Lib.rs

› Science › Machine learning › Burn
#deep-learning #machine-learning #data

no-std burn-compute

Compute crate that helps creating high performance async backends

by Nathaniel Simard, Louis Fortier-Dubois and 90 contributors

7 unstable releases (3 breaking)

new 0.13.1 Apr 26, 2024
0.13.0 Apr 12, 2024
0.12.1 Feb 1, 2024
0.12.0 Jan 31, 2024
0.10.0 Oct 24, 2023

#877 in Machine learning

Download history 1943/week @ 2024-01-05 1406/week @ 2024-01-12 1272/week @ 2024-01-19 1344/week @ 2024-01-26 1136/week @ 2024-02-02 1114/week @ 2024-02-09 1093/week @ 2024-02-16 1087/week @ 2024-02-23 1074/week @ 2024-03-01 1225/week @ 2024-03-08 1140/week @ 2024-03-15 4425/week @ 2024-03-22 4164/week @ 2024-03-29 2260/week @ 2024-04-05 3385/week @ 2024-04-12 2353/week @ 2024-04-19

12,742 downloads per month
Used in 11 crates (2 directly)

MIT/Apache

86KB
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

~3–13MB
~118K SLoC