Skip to content

docs: update readme about differentiation #26

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ZedongPeng merged 1 commit into from
Jun 2, 2025
Merged

docs: update readme about differentiation #26

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,11 +12,11 @@
MPAX is a hardware-accelerated, differentiable, batchable, and distributable solver for mathematical programming in JAX, designed to integrate with modern computational and deep learning workflows:

- **Hardware accelerated**: executes on multiple architectures including CPUs, GPUs and TPUs.
- **Differentiable**: computes derivatives of solutions with respect to inputs through implicit or unrolled differentiation.
- **Differentiable**: computes derivatives of solutions with respect to inputs through unrolled differentiation.
- **Batchable**: solves multiple problem instances of the same shape simultaneously.
- **Distributed**: executes in parallel across multiple devices, such as several GPUs.

MPAX's primary motivation is to integrate mathematical programming with deep learning pipelines. To achieve this, MPAX aligns its algorithms and implementations with the requirements of deep learning hardware, ensuring compatibility with GPUs and TPUs. By being differentiable, MPAX can integrate directly into the backpropagation process of neural network training. Its batchability and distributability further enable scalable deployment in large-scale applications.
MPAX's primary motivation is to integrate mathematical programming with deep learning pipelines. To achieve this, MPAX aligns its algorithms and implementations with the requirements of deep learning hardware, ensuring compatibility with GPUs and TPUs. By being differentiable, MPAX can integrate directly into the backpropagation process of neural network training. Its batchability and distributability further enable scalable deployment in large-scale applications.

Currently, MPAX supports **linear programming (LP)** and **quadratic programming (QP)**, the foundational problems in mathematical programming. Future releases will expand support to include other problem classes of mathematical programming.

Expand Down