CountFilterEntry¶

class paddle.distributed. CountFilterEntry ( count_filter ) [source]

Examples

>>> import paddle
>>> paddle.enable_static()

>>> sparse_feature_dim = 1024
>>> embedding_size = 64

>>> entry = paddle.distributed.CountFilterEntry(10)

>>> input = paddle.static.data(name='ins', shape=[1], dtype='int64')

>>> emb = paddle.static.nn.sparse_embedding(
...     input=input,
...     size=[sparse_feature_dim, embedding_size],
...     is_test=False,
...     entry=entry,
...     param_attr=paddle.ParamAttr(
...         name="SparseFeatFactors",
...         initializer=paddle.nn.initializer.Uniform()
...     )
... )