Hi @LinuxMan,
Thanks for sending this information.
We are not aware of this error and have not seen it yet.
The tutorial that you are following to build the BSP is will not work out of the box for Verdin iMX95. It is tested on Verdin iMX8MP, and it works there as you found.
Can you try the following steps:
$ mkdir -p ~/yocto-ml-build/bsp-toradex
$ cd ~/yocto-ml-build/bsp-toradex
$ repo init -u git://git.toradex.com/toradex-manifest.git -b refs/tags/7.4.0 -m tdxref/default.xml
$ repo sync
$ . export
$ cd ~/yocto-ml-build/bsp-toradex/build
$ git clone --depth 1 -b scarthgap-6.6.52-2.2.1 https://github.com/nxp-imx/meta-imx.git ../meta-imx
$ bitbake-layers create-layer ../layers/meta-imx-ml
$ bitbake-layers add-layer ../layers/meta-imx-ml
$ rm -rf ../layers/meta-imx-ml/recipes-example
$ cp -r ../meta-imx/meta-imx-ml/recipes-* ../layers/meta-imx-ml/
# In local.conf file:
IMAGE_INSTALL:append = " tensorflow-lite tensorflow-lite-neutron-delegate opencv python3-pillow adwaita-icon-theme "
MACHINE ?= "verdin-imx95"
ACCEPT_FSL_EULA = "1"
# In layers/meta-toradex-nxp/conf/machine/verdin-imx95.conf:
MACHINEOVERRIDES =. "tdx:mx95:mx95-nxp-bsp"
# Remove the line *PACKAGECONFIG_GPU_DELEGATE:mx95-nxp-bsp = "gpu-delegate"* from the file layers/meta-imx-ml/recipes-libraries/tensorflow-lite/tensorflow-lite_2.16.2.bb.
#PACKAGECONFIG_GPU_DELEGATE:mx95-nxp-bsp = "gpu-delegate"
# Build the image:
bitbake tdx-reference-multimedia-image