| commit | d16b02f6a130fe486b51932b0dce75937dfb0076 | [log] [tgz] |
|---|---|---|
| author | harsh-nod <harsh@nod-labs.com> | Wed Dec 14 16:26:51 2022 -0800 |
| committer | GitHub <noreply@github.com> | Thu Dec 15 00:26:51 2022 +0000 |
| tree | 6f6b940fd29870e7a2b42af507ce12dc858aeb1e | |
| parent | 0ca91c1da5d05231697e97c5fc3dde10f1998b0c [diff] |
Add Winograd support for NCHW convolutions (#11475) This patch adds support for NCHW convolutions. In order to avoid the cost of additional transposes, the winograd input op is modified so that regardless of whether the input is (N, H, W, C) or (N, C, H, W), the output is always (T, T, N, H, W, C). Similarly, the winograd output op is modified to have the input always be (T, T, N, H, W, C), but have the output be either (N, H, W, C) or (N, C, H, W) depending on the specified image dimensions. The tiling and decomposition passes are also modified to account for this.
IREE (Intermediate Representation Execution Environment, pronounced as “eerie”) is an MLIR-based end-to-end compiler and runtime that lowers Machine Learning (ML) models to a unified IR that scales up to meet the needs of the datacenter and down to satisfy the constraints and special considerations of mobile and edge deployments.
See our website for project details, user guides, and instructions on building from source.
IREE is still in its early phase. We have settled down on the overarching infrastructure and are actively improving various software components as well as project logistics. It is still quite far from ready for everyday use and is made available without any support at the moment. With that said, we welcome any kind of feedback on any communication channels!
See our website for more information.
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.