TVM: Tensor IR Stack for Deep Learning Systems

Installation | Documentation | Tutorials | Operator Inventory | FAQ | Contributors | Release Notes

TVM is a Tensor intermediate representation(IR) 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 our announcement for more details.

License

© Contributors, 2017. 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.

• Contributor Guide
• Please add your name to CONTRIBUTORS.md
• Please also update NEWS.md on changes and improvements in API and codes.

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.