I have seen in the pre-trained models that different networks provide different kinds of outputs. Such as classification outputs, detection output, blob etc. How to handle these different kinds of output. Is there any reference/documentation? Do I have to check all the samples how they have handled different types of output.
Shovan Kumar Paul
Hi Shovan,
You can refer the samples to understand how the output is handled for different types of models.
Refer Object_detection_sample_ssd_async for the object detection models.
Classification_sample for the Classification models.
Segmentation_demo for the segmentation models.
Face_recognition_demo for the object recognition models.
Hope this helps.
Regards,
Ram prasad
For more complete information about compiler optimizations, see our Optimization Notice.