Whether this whl package can be used to run the model on AMD or Hygon GPU(ROCm).
Returns (bool): True if ROCm is currently available, otherwise False.
Examples
import paddle support_gpu = paddle.device.is_compiled_with_rocm()
