From 99fcb89f6a6c09b5677a5323abdee5714ac5117a Mon Sep 17 00:00:00 2001 From: Michael McGeagh Date: Thu, 27 Aug 2020 14:48:44 +0100 Subject: MLBEDSW-2839 Update release notes for 1.2.0 Signed-off-by: Michael McGeagh Change-Id: I3aa686bd0ece68af5840bcabbdb5702ba047d3de --- RELEASES.md | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'RELEASES.md') diff --git a/RELEASES.md b/RELEASES.md index abab754a..8498fd6a 100644 --- a/RELEASES.md +++ b/RELEASES.md @@ -5,6 +5,29 @@ 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 1.2.0 - 31/08/2020 + +**Main feature changes:** + +* New Ethos-U operator support: SOFTMAX and QUANTIZE (requantise only) +* Improved Ethos-U operator support: RESIZE_BILINEAR and LEAKY_RELU +* Improved pass-through of operators not supported by Ethos-U +* Preserve TensorFlow Lite (.tflite) file metadata +* Improved network performance +* Performance estimation statistics improved +* New external API + +**Interface changes:** + +* Addition of CLI options: `--weight-estimation-scaling` +* Addition of External APIs: Weight compression and Bias & scale encoding + +**Reported defect fixes:** + +* Custom operator not passing through Vela (MLCE-223) +* Flash usage increases in 1.1.0 (MLCE-227) +* Vela fails with optional bias tensor (MLCE-231) + ## Release 1.1.0 - 30/06/2020 **Main feature changes:** -- cgit v1.2.1