torch.cuda.memory.caching_allocator_enable# torch.cuda.memory.caching_allocator_enable(value=True)[source]# Enable or disable the CUDA memory allocator. On by default.