From 7ad76355af156f046008cd7d2687cfd7be84b965 Mon Sep 17 00:00:00 2001
From: zhaojiaying01 <zhaojiaying01@baidu.com>
Date: Wed, 24 Oct 2018 19:08:57 +0800
Subject: [PATCH] update conv3x3 and fix CLImageConverterNWBlock

---
 src/framework/cl/cl_image_converter.cpp             | 8 ++++----
 src/operators/kernel/cl/conv_add_bn_relu_kernel.cpp | 4 ++--
 src/operators/kernel/cl/conv_add_kernel.cpp         | 2 +-
 3 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/src/framework/cl/cl_image_converter.cpp b/src/framework/cl/cl_image_converter.cpp
index ad3aec5b86..ebcfd0d675 100644
--- a/src/framework/cl/cl_image_converter.cpp
+++ b/src/framework/cl/cl_image_converter.cpp
@@ -248,8 +248,8 @@ void CLImageConverterNWBlock::NCHWToImage(float *tensor, half_t *image,
     for (int c = 0; c < C; c++) {
       for (int h = 0; h < H; ++h) {
         for (int w = 0; w < W; ++w) {
-          int index = 4 * c * (width * H) + 4 * (n / 4) * H * W + h * W * 4 +
-                      w * 4 + (n % 4);
+          int index = 4 * c * (width * H) + 4 * h * width + 4 * W * (n / 4) +
+                      w * 4 + n % 4;
           if (n < N) {
             image[index] = Float2Half(*p);
             p++;
@@ -283,8 +283,8 @@ void CLImageConverterNWBlock::ImageToNCHW(half_t *image, float *tensor,
     for (int c = 0; c < C; c++) {
       for (int h = 0; h < H; ++h) {
         for (int w = 0; w < W; ++w) {
-          int index = 4 * c * (width * H) + 4 * (n / 4) * H * W + h * W * 4 +
-                      w * 4 + (n % 4);
+          int index = 4 * c * (width * H) + 4 * h * width + 4 * W * (n / 4) +
+                      w * 4 + n % 4;
           *p = Half2Float(image[index]);
           p++;
           if (index >= (width * height * 4)) {
diff --git a/src/operators/kernel/cl/conv_add_bn_relu_kernel.cpp b/src/operators/kernel/cl/conv_add_bn_relu_kernel.cpp
index 29f1b15199..c990d64f67 100644
--- a/src/operators/kernel/cl/conv_add_bn_relu_kernel.cpp
+++ b/src/operators/kernel/cl/conv_add_bn_relu_kernel.cpp
@@ -150,8 +150,8 @@ bool ConvAddBNReluKernel<GPU_CL, float>::Init(
 
   } else if (param->Filter()->dims()[2] == 3 &&
              param->Filter()->dims()[3] == 3) {
-    param->Filter()->InitCLImage(cl_helper_.CLContext(),
-                                 cl_helper_.CLCommandQueue());
+    param->Filter()->InitNImage(cl_helper_.CLContext(),
+                                cl_helper_.CLCommandQueue());
 
     this->cl_helper_.AddKernel("conv_3x3", "conv_add_bn_relu_kernel.cl");
     DLOG << " conv add bn relu conv_3x3";
diff --git a/src/operators/kernel/cl/conv_add_kernel.cpp b/src/operators/kernel/cl/conv_add_kernel.cpp
index d8064fd50e..2a21391c5e 100644
--- a/src/operators/kernel/cl/conv_add_kernel.cpp
+++ b/src/operators/kernel/cl/conv_add_kernel.cpp
@@ -45,7 +45,7 @@ bool ConvAddKernel<GPU_CL, float>::Init(FusionConvAddParam<GPU_CL> *param) {
 
   } else if (param->Filter()->dims()[2] == 3 &&
              param->Filter()->dims()[3] == 3) {
-    param->Filter()->InitCLImage(cl_helper_.CLContext(),
+    param->Filter()->InitNImage(cl_helper_.CLContext(),
                                  cl_helper_.CLCommandQueue());
 
     this->cl_helper_.AddKernel("conv_3x3", "conv_add_kernel.cl");
-- 
GitLab