Hdf5 with num_workers > 0

Hello, my hdf5 version is 1.12.0. I open the hdf5 file by using hf5 = h5py(‘path’, r), and give this class as an argument to my Dataset. so that i dont need to open hdf5 file every time in getitem(). I searched something online, So, it is possible now that the multi-processing read the same hdf5 file (no change, only read mode)? but i get a warning at the end of one epoch: Leaking Caffee2 thread-pool after fork. (function pthreadpool). my dataloader setting is num_worker =16, pin_memory = true. This warning has nothing to do with hdf5 and i can just ignore this warning, right ? what is the cause of this warning ?