LoreonLabsPlatform

Overview

Intelligence

Markets
Builders
Research
Ecosystems
Launchpads

Search

Rust

burn-mlx

MLX backend for Burn - native Apple Silicon GPU acceleration

RustEmerging

Stars

—

Forks

—

Contributors

3

Last push

2mo ago

Recent commits

Latest commits.

Update README.md
d76a9f3Marcin Nowak-Liebiediew2mo ago
fix remaining upstream rename cleanup
23b9b9aMarcin Nowak Liebiediew2mo ago
rename crate surface to burn-mlx
7ea9c51Marcin Nowak Liebiediew2mo ago
burn-0-21 compat + upstream backend tests
0d5c1f8Marcin Nowak Liebiediew2mo ago
Cast dequantize and quantized_matmul results to backend float type
beb4a1bMike Marcacci4mo ago

Optimize q_reshape/q_swap_dims/q_expand to avoid unnecessary dequant/requant cycles

c3e19bfMike Marcacci4mo ago

Implement native MLX quantization support via mlx_rs ops

4c541fcMike Marcacci4mo ago

Make Mlx backend generic over float precision (f32, f16, bf16, f64)

a3c1ccaMike Marcacci4mo ago

Top contributors

Builders behind this project.

ravituringworks