ad-delcont: Reverse-mode automatic differentiation with delimited continuations
Reverse-mode automatic differentiation using delimited continuations (shift
/reset
), inspired by the papers
F. Wang et al, Backpropagation with Continuation Callbacks : Foundations for Efficient and Expressive Differentiable Programming, NeurIPS 2018 - https://papers.nips.cc/paper/2018/file/34e157766f31db3d2099831d348a7933-Paper.pdf
F. Wang et al, Demystifying Differentiable Programming : Shift/Reset the Penultimate Backpropagator, ICFP 2019 - https://www.cs.purdue.edu/homes/rompf/papers/wang-icfp19.pdf
The package exposes a small and easily extensible user interface to automatic differentiation combinators. It also manages to be very lightweight, as it only depends on base
and transformers
.
To use the library, import Numeric.AD.DelCont, which also contains all documentation.
[Skip to Readme]
Downloads
- ad-delcont-0.1.0.0.tar.gz [browse] (Cabal source package)
- Package description (as included in the package)
Maintainer's Corner
For package maintainers and hackage trustees
Candidates
- No Candidates
Versions [RSS] | 0.1.0.0, 0.2.0.0, 0.3.0.0, 0.5.0.0 |
---|---|
Change log | Changelog.md |
Dependencies | base (>=4.7 && <5), transformers (>=0.5) [details] |
License | BSD-3-Clause |
Copyright | (c) 2021 Marco Zocca |
Author | Marco Zocca |
Maintainer | ocramz |
Category | Math, Numeric, Machine Learning, Optimization, Optimisation |
Home page | https://github.com/ocramz/ad-delcont |
Source repo | head: git clone https://github.com/ocramz/ad-delcont |
Uploaded | by ocramz at 2021-07-18T16:21:35Z |
Distributions | LTSHaskell:0.5.0.0, NixOS:0.5.0.0, Stackage:0.5.0.0 |
Downloads | 644 total (21 in the last 30 days) |
Rating | 2.25 (votes: 2) [estimated by Bayesian average] |
Your Rating | |
Status | Docs available [build log] Last success reported on 2021-07-18 [all 1 reports] |