Feed training/inference data with DataFeeder¶

Fluid provides the DataFeeder class, which converts data types such as numpy array into a LoDTensor type to feed the training/inference network.

To create a DataFeeder object:

import paddle.fluid as fluid

image = fluid.layers.data(name='image', shape=[-1, 3, 224, 224], dtype='float32')
label = fluid.layers.data(name='label', shape=[-1, 1], dtype='int64')
place = fluid.CUDAPlace(0) if fluid.core.is_compiled_with_cuda() else fluid.CPUPlace()
feeder = fluid.DataFeeder(feed_list=[image, label], place=place)

The feed_list parameter is a list of variables created by fluid.layers.data() . The place parameter indicates that data such as numpy array passed in from the Python side should be converted to GPU or CPU LoDTensor. After creating the DataFeeder object, the user can call the feed(iterable) method to convert iterable data given by user into LoDTensor .

iterable should be a object of Python List or a Tuple type, and each element in iterable is a Python List of length N or Tuple type object, where N is the number of feed_list variables passed in when the DataFeeder object is created.

The concrete format of iterable is:

iterable = [
    (image_1, label_1),
    (image_2, label_2),
    ...
    (image_n, label_n)
]

image_i and label_i are both numpy array data. If the dimension of the input data is [1], such as label_i, you can feed Python int, float, and other types of data. The data types and dimensions of image_i and label_i are not necessarily the same as dtype and shape specified at fluid.layers.data(). DataFeeder internally performs the conversion of data types and dimensions. If the lod_level of the variable in feed_list is not zero, in Fluid, the 0th dimension of each row in the dimensionally converted iterable will be returned as LoD .

Read api_fluid_DataFeeder for specific usage.