Neural Magic Documentation

Neural Magic’s Deep Sparse Platform provides a suite of software components to select, build, and run performant deep learning models on CPU resources.

To learn more about specific projects founded by Neural Magic, check out documentation links below.

To file a docs issue for a specific software component, use its repo issue tracker.

Docs	GitHub	Issue Tracker	Releases	License
SparseML	Repo	Queue	Index	Apache 2.0
Sparsify	Repo	Queue	Index	Apache 2.0
SparseZoo	Repo	Queue	Index	Apache 2.0
DeepSparse Engine	Repo	Queue	Index	Neural Magic Engine

We also have a Deep Sparse Community Discourse Forum and Slack . We are growing the community member by member and happy to see you there.

To contribute to Neural Magic documentation, please:

• Open a GitHub issue and give it the documentation label.

• Submit a pull request with your proposed documentation changes.

• Get started with writing and formatting on GitHub

To learn more about Neural Magic, check our website .

Get Started

• Get Started
  • Overview
  • Sparsification
  • Software Components
  • Supported Architectures & Frameworks
    • Notes
  • Try it Out

Sparse Models

• ResNet-50
  • Sparse Models
  • Sparse Inference
  • Sparse Training
• MobileNetV1
  • Sparse Models
  • Sparse Inference
  • Sparse Training
• YOLOv5
  • Sparse Models
  • Sparse Inference
  • Sparse Training
• YOLOv3
  • Sparse Models
  • Sparse Inference
  • Sparse Training
• BERT
  • Sparse Models
  • Sparse Inference
  • Sparse Training

Details

• Technical Papers
• FAQs
  • General Product FAQs
  • Benchmarking FAQs
  • Infrastructure FAQs
  • Model Compression FAQs
  • Runtime FAQs
• Glossary

Connect Online

• Bugs, Feature Requests
• Support, General Q&A Forums
• Deep Sparse Community
• Neural Magic GitHub
• Neural Magic Docs