wait_fetch_complete

TensorTransportManager.wait_fetch_complete(fetch_request: FetchRequest, timeout: float = -1) → List[Any]

Wait for a previously initiated fetch to complete and return the tensors.

The default implementation returns the tensors stored in the FetchRequest directly, since the default fetch_multiple_tensors is synchronous.

Parameters:

• fetch_request – The FetchRequest returned by fetch_multiple_tensors.

• timeout – Maximum time in seconds to wait. -1 means wait indefinitely. 0 means return immediately if not ready.

Returns:

The received tensors.

Raises:

TimeoutError – If timeout is exceeded.