Merge pull request #10602 from pengli:dnn

modules/dnn/src/layers/batch_norm_layer.cpp

@@ -12,6 +12,7 @@ Implementation of Batch Normalization layer.
 #include "../precomp.hpp"
 #include "op_halide.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
+#include "opencl_kernels_dnn.hpp"
 
 namespace cv
 {
@@ -22,7 +23,7 @@ class BatchNormLayerImpl : public BatchNormLayer
 {
 public:
     Mat weights_, bias_;
-    Mat weightMat, biasMat;
+    UMat umat_weight, umat_bias;
 
     BatchNormLayerImpl(const LayerParams& params)
     {
@@ -80,6 +81,9 @@ public:
             dstWeightsData[i] = w;
             dstBiasData[i] = (hasBias ? biasData[i] : 0.0f) - w * meanData[i] * varMeanScale;
         }
+
+        umat_weight = weights_.getUMat(ACCESS_READ);
+        umat_bias = bias_.getUMat(ACCESS_READ);
     }
 
     void getScaleShift(Mat& scale, Mat& shift) const
@@ -97,25 +101,6 @@ public:
         return true;
     }
 
-    void finalize(const std::vector<Mat*> &inputs, std::vector<Mat> &outputs)
-    {
-        if (inputs[0]->dims == 4)
-        {
-            int groups = inputs[0]->size[0];
-            int channels = inputs[0]->size[1];
-            int rows = inputs[0]->size[2];
-            int cols = inputs[0]->size[3];
-            MatShape s = shape(groups * channels, rows * cols);
-            weightMat = Mat(s[0], s[1], CV_32FC1);
-            biasMat = Mat(s[0], s[1], CV_32FC1);
-            for (int n = 0; n < s[0]; n++)
-            {
-                weightMat.row(n).setTo(weights_.at<float>(n % channels));
-                biasMat.row(n).setTo(bias_.at<float>(n % channels));
-            }
-        }
-    }
-
     virtual bool supportBackend(int backendId)
     {
         return backendId == DNN_BACKEND_DEFAULT ||
@@ -155,8 +140,23 @@ public:
                 MatShape s = shape(groups * channels, rows * cols);
                 UMat src = inputs[ii].reshape(1, s.size(), &s[0]);
                 UMat dst = outputs[ii].reshape(1, s.size(), &s[0]);
-                multiply(src, weightMat, dst);
-                add(dst, biasMat, dst);
+                int number = (s[1] % 8 == 0) ? 8 : ((s[1] % 4 == 0) ? 4 : 1);
+                String buildopt = format("-DNUM=%d ", number);
+                String kname = format("batch_norm%d", number);
+                ocl::Kernel kernel(kname.c_str(), ocl::dnn::batchnorm_oclsrc, buildopt);
+                if (kernel.empty())
+                    return false;
+                size_t global[] = { (size_t)s[0], (size_t)(s[1] / number) };
+                kernel.set(0, ocl::KernelArg::PtrReadOnly(src));
+                kernel.set(1, (int)s[0]);
+                kernel.set(2, (int)s[1]);
+                kernel.set(3, (int)channels);
+                kernel.set(4, ocl::KernelArg::PtrReadOnly(umat_weight));
+                kernel.set(5, ocl::KernelArg::PtrReadOnly(umat_bias));
+                kernel.set(6, ocl::KernelArg::PtrWriteOnly(dst));
+                bool ret = kernel.run(2, global, NULL, false);
+                if (!ret)
+                    return false;
             }
         }
         return true;

modules/dnn/src/layers/elementwise_layers.cpp

@@ -267,7 +267,6 @@ struct ReLUFunctor
 
     bool applyOCL(InputArrayOfArrays inps, OutputArrayOfArrays outs, OutputArrayOfArrays internals)
     {
-        size_t wgSize = ocl::Device::getDefault().maxWorkGroupSize();
         std::vector<UMat> inputs;
         std::vector<UMat> outputs;
 
@@ -287,7 +286,7 @@ struct ReLUFunctor
             kernel.set(2, ocl::KernelArg::PtrWriteOnly(dst));
 
             size_t gSize = src.total();
-            CV_Assert(kernel.run(1, &gSize, &wgSize, false));
+            CV_Assert(kernel.run(1, &gSize, NULL, false));
         }
 
         return true;
@@ -395,8 +394,28 @@ struct TanHFunctor
 #ifdef HAVE_OPENCL
     bool applyOCL(InputArrayOfArrays inps, OutputArrayOfArrays outs, OutputArrayOfArrays internals)
     {
-        // TODO: implement OCL version
-        return false;
+        std::vector<UMat> inputs;
+        std::vector<UMat> outputs;
+
+        inps.getUMatVector(inputs);
+        outs.getUMatVector(outputs);
+        String buildopt = oclGetTMacro(inputs[0]);
+
+        for (size_t i = 0; i < inputs.size(); i++)
+        {
+            UMat& src = inputs[i];
+            UMat& dst = outputs[i];
+
+            ocl::Kernel kernel("TanHForward", ocl::dnn::activations_oclsrc, buildopt);
+            kernel.set(0, (int)src.total());
+            kernel.set(1, ocl::KernelArg::PtrReadOnly(src));
+            kernel.set(2, ocl::KernelArg::PtrWriteOnly(dst));
+
+            size_t gSize = src.total();
+            CV_Assert(kernel.run(1, &gSize, NULL, false));
+        }
+
+        return true;
     }
 #endif
 
@@ -594,8 +613,31 @@ struct PowerFunctor
 #ifdef HAVE_OPENCL
     bool applyOCL(InputArrayOfArrays inps, OutputArrayOfArrays outs, OutputArrayOfArrays internals)
     {
-        // TODO: implement OCL version
-        return false;
+        std::vector<UMat> inputs;
+        std::vector<UMat> outputs;
+
+        inps.getUMatVector(inputs);
+        outs.getUMatVector(outputs);
+        String buildopt = oclGetTMacro(inputs[0]);
+
+        for (size_t i = 0; i < inputs.size(); i++)
+        {
+            UMat& src = inputs[i];
+            UMat& dst = outputs[i];
+
+            ocl::Kernel kernel("PowForward", ocl::dnn::activations_oclsrc, buildopt);
+            kernel.set(0, (int)src.total());
+            kernel.set(1, ocl::KernelArg::PtrReadOnly(src));
+            kernel.set(2, ocl::KernelArg::PtrWriteOnly(dst));
+            kernel.set(3, (float)power);
+            kernel.set(4, (float)scale);
+            kernel.set(5, (float)shift);
+
+            size_t gSize = src.total();
+            CV_Assert(kernel.run(1, &gSize, NULL, false));
+        }
+
+        return true;
     }
 #endif
 
@@ -624,9 +666,11 @@ struct ChannelsPReLUFunctor
 {
     typedef ChannelsPReLULayer Layer;
     Mat scale;
+    UMat scale_umat;
 
     explicit ChannelsPReLUFunctor(const Mat& scale_=Mat()) : scale(scale_)
     {
+        scale_umat = scale.getUMat(ACCESS_READ);
     }
 
     void apply(const float* srcptr, float* dstptr, int len, size_t planeSize, int cn0, int cn1) const
@@ -669,8 +713,31 @@ struct ChannelsPReLUFunctor
 #ifdef HAVE_OPENCL
     bool applyOCL(InputArrayOfArrays inps, OutputArrayOfArrays outs, OutputArrayOfArrays internals)
     {
-        // TODO: implement OCL version
-        return false;
+        std::vector<UMat> inputs;
+        std::vector<UMat> outputs;
+
+        inps.getUMatVector(inputs);
+        outs.getUMatVector(outputs);
+        String buildopt = oclGetTMacro(inputs[0]);
+
+        for (size_t i = 0; i < inputs.size(); i++)
+        {
+            UMat& src = inputs[i];
+            UMat& dst = outputs[i];
+
+            ocl::Kernel kernel("PReLUForward", ocl::dnn::activations_oclsrc, buildopt);
+            kernel.set(0, (int)src.total());
+            kernel.set(1, (int)src.size[1]);
+            kernel.set(2, (int)total(shape(src), 2));
+            kernel.set(3, ocl::KernelArg::PtrReadOnly(src));
+            kernel.set(4, ocl::KernelArg::PtrWriteOnly(dst));
+            kernel.set(5, ocl::KernelArg::PtrReadOnly(scale_umat));
+
+            size_t gSize = src.total();
+            CV_Assert(kernel.run(1, &gSize, NULL, false));
+        }
+
+        return true;
     }
 #endif
 

modules/dnn/src/layers/mvn_layer.cpp

@@ -43,6 +43,8 @@
 #include "../precomp.hpp"
 #include "layers_common.hpp"
 #include <opencv2/dnn/shape_utils.hpp>
+#include "math_functions.hpp"
+#include "opencl_kernels_dnn.hpp"
 
 namespace cv
 {
@@ -60,11 +62,93 @@ public:
         eps = params.get<double>("eps", 1e-9);
     }
 
+#ifdef HAVE_OPENCL
+    bool forward_ocl(InputArrayOfArrays inputs_, OutputArrayOfArrays outputs_, OutputArrayOfArrays internals_)
+    {
+        std::vector<UMat> inputs;
+        std::vector<UMat> outputs;
+
+        inputs_.getUMatVector(inputs);
+        outputs_.getUMatVector(outputs);
+
+        for (size_t inpIdx = 0; inpIdx < inputs.size(); inpIdx++)
+        {
+            UMat &inpBlob = inputs[inpIdx];
+            UMat &outBlob = outputs[inpIdx];
+
+            int splitDim = (acrossChannels) ? 1 : 2;
+            int i, newRows = 1;
+            for( i = 0; i < splitDim; i++ )
+                newRows *= inpBlob.size[i];
+
+            MatShape s = shape(newRows, inpBlob.total() / newRows);
+            UMat& inpMat = inpBlob;
+            UMat& outMat = outBlob;
+            UMat oneMat = UMat::ones(s[1], 1, CV_32F);
+            UMat meanMat = UMat(s[0], 1, CV_32F);
+            UMat devMat  = UMat(s[0], 1, CV_32F);
+            UMat tmpMat  = UMat(s[0], s[1], CV_32F);
+            float alpha = 1.0f / s[1];
+
+            bool ret = ocl4dnn::ocl4dnnGEMV<float>(ocl4dnn::CblasNoTrans, s[0], s[1], alpha,
+                                                   inpMat, 0, oneMat, 0, 0.0f, meanMat, 0);
+            if (!ret)
+                return false;
+
+            int number = (s[1] % 8 == 0) ? 8 : ((s[1] % 4 == 0) ? 4 : 1);
+            String buildopt = format("-DNUM=%d ", number);
+            String kname = format("calc_mean%d", number);
+            ocl::Kernel kernel(kname.c_str(), ocl::dnn::mvn_oclsrc, buildopt);
+            if (kernel.empty())
+                return false;
+            size_t global[] = { (size_t)s[0], (size_t)(s[1] / number) };
+            kernel.set(0, ocl::KernelArg::PtrReadOnly(inpMat));
+            kernel.set(1, (int)s[0]);
+            kernel.set(2, (int)s[1]);
+            kernel.set(3, ocl::KernelArg::PtrReadOnly(meanMat));
+            kernel.set(4, ocl::KernelArg::PtrWriteOnly(tmpMat));
+            ret = kernel.run(2, global, NULL, false);
+            if (!ret)
+                return false;
+
+            if (normVariance)
+            {
+                ret = ocl4dnn::ocl4dnnGEMV<float>(ocl4dnn::CblasNoTrans, s[0], s[1], alpha,
+                                                  tmpMat, 0, oneMat, 0, 0.0f, devMat, 0);
+                if (!ret)
+                    return false;
+            }
+
+            kname = format("mvn%d", number);
+            if (normVariance)
+                buildopt += "-DNORM_VARIANCE";
+            ocl::Kernel kernel1(kname.c_str(), ocl::dnn::mvn_oclsrc, buildopt);
+            if (kernel1.empty())
+                return false;
+            kernel1.set(0, ocl::KernelArg::PtrReadOnly(inpMat));
+            kernel1.set(1, (int)s[0]);
+            kernel1.set(2, (int)s[1]);
+            kernel1.set(3, (float)eps);
+            kernel1.set(4, ocl::KernelArg::PtrReadOnly(meanMat));
+            kernel1.set(5, ocl::KernelArg::PtrReadOnly(devMat));
+            kernel1.set(6, ocl::KernelArg::PtrWriteOnly(outMat));
+            ret = kernel1.run(2, global, NULL, false);
+            if (!ret)
+                return false;
+        }
+        return true;
+    }
+#endif
+
     void forward(InputArrayOfArrays inputs_arr, OutputArrayOfArrays outputs_arr, OutputArrayOfArrays internals_arr)
     {
         CV_TRACE_FUNCTION();
         CV_TRACE_ARG_VALUE(name, "name", name.c_str());
 
+        CV_OCL_RUN((preferableTarget == DNN_TARGET_OPENCL) &&
+                   OCL_PERFORMANCE_CHECK(ocl::Device::getDefault().isIntel()),
+                   forward_ocl(inputs_arr, outputs_arr, internals_arr))
+
         Layer::forward_fallback(inputs_arr, outputs_arr, internals_arr);
     }
 

modules/dnn/src/opencl/activations.cl

@@ -54,6 +54,15 @@ __kernel void ReLUForward(const int count, __global const T* in, __global T* out
 #endif
 }
 
+__kernel void PReLUForward(const int count, const int channels, const int plane_size,
+                           __global const T* in, __global T* out, __global const T* slope_data)
+{
+  int index = get_global_id(0);
+  int c = (index / plane_size) % channels;
+  if(index < count)
+  out[index] = in[index] > 0 ? in[index] : in[index] * slope_data[c];
+}
+
 __kernel void TanHForward(const int count, __global T* in, __global T* out) {
   int index = get_global_id(0);
   if(index < count)

modules/dnn/src/opencl/batchnorm.cl

@@ -1,26 +1,84 @@
+/*M///////////////////////////////////////////////////////////////////////////////////////
+//
+//  IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
+//
+//  By downloading, copying, installing or using the software you agree to this license.
+//  If you do not agree to this license, do not download, install,
+//  copy or use the software.
+//
+//
+//                           License Agreement
+//                For Open Source Computer Vision Library
+//
+// Copyright (C) 2017, Intel Corporation, all rights reserved.
+// Copyright (c) 2016-2017 Fabian David Tschopp, all rights reserved.
+// Third party copyrights are property of their respective owners.
+//
+// Redistribution and use in source and binary forms, with or without modification,
+// are permitted provided that the following conditions are met:
+//
+//   * Redistribution's of source code must retain the above copyright notice,
+//     this list of conditions and the following disclaimer.
+//
+//   * Redistribution's in binary form must reproduce the above copyright notice,
+//     this list of conditions and the following disclaimer in the documentation
+//     and/or other materials provided with the distribution.
+//
+//   * The name of the copyright holders may not be used to endorse or promote products
+//     derived from this software without specific prior written permission.
+//
+// This software is provided by the copyright holders and contributors "as is" and
+// any express or implied warranties, including, but not limited to, the implied
+// warranties of merchantability and fitness for a particular purpose are disclaimed.
+// In no event shall the Intel Corporation or contributors be liable for any direct,
+// indirect, incidental, special, exemplary, or consequential damages
+// (including, but not limited to, procurement of substitute goods or services;
+// loss of use, data, or profits; or business interruption) however caused
+// and on any theory of liability, whether in contract, strict liability,
+// or tort (including negligence or otherwise) arising in any way out of
+// the use of this software, even if advised of the possibility of such damage.
+//
+//M*/
 
-__kernel void batchnorm(__global const T *src, int src_offset,
-                        __global const float *meanMat,
-                        float varMeanScale,
-                        __global const float *invStdMat,
-                        __global const float *weight,
-                        __global const float *bias,
-                        int hasWeight, int hasBias,
-                        int width, int height, int channel,
-                        __global T *dst, int dst_offset)
+#define Dtype float
+#define Dtype4 float4
+#define Dtype8 float8
+
+#if NUM == 8
+    #define load(src, index) vload8(0, src + index)
+    #define store(vec, dst, index) vstore8(vec, 0, dst + index)
+    #define vec_type Dtype8
+    #define BATCH_NORM batch_norm8
+#elif NUM == 4
+    #define load(src, index) vload4(0, src + index)
+    #define store(vec, dst, index) vstore4(vec, 0, dst + index)
+    #define vec_type Dtype4
+    #define BATCH_NORM batch_norm4
+#elif NUM == 1
+    #define load(src, index) src[index]
+    #define store(vec, dst, index) dst[index] = vec
+    #define vec_type Dtype
+    #define BATCH_NORM batch_norm1
+#endif
+
+__kernel void BATCH_NORM(__global const Dtype* src,
+                         const int rows,
+                         const int cols,
+                         const int channels,
+                         __global const Dtype* weight,
+                         __global const Dtype* bias,
+                         __global Dtype* dst)
 {
     int x = get_global_id(0);
-    int y = get_global_id(1);
-    int c = get_global_id(2);
+    int y = get_global_id(1) * NUM;
+    int index = x * cols + y;
 
-    if (x >= width || y >= height || c >= channel)
+    if (x >= rows || y >= cols)
         return;
 
-    float mean = meanMat[c] * varMeanScale;
-    float invstd = invStdMat[c];
-    float w = hasWeight ? weight[c] : 1;
-    float b = hasBias ? bias[c] : 0;
-    int index = y * width + x + c * width * height;
-    T val = (src[index + src_offset] - mean) * w * invstd + b;
-    dst[index + dst_offset] = val;
+    Dtype w = weight[x % channels];
+    Dtype b = bias[x % channels];
+    vec_type src_vec = load(src, index);
+    vec_type dst_vec = src_vec * w + (vec_type)b;
+    store(dst_vec, dst, index);
 }

modules/dnn/src/opencl/mvn.cl

@@ -0,0 +1,112 @@
+/*M///////////////////////////////////////////////////////////////////////////////////////
+//
+//  IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
+//
+//  By downloading, copying, installing or using the software you agree to this license.
+//  If you do not agree to this license, do not download, install,
+//  copy or use the software.
+//
+//
+//                           License Agreement
+//                For Open Source Computer Vision Library
+//
+// Copyright (C) 2017, Intel Corporation, all rights reserved.
+// Copyright (c) 2016-2017 Fabian David Tschopp, all rights reserved.
+// Third party copyrights are property of their respective owners.
+//
+// Redistribution and use in source and binary forms, with or without modification,
+// are permitted provided that the following conditions are met:
+//
+//   * Redistribution's of source code must retain the above copyright notice,
+//     this list of conditions and the following disclaimer.
+//
+//   * Redistribution's in binary form must reproduce the above copyright notice,
+//     this list of conditions and the following disclaimer in the documentation
+//     and/or other materials provided with the distribution.
+//
+//   * The name of the copyright holders may not be used to endorse or promote products
+//     derived from this software without specific prior written permission.
+//
+// This software is provided by the copyright holders and contributors "as is" and
+// any express or implied warranties, including, but not limited to, the implied
+// warranties of merchantability and fitness for a particular purpose are disclaimed.
+// In no event shall the Intel Corporation or contributors be liable for any direct,
+// indirect, incidental, special, exemplary, or consequential damages
+// (including, but not limited to, procurement of substitute goods or services;
+// loss of use, data, or profits; or business interruption) however caused
+// and on any theory of liability, whether in contract, strict liability,
+// or tort (including negligence or otherwise) arising in any way out of
+// the use of this software, even if advised of the possibility of such damage.
+//
+//M*/
+
+#define Dtype float
+#define Dtype4 float4
+#define Dtype8 float8
+
+#if NUM == 8
+    #define load(src, index) vload8(0, src + index)
+    #define store(vec, dst, index) vstore8(vec, 0, dst + index)
+    #define vec_type Dtype8
+    #define CALC_MEAN calc_mean8
+    #define MVN mvn8
+#elif NUM == 4
+    #define load(src, index) vload4(0, src + index)
+    #define store(vec, dst, index) vstore4(vec, 0, dst + index)
+    #define vec_type Dtype4
+    #define CALC_MEAN calc_mean4
+    #define MVN mvn4
+#elif NUM == 1
+    #define load(src, index) src[index]
+    #define store(vec, dst, index) dst[index] = vec
+    #define vec_type Dtype
+    #define CALC_MEAN calc_mean1
+    #define MVN mvn1
+#endif
+
+__kernel void CALC_MEAN(__global const Dtype* src,
+                        const int rows,
+                        const int cols,
+                        __global Dtype* mean,
+                        __global Dtype* dst)
+{
+    int x = get_global_id(0);
+    int y = get_global_id(1) * NUM;
+    int index = x * cols + y;
+
+    if (x >= rows || y >= cols)
+        return;
+
+    Dtype mean_val = mean[x];
+    vec_type src_vec = load(src, index);
+    vec_type dst_vec = pow(src_vec - (vec_type)mean_val, 2);
+    store(dst_vec, dst, index);
+}
+
+__kernel void MVN(__global const Dtype* src,
+                  const int rows,
+                  const int cols,
+                  const Dtype eps,
+                  __global const Dtype* mean,
+                  __global const Dtype* dev,
+                  __global Dtype* dst)
+{
+    int x = get_global_id(0);
+    int y = get_global_id(1) * NUM;
+    int index = x * cols + y;
+
+    if (x >= rows || y >= cols)
+        return;
+
+    Dtype mean_val = mean[x];
+    Dtype dev_val = sqrt(dev[x]);
+    Dtype alpha;
+#ifdef NORM_VARIANCE
+    alpha = 1 / (eps + dev_val);
+#else
+    alpha = 1;
+#endif
+    vec_type src_vec = load(src, index) - (vec_type)mean_val;
+    vec_type dst_vec = src_vec * alpha;
+    store(dst_vec, dst, index);
+}

modules/dnn/test/test_layers.cpp

@@ -202,6 +202,11 @@ TEST(Layer_Test_MVN, Accuracy)
     testLayerUsingCaffeModels("layer_mvn");
 }
 
+OCL_TEST(Layer_Test_MVN, Accuracy)
+{
+    testLayerUsingCaffeModels("layer_mvn", DNN_TARGET_OPENCL);
+}
+
 void testReshape(const MatShape& inputShape, const MatShape& targetShape,
                  int axis = 0, int num_axes = -1,
                  MatShape mask = MatShape())
@@ -331,6 +336,12 @@ TEST(Layer_Test_PReLU, Accuracy)
     testLayerUsingCaffeModels("layer_prelu_fc", DNN_TARGET_CPU, true, false);
 }
 
+OCL_TEST(Layer_Test_PReLU, Accuracy)
+{
+    testLayerUsingCaffeModels("layer_prelu", DNN_TARGET_OPENCL, true);
+    testLayerUsingCaffeModels("layer_prelu_fc", DNN_TARGET_OPENCL, true, false);
+}
+
 //template<typename XMat>
 //static void test_Layer_Concat()
 //{