PyPI page
Home page
Author:
Shuo Yang
Summary:
High-performance ML primitives, applications, and informative cost API — Triton + CuteDSL kernels for NVIDIA GPUs.
Latest version:
0.4.0
Required dependencies:
numpy
|
torch
|
tqdm
|
triton
|
triton-windows
Optional dependencies:
flashlib
|
matplotlib
|
nvidia-cutlass-dsl
|
pytest
|
scikit-learn
|
tqdm
Downloads last day:
518
Downloads last week:
3,330
Downloads last month:
27,489
7-day smoothing