From aa5996d58e68edfbefe51061856aecd549dd09c4 Mon Sep 17 00:00:00 2001
From: Joseph Redmon <pjreddie@gmail.com>
Date: Tue, 13 Jan 2015 01:27:08 +0000
Subject: [PATCH] Faster
---
src/convolutional_layer.c | 24 ++++++++++++++++--------
1 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/src/convolutional_layer.c b/src/convolutional_layer.c
index bb2135f..fc5cb0e 100644
--- a/src/convolutional_layer.c
+++ b/src/convolutional_layer.c
@@ -62,12 +62,12 @@
layer->biases = calloc(n, sizeof(float));
layer->bias_updates = calloc(n, sizeof(float));
- float scale = 1./(size*size*c);
- scale = .01;
+ float scale = 1./sqrt(size*size*c);
+ //scale = .05;
for(i = 0; i < c*n*size*size; ++i) layer->filters[i] = scale*rand_normal();
for(i = 0; i < n; ++i){
//layer->biases[i] = rand_normal()*scale + scale;
- layer->biases[i] = .01;
+ layer->biases[i] = scale;
}
int out_h = convolutional_out_height(*layer);
int out_w = convolutional_out_width(*layer);
@@ -170,7 +170,9 @@
int n = layer.size*layer.size*layer.c;
int k = convolutional_out_height(layer)*
convolutional_out_width(layer);
+
gradient_array(layer.output, m*k*layer.batch, layer.activation, layer.delta);
+
learn_bias_convolutional_layer(layer);
if(delta) memset(delta, 0, layer.batch*layer.h*layer.w*layer.c*sizeof(float));
@@ -264,13 +266,18 @@
}
#ifdef GPU
+#define BLOCK 32
+
+#define STR_HELPER(x) #x
+#define STR(x) STR_HELPER(x)
+
cl_kernel get_convolutional_learn_bias_kernel()
{
static int init = 0;
static cl_kernel kernel;
if(!init){
- kernel = get_kernel("src/convolutional_layer.cl", "learn_bias", 0);
+ kernel = get_kernel("src/convolutional_layer.cl", "learn_bias", "-D BLOCK=" STR(BLOCK));
init = 1;
}
return kernel;
@@ -291,9 +298,10 @@
cl.error = clSetKernelArg(kernel, i++, sizeof(layer.bias_updates_cl), (void*) &layer.bias_updates_cl);
check_error(cl);
- const size_t global_size[] = {layer.n};
+ const size_t global_size[] = {layer.n*BLOCK};
+ const size_t local_size[] = {BLOCK};
- cl.error = clEnqueueNDRangeKernel(queue, kernel, 1, 0, global_size, 0, 0, 0, 0);
+ cl.error = clEnqueueNDRangeKernel(queue, kernel, 1, 0, global_size, local_size, 0, 0, 0);
check_error(cl);
}
@@ -302,7 +310,7 @@
static int init = 0;
static cl_kernel kernel;
if(!init){
- kernel = get_kernel("src/convolutional_layer.cl", "bias", 0);
+ kernel = get_kernel("src/convolutional_layer.cl", "bias", "-D BLOCK=" STR(BLOCK));
init = 1;
}
return kernel;
@@ -410,7 +418,7 @@
axpy_ongpu(size, -layer.decay, layer.filters_cl, 1, layer.filter_updates_cl, 1);
axpy_ongpu(size, layer.learning_rate, layer.filter_updates_cl, 1, layer.filters_cl, 1);
scal_ongpu(size, layer.momentum, layer.filter_updates_cl, 1);
- pull_convolutional_layer(layer);
+ //pull_convolutional_layer(layer);
}
--
Gitblit v1.10.0