From 5b6be00d4b1ffd671c20c4c72d2239c924eaa3d4 Mon Sep 17 00:00:00 2001
From: AlexeyAB <alexeyab84@gmail.com>
Date: Thu, 23 Aug 2018 12:28:34 +0000
Subject: [PATCH] Added yolov3-tiny_xnor.cfg
---
src/maxpool_layer.c | 21 ++++++++++++++-------
1 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/src/maxpool_layer.c b/src/maxpool_layer.c
index d1fbacb..5ff7e9d 100644
--- a/src/maxpool_layer.c
+++ b/src/maxpool_layer.c
@@ -1,5 +1,6 @@
#include "maxpool_layer.h"
#include "cuda.h"
+#include "gemm.h"
#include <stdio.h>
image get_maxpool_image(maxpool_layer l)
@@ -27,8 +28,8 @@
l.w = w;
l.c = c;
l.pad = padding;
- l.out_w = (w + 2*padding - size + 1)/stride + 1;
- l.out_h = (h + 2*padding - size + 1)/stride + 1;
+ l.out_w = (w + padding - size) / stride + 1;
+ l.out_h = (h + padding - size) / stride + 1;
l.out_c = c;
l.outputs = l.out_h * l.out_w * l.out_c;
l.inputs = h*w*c;
@@ -47,7 +48,8 @@
l.output_gpu = cuda_make_array(l.output, output_size);
l.delta_gpu = cuda_make_array(l.delta, output_size);
#endif
- fprintf(stderr, "max %d x %d / %d %4d x%4d x%4d -> %4d x%4d x%4d\n", size, size, stride, w, h, c, l.out_w, l.out_h, l.out_c);
+ l.bflops = (l.size*l.size*l.c * l.out_h*l.out_w) / 1000000000.;
+ fprintf(stderr, "max %d x %d / %d %4d x%4d x%4d -> %4d x%4d x%4d %5.3f BF\n", size, size, stride, w, h, c, l.out_w, l.out_h, l.out_c, l.bflops);
return l;
}
@@ -57,8 +59,8 @@
l->w = w;
l->inputs = h*w*l->c;
- l->out_w = (w + 2*l->pad - l->size + 1)/l->stride + 1;
- l->out_h = (h + 2*l->pad - l->size + 1)/l->stride + 1;
+ l->out_w = (w + l->pad - l->size) / l->stride + 1;
+ l->out_h = (h + l->pad - l->size) / l->stride + 1;
l->outputs = l->out_w * l->out_h * l->c;
int output_size = l->outputs * l->batch;
@@ -78,9 +80,14 @@
void forward_maxpool_layer(const maxpool_layer l, network_state state)
{
+ if (!state.train) {
+ forward_maxpool_layer_avx(state.input, l.output, l.indexes, l.size, l.w, l.h, l.out_w, l.out_h, l.c, l.pad, l.stride, l.batch);
+ return;
+ }
+
int b,i,j,k,m,n;
- int w_offset = -l.pad;
- int h_offset = -l.pad;
+ int w_offset = -l.pad / 2;
+ int h_offset = -l.pad / 2;
int h = l.out_h;
int w = l.out_w;
--
Gitblit v1.10.0