Revised deconv implementations according to luotao1

paddle/gserver/layers/ConvBaseLayer.cpp

@@ -22,9 +22,9 @@ bool ConvBaseLayer::init(const LayerMap& layerMap,
   Layer::init(layerMap, parameterMap);
 
   if (config_.type() == "exconv" || config_.type() == "cudnn_conv") {
-    isConv_ = true;
+    isDeconv_ = false;
   } else {
-    isConv_ = false;
+    isDeconv_ = true;
   }
 
   /* Initialize the convolutional layer parameter */

paddle/gserver/layers/ConvBaseLayer.h

@@ -28,8 +28,8 @@ class ConvBaseLayer : public Layer {
 protected:
   typedef std::vector<int> IntV;
 
-  /// True if it's convolution layer, false if it's deconv layer
-  bool isConv_;
+  /// True if it's deconv layer, false if it's convolution layer
+  bool isDeconv_;
 
   /// The number of filters.
   int numFilters_;

paddle/gserver/layers/ExpandConvBaseLayer.cpp

@@ -13,11 +13,12 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 
+#include "ExpandConvBaseLayer.h"
+
 #include "paddle/utils/Logging.h"
-#include "ConvBaseLayerCpu.h"
 namespace paddle {
 
-bool ConvBaseLayerCpu::init(const LayerMap &layerMap,
+bool ExpandConvBaseLayer::init(const LayerMap &layerMap,
                            const ParameterMap &parameterMap) {
   /* Initialize the basic convolutional parent class */
   ConvBaseLayer::init(layerMap, parameterMap);
@@ -34,10 +35,10 @@ bool ConvBaseLayerCpu::init(const LayerMap &layerMap,
   /* Initialize the projection */
   for (auto &inputConfig : config_.inputs()) {
     const ConvConfig &conf = inputConfig.conv_conf();
-    nf = isConv_ ? numFilters_ : conf.channels();
+    nf = (!isDeconv_) ? numFilters_ : conf.channels();
     subM_.push_back(nf / conf.groups());
     subN_.push_back(conf.output_x() * conf.output_x());
-    channel = isConv_ ? conf.channels() : numFilters_;
+    channel = (!isDeconv_) ? conf.channels() : numFilters_;
     subK_.push_back(channel * conf.filter_size() * conf.filter_size() /
                     conf.groups());
     /* Consistent caffe mode for multiple input */
@@ -47,11 +48,11 @@ bool ConvBaseLayerCpu::init(const LayerMap &layerMap,
   return true;
 }
 
-void ConvBaseLayerCpu::resetExpandInput(size_t height, size_t width) {
+void ExpandConvBaseLayer::resetExpandInput(size_t height, size_t width) {
   Matrix::resizeOrCreate(expandInput_, height, width, false, useGpu_);
 }
 
-void ConvBaseLayerCpu::addSharedBias() {
+void ExpandConvBaseLayer::addSharedBias() {
   size_t mapW = getSize() / numFilters_;
   size_t mapH = getOutputValue()->getElementCnt() / mapW;
   MatrixPtr out =
@@ -75,7 +76,7 @@ void ConvBaseLayerCpu::addSharedBias() {
   bias->clear();
 }
 
-void ConvBaseLayerCpu::addUnsharedBias() {
+void ExpandConvBaseLayer::addUnsharedBias() {
   MatrixPtr outValue = getOutputValue();
   MatrixPtr bias =
       Matrix::create(biases_->getW()->getData(), 1,
@@ -84,9 +85,9 @@ void ConvBaseLayerCpu::addUnsharedBias() {
 }
 
 
-void ConvBaseLayerCpu::expandOneFrame(MatrixPtr image, size_t startIdx,
+void ExpandConvBaseLayer::expandOneFrame(MatrixPtr image, size_t startIdx,
                                      int inIdx) {
-  int channel = isConv_ ? channels_[inIdx] : numFilters_;
+  int channel = (!isDeconv_) ? channels_[inIdx] : numFilters_;
 
   resetExpandInput(subK_[inIdx] * groups_[inIdx], subN_[inIdx]);
   real *imgData = image->getData() + startIdx * image->getWidth();
@@ -101,7 +102,7 @@ void ConvBaseLayerCpu::expandOneFrame(MatrixPtr image, size_t startIdx,
   imageTmp->clear();
 }
 
-void ConvBaseLayerCpu::expandFwdOnce(MatrixPtr image, MatrixPtr out,
+void ExpandConvBaseLayer::expandFwdOnce(MatrixPtr image, MatrixPtr out,
                                      int inIdx, int startIdx) {
   int subM = subM_[inIdx];
   int subN = subN_[inIdx];
@@ -109,7 +110,7 @@ void ConvBaseLayerCpu::expandFwdOnce(MatrixPtr image, MatrixPtr out,
 
   expandOneFrame(image, startIdx, inIdx);
 
-  int nf = isConv_ ? numFilters_ : channels_[inIdx];
+  int nf = (!isDeconv_) ? numFilters_ : channels_[inIdx];
 
   real *outData =
       out->getData() + startIdx * subN * nf;
@@ -132,8 +133,9 @@ void ConvBaseLayerCpu::expandFwdOnce(MatrixPtr image, MatrixPtr out,
   }
 }
 
-void ConvBaseLayerCpu::bpropActs(MatrixPtr out, MatrixPtr image, int inpIdx) {
-  int channel = isConv_ ? channels_[inpIdx] : numFilters_;
+void ExpandConvBaseLayer::bpropActs(MatrixPtr out, MatrixPtr image,
+                                    int inpIdx) {
+  int channel = (!isDeconv_) ? channels_[inpIdx] : numFilters_;
 
   int subM = subM_[inpIdx];
   int subN = subN_[inpIdx];
@@ -186,7 +188,7 @@ void ConvBaseLayerCpu::bpropActs(MatrixPtr out, MatrixPtr image, int inpIdx) {
   }
 }
 
-void ConvBaseLayerCpu::bpropWeights(MatrixPtr image, MatrixPtr out,
+void ExpandConvBaseLayer::bpropWeights(MatrixPtr image, MatrixPtr out,
                                     int inpIdx) {
   MatrixPtr weightGrad = weights_[inpIdx]->getWGrad();
 
@@ -221,7 +223,7 @@ void ConvBaseLayerCpu::bpropWeights(MatrixPtr image, MatrixPtr out,
   }
 }
 
-void ConvBaseLayerCpu::bpropSharedBias(MatrixPtr biases, MatrixPtr v) {
+void ExpandConvBaseLayer::bpropSharedBias(MatrixPtr biases, MatrixPtr v) {
   size_t mapW = getSize() / numFilters_;
   size_t mapH = v->getElementCnt() / mapW;
   MatrixPtr vTmp = Matrix::create(v->getData(), mapH, mapW, false, useGpu_);
@@ -234,7 +236,7 @@ void ConvBaseLayerCpu::bpropSharedBias(MatrixPtr biases, MatrixPtr v) {
   biases->collectBias(*transOutValue_, 1.0f);
 }
 
-void ConvBaseLayerCpu::bpropBiases(MatrixPtr v) {
+void ExpandConvBaseLayer::bpropBiases(MatrixPtr v) {
   MatrixPtr biases =
       Matrix::create(biases_->getWGrad()->getData(), 1,
                      biases_->getWGrad()->getElementCnt(), false, useGpu_);

paddle/gserver/layers/ExpandConvBaseLayer.h

@@ -25,7 +25,7 @@ namespace paddle {
  * @brief A subclass of ConvBaseLayer that is a superclass of both
  * ExpandConvLayer and ExpandConvTransLayer
  */
-class ConvBaseLayerCpu : public ConvBaseLayer {
+class ExpandConvBaseLayer : public ConvBaseLayer {
 protected:
   /// For expand convolution.
   /// subM_ = numFilters_ / groups_.
@@ -43,18 +43,19 @@ protected:
   /// The spatial dimensions of width of output feature map.
   IntV outputW_;
 
-  /*The expandInput_ and transOutValue_ are used for CPU expand conv calc*/
-  /// Expand one sample at a time. shape:
-  /// (numChannels * filterPixels_, outputSizeH * outputSizeW)
+  /*The expandInput_ and transOutValue_ are used for CPU expand conv calc
+   * Expand one sample at a time. shape:
+   * (numChannels * filterPixels_, outputSizeH * outputSizeW)
+   * */
   MatrixPtr expandInput_;
   /// The transpose of output, which is an auxiliary matrix.
   MatrixPtr transOutValue_;
 
 public:
-  explicit ConvBaseLayerCpu(const LayerConfig& config)
+  explicit ExpandConvBaseLayer(const LayerConfig& config)
     : ConvBaseLayer(config) {}
 
-  ~ConvBaseLayerCpu() {}
+  ~ExpandConvBaseLayer() {}
 
   bool init(const LayerMap& layerMap, const ParameterMap& parameterMap);
 

paddle/gserver/layers/ExpandConvLayer.cpp

@@ -24,7 +24,7 @@ REGISTER_LAYER(exconv, ExpandConvLayer);
 bool ExpandConvLayer::init(const LayerMap &layerMap,
                            const ParameterMap &parameterMap) {
   /* Initialize the basic convolutional parent class */
-  ConvBaseLayerCpu::init(layerMap, parameterMap);
+  ExpandConvBaseLayer::init(layerMap, parameterMap);
   return true;
 }
 
@@ -49,16 +49,17 @@ void ExpandConvLayer::forward(PassType passType) {
   resetOutput(batchSize, getOutputSize());
 
   MatrixPtr image = nullptr;
-  for (size_t i = 0; i != inputLayers_.size(); ++i) {
+  MatrixPtr outV = getOutputValue();
+  for (size_t i = 0; i < inputLayers_.size(); ++i) {
     LayerPtr prevLayer = getPrev(i);
     image = prevLayer->getOutputValue();
     for (size_t off = 0; off < image->getHeight(); off++) {
       REGISTER_TIMER_INFO("expandFwdOnce", getName().c_str());
-      expandFwdOnce(image, getOutputValue(), i, off);
+      expandFwdOnce(image, outV, i, off);
     }
   }
   /* add the bias-vector */
-  if (biases_.get() != NULL) {
+  if (biases_.get()) {
     if (sharedBiases_) {
       addSharedBias();
     } else {
@@ -81,9 +82,9 @@ void ExpandConvLayer::backward(const UpdateCallback &callback) {
     biases_->getParameterPtr()->incUpdate(callback);
   }
 
-  for (size_t i = 0; i != inputLayers_.size(); ++i) {
+  for (size_t i = 0; i < inputLayers_.size(); ++i) {
     /* First, calculate the input layers error */
-    if (NULL != getPrev(i)->getOutputGrad()) {
+    if (getPrev(i)->getOutputGrad()) {
       bpropActs(outGrad, getPrev(i)->getOutputGrad(), i);
     }
     if (weights_[i]->getWGrad()) {

paddle/gserver/layers/ExpandConvLayer.h

@@ -15,9 +15,9 @@ limitations under the License. */
 
 #pragma once
 
-#include "ConvBaseLayerCpu.h"
 #include "paddle/math/Matrix.h"
 #include <vector>
+#include "ExpandConvBaseLayer.h"
 
 namespace paddle {
 
@@ -29,10 +29,10 @@ namespace paddle {
  * The config file api is img_conv_layer.
  */
 
-class ExpandConvLayer : public ConvBaseLayerCpu {
+class ExpandConvLayer : public ExpandConvBaseLayer {
 public:
   explicit ExpandConvLayer(const LayerConfig& config) :
-    ConvBaseLayerCpu(config) {}
+    ExpandConvBaseLayer(config) {}
 
   ~ExpandConvLayer() {}
 

paddle/gserver/layers/ExpandConvTransLayer.cpp

@@ -29,7 +29,7 @@ REGISTER_LAYER(exconvt, ExpandConvTransLayer);
 bool ExpandConvTransLayer::init(const LayerMap &layerMap,
                            const ParameterMap &parameterMap) {
   /* Initialize the basic convolutional parent class */
-  ConvBaseLayerCpu::init(layerMap, parameterMap);
+  ExpandConvBaseLayer::init(layerMap, parameterMap);
 
   return true;
 }
@@ -72,7 +72,7 @@ void ExpandConvTransLayer::forward(PassType passType) {
   resetOutput(batchSize, getSize());
 
   MatrixPtr output = nullptr;
-  for (size_t i = 0; i != inputLayers_.size(); ++i) {
+  for (size_t i = 0; i < inputLayers_.size(); ++i) {
     LayerPtr prevLayer = getPrev(i);
     output = prevLayer->getOutputValue();
     REGISTER_TIMER_INFO("shrinkFwd", getName().c_str());
@@ -80,7 +80,7 @@ void ExpandConvTransLayer::forward(PassType passType) {
   }
 
   /* add the bias-vector */
-  if (biases_.get() != NULL) {
+  if (biases_.get()) {
     if (sharedBiases_) {
       addSharedBias();
     } else {
@@ -102,10 +102,10 @@ void ExpandConvTransLayer::backward(const UpdateCallback &callback) {
     biases_->getParameterPtr()->incUpdate(callback);
   }
 
-  for (size_t i = 0; i != inputLayers_.size(); ++i) {
+  for (size_t i = 0; i < inputLayers_.size(); ++i) {
     /* First, calculate the input layers error */
     for (size_t off = 0; off < imageGrad->getHeight(); off++) {
-      if (NULL != getPrev(i)->getOutputGrad()) {
+      if (getPrev(i)->getOutputGrad()) {
         expandFwdOnce(imageGrad, getPrev(i)->getOutputGrad(), i, off);
       }
     }

paddle/gserver/layers/ExpandConvTransLayer.h

@@ -15,9 +15,9 @@ limitations under the License. */
 
 #pragma once
 
-#include "ConvBaseLayerCpu.h"
 #include "paddle/math/Matrix.h"
 #include <vector>
+#include "ExpandConvBaseLayer.h"
 
 namespace paddle {
 
@@ -28,10 +28,10 @@ namespace paddle {
  *
  * The config file api is img_convTrans_layer.
  */
-class ExpandConvTransLayer : public ConvBaseLayerCpu {
+class ExpandConvTransLayer : public ExpandConvBaseLayer {
 public:
   explicit ExpandConvTransLayer(const LayerConfig& config) :
-    ConvBaseLayerCpu(config) {}
+    ExpandConvBaseLayer(config) {}
 
   ~ExpandConvTransLayer() {}
 

python/paddle/trainer/config_parser.py

@@ -1107,7 +1107,7 @@ def parse_conv(conv, input_layer_name, conv_conf):
                                          conv_conf.caffe_mode)
 
 
-def parse_convt(conv, input_layer_name, conv_conf, num_filters):
+def parse_conv_trans(conv, input_layer_name, conv_conf, num_filters):
     conv_conf.filter_size = conv.filter_size
     conv_conf.filter_size_y = conv.filter_size_y
     conv_conf.channels = conv.channels
@@ -1683,7 +1683,7 @@ class ConvTransLayerBase(LayerBase):
 
         for input_index in xrange(len(self.inputs)):
             input_layer = self.get_input_layer(input_index)
-            parse_convt(
+            parse_conv_trans(
                 self.inputs[input_index].conv,
                 input_layer.name,
                 self.config.inputs[input_index].conv_conf, num_filters)

python/paddle/trainer_config_helpers/layers.py

@@ -1515,7 +1515,8 @@ def img_conv_layer(input, filter_size, num_filters,
                    name=None, num_channels=None,
                    act=None, groups=1, stride=1, padding=0, bias_attr=None,
                    param_attr=None, shared_biases=True, layer_attr=None,
-                   filter_size_y=None, stride_y=None, padding_y=None):
+                   filter_size_y=None, stride_y=None, padding_y=None,
+                   trans=False):
     """
     Convolution layer for image. Paddle only support square input currently and
     thus input image's width equals height.
@@ -1523,120 +1524,7 @@ def img_conv_layer(input, filter_size, num_filters,
     The details of convolution layer, please refer UFLDL's `convolution
     <http://ufldl.stanford.edu/tutorial/supervised/
     FeatureExtractionUsingConvolution/>`_ .
-
-    The num_channel means input image's channel number. It may be 1 or 3 when
-    input is raw pixels of image(mono or RGB), or it may be the previous layer's
-    num_filters * num_group.
-
-    There are several group of filter in PaddlePaddle implementation.
-    Each group will process some channel of the inputs. For example, if an input
-    num_channel = 256, group = 4, num_filter=32, the PaddlePaddle will create
-    32*4 = 128 filters to process inputs. The channels will be split into 4
-    pieces. First 256/4 = 64 channels will process by first 32 filters. The
-    rest channels will be processed by rest group of filters.
-
-    :param name: Layer name.
-    :type name: basestring
-    :param input: Layer Input.
-    :type input: LayerOutput
-    :param filter_size: The x dimension of a filter kernel. Or input a tuple for
-                        two image dimension.
-    :type filter_size: int|tuple|list
-    :param filter_size_y: The y dimension of a filter kernel. Since PaddlePaddle
-                        currently supports rectangular filters, the filter's
-                        shape will be (filter_size, filter_size_y).
-    :type filter_size_y: int|None
-    :param num_filters: Each filter group's number of filter
-    :param act: Activation type. Default is tanh
-    :type act: BaseActivation
-    :param groups: Group size of filters.
-    :type groups: int
-    :param stride: The x dimension of the stride. Or input a tuple for two image
-                   dimension.
-    :type stride: int|tuple|list
-    :param stride_y: The y dimension of the stride.
-    :type stride_y: int
-    :param padding: The x dimension of the padding. Or input a tuple for two
-                    image dimension
-    :type padding: int|tuple|list
-    :param padding_y: The y dimension of the padding.
-    :type padding_y: int
-    :param bias_attr: Convolution bias attribute. None means default bias.
-                      False means no bias.
-    :type bias_attr: ParameterAttribute|False
-    :param num_channels: number of input channels. If None will be set
-                        automatically from previous output.
-    :type num_channels: int
-    :param param_attr: Convolution param attribute. None means default attribute
-    :type param_attr: ParameterAttribute
-    :param shared_biases: Is biases will be shared between filters or not.
-    :type shared_biases: bool
-    :param layer_attr: Layer Extra Attribute.
-    :type layer_attr: ExtraLayerAttribute
-    :return: LayerOutput object.
-    :rtype: LayerOutput
-    """
-    if num_channels is None:
-        assert input.num_filters is not None
-        num_channels = input.num_filters
-
-    if filter_size_y is None:
-        if isinstance(filter_size, collections.Sequence):
-            assert len(filter_size) == 2
-            filter_size, filter_size_y = filter_size
-        else:
-            filter_size_y = filter_size
-
-    if stride_y is None:
-        if isinstance(stride, collections.Sequence):
-            assert len(stride) == 2
-            stride, stride_y = stride
-        else:
-            stride_y = stride
-
-    if padding_y is None:
-        if isinstance(padding, collections.Sequence):
-            assert len(padding) == 2
-            padding, padding_y = padding
-        else:
-            padding_y = padding
-
-    if param_attr.attr.get('initial_smart'):
-        # special initial for conv layers.
-        init_w = (2.0 / (filter_size ** 2 * num_channels)) ** 0.5
-        param_attr.attr["initial_mean"] = 0.0
-        param_attr.attr["initial_std"] = init_w
-        param_attr.attr["initial_strategy"] = 0
-        param_attr.attr["initial_smart"] = False
-    Layer(
-        name=name,
-        inputs=Input(input.name, conv=Conv(
-            filter_size=filter_size, padding=padding, stride=stride,
-            channels=num_channels, groups=groups,
-            filter_size_y=filter_size_y, padding_y=padding_y,
-            stride_y=stride_y),
-                     **param_attr.attr),
-        active_type=act.name,
-        num_filters=num_filters,
-        bias=ParamAttr.to_bias(bias_attr),
-        shared_biases=shared_biases,
-        type=LayerType.CONV_LAYER,
-        **ExtraLayerAttribute.to_kwargs(layer_attr)
-    )
-    return LayerOutput(name, LayerType.CONV_LAYER, parents=[input],
-                       activation=act, num_filters=num_filters)
-
-@wrap_name_default("convt")
-@wrap_param_attr_default()
-@wrap_bias_attr_default()
-@wrap_act_default(act=ReluActivation())
-@layer_support(DROPOUT)
-def img_convTrans_layer(input, filter_size, num_filters,
-                   name=None, num_channels=None,
-                   act=None, groups=1, stride=1, padding=0, bias_attr=None,
-                   param_attr=None, shared_biases=True, layer_attr=None,
-                   filter_size_y=None, stride_y=None, padding_y=None):
-    """
+    
     Convolution Transpose (deconv) layer for image. Paddle only support square 
     input currently and thus input image's width equals height.
 
@@ -1644,7 +1532,6 @@ def img_convTrans_layer(input, filter_size, num_filters,
     please refer to the following explanation and references therein
     <http://datascience.stackexchange.com/questions/6107/
     what-are-deconvolutional-layers/>`_ .
-
     The num_channel means input image's channel number. It may be 1 or 3 when
     input is raw pixels of image(mono or RGB), or it may be the previous layer's
     num_filters * num_group.
@@ -1694,6 +1581,8 @@ def img_convTrans_layer(input, filter_size, num_filters,
     :type shared_biases: bool
     :param layer_attr: Layer Extra Attribute.
     :type layer_attr: ExtraLayerAttribute
+    :param trans: true if it is a convTransLayer, false if it is a convLayer
+    :type trans: bool
     :return: LayerOutput object.
     :rtype: LayerOutput
     """
@@ -1729,6 +1618,12 @@ def img_convTrans_layer(input, filter_size, num_filters,
         param_attr.attr["initial_std"] = init_w
         param_attr.attr["initial_strategy"] = 0
         param_attr.attr["initial_smart"] = False
+    
+    if trans:
+        lt = LayerType.CONVTRANS_LAYER
+    else:
+        lt = LayerType.CONV_LAYER
+    
     Layer(
         name=name,
         inputs=Input(input.name, conv=Conv(
@@ -1741,14 +1636,13 @@ def img_convTrans_layer(input, filter_size, num_filters,
         num_filters=num_filters,
         bias=ParamAttr.to_bias(bias_attr),
         shared_biases=shared_biases,
-        type=LayerType.CONVTRANS_LAYER,
+        type=lt,
         **ExtraLayerAttribute.to_kwargs(layer_attr)
     )
-    return LayerOutput(name, LayerType.CONVTRANS_LAYER, parents=[input],
+    return LayerOutput(name, lt, parents=[input],
                        activation=act, num_filters=num_filters)
 
 
-
 @wrap_name_default("pool")
 @layer_support()
 def img_pool_layer(input, pool_size, name=None,