roastduck/FreeTensor

A language and compiler for irregular tensor programs.

38
/ 100
Emerging

This tool helps developers write high-performance numerical programs in Python by compiling their loop-based tensor code into optimized native code. It takes Python code defining tensor operations and outputs highly efficient executables, making it useful for machine learning engineers, scientific programmers, or anyone needing to accelerate computational tasks.

152 stars. No commits in the last 6 months.

Use this if you need to significantly speed up your Python-based tensor computations, especially those with dynamic shapes or irregular access patterns, by compiling them to native code and applying explicit optimizations like parallelization.

Not ideal if you primarily work with existing high-level deep learning frameworks like TensorFlow or PyTorch and don't require low-level control over loop transformations and memory management.

numerical-computing high-performance-computing scientific-programming machine-learning-engineering compiler-development
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

152

Forks

12

Language

C++

License

Apache-2.0

Last pushed

Nov 29, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ai-coding/roastduck/FreeTensor"

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