#deep-learning #machine-learning #data

no-std burn-compute

Compute crate that helps creating high performance async backends

5 unstable releases

0.12.1 Feb 1, 2024
0.12.0 Jan 31, 2024
0.11.1 Dec 4, 2023
0.11.0 Dec 1, 2023
0.10.0 Oct 24, 2023

#692 in Machine learning

Download history 102/week @ 2023-10-27 101/week @ 2023-11-03 106/week @ 2023-11-10 110/week @ 2023-11-17 156/week @ 2023-11-24 327/week @ 2023-12-01 299/week @ 2023-12-08 408/week @ 2023-12-15 2690/week @ 2023-12-22 2921/week @ 2023-12-29 2004/week @ 2024-01-05 1467/week @ 2024-01-12 1324/week @ 2024-01-19 1440/week @ 2024-01-26 1215/week @ 2024-02-02 1032/week @ 2024-02-09

5,186 downloads per month
Used in 9 crates (via burn-wgpu)

MIT/Apache

82KB
2K SLoC

Burn Compute

This crate helps creating high performance async backends.

  • Asynchronous kernel executions
  • Memory allocation management
  • Autotuning

lib.rs:

Burn compute crate that helps creating high performance async backends.

Dependencies

~3–12MB
~116K SLoC