commit | a6bf65ca220c448b2f73454a66da179ee6c2c414 | [log] [tgz] |
---|---|---|
author | Stella Laurenzo <stellaraccident@gmail.com> | Tue Jan 11 20:47:51 2022 -0800 |
committer | GitHub <noreply@github.com> | Tue Jan 11 20:47:51 2022 -0800 |
tree | 1273ad0fd1ec3865861d40e9b11cda1b985ac0a2 | |
parent | 5dc5d75c47c395710317515da94fa6a32f6c9779 [diff] |
Uncouples integrations/tensorflow from the CMake build. (#8078) * Removes CMake files and options. * Makes real Python packages for the projects. * Moves a high value subset of the existing TF test suite to be lit-based with no build system integration: * Uncategorized tests (includes mobilebert) * Math tests * Layer tests * Excludes big, whole model TF tests. We will replace them with corresponding tflite tests in an (immediate) followup (just would like to not add new features to such a large patch). * Updates CI to build/test the new configuration. * Will require a bit of work on the release pipeline after this. Will do in a follow-up. * Add test timeout * Remove bindings/python dir * Fix benchmark build so it can find iree-import-tflite. * Delete e2e directory * Deflake mobile bert squad test by not writing to global tmp dir * Enables a bunch of tests that were asserting in debug builds (but passing in release builds) once restricting a buggy pattern to static shapes (which is all it was coded to handle). See: https://github.com/google/iree/issues/8083
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.