Currently, I'm using the IMX8p environment of linux5.4. I want to use the npu to reason the yolo model. The glibc version of the linux environment is relatively low, 2.25. Currently, there are relevant versions for the npu. For example, how does the source code of the ovxlib library support the use of npu in C++ on the current board environment?
Hi,
We recommend you not use early verison, the performance and model
compatibility of NPU driver is not good. You can use L5.15.71 or L6.6.52.
Please download document from here, open the i.MX Machine Learning User's Guide
Best Regards,
Zhiming