INNER CODE UNIT · Python
_to_device
onnxsim/onnxsim · onnxsim/rpc/server.py:145
def _to_device(self, inputs: Dict[str, np.ndarray]):
return {
k: self._Tensor(v, device=self.device).realize() for k, v in inputs.items()
}
def run(self, inputs: Dict[str, np.ndarray]) -> Dict[str, np.ndarray]:
outs = self._forward(self._to_device(inputs))
return {name: out.numpy() for name, out in zip(self.output_names, outs)}
def time(self, inputs: Dict[str, np.ndarray], number: int, repeat: int):
"""(per-call seconds, statistics), measuring steady-state ``TinyJit`` replays.
Replaying a captured graph removes the Python ONNX interpreter's per-call cost, so the
numbers reflect the generated kernels. ``stats`` also carries the eager (interpreted) call
time and, from one replay under ``DEBUG=2``, device kernel time, FLOPs and bytes moved.
"""
import contextlib
import io