mirror/darknet
1
0
Fork 0
mirror of https://github.com/pjreddie/darknet.git synced 2023-08-10 21:13:14 +03:00
Code Issues Projects Releases Wiki Activity

Refactored connected to use blas

Browse Source
This commit is contained in:
Joseph Redmon 2014-10-13 22:31:48 -07:00
parent 787d534560
commit 7756cccb79
5 changed files with 24 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
Makefile
View File

@ -1,5 +1,5 @@
CC=gcc
GPU=1
GPU=0
COMMON=-Wall -Wfatal-errors `pkg-config --cflags opencv` -I/usr/local/cuda/include/
ifeq ($(GPU), 1)
COMMON+=-DGPU
@ -7,6 +7,7 @@ else
endif
UNAME = $(shell uname)
OPTS=-Ofast -flto
OPTS=-Ofast -flto
ifeq ($(UNAME), Darwin)
COMMON+= -isystem /usr/local/Cellar/opencv/2.4.6.1/include/opencv -isystem /usr/local/Cellar/opencv/2.4.6.1/include
ifeq ($(GPU), 1)

4
src/cnn.c
View File

@ -916,8 +916,8 @@ int main(int argc, char *argv[])
//test_ensemble();
//test_nist_single();
//test_nist();
//train_nist();
test_convolutional_layer();
train_nist();
//test_convolutional_layer();
//test_col2im();
//test_cifar10();
//train_cifar10();

23
src/connected_layer.c
View File

@ -26,7 +26,6 @@ connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVA
layer->weight_updates = calloc(inputs*outputs, sizeof(float));
//layer->weight_adapt = calloc(inputs*outputs, sizeof(float));
layer->weight_momentum = calloc(inputs*outputs, sizeof(float));
layer->weights = calloc(inputs*outputs, sizeof(float));
float scale = 1./inputs;
scale = .05;
@ -35,7 +34,6 @@ connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVA
layer->bias_updates = calloc(outputs, sizeof(float));
//layer->bias_adapt = calloc(outputs, sizeof(float));
layer->bias_momentum = calloc(outputs, sizeof(float));
layer->biases = calloc(outputs, sizeof(float));
for(i = 0; i < outputs; ++i){
//layer->biases[i] = rand_normal()*scale + scale;
@ -50,24 +48,19 @@ connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVA
void update_connected_layer(connected_layer layer)
{
int i;
for(i = 0; i < layer.outputs; ++i){
layer.bias_momentum[i] = layer.learning_rate*(layer.bias_updates[i]) + layer.momentum*layer.bias_momentum[i];
layer.biases[i] += layer.bias_momentum[i];
}
for(i = 0; i < layer.outputs*layer.inputs; ++i){
layer.weight_momentum[i] = layer.learning_rate*(layer.weight_updates[i] - layer.decay*layer.weights[i]) + layer.momentum*layer.weight_momentum[i];
layer.weights[i] += layer.weight_momentum[i];
}
memset(layer.bias_updates, 0, layer.outputs*sizeof(float));
memset(layer.weight_updates, 0, layer.outputs*layer.inputs*sizeof(float));
axpy_cpu(layer.outputs, layer.learning_rate, layer.bias_updates, 1, layer.biases, 1);
scal_cpu(layer.outputs, layer.momentum, layer.bias_updates, 1);
scal_cpu(layer.inputs*layer.outputs, 1.-layer.learning_rate*layer.decay, layer.weights, 1);
axpy_cpu(layer.inputs*layer.outputs, layer.learning_rate, layer.weight_updates, 1, layer.weights, 1);
scal_cpu(layer.inputs*layer.outputs, layer.momentum, layer.weight_updates, 1);
}
void forward_connected_layer(connected_layer layer, float *input)
{
int i;
for(i = 0; i < layer.batch; ++i){
memcpy(layer.output+i*layer.outputs, layer.biases, layer.outputs*sizeof(float));
copy_cpu(layer.outputs, layer.biases, 1, layer.output + i*layer.outputs, 1);
}
int m = layer.batch;
int k = layer.inputs;
@ -82,8 +75,8 @@ void forward_connected_layer(connected_layer layer, float *input)
void backward_connected_layer(connected_layer layer, float *input, float *delta)
{
int i;
gradient_array(layer.output, layer.outputs*layer.batch, layer.activation, layer.delta);
for(i = 0; i < layer.outputs*layer.batch; ++i){
layer.delta[i] *= gradient(layer.output[i], layer.activation);
layer.bias_updates[i%layer.outputs] += layer.delta[i];
}
int m = layer.inputs;

3
src/connected_layer.h
View File

@ -21,9 +21,6 @@ typedef struct{
float *weight_adapt;
float *bias_adapt;
float *weight_momentum;
float *bias_momentum;
float *output;
float *delta;

12
src/network.c
View File

@ -229,6 +229,8 @@ float *get_network_output_layer(network net, int i)
return layer.output;
} else if(net.types[i] == DROPOUT){
return get_network_output_layer(net, i-1);
} else if(net.types[i] == FREEWEIGHT){
return get_network_output_layer(net, i-1);
} else if(net.types[i] == CONNECTED){
connected_layer layer = *(connected_layer *)net.layers[i];
return layer.output;
@ -258,6 +260,8 @@ float *get_network_delta_layer(network net, int i)
return layer.delta;
} else if(net.types[i] == DROPOUT){
return get_network_delta_layer(net, i-1);
} else if(net.types[i] == FREEWEIGHT){
return get_network_delta_layer(net, i-1);
} else if(net.types[i] == CONNECTED){
connected_layer layer = *(connected_layer *)net.layers[i];
return layer.delta;
@ -424,6 +428,10 @@ int get_network_input_size_layer(network net, int i)
dropout_layer layer = *(dropout_layer *) net.layers[i];
return layer.inputs;
}
else if(net.types[i] == FREEWEIGHT){
freeweight_layer layer = *(freeweight_layer *) net.layers[i];
return layer.inputs;
}
else if(net.types[i] == SOFTMAX){
softmax_layer layer = *(softmax_layer *)net.layers[i];
return layer.inputs;
@ -451,6 +459,10 @@ int get_network_output_size_layer(network net, int i)
dropout_layer layer = *(dropout_layer *) net.layers[i];
return layer.inputs;
}
else if(net.types[i] == FREEWEIGHT){
freeweight_layer layer = *(freeweight_layer *) net.layers[i];
return layer.inputs;
}
else if(net.types[i] == SOFTMAX){
softmax_layer layer = *(softmax_layer *)net.layers[i];
return layer.inputs;