Hi,
HW: imx8mp-evk.
SW: LF_v5.10.72-2.2.0_images_IMX8MPEVK.
Reference document: i.MX_Machine_Learning_User's_Guide.pdf
1) I tried updating LF-v5.15.71-2.2.2_images_iMX8MPEVK, but it failed to run with the error 'nninferencemo' not found.
2) We want to add our own plugin. Could you share the source code of "nninferencemo" with me? I want to make modifications based on this plugin, which would be very helpful for us.
Thanks,
Joshua
I discussed with the AE team, please refer to the following update from them.
There are two questions.
The first question is, did the customer encounter the issues of nninferencemo' not found when back porting vx_delegate or some LIBs from LF-v5.15.71-2.2.2 to LF_v5.10.72-2.2.0? Or, the customer met issue on the LF-v5.15.71-2.2.2.
If the former, then there may be compatibility issues with the libraries required by the customer during the migration process. You need to check that all dependent libraries have been updated. If only some libraries are updated, the above issues may occur. Please provide the customer with detailed usage scenarios to facilitate further issue analysis.
For the second question, the source code of nninferencemo is linked .
Note : Please refer to the i.MX Yocto Project User’s Guide (UG10164) for detailed information how to generate Yocto SDK environment for cross-compiling
In versions after 5.15.5_1.0.0, there is not support this library. So when updating to LF-v5.15.71-2.2.2, an error of 'nninferencemo' not found will be reported.
But this does not affect of use the nnstreamer or gstreamer pipeline. For examples of using nnstreamer, please refer to the code examples in nxp gopoint. https://github.com/nxp-imx-support/nxp-demo-experience-demos-list/tree/lf-6.6.3_1.0.0/scripts/machin... There are python and C++ user case about how to using the nnstreamer and gstreamer pipeline.
And ,the source code of nninferencedmo is linked
Note : Please refer to the i.MX Yocto Project User’s Guide (UG10164) for detailed information how to generate Yocto SDK environment for cross-compiling