torch.accelerator.memory.max_memory_allocated#

torch.accelerator.memory.max_memory_allocated(device_index=None,/)[source]#

Return the currentaccelerator maximum device memory occupied by tensorsin bytes for a given device index.

By default, this returns the peak allocated memory since the beginning ofthis program.reset_peak_memory_stats() can be used toreset the starting point in tracking this metric.

Parameters

device_index (torch.device, str, int, optional) – the index of the device to target.If not given, usetorch.accelerator.current_device_index() by default.If atorch.device or str is provided, its type must match the currentaccelerator device type.

Return type

int