Open deep learning compiler stack for cpu, gpu and specialized accelerators
Switch branches/tags
Clone or download
Permalink
Failed to load latest commit information.
.github [DOCS] Improve review guide, improve cmake llvm build (#1295) Jun 17, 2018
HalideIR @ cf6090a [NODE] Node base system refactor (#1739) Sep 20, 2018
apps [RUNTIME] Improve memory usage for RPC (#1741) Sep 20, 2018
cmake [RUNTIME] Refactor to enable stackvm in runtime. (#1588) Aug 11, 2018
conda [VERSION] Update to 0.5.dev (#1623) Aug 20, 2018
dlpack @ bee4d1d [SUBMODULE] update submodule to latest (#1728) Sep 19, 2018
dmlc-core @ 4f0564e update dmlc-core for security reason (#1584) Aug 10, 2018
docker [DOCKER] Golang CI recipe. (#1759) Sep 23, 2018
docs [High level OPT][RFC] NNVMv2 IR - Relay (#1672) Sep 19, 2018
include/tvm [RELAY] IR Wellform Checker (#1748) Sep 22, 2018
jvm Vulkan TVM Android Support (#1571) Aug 10, 2018
nnvm [FRONTEND][TENSORFLOW] NCHW layout support (Resnet V1/V2). (#1743) Sep 24, 2018
python Heterogeneous Runtime (#1695) Sep 22, 2018
src fix buffer elem_offset calculation (#1762) Sep 24, 2018
tests fix buffer elem_offset calculation (#1762) Sep 24, 2018
topi [NNVM] Recover reduction behavir, fix CI (#1740) Sep 19, 2018
tutorials [FRONTEND][TENSORFLOW] GPU support for tensorflow models. (#1718) Sep 20, 2018
verilog Remove leading "./" from include paths (#1640) Aug 23, 2018
vta Fix VTA Tutorial for more strict graphrt check (#1737) Sep 19, 2018
web [VERSION] Update to 0.5.dev (#1623) Aug 20, 2018
.gitignore [AUTOTVM] TOPI integration for ARM CPU (#1487) Aug 2, 2018
.gitmodules [SUBMODULE] switch to https (#341) Aug 17, 2017
.travis.yml Remove linux from travis (#156) May 23, 2017
CMakeLists.txt [High level OPT][RFC] NNVMv2 IR - Relay (#1672) Sep 19, 2018
CONTRIBUTORS.md [TEAM] siju-samuel -> Reviewer (#1745) Sep 20, 2018
Jenkinsfile [RUNTIME] Refactor to enable stackvm in runtime. (#1588) Aug 11, 2018
LICENSE [DOC/LICENSE] Make doc and license consistent, opensource repo when w… May 10, 2017
Makefile [tvm4j] fix java build (#1471) Jul 22, 2018
NEWS.md [VERSION] Update to 0.5.dev (#1623) Aug 20, 2018
README.md Update ci to new location (#1552) Aug 3, 2018

README.md

Open Deep Learning Compiler Stack

GitHub license Build Status

Documentation | Contributors | Community | Release Notes

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. Checkout the tvm stack homepage for more information.

License

© Contributors Licensed under an Apache-2.0 license.

Contribute to TVM

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

Acknowledgement

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

  • Halide: TVM uses HalideIR as data structure for arithematic simplification and low level lowering. We also learnt 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.