A source-native, automated machine learning workflow
Project description
DeepUtilities
What is it?
Pipeline for developing, training, and diagnosing machine learning algorithms
Why do we need it? Why is it useful?
- There are standard methods for pre-processing, training, and diagnosing algorithms; why should everyone reinvent the wheel every time?
- Let's standardize and grow the library of diagnostics for science analyses with machine learning!
General Features
- modularity - replaceable/extensible pipeline elements
- fully connected pipeline elements
- good for both TensorFlow and PyTorch
Requirements
- python 3.x
- more, but not sure yet.
Example
Pipeline Flow
- Ingest
- Pre-process
- Train
- Test
- Evalute
- Report
Original Development Team
- Praveen Balaji
- Joao Caldeira
- Callista Christ
- Aleksandra Ciprijanovic
- Anand Jain
- Francois Lanusse
- Brian Nord
- Marwah Roussi
- Amanda Whaley
How to contribute
I'm really glad you're reading this, because we need volunteer developers to help this project come to fruition.
Submitting changes
Please send a GitHub Pull Request to DeepUtilities with a clear list of what you've done (read more about pull requests). When you send a pull request, we will love you forever if you include examples. We can always use more test coverage. Please follow our coding conventions (below) and make sure all of your commits are atomic (one feature per commit).
Always write a clear log message for your commits. One-line messages are fine for small changes, but bigger changes should look like this:
$ git commit -m "A brief summary of the commit
>
> A paragraph describing what changed and its impact."
Coding conventions
Start reading our code and you'll get the hang of it. We optimize for readability:
- We indent using tabs
- We ALWAYS put spaces after list items and method parameters (
[1, 2, 3]
, not[1,2,3]
), around operators (x += 1
, notx+=1
), and around hash arrows. - This is open source software. Consider the people who will read your code, and make it look nice for them. It's sort of like driving a car: Perhaps you love doing donuts when you're alone, but with passengers the goal is to make the ride as smooth as possible.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for deeputilities-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | ae2510b499ccc574eec76cbe540e99f32e39edb7971bfb1dfb9c152df5e3b7d4 |
|
MD5 | 57805c7e4da2decf0ab425e40a688da7 |
|
BLAKE2b-256 | e19a820a07728567219e8652f31e0b55bcf3243451d8c3b4c4b85973a241e2cd |