Skip to content

jelite/mlsysUtils

Repository files navigation

MyPackage

This package provides functionality for CUDA Graph capturing in PyTorch. It includes a warm-up function and CUDA Graph capture function to optimize inference execution.

Installation

To install this package, run the following command:

pip install mlsysUtil

Or install from the local build:

pip install dist/mlsysUtils-0.1.0-py3-none-any.whl

Usage

import torch
from mlsysUtils.graph_capture import get_graph

model = torch.nn.Linear(10, 10).cuda()
input_tensor = torch.randn(1, 10).cuda()

graph, static_output = get_graph(model, input_tensor)

License

This project is licensed under the MIT License.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages