From b04bc774e6d79ff4e9b2805cf150025f8082b716 Mon Sep 17 00:00:00 2001 From: Pratham Deshmukh Date: Thu, 14 May 2026 11:36:01 +0530 Subject: [PATCH] feat(linux): Add AM62D benchmarks to performance table Add benchmarking results for AM62D SoC in Table 3.2, including both CPU-only and XNNPACK delegate performance measurements using the ssd_mobilenet_v2_coco.tflite model. 62D benchmarking results: https://gist.github.com/PrathamTI/2c3b2d9dad8b3083db67d04845c41adb Signed-off-by: Pratham Deshmukh --- .../linux/Foundational_Components/Machine_Learning/tflite.rst | 2 ++ 1 file changed, 2 insertions(+) diff --git a/source/linux/Foundational_Components/Machine_Learning/tflite.rst b/source/linux/Foundational_Components/Machine_Learning/tflite.rst index 9c90bcea1..8de0f5b82 100644 --- a/source/linux/Foundational_Components/Machine_Learning/tflite.rst +++ b/source/linux/Foundational_Components/Machine_Learning/tflite.rst @@ -169,6 +169,8 @@ The following performance numbers are captured with :command:`benchmark_model` o "", "XNNPACK", "0.740743", "885.636", "150.484" "AM62L", "CPU only", "1.3708", "807.076", "111.152" "", "XNNPACK", "0.930577", "769.145", "150.496" + "AM62D", "CPU only", "1.10024", "127.263", "108.918" + "", "XNNPACK", "0.264193", "532.539", "151.066" Based on the above data, using the XNNPACK delegate significantly improves inference times across all SoCs, though it generally increases initialization time and overall memory footprint.