backprop
Heterogeneous automatic differentiation ("backpropagation") in Haskell (by mstksg)
accelerate-ad
Combinatory automatic differentiation in Haskell for heterogeneous computing. (by ajknapp)
Our great sponsors
backprop | accelerate-ad | |
---|---|---|
2 | 1 | |
177 | 21 | |
- | - | |
3.8 | 10.0 | |
6 months ago | over 5 years ago | |
Haskell | Haskell | |
BSD 3-clause "New" or "Revised" License | BSD 3-clause "New" or "Revised" License |
The number of mentions indicates the total number of mentions that we've tracked plus the number of user suggested alternatives.
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
backprop
Posts with mentions or reviews of backprop.
We have used some of these posts to build our list of alternatives
and similar projects. The last one was on 2023-01-17.
accelerate-ad
Posts with mentions or reviews of accelerate-ad.
We have used some of these posts to build our list of alternatives
and similar projects. The last one was on 2023-01-17.
-
Backpropagation and Accelerate
Now back to your question. I have a bit of experience with backprop and accelerate but it's neither recent nor with both of them at once. Accelerate has two layers of abstraction. There are Exp and Acc that build an AST. After compiling them with llvm-native or llvm-ptx backend you enter another layer of abstraction – functions Array -> Array -> ... -> Array. How much automatic you want AD to be? Automatic differentiation AST of Exps and Accs is going to be hard and backprop has nothing to help you here. There was a google summer of code project on this topic. As I understand, it ran short of completion.
What are some alternatives?
When comparing backprop and accelerate-ad you can also consider the following projects:
forecast-io - A Haskell library for working with forecast.io data.
swagger2 - Swagger 2.0 data model.
inquire
json - Haskell JSON library
listenbrainz-client - A client to the ListenBrainz project
clock-extras
android-lint-summary - Prettier display of Android Lint issues
hriemann - A Riemann Client for Haskell
ghcjs-codemirror - In-browser code editor
language-elm - A haskell library for generating elm source code from an ast.
urlencoded - Haskell urlencoded package
url - A URL package for Haskell