From 51bb3379a24139e2af234548e56ff063728f8e7e Mon Sep 17 00:00:00 2001 From: James Peet Date: Fri, 11 Feb 2022 12:21:44 +0000 Subject: MLBEDSW-6184: Release notes for version 3.3.0 - Bump minor release version and add release notes - Update README and SUPPORTED_OPS versions Change-Id: Ic14d028483c12d281e69515b25f66346d9a3afeb Signed-off-by: James Peet Signed-off-by: Tim Hall --- README.md | 2 ++ RELEASES.md | 25 +++++++++++++++++++++++++ SUPPORTED_OPS.md | 2 +- setup.py | 2 +- 4 files changed, 29 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index e223e884..b5f352eb 100644 --- a/README.md +++ b/README.md @@ -44,6 +44,8 @@ Vela runs on Linux and Microsoft Windows 10 operating systems. The following should be installed prior to the installation of Vela: * Python 3.8 or compatible + - Development version containing the Python/C API header files + - e.g. `apt install python3.8-dev` or `yum install python38-devel` * Pip3 * A C99 capable compiler and associated toolchain - For Linux operating systems, a GNU toolchain is recommended. diff --git a/RELEASES.md b/RELEASES.md index 6011d046..fec7824e 100644 --- a/RELEASES.md +++ b/RELEASES.md @@ -5,6 +5,31 @@ main feature changes, interface changes and reported defects that have been fixed. The version numbering adheres to the [semantic versioning](https://semver.org/) scheme. +## Release 3.3.0 - 25/02/2022 + +**Main feature changes:** + +* Upgrade TensorFlow Lite support to version 2.7 + * Increases Python requirement to at least version 3.8 +* Scheduling algorithm improvements + * Improved spilling reduces DRAM bandwidth for Ethos-U65 Dedicated SRAM + * Improved weight buffering increases performance for Ethos-U65 512-MAC + * Optimised tensor allocation reduces compilation time for large networks +* Extended operator support: + * RESIZE_BILINEAR: Adds support for upscaling on NPU by factors of 4x4 and 8x8 + +**Interface changes:** + +* None + +**Reported defect fixes:** + +* Memory corruption with custom operator and zero concatenation (MLCE-678) +* Crash when creating explicit padding (MLCE-684) +* Fix assert when setting address on identical LUT tensors (MLCE-691) +* Reduce SRAM usage for some elementwise operations (MLCE-750) +* Issue when running a model with Padding (MLCE-768) + ## Release 3.2.0 - 26/11/2021 **Main feature changes:** diff --git a/SUPPORTED_OPS.md b/SUPPORTED_OPS.md index f1e59aa5..de0adf92 100644 --- a/SUPPORTED_OPS.md +++ b/SUPPORTED_OPS.md @@ -1,7 +1,7 @@ # Supported Ops This file was automatically generated by Vela using the `--supported-ops-report` parameter. -Vela version: `3.2.1.dev26+gd13cd94.d20220215` +Vela version: `3.3.0` This file complies with [**Gitiles Markdown syntax**](https://github.com/google/gitiles/blob/master/Documentation/markdown.md) diff --git a/setup.py b/setup.py index 4982bb34..488d2966 100644 --- a/setup.py +++ b/setup.py @@ -40,7 +40,7 @@ class BuildExtension(build_ext): this_directory = os.path.abspath(os.path.dirname(__file__)) with open(os.path.join(this_directory, "README.md"), encoding="utf-8") as f: long_description = f.read() - tag = "3.2.0" + tag = "3.3.0" url = f"https://review.mlplatform.org/plugins/gitiles/ml/ethos-u/ethos-u-vela/+/refs/tags/{tag}/" # Find all markdown links that match the format: [text](link) for match, link in re.findall(r"(\[.+?\]\((.+?)\))", long_description): -- cgit v1.2.1