fluid.nets

glu

paddle.fluid.nets.glu(input, dim=-1)[source]

The Gated Linear Units(GLU) composed by split , sigmoid and elementwise_mul . Specifically, GLU will plit the input into two equal-sized parts, \(a\) and \(b\), along the given dimension and then compute as following:

\[{GLU}(a, b)= a \otimes \sigma(b)\]

Refer to Language Modeling with Gated Convolutional Networks.

Parameters
  • input (Variable) – The input variable which is a Tensor or LoDTensor. The supported data types include float32, float64 and float16 (only for GPU).

  • dim (int, optional) – The dimension along which to split. If \(dim < 0\), the dimension to split along is \(rank(input) + dim\). Default -1.

Returns

Variable with half the size and same data type of input.

Return type

Variable

Examples

import paddle.fluid as fluid
data = fluid.data(
    name="words", shape=[-1, 6, 3, 9], dtype="float32")
# shape of output: [-1, 3, 3, 9]
output = fluid.nets.glu(input=data, dim=1)

img_conv_group

paddle.fluid.nets.img_conv_group(input, conv_num_filter, pool_size, conv_padding=1, conv_filter_size=3, conv_act=None, param_attr=None, conv_with_batchnorm=False, conv_batchnorm_drop_rate=0.0, pool_stride=1, pool_type='max', use_cudnn=True)[source]

The Image Convolution Group is composed of Convolution2d, BatchNorm, DropOut, and Pool2d. According to the input arguments, img_conv_group will do serials of computation for Input using Convolution2d, BatchNorm, DropOut, and pass the last result to Pool2d.

Parameters
  • input (Variable) – The input is 4-D Tensor with shape [N, C, H, W], the data type of input is float32 or float64.

  • conv_num_filter (list|tuple) – Indicates the numbers of filter of this group.

  • pool_size (int|list|tuple) – The pooling size of Pool2d Layer. If pool_size is a list or tuple, it must contain two integers, (pool_size_height, pool_size_width). Otherwise, the pool_size_height = pool_size_width = pool_size.

  • conv_padding (int|list|tuple) – The padding size of the Conv2d Layer. If padding is a list or tuple, its length must be equal to the length of conv_num_filter. Otherwise the conv_padding of all Conv2d Layers are the same. Default 1.

  • conv_filter_size (int|list|tuple) – The filter size. If filter_size is a list or tuple, its length must be equal to the length of conv_num_filter. Otherwise the conv_filter_size of all Conv2d Layers are the same. Default 3.

  • conv_act (str) – Activation type for Conv2d Layer that is not followed by BatchNorm. Default: None.

  • param_attr (ParamAttr) – The parameters to the Conv2d Layer. Default: None

  • conv_with_batchnorm (bool|list) – Indicates whether to use BatchNorm after Conv2d Layer. If conv_with_batchnorm is a list, its length must be equal to the length of conv_num_filter. Otherwise, conv_with_batchnorm indicates whether all the Conv2d Layer follows a BatchNorm. Default False.

  • conv_batchnorm_drop_rate (float|list) – Indicates the drop_rate of Dropout Layer after BatchNorm. If conv_batchnorm_drop_rate is a list, its length must be equal to the length of conv_num_filter. Otherwise, drop_rate of all Dropout Layers is conv_batchnorm_drop_rate. Default 0.0.

  • pool_stride (int|list|tuple) – The pooling stride of Pool2d layer. If pool_stride is a list or tuple, it must contain two integers, (pooling_stride_H, pooling_stride_W). Otherwise, the pooling_stride_H = pooling_stride_W = pool_stride. Default 1.

  • pool_type (str) – Pooling type can be \(max\) for max-pooling and \(avg\) for average-pooling. Default \(max\).

  • use_cudnn (bool) – Use cudnn kernel or not, it is valid only when the cudnn library is installed. Default: True

Returns

A Variable holding Tensor representing the final result after serial computation using Convolution2d, BatchNorm, DropOut, and Pool2d, whose data type is the same with input.

Examples

import paddle.fluid as fluid
img = fluid.data(name='img', shape=[None, 1, 28, 28], dtype='float32')
conv_pool = fluid.nets.img_conv_group(input=img,
                                      conv_padding=1,
                                      conv_num_filter=[3, 3],
                                      conv_filter_size=3,
                                      conv_act="relu",
                                      pool_size=2,
                                      pool_stride=2)

scaled_dot_product_attention

paddle.fluid.nets.scaled_dot_product_attention(queries, keys, values, num_heads=1, dropout_rate=0.0)[source]

This interface Multi-Head Attention using scaled dot product. Attention mechanism can be seen as mapping a query and a set of key-value pairs to an output. Multi-Head Attention performs attention using multi-head parallel, and the inputs of attention would be transformed by linear projection. The formula is as follows:

\[ \begin{align}\begin{aligned}MultiHead(Q, K, V ) & = Concat(head_1, ..., head_h)\\where \ head_i & = Attention(QW_i^Q , KW_i^K , VW_i^V )\\Attention(Q, K, V) & = softmax (\frac{QK^\mathrm{T}}{\sqrt{d_k}}) V\end{aligned}\end{align} \]

For more details, please refer to Attention Is All You Need .

Note that the implementation is adapted to batch, and all matrix multiplication in \(Attention(Q, K, V)\) is batched matrix multiplication. Refer to matmul .

Parameters
  • queries (Variable) – A 3-D Tensor with shape \([N, L_q, d_k \times h]\) , where \(N\) stands for batch size, \(L_q\) for the sequence length of query, \(d_k \times h\) for the feature size of query, \(h\) for head number. The data type should be float32 or float64.

  • keys (Variable) – A 3-D Tensor with shape \([N, L_k, d_k \times h]\) , where \(N\) stands for batch size, \(L_k\) for the sequence length of key, \(d_k \times h\) for the feature size of key, \(h\) for head number. The data type should be the same as queries .

  • values (Variable) – A 3-D Tensor with shape \([N, L_k, d_v \times h]\) , where \(N\) stands for batch size, \(L_k\) for the sequence length of key, \(d_v \times h\) for the feature size of value, \(h\) for head number. The data type should be the same as queries .

  • num_heads (int, optional) – Indicate the number of head. If the numher is 1, linear projection would not be performed on inputs. Default: 1.

  • dropout_rate (float, optional) – The rate to drop the attention weight. Default: 0.0, which means no dropout.

Returns

A 3-D Tensor with shape \([N, L_q, d_v \times h]\) , where \(N\) stands for batch size, \(L_q\) for the sequence length of query, \(d_v \times h\) for the feature size of value. It has the same data type with inputs, representing the output of Multi-Head Attention.

Return type

Variable

Raises
  • ValueError – Inputs quries, keys and values should all be 3-D tensors.

  • ValueError – The hidden size of queries and keys should be the same.

  • ValueError – The max sequence length in query batch and in key batch should be the same.

  • ValueError – he hidden size of keys must be divisible by the number of attention heads.

  • ValueError – he hidden size of values must be divisible by the number of attention heads.

Examples

import paddle.fluid as fluid

queries = fluid.data(name="queries", shape=[3, 5, 9], dtype="float32")
keys = fluid.data(name="keys", shape=[3, 6, 9], dtype="float32")
values = fluid.data(name="values", shape=[3, 6, 10], dtype="float32")
contexts = fluid.nets.scaled_dot_product_attention(queries, keys, values)
contexts.shape  # [3, 5, 10]

sequence_conv_pool

paddle.fluid.nets.sequence_conv_pool(input, num_filters, filter_size, param_attr=None, act='sigmoid', pool_type='max', bias_attr=None)[source]

This api takes input as an LoDTensor. If input is a Tensor, please use simple_img_conv_pool instead

The sequence_conv_pool is composed of sequence_conv and sequence_pool .

Parameters
  • input (Variable) – 2-D LoDTensor, the input of sequence_conv, which supports variable-time length input sequence. The underlying of input is a matrix with shape (T, N), where T is the total time steps in this mini-batch and N is the input_hidden_size. The data type is float32 or float64.

  • num_filters (int) – The number of filter.

  • filter_size (int) – The filter size.

  • param_attr (ParamAttr) – The parameters of the sequence_conv Layer. Default: None.

  • act (str|None) – Activation type for Sequence_conv Layer. If set to None, no activation will be applied. Default: “sigmoid”.

  • pool_type (str) – Pooling type can be \(max\) for max-pooling, \(average\) for average-pooling, \(sum\) for sum-pooling, \(sqrt\) for sqrt-pooling. Default \(max\).

  • bias_attr (ParamAttr|bool|None) – The parameter attribute for the bias of sequence_conv. If it is set to False, no bias will be added to the output units. If it is set to None or one attribute of ParamAttr, sequence_conv will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not set, the bias is initialized zero. Default: None.

Returns

The final result after sequence_conv and sequence_pool. It is a 2-D Tensor, with the same data type as input

Return Type:

Variable

Examples

import paddle.fluid as fluid
input_dim = 100 #len(word_dict)
emb_dim = 128
hid_dim = 512
data = fluid.data(name="words", shape=[None, 1], dtype="int64", lod_level=1)
emb = fluid.layers.embedding(input=data, size=[input_dim, emb_dim], is_sparse=True)
seq_conv = fluid.nets.sequence_conv_pool(input=emb,
                                         num_filters=hid_dim,
                                         filter_size=3,
                                         act="tanh",
                                         pool_type="sqrt")

simple_img_conv_pool

paddle.fluid.nets.simple_img_conv_pool(input, num_filters, filter_size, pool_size, pool_stride, pool_padding=0, pool_type='max', global_pooling=False, conv_stride=1, conv_padding=0, conv_dilation=1, conv_groups=1, param_attr=None, bias_attr=None, act=None, use_cudnn=True)[source]

The simple_img_conv_pool api is composed of conv2d and pool2d .

Parameters
  • input (Variable) – 4-D Tensor, shape is [N, C, H, W], data type can be float32 or float64.

  • num_filters (int) – The number of filters. It is the same as the output channels.

  • filter_size (int|list|tuple) – The filter size. If filter_size is a list or tuple, it must contain two integers, (filter_size_H, filter_size_W). Otherwise, the filter_size_H = filter_size_W = filter_size.

  • pool_size (int|list|tuple) – The pooling size of pool2d layer. If pool_size is a list or tuple, it must contain two integers, (pool_size_H, pool_size_W). Otherwise, the pool_size_H = pool_size_W = pool_size.

  • pool_stride (int|list|tuple) – The pooling stride of pool2d layer. If pool_stride is a list or tuple, it must contain two integers, (pooling_stride_H, pooling_stride_W). Otherwise, the pooling_stride_H = pooling_stride_W = pool_stride.

  • pool_padding (int|list|tuple) – The padding of pool2d layer. If pool_padding is a list or tuple, it must contain two integers, (pool_padding_H, pool_padding_W). Otherwise, the pool_padding_H = pool_padding_W = pool_padding. Default 0.

  • pool_type (str) – Pooling type can be \(max\) for max-pooling or \(avg\) for average-pooling. Default \(max\).

  • global_pooling (bool) – Whether to use the global pooling. If global_pooling = true, pool_size and pool_padding while be ignored. Default False

  • conv_stride (int|list|tuple) – The stride size of the conv2d Layer. If stride is a list or tuple, it must contain two integers, (conv_stride_H, conv_stride_W). Otherwise, the conv_stride_H = conv_stride_W = conv_stride. Default: conv_stride = 1.

  • conv_padding (int|list|tuple) – The padding size of the conv2d Layer. If padding is a list or tuple, it must contain two integers, (conv_padding_H, conv_padding_W). Otherwise, the conv_padding_H = conv_padding_W = conv_padding. Default: conv_padding = 0.

  • conv_dilation (int|list|tuple) – The dilation size of the conv2d Layer. If dilation is a list or tuple, it must contain two integers, (conv_dilation_H, conv_dilation_W). Otherwise, the conv_dilation_H = conv_dilation_W = conv_dilation. Default: conv_dilation = 1.

  • conv_groups (int) – The groups number of the conv2d Layer. According to grouped convolution in Alex Krizhevsky’s Deep CNN paper: when group=2, the first half of the filters is only connected to the first half of the input channels, while the second half of the filters is only connected to the second half of the input channels. Default: groups=1.

  • param_attr (ParamAttr|None) – The parameter attribute for learnable parameters/weights of conv2d. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as param_attr. If the Initializer of the param_attr is not set, the parameter is initialized with \(Normal(0.0, std)\), and the \(std\) is \((\frac{2.0 }{filter\_elem\_num})^{0.5}\). Default: None.

  • bias_attr (ParamAttr|bool|None) – The parameter attribute for the bias of conv2d. If it is set to False, no bias will be added to the output units. If it is set to None or one attribute of ParamAttr, conv2d will create ParamAttr as bias_attr. If the Initializer of the bias_attr is not set, the bias is initialized zero. Default: None.

  • act (str) – Activation type for conv2d, if it is set to None, activation is not appended. Default: None.

  • use_cudnn (bool) – Use cudnn kernel or not, it is valid only when the cudnn library is installed. Default: True

Returns

4-D Tensor, the result of input after conv2d and pool2d, with the same data type as input

Return Type:

Variable

Examples

import paddle.fluid as fluid
img = fluid.data(name='img', shape=[100, 1, 28, 28], dtype='float32')
conv_pool = fluid.nets.simple_img_conv_pool(input=img,
                                            filter_size=5,
                                            num_filters=20,
                                            pool_size=2,
                                            pool_stride=2,
                                            act="relu")