tvm

Open Machine Learning Compiler Framework

What is tvm?

Open Machine Learning Compiler Framework

Overview of tvm

What it is

Apache TVM is an open machine learning compilation framework. It follows principles of Python-first development for quick customization of machine learning compiler pipelines and universal deployment to bring models into minimum deployable modules. The project started as a research project for deep learning compilation and has evolved into a cross-level design with TensorIR as the tensor-level representation and Relax as the graph-level representation.

Key features

  • Python-first development that enables quick customization of machine learning compiler pipelines.
  • Universal deployment to bring models into minimum deployable modules.
  • Cross-level design with TensorIR (tensor-level) and Relax (graph-level) representations.
  • Python-first transformations enabling most customizations in Python.
  • Joint optimization of computational graphs, tensor programs, and libraries.
  • Foundation infrastructure for building Python-first vertical compilers for domains such as LLMs.

Use cases

  • Compiling deep learning models for efficient deployment across various hardware.
  • Jointly optimizing computational graphs, tensor programs, and libraries to improve performance.
  • Building custom vertical compilers for specific domains, such as large language models (LLMs).

Getting started

See the repository README or the TVM Documentation site for installation instructions, tutorials, and examples. The Getting Started with TVM tutorial is a recommended starting point.

License and pricing

TVM is licensed under the Apache-2.0 license. It is an open-source project with no stated pricing.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/apachetvm.svg)](https://yunodb.com/tools/apachetvm?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/apachetvm?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/apachetvm.svg" alt="Featured on YunoDB" height="20"></a>