Running and Calibrating TensorRT engine using jpeg images

According to sample shown here (from 373-406) is preprocessing input image for inference. But that function works only for images with ppm extension.

If I like to calibrate or do inference with jpeg images, how can I use the function, it has segmentation fault at line 387
resizePPM(ppms[i], inputW, inputH)
How can I solve the issue?

Say if I set batch size is 16, how i can input multiple images to program?

Please refer to below link:


The key answer link is broken. The key answer is here, but broken.

Please refer below link


1 Like