Open deep learning compiler stack for cpu, gpu and specialized accelerators

Clone this repo:
  1. c9d87ef [Relax][Bugfix] Annotate ComputePrimValue output as host function (#17032) by Eric Lunderberg · 8 hours ago main
  2. b2c6116 [Relax][Bugfix] Bind symbolic variables in R.match_cast (#17034) by Eric Lunderberg · 9 hours ago
  3. d4b096f [Web] Fix string to uint8 array for special characters (#17031) by Charlie Ruan · 15 hours ago
  4. cab54e0 [SME][TOPI] Add conv2d NHWC SME fp32 schedule (#17003) by Andrei Hutu · 16 hours ago
  5. 430e02f [SME] Add scalable fp16->fp32 dense schedule (#16981) by Luke Hutton · 18 hours ago

Open Deep Learning Compiler Stack

Documentation | Contributors | Community | Release Notes

Build Status WinMacBuild

Apache TVM is a compiler stack for deep learning systems. It is designed to close the gap between the productivity-focused deep learning frameworks, and the performance- and efficiency-focused hardware backends. TVM works with deep learning frameworks to provide end to end compilation to different backends.


TVM is licensed under the Apache-2.0 license.

Getting Started

Check out the TVM Documentation site for installation instructions, tutorials, examples, and more. The Getting Started with TVM tutorial is a great place to start.

Contribute to TVM

TVM adopts apache committer model, we aim to create an open source project that is maintained and owned by the community. Check out the Contributor Guide.


We learned a lot from the following projects when building TVM.

  • Halide: Part of TVM's TIR and arithmetic simplification module originates from Halide. We also learned and adapted some part of lowering pipeline from Halide.
  • Loopy: use of integer set analysis and its loop transformation primitives.
  • Theano: the design inspiration of symbolic scan operator for recurrence.