From a2f3644e87c11c222f86aa2aeb0e31d858ebb024 Mon Sep 17 00:00:00 2001
From: Alexey <AlexeyAB@users.noreply.github.com>
Date: Sat, 21 Apr 2018 20:56:46 +0000
Subject: [PATCH] Merge pull request #675 from IlyaOvodov/Flip_Fix
---
src/convolutional_layer.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/src/convolutional_layer.c b/src/convolutional_layer.c
index 91c5b3c..9a76bdf 100644
--- a/src/convolutional_layer.c
+++ b/src/convolutional_layer.c
@@ -178,6 +178,8 @@
// batch norm
cudnnSetTensor4dDescriptor(l->normTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, 1, l->out_c, 1, 1);
cudnnSetTensor4dDescriptor(l->normDstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, l->batch, l->out_c, l->out_h, l->out_w);
+
+ cudnnSetTensor4dDescriptor(l->normDstTensorDescF16, CUDNN_TENSOR_NCHW, data_type, l->batch, l->out_c, l->out_h, l->out_w);
#if(CUDNN_MAJOR >= 6)
cudnnSetConvolution2dDescriptor(l->convDesc, l->pad, l->pad, l->stride, l->stride, 1, 1, CUDNN_CROSS_CORRELATION, CUDNN_DATA_FLOAT); // cudnn >= 6.0
#else
@@ -379,6 +381,7 @@
}
#ifdef CUDNN
cudnnCreateTensorDescriptor(&l.normDstTensorDesc);
+ cudnnCreateTensorDescriptor(&l.normDstTensorDescF16);
cudnnCreateTensorDescriptor(&l.normTensorDesc);
cudnnCreateTensorDescriptor(&l.srcTensorDesc);
cudnnCreateTensorDescriptor(&l.dstTensorDesc);
--
Gitblit v1.10.0