ise-uiuc/NablaFuzz

Fuzzing Automatic Differentiation in Deep-Learning Libraries (ICSE'23)

27
/ 100
Experimental

This tool helps deep learning library developers find bugs in the automatic differentiation (AD) features of their frameworks. It takes your existing deep learning library (like PyTorch or TensorFlow) as input and automatically generates tests to expose inconsistencies in how gradients are computed. The output is a list of potential bugs, pinpointing where the library might be calculating gradients incorrectly. Developers of deep learning frameworks are the primary users.

No commits in the last 6 months.

Use this if you are a developer or researcher working on deep learning libraries and need to rigorously test the correctness of their automatic differentiation implementations.

Not ideal if you are an end-user building or training deep learning models and not directly involved in the development or maintenance of the underlying deep learning framework.

deep-learning-library-development automatic-differentiation gradient-testing software-quality-assurance framework-validation
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 8 / 25
Community 12 / 25

How are scores calculated?

Stars

27

Forks

4

Language

Python

License

Last pushed

Mar 02, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/ise-uiuc/NablaFuzz"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.