compatibility

DifferentiationInterface.jl itself is compatible with Julia 1.6, the Long Term Support (LTS) version of the language. However, we were only able to test the following backends on Julia 1.6:

  • AutoFiniteDifferences
  • AutoForwardDiff
  • AutoReverseDiff
  • AutoTracker
  • AutoZygote

We strongly recommend that users upgrade to Julia 1.10 or above, where all backends are tested.

Features

Given a backend object, you can use:

  • check_available to know whether the required AD package is loaded
  • check_inplace to know whether the backend supports in-place functions (all backends support out-of-place functions)
BackendAvailabilityIn-place functions
AutoChainRules✅❌
AutoDiffractor✅❌
AutoEnzyme✅✅
AutoFastDifferentiation✅✅
AutoFiniteDiff✅✅
AutoFiniteDifferences✅❌
AutoForwardDiff✅✅
AutoPolyesterForwardDiff✅✅
AutoReverseDiff✅✅
AutoSymbolics✅✅
AutoTapir✅✅
AutoTracker✅❌
AutoZygote✅❌

Backend switch

The wrapper DifferentiateWith allows you to switch between backends. It takes a function f and specifies that f should be differentiated with the backend of your choice, instead of whatever other backend the code is trying to use. In other words, when someone tries to differentiate dw = DifferentiateWith(f, backend1) with backend2, then backend1 steps in and backend2 does nothing. At the moment, DifferentiateWith only works when backend2 supports ChainRules.jl.