From d7286c273211ffeb1f56594f863d1ee9922be6d4 Mon Sep 17 00:00:00 2001
From: Joseph Redmon <pjreddie@gmail.com>
Date: Thu, 07 Nov 2013 00:09:41 +0000
Subject: [PATCH] Loading may or may not work. But probably.
---
src/network.c | 50 ++++++++++++
.gitignore | 1
src/maxpool_layer.h | 2
src/network.h | 6 +
Makefile | 2
src/connected_layer.c | 34 ++++----
src/connected_layer.h | 2
src/convolutional_layer.h | 5 +
src/convolutional_layer.c | 20 ++--
src/activations.h | 9 ++
src/activations.c | 11 ++
src/tests.c | 61 ++++++++++----
src/image.h | 1
src/maxpool_layer.c | 8 +-
14 files changed, 155 insertions(+), 57 deletions(-)
diff --git a/.gitignore b/.gitignore
index d15b2e8..9153fbb 100644
--- a/.gitignore
+++ b/.gitignore
@@ -4,6 +4,7 @@
images/
opencv/
convnet/
+decaf/
cnn
# OS Generated #
diff --git a/Makefile b/Makefile
index 4cddfd5..bf7cfd3 100644
--- a/Makefile
+++ b/Makefile
@@ -4,7 +4,7 @@
LDFLAGS=`pkg-config --libs opencv` -lm
VPATH=./src/
-OBJ=network.o image.o tests.o convolutional_layer.o connected_layer.o maxpool_layer.o activations.o
+OBJ=network.o image.o tests.o convolutional_layer.o connected_layer.o maxpool_layer.o activations.o list.o option_list.o parser.o utils.o
all: cnn
diff --git a/src/activations.c b/src/activations.c
index aef21cb..a128029 100644
--- a/src/activations.c
+++ b/src/activations.c
@@ -1,6 +1,17 @@
#include "activations.h"
#include <math.h>
+#include <stdio.h>
+#include <string.h>
+
+ACTIVATION get_activation(char *s)
+{
+ if (strcmp(s, "sigmoid")==0) return SIGMOID;
+ if (strcmp(s, "relu")==0) return RELU;
+ if (strcmp(s, "identity")==0) return IDENTITY;
+ fprintf(stderr, "Couldn't find activation function %s, going with ReLU\n", s);
+ return RELU;
+}
double identity_activation(double x)
{
diff --git a/src/activations.h b/src/activations.h
index 294cf28..09584cc 100644
--- a/src/activations.h
+++ b/src/activations.h
@@ -1,10 +1,17 @@
+#ifndef ACTIVATIONS_H
+#define ACTIVATIONS_H
+
typedef enum{
SIGMOID, RELU, IDENTITY
-}ACTIVATOR_TYPE;
+}ACTIVATION;
+ACTIVATION get_activation(char *s);
double relu_activation(double x);
double relu_gradient(double x);
double sigmoid_activation(double x);
double sigmoid_gradient(double x);
double identity_activation(double x);
double identity_gradient(double x);
+
+#endif
+
diff --git a/src/connected_layer.c b/src/connected_layer.c
index 11143b9..9fafc38 100644
--- a/src/connected_layer.c
+++ b/src/connected_layer.c
@@ -4,34 +4,34 @@
#include <stdlib.h>
#include <string.h>
-connected_layer make_connected_layer(int inputs, int outputs, ACTIVATOR_TYPE activator)
+connected_layer *make_connected_layer(int inputs, int outputs, ACTIVATION activator)
{
int i;
- connected_layer layer;
- layer.inputs = inputs;
- layer.outputs = outputs;
+ connected_layer *layer = calloc(1, sizeof(connected_layer));
+ layer->inputs = inputs;
+ layer->outputs = outputs;
- layer.output = calloc(outputs, sizeof(double*));
+ layer->output = calloc(outputs, sizeof(double*));
- layer.weight_updates = calloc(inputs*outputs, sizeof(double));
- layer.weights = calloc(inputs*outputs, sizeof(double));
+ layer->weight_updates = calloc(inputs*outputs, sizeof(double));
+ layer->weights = calloc(inputs*outputs, sizeof(double));
for(i = 0; i < inputs*outputs; ++i)
- layer.weights[i] = .5 - (double)rand()/RAND_MAX;
+ layer->weights[i] = .5 - (double)rand()/RAND_MAX;
- layer.bias_updates = calloc(outputs, sizeof(double));
- layer.biases = calloc(outputs, sizeof(double));
+ layer->bias_updates = calloc(outputs, sizeof(double));
+ layer->biases = calloc(outputs, sizeof(double));
for(i = 0; i < outputs; ++i)
- layer.biases[i] = (double)rand()/RAND_MAX;
+ layer->biases[i] = (double)rand()/RAND_MAX;
if(activator == SIGMOID){
- layer.activation = sigmoid_activation;
- layer.gradient = sigmoid_gradient;
+ layer->activation = sigmoid_activation;
+ layer->gradient = sigmoid_gradient;
}else if(activator == RELU){
- layer.activation = relu_activation;
- layer.gradient = relu_gradient;
+ layer->activation = relu_activation;
+ layer->gradient = relu_gradient;
}else if(activator == IDENTITY){
- layer.activation = identity_activation;
- layer.gradient = identity_gradient;
+ layer->activation = identity_activation;
+ layer->gradient = identity_gradient;
}
return layer;
diff --git a/src/connected_layer.h b/src/connected_layer.h
index 4f0e42c..eaea306 100644
--- a/src/connected_layer.h
+++ b/src/connected_layer.h
@@ -16,7 +16,7 @@
double (* gradient)();
} connected_layer;
-connected_layer make_connected_layer(int inputs, int outputs, ACTIVATOR_TYPE activator);
+connected_layer *make_connected_layer(int inputs, int outputs, ACTIVATION activator);
void run_connected_layer(double *input, connected_layer layer);
void learn_connected_layer(double *input, connected_layer layer);
diff --git a/src/convolutional_layer.c b/src/convolutional_layer.c
index 8053133..7478158 100644
--- a/src/convolutional_layer.c
+++ b/src/convolutional_layer.c
@@ -10,20 +10,20 @@
return (x>=0);
}
-convolutional_layer make_convolutional_layer(int h, int w, int c, int n, int size, int stride)
+convolutional_layer *make_convolutional_layer(int h, int w, int c, int n, int size, int stride)
{
int i;
- convolutional_layer layer;
- layer.n = n;
- layer.stride = stride;
- layer.kernels = calloc(n, sizeof(image));
- layer.kernel_updates = calloc(n, sizeof(image));
+ convolutional_layer *layer = calloc(1, sizeof(convolutional_layer));
+ layer->n = n;
+ layer->stride = stride;
+ layer->kernels = calloc(n, sizeof(image));
+ layer->kernel_updates = calloc(n, sizeof(image));
for(i = 0; i < n; ++i){
- layer.kernels[i] = make_random_kernel(size, c);
- layer.kernel_updates[i] = make_random_kernel(size, c);
+ layer->kernels[i] = make_random_kernel(size, c);
+ layer->kernel_updates[i] = make_random_kernel(size, c);
}
- layer.output = make_image((h-1)/stride+1, (w-1)/stride+1, n);
- layer.upsampled = make_image(h,w,n);
+ layer->output = make_image((h-1)/stride+1, (w-1)/stride+1, n);
+ layer->upsampled = make_image(h,w,n);
return layer;
}
diff --git a/src/convolutional_layer.h b/src/convolutional_layer.h
index 2428715..75be04b 100644
--- a/src/convolutional_layer.h
+++ b/src/convolutional_layer.h
@@ -12,9 +12,12 @@
image output;
} convolutional_layer;
-convolutional_layer make_convolutional_layer(int w, int h, int c, int n, int size, int stride);
+convolutional_layer *make_convolutional_layer(int h, int w, int c, int n, int size, int stride);
void run_convolutional_layer(const image input, const convolutional_layer layer);
void learn_convolutional_layer(image input, convolutional_layer layer);
+void update_convolutional_layer(convolutional_layer layer, double step);
+void backpropagate_convolutional_layer(image input, convolutional_layer layer);
+void backpropagate_convolutional_layer_convolve(image input, convolutional_layer layer);
#endif
diff --git a/src/image.h b/src/image.h
index 85306f0..e2fe8c0 100644
--- a/src/image.h
+++ b/src/image.h
@@ -14,6 +14,7 @@
void threshold_image(image p, double t);
void zero_image(image m);
void rotate_image(image m);
+void subtract_image(image a, image b);
void show_image(image p, char *name);
void show_image_layers(image p, char *name);
diff --git a/src/maxpool_layer.c b/src/maxpool_layer.c
index 38ac582..6f7d2a2 100644
--- a/src/maxpool_layer.c
+++ b/src/maxpool_layer.c
@@ -1,10 +1,10 @@
#include "maxpool_layer.h"
-maxpool_layer make_maxpool_layer(int h, int w, int c, int stride)
+maxpool_layer *make_maxpool_layer(int h, int w, int c, int stride)
{
- maxpool_layer layer;
- layer.stride = stride;
- layer.output = make_image((h-1)/stride+1, (w-1)/stride+1, c);
+ maxpool_layer *layer = calloc(1, sizeof(maxpool_layer));
+ layer->stride = stride;
+ layer->output = make_image((h-1)/stride+1, (w-1)/stride+1, c);
return layer;
}
diff --git a/src/maxpool_layer.h b/src/maxpool_layer.h
index 077bcfa..4d7726d 100644
--- a/src/maxpool_layer.h
+++ b/src/maxpool_layer.h
@@ -8,7 +8,7 @@
image output;
} maxpool_layer;
-maxpool_layer make_maxpool_layer(int h, int w, int c, int stride);
+maxpool_layer *make_maxpool_layer(int h, int w, int c, int stride);
void run_maxpool_layer(const image input, const maxpool_layer layer);
#endif
diff --git a/src/network.c b/src/network.c
index 0a74b63..53184d9 100644
--- a/src/network.c
+++ b/src/network.c
@@ -5,6 +5,15 @@
#include "convolutional_layer.h"
#include "maxpool_layer.h"
+network make_network(int n)
+{
+ network net;
+ net.n = n;
+ net.layers = calloc(net.n, sizeof(void *));
+ net.types = calloc(net.n, sizeof(LAYER_TYPE));
+ return net;
+}
+
void run_network(image input, network net)
{
int i;
@@ -84,9 +93,9 @@
}
}
-double *get_network_output(network net)
+
+double *get_network_output_layer(network net, int i)
{
- int i = net.n-1;
if(net.types[i] == CONVOLUTIONAL){
convolutional_layer layer = *(convolutional_layer *)net.layers[i];
return layer.output.data;
@@ -101,6 +110,43 @@
}
return 0;
}
+
+int get_network_output_size_layer(network net, int i)
+{
+ if(net.types[i] == CONVOLUTIONAL){
+ convolutional_layer layer = *(convolutional_layer *)net.layers[i];
+ return layer.output.h*layer.output.w*layer.output.c;
+ }
+ else if(net.types[i] == MAXPOOL){
+ maxpool_layer layer = *(maxpool_layer *)net.layers[i];
+ return layer.output.h*layer.output.w*layer.output.c;
+ }
+ else if(net.types[i] == CONNECTED){
+ connected_layer layer = *(connected_layer *)net.layers[i];
+ return layer.outputs;
+ }
+ return 0;
+}
+
+double *get_network_output(network net)
+{
+ int i = net.n-1;
+ return get_network_output_layer(net, i);
+}
+
+image get_network_image_layer(network net, int i)
+{
+ if(net.types[i] == CONVOLUTIONAL){
+ convolutional_layer layer = *(convolutional_layer *)net.layers[i];
+ return layer.output;
+ }
+ else if(net.types[i] == MAXPOOL){
+ maxpool_layer layer = *(maxpool_layer *)net.layers[i];
+ return layer.output;
+ }
+ return make_image(0,0,0);
+}
+
image get_network_image(network net)
{
int i;
diff --git a/src/network.h b/src/network.h
index 2fb9225..ad2b1dc 100644
--- a/src/network.h
+++ b/src/network.h
@@ -16,11 +16,15 @@
LAYER_TYPE *types;
} network;
+network make_network(int n);
void run_network(image input, network net);
-double *get_network_output(network net);
void learn_network(image input, network net);
void update_network(network net, double step);
+double *get_network_output(network net);
+double *get_network_output_layer(network net, int i);
+int get_network_output_size_layer(network net, int i);
image get_network_image(network net);
+image get_network_image_layer(network net, int i);
#endif
diff --git a/src/tests.c b/src/tests.c
index f2b50dc..0e639be 100644
--- a/src/tests.c
+++ b/src/tests.c
@@ -3,6 +3,7 @@
#include "maxpool_layer.h"
#include "network.h"
#include "image.h"
+#include "parser.h"
#include <time.h>
#include <stdlib.h>
@@ -39,7 +40,7 @@
int n = 3;
int stride = 1;
int size = 3;
- convolutional_layer layer = make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
+ convolutional_layer layer = *make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
char buff[256];
for(i = 0; i < n; ++i) {
sprintf(buff, "Kernel %d", i);
@@ -47,7 +48,7 @@
}
run_convolutional_layer(dog, layer);
- maxpool_layer mlayer = make_maxpool_layer(layer.output.h, layer.output.w, layer.output.c, 2);
+ maxpool_layer mlayer = *make_maxpool_layer(layer.output.h, layer.output.w, layer.output.c, 2);
run_maxpool_layer(layer.output,mlayer);
show_image_layers(mlayer.output, "Test Maxpool Layer");
@@ -112,25 +113,25 @@
int n = 48;
int stride = 4;
int size = 11;
- convolutional_layer cl = make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
- maxpool_layer ml = make_maxpool_layer(cl.output.h, cl.output.w, cl.output.c, 2);
+ convolutional_layer cl = *make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
+ maxpool_layer ml = *make_maxpool_layer(cl.output.h, cl.output.w, cl.output.c, 2);
n = 128;
size = 5;
stride = 1;
- convolutional_layer cl2 = make_convolutional_layer(ml.output.h, ml.output.w, ml.output.c, n, size, stride);
- maxpool_layer ml2 = make_maxpool_layer(cl2.output.h, cl2.output.w, cl2.output.c, 2);
+ convolutional_layer cl2 = *make_convolutional_layer(ml.output.h, ml.output.w, ml.output.c, n, size, stride);
+ maxpool_layer ml2 = *make_maxpool_layer(cl2.output.h, cl2.output.w, cl2.output.c, 2);
n = 192;
size = 3;
- convolutional_layer cl3 = make_convolutional_layer(ml2.output.h, ml2.output.w, ml2.output.c, n, size, stride);
- convolutional_layer cl4 = make_convolutional_layer(cl3.output.h, cl3.output.w, cl3.output.c, n, size, stride);
+ convolutional_layer cl3 = *make_convolutional_layer(ml2.output.h, ml2.output.w, ml2.output.c, n, size, stride);
+ convolutional_layer cl4 = *make_convolutional_layer(cl3.output.h, cl3.output.w, cl3.output.c, n, size, stride);
n = 128;
- convolutional_layer cl5 = make_convolutional_layer(cl4.output.h, cl4.output.w, cl4.output.c, n, size, stride);
- maxpool_layer ml3 = make_maxpool_layer(cl5.output.h, cl5.output.w, cl5.output.c, 4);
- connected_layer nl = make_connected_layer(ml3.output.h*ml3.output.w*ml3.output.c, 4096, RELU);
- connected_layer nl2 = make_connected_layer(4096, 4096, RELU);
- connected_layer nl3 = make_connected_layer(4096, 1000, RELU);
+ convolutional_layer cl5 = *make_convolutional_layer(cl4.output.h, cl4.output.w, cl4.output.c, n, size, stride);
+ maxpool_layer ml3 = *make_maxpool_layer(cl5.output.h, cl5.output.w, cl5.output.c, 4);
+ connected_layer nl = *make_connected_layer(ml3.output.h*ml3.output.w*ml3.output.c, 4096, RELU);
+ connected_layer nl2 = *make_connected_layer(4096, 4096, RELU);
+ connected_layer nl3 = *make_connected_layer(4096, 1000, RELU);
net.layers[0] = &cl;
net.layers[1] = &ml;
@@ -164,7 +165,7 @@
image dog = load_image("dog.jpg");
show_image(dog, "Test Backpropagate Input");
image dog_copy = copy_image(dog);
- convolutional_layer cl = make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
+ convolutional_layer cl = *make_convolutional_layer(dog.h, dog.w, dog.c, n, size, stride);
run_convolutional_layer(dog, cl);
show_image(cl.output, "Test Backpropagate Output");
int i;
@@ -196,9 +197,9 @@
net.types[1] = CONNECTED;
net.types[2] = CONNECTED;
- connected_layer nl = make_connected_layer(1, 20, RELU);
- connected_layer nl2 = make_connected_layer(20, 20, RELU);
- connected_layer nl3 = make_connected_layer(20, 1, RELU);
+ connected_layer nl = *make_connected_layer(1, 20, RELU);
+ connected_layer nl2 = *make_connected_layer(20, 20, RELU);
+ connected_layer nl3 = *make_connected_layer(20, 1, RELU);
net.layers[0] = &nl;
net.layers[1] = &nl2;
@@ -225,10 +226,34 @@
}
+void test_parser()
+{
+ network net = parse_network_cfg("test.cfg");
+ image t = make_image(1,1,1);
+ int count = 0;
+
+ double avgerr = 0;
+ while(1){
+ double v = ((double)rand()/RAND_MAX);
+ double truth = v*v;
+ set_pixel(t,0,0,0,v);
+ run_network(t, net);
+ double *out = get_network_output(net);
+ double err = pow((out[0]-truth),2.);
+ avgerr = .99 * avgerr + .01 * err;
+ //if(++count % 100000 == 0) printf("%f\n", avgerr);
+ if(++count % 100000 == 0) printf("%f %f :%f AVG %f \n", truth, out[0], err, avgerr);
+ out[0] = truth - out[0];
+ learn_network(t, net);
+ update_network(net, .001);
+ }
+}
+
int main()
{
+ test_parser();
//test_backpropagate();
- test_ann();
+ //test_ann();
//test_convolve();
//test_upsample();
//test_rotate();
--
Gitblit v1.10.0