torch.mtia.attach_out_of_memory_observer# torch.mtia.attach_out_of_memory_observer(observer)[source]# Attach an out-of-memory observer to MTIA memory allocator
