From e4fc92b9d3e80a0ae41270609b831273da876e06 Mon Sep 17 00:00:00 2001 From: Devasena Kirthika S Date: Thu, 14 Nov 2024 11:34:13 +0530 Subject: [PATCH] Update to BUILD SOURCE file - Specifying numpy version required for different python builds Signed-off-by: Devasena Kirthika S Change-Id: I6a41f08ebe52a65cdd2833e61519b02651c68086 Reviewed-on: http://gerrit-git.amd.com/c/amd/ec/ZenDNN_ONNXRuntime/+/1152372 Reviewed-by: Kiriti Pendyala Reviewed-by: Rahul Jauhari Tested-by: Kiriti Pendyala (cherry picked from commit 60c220041ff62f2e08869a1a15cdf53bb04a49a0) Reviewed-on: http://gerrit-git.amd.com/c/amd/ec/ZenDNN_ONNXRuntime/+/1152223 --- BUILD_SOURCE.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/BUILD_SOURCE.md b/BUILD_SOURCE.md index bbc6110651..483562e36e 100644 --- a/BUILD_SOURCE.md +++ b/BUILD_SOURCE.md @@ -9,11 +9,14 @@ Create and activate a conda environment ``` $ conda create -n onnxrt-v1.19.2-zendnn-5.0-rel-env python=3.8 -y $ conda activate onnxrt-v1.19.2-zendnn-5.0-rel-env +> Note: Here we use python 3.8 environment as an example ``` Install necessary dependencies ``` $ pip install -U cmake onnx==1.16.1 pytest $ pip install psutil coloredlogs torch==2.1.2 torchvision transformers==4.39.3 sympy +$ pip install numpy==1.24.4 +> Note: If you are building with Python version 3.9/3.10/3.11, use pip install numpy==1.26.4 ``` ### Download the AMD ZenDNN ONNXRUNTIME source code