How can I use SVO files for neural network batch loading? (Memory Leak)

I’m saving data in the form of SVO files, and would like to load multiple SVO files to get batches of this data to train a neural network.

However, when I do this I get a memory leak that eventually (rather quickly) causes my program to crash.

Has anyone encountered this issue before / does anyone have a solution?

The memory leak seems to be discussed here and have been around for a while: Memory leak in SVO reader · Issue #111 · stereolabs/zed-python-api · GitHub

I am using the most recent SDK which claims to have solved the SVO memory leak.

Hi, this post a a duplicate from here : Memory Leak In SVO Reading · Issue #534 · stereolabs/zed-examples · GitHub

Since the thread is older on GitHub and has people working on it, I’ll let you continue your conversion there.

Antoine