Hello,
I am using PSDKRA 6.0.1.0/TIDL 0.9.1.0, and attempting to visualize the outputs of semantic segmentation and object detection networks.
For J6 platform, TIDL used to generate WxHx1 (W=input_width, H=input_height) bytes of output for semantic segmentation networks, and Kx7x4 (K=number of detections) bytes of output for single-class object detection networks.
What is the format of the output generated by TIDL for J7?
And how does the format change if I use multi-class networks?
Thank you.
