From 4f0e0cae4716fa2475cfb6c99296feb274183d2a Mon Sep 17 00:00:00 2001 From: Isak Jakobsson Date: Tue, 2 Dec 2025 09:41:57 +0100 Subject: [PATCH] bugfix: Mention A9 when talking about multiple SoCs --- axoverlay/README.md | 2 +- vdo-larod/README.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/axoverlay/README.md b/axoverlay/README.md index 673dd036..8374b373 100644 --- a/axoverlay/README.md +++ b/axoverlay/README.md @@ -37,7 +37,7 @@ axoverlay ### Limitations -- ARTPEC-8, ARTPEC-7 and ARTPEC-6 based devices. +- ARTPEC-9, ARTPEC-8, ARTPEC-7 and ARTPEC-6 based devices. - It is not possible to combine different color spaces for ARTPEC-6. ### How to run the code diff --git a/vdo-larod/README.md b/vdo-larod/README.md index 0068fac6..36992449 100644 --- a/vdo-larod/README.md +++ b/vdo-larod/README.md @@ -456,7 +456,7 @@ improve performance, but with some added complexity to the program. - This is an example of test data, which is dependent on selected device and chip. - One full-screen banana has been used for testing. -- Running inference is much faster on ARTPEC-8 and Google TPU in comparison to CPU. +- Running inference is much faster on a DLPU in comparison to CPU. - Converting images takes almost the same time on all chips. - Objects with score less than 60% are generally not good enough to be used as classification results.