mindspore.dataset.Dataset.recv

Dataset.recv(src=0, group=None)

The dataset communication interface receives data sent by the source Dataset using mindspore.dataset.Dataset.send .

Each call to the recv operation only receives data once.

Note

This is an experimental API that is subject to change or deletion.

Parameters

• src (Union[int, list[int]], optional) – List of the src rank id(s) to receive. If the Rank ID of the current process is specified, data will be obtained directly from itself. Default: 0 , which indicates receive data from src rank 0.

• group (str, optional) – The communication group to work on. The group is created by mindspore.communication.create_group() . Default: None, which indicates GlobalComm.WORLD_COMM_GROUP .

Returns

Union[Tensor, list[Tensor]], List of the Tensor(s) received.

Examples

>>> from mindspore.mint.distributed import init_process_group
>>> from mindspore.mint.distributed import get_rank
>>> import mindspore.dataset as ds
>>>
>>> # Launch 8 processes by msrun --worker_num=8 --local_worker_num=8 script.py
>>> init_process_group()
>>> this_rank = get_rank()
>>>
>>> # Create a dataset with 3 columns
>>> input_columns = ["column1", "column2", "column3"]
>>> dataset = ds.GeneratorDataset([(1, 2, 3), (3, 4, 5), (5, 6, 7)], column_names=input_columns)
>>>
>>> # Send data from rank: 0 to rank: [1, 2, 3, 4, 5, 6, 7]
>>> if this_rank == 0:
>>>     dataset.send(dst=[1, 2, 3, 4, 5, 6, 7])
>>> if this_rank in [1, 2, 3, 4, 5, 6, 7]:
>>>     data1 = dataset.recv()
>>>
>>> # Receive a data from the current dataset
>>> data2 = dataset.recv(src=get_rank())
>>>
>>> # Send data from rank: [1, 2, 3, 4, 5, 6, 7] to rank: 0
>>> if this_rank in [1, 2, 3, 4, 5, 6, 7]:
>>>     dataset.send()
>>> if this_rank == 0:
>>>     data3 = dataset.recv(src=[1, 2, 3, 4, 5, 6, 7])