So there WAS this huge bug. Gone now

Makefile

@@ -1,18 +1,19 @@
 CC=gcc
-GPU=1
+GPU=0
 COMMON=-Wall -Werror -Wfatal-errors `pkg-config --cflags opencv` -I/usr/local/cuda/include/
 ifeq ($(GPU), 1) 
 COMMON+=-DGPU
 else
 endif
 UNAME = $(shell uname)
-OPTS=-O3 -flto
+OPTS=-Ofast -flto
 ifeq ($(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 ($(GPU), 1)
 LDFLAGS= -framework OpenCL
 endif
 else
+OPTS+= -march=native
 ifeq ($(GPU), 1)
 LDFLAGS= -lOpenCL
 endif

src/activations.c

@@ -2,6 +2,7 @@
 
 #include <math.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
 
 char *get_activation_string(ACTIVATION a)
@@ -40,27 +41,29 @@ float relu_activate(float x){return x*(x>0);}
 float ramp_activate(float x){return x*(x>0)+.1*x;}
 float tanh_activate(float x){return (exp(2*x)-1)/(exp(2*x)+1);}
 
-float activate(float x, ACTIVATION a){
+float activate(float x, ACTIVATION a, float dropout)
+{
+    if((float)rand()/RAND_MAX < dropout) return 0;
     switch(a){
         case LINEAR:
-            return linear_activate(x);
+            return linear_activate(x)/(1-dropout);
         case SIGMOID:
-            return sigmoid_activate(x);
+            return sigmoid_activate(x)/(1-dropout);
         case RELU:
-            return relu_activate(x);
+            return relu_activate(x)/(1-dropout);
         case RAMP:
-            return ramp_activate(x);
+            return ramp_activate(x)/(1-dropout);
         case TANH:
-            return tanh_activate(x);
+            return tanh_activate(x)/(1-dropout);
     }
     return 0;
 }
 
-void activate_array(float *x, const int n, const ACTIVATION a)
+void activate_array(float *x, const int n, const ACTIVATION a, float dropout)
 {
     int i;
     for(i = 0; i < n; ++i){
-        x[i] = activate(x[i], a);
+        x[i] = activate(x[i], a, dropout);
     }
 }
 
@@ -89,3 +92,40 @@ void gradient_array(const float *x, const int n, const ACTIVATION a, float *delt
     }
 } 
 
+#ifdef GPU
+
+#include "opencl.h"
+#include <math.h>
+
+cl_kernel get_activation_kernel()
+{
+    static int init = 0;
+    static cl_kernel kernel;
+    if(!init){
+        kernel = get_kernel("src/activations.cl", "activate_array", 0);
+        init = 1;
+    }
+    return kernel;
+}
+
+
+void activate_array_ongpu(cl_mem x, int n, ACTIVATION a, float dropout) 
+{
+    cl_setup();
+    cl_kernel kernel = get_activation_kernel();
+    cl_command_queue queue = cl.queue;
+
+    cl_uint i = 0;
+    cl.error = clSetKernelArg(kernel, i++, sizeof(x), (void*) &x);
+    cl.error = clSetKernelArg(kernel, i++, sizeof(n), (void*) &n);
+    cl.error = clSetKernelArg(kernel, i++, sizeof(a), (void*) &a);
+    cl.error = clSetKernelArg(kernel, i++, sizeof(dropout), 
+        (void*) &dropout);
+    check_error(cl);
+
+    size_t gsize = n;
+
+    clEnqueueNDRangeKernel(queue, kernel, 1, 0, &gsize, 0, 0, 0, 0);
+    check_error(cl);
+}
+#endif

src/activations.cl

@@ -0,0 +1,28 @@
+typedef enum{
+    SIGMOID, RELU, LINEAR, RAMP, TANH
+}ACTIVATION;
+
+float activate(float x, ACTIVATION a, float dropout)
+{
+    //if((float)rand()/RAND_MAX < dropout) return 0;
+    switch(a){
+        case LINEAR:
+            return linear_activate(x)/(1-dropout);
+        case SIGMOID:
+            return sigmoid_activate(x)/(1-dropout);
+        case RELU:
+            return relu_activate(x)/(1-dropout);
+        case RAMP:
+            return ramp_activate(x)/(1-dropout);
+        case TANH:
+            return tanh_activate(x)/(1-dropout);
+    }
+    return 0;
+}
+
+__kernel void activate_array(__global float *x,
+    const int n, const ACTIVATION a, const float dropout)
+{
+    int i = get_global_id(0);
+    x[i] = activate(x[i], a, dropout);
+}

src/activations.h

@@ -1,3 +1,4 @@
+#include "opencl.h"
 #ifndef ACTIVATIONS_H
 #define ACTIVATIONS_H
 
@@ -8,10 +9,13 @@ typedef enum{
 ACTIVATION get_activation(char *s);
 
 char *get_activation_string(ACTIVATION a);
-float activate(float x, ACTIVATION a);
+float activate(float x, ACTIVATION a, float dropout);
 float gradient(float x, ACTIVATION a);
 void gradient_array(const float *x, const int n, const ACTIVATION a, float *delta);
-void activate_array(float *x, const int n, const ACTIVATION a);
+void activate_array(float *x, const int n, const ACTIVATION a, float dropout);
+#ifdef GPU
+void activate_array_ongpu(cl_mem x, int n, ACTIVATION a, float dropout);
+#endif
 
 #endif
 

src/axpy.c

@@ -0,0 +1,14 @@
+#include "mini_blas.h"
+
+void axpy_cpu(int N, float ALPHA, float *X, int INCX, float *Y, int INCY)
+{
+    int i;
+    for(i = 0; i < N; ++i) Y[i*INCY] += ALPHA*X[i*INCX];
+}
+
+void scal_cpu(int N, float ALPHA, float *X, int INCX)
+{
+    int i;
+    for(i = 0; i < N; ++i) X[i*INCX] *= ALPHA;
+}
+

src/connected_layer.c

@@ -7,7 +7,7 @@
 #include <stdlib.h>
 #include <string.h>
 
-connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVATION activation)
+connected_layer *make_connected_layer(int batch, int inputs, int outputs, float dropout, ACTIVATION activation)
 {
     fprintf(stderr, "Connected Layer: %d inputs, %d outputs\n", inputs, outputs);
     int i;
@@ -15,6 +15,7 @@ connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVA
     layer->inputs = inputs;
     layer->outputs = outputs;
     layer->batch=batch;
+    layer->dropout = dropout;
 
     layer->output = calloc(batch*outputs, sizeof(float*));
     layer->delta = calloc(batch*outputs, sizeof(float*));
@@ -54,9 +55,9 @@ void update_connected_layer(connected_layer layer, float step, float momentum, f
     memset(layer.weight_updates, 0, layer.outputs*layer.inputs*sizeof(float));
 }
 
-void forward_connected_layer(connected_layer layer, float *input)
+void forward_connected_layer(connected_layer layer, float *input, int train)
 {
-    int i;
+    if(!train) layer.dropout = 0;
     memcpy(layer.output, layer.biases, layer.outputs*sizeof(float));
     int m = layer.batch;
     int k = layer.inputs;
@@ -65,17 +66,15 @@ void forward_connected_layer(connected_layer layer, float *input)
     float *b = layer.weights;
     float *c = layer.output;
     gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
-    for(i = 0; i < layer.outputs*layer.batch; ++i){
-        layer.output[i] = activate(layer.output[i], layer.activation);
-    }
+    activate_array(layer.output, layer.outputs*layer.batch, layer.activation, layer.dropout);
 }
 
-void learn_connected_layer(connected_layer layer, float *input)
+void backward_connected_layer(connected_layer layer, float *input, float *delta)
 {
     int i;
     for(i = 0; i < layer.outputs*layer.batch; ++i){
         layer.delta[i] *= gradient(layer.output[i], layer.activation);
-        layer.bias_updates[i%layer.batch] += layer.delta[i]/layer.batch;
+        layer.bias_updates[i%layer.batch] += layer.delta[i];
     }
     int m = layer.inputs;
     int k = layer.batch;
@@ -84,18 +83,15 @@ void learn_connected_layer(connected_layer layer, float *input)
     float *b = layer.delta;
     float *c = layer.weight_updates;
     gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
-}
-
-void backward_connected_layer(connected_layer layer, float *input, float *delta)
-{
-    int m = layer.inputs;
-    int k = layer.outputs;
-    int n = layer.batch;
-
-    float *a = layer.weights;
-    float *b = layer.delta;
-    float *c = delta;
-
-    gemm(0,0,m,n,k,1,a,k,b,n,0,c,n);
+
+    m = layer.inputs;
+    k = layer.outputs;
+    n = layer.batch;
+
+    a = layer.weights;
+    b = layer.delta;
+    c = delta;
+
+    if(c) gemm(0,0,m,n,k,1,a,k,b,n,0,c,n);
 }
 

src/connected_layer.h

@@ -22,15 +22,16 @@ typedef struct{
     float *output;
     float *delta;
     
+    float dropout;
+
     ACTIVATION activation;
 
 } connected_layer;
 
-connected_layer *make_connected_layer(int batch, int inputs, int outputs, ACTIVATION activation);
+connected_layer *make_connected_layer(int batch, int inputs, int outputs, float dropout, ACTIVATION activation);
 
-void forward_connected_layer(connected_layer layer, float *input);
+void forward_connected_layer(connected_layer layer, float *input, int train);
 void backward_connected_layer(connected_layer layer, float *input, float *delta);
-void learn_connected_layer(connected_layer layer, float *input);
 void update_connected_layer(connected_layer layer, float step, float momentum, float decay);
 
 

src/convolutional_layer.c

@@ -55,7 +55,7 @@ convolutional_layer *make_convolutional_layer(int batch, int h, int w, int c, in
     for(i = 0; i < c*n*size*size; ++i) layer->filters[i] = scale*(rand_uniform());
     for(i = 0; i < n; ++i){
         //layer->biases[i] = rand_normal()*scale + scale;
-        layer->biases[i] = 0;
+        layer->biases[i] = .5;
     }
     int out_h = convolutional_out_height(*layer);
     int out_w = convolutional_out_width(*layer);
@@ -63,6 +63,8 @@ convolutional_layer *make_convolutional_layer(int batch, int h, int w, int c, in
     layer->col_image = calloc(layer->batch*out_h*out_w*size*size*c, sizeof(float));
     layer->output = calloc(layer->batch*out_h * out_w * n, sizeof(float));
     layer->delta  = calloc(layer->batch*out_h * out_w * n, sizeof(float));
+    #ifdef GPU
+    #endif
     layer->activation = activation;
 
     fprintf(stderr, "Convolutional Layer: %d x %d x %d image, %d filters -> %d x %d x %d image\n", h,w,c,n, out_h, out_w, n);
@@ -87,42 +89,70 @@ void resize_convolutional_layer(convolutional_layer *layer, int h, int w, int c)
                                 layer->batch*out_h * out_w * layer->n*sizeof(float));
 }
 
+void bias_output(const convolutional_layer layer)
+{
+    int i,j;
+    int out_h = convolutional_out_height(layer);
+    int out_w = convolutional_out_width(layer);
+    for(i = 0; i < layer.n; ++i){
+        for(j = 0; j < out_h*out_w; ++j){
+            layer.output[i*out_h*out_w + j] = layer.biases[i];
+        }
+    }
+}
+
 void forward_convolutional_layer(const convolutional_layer layer, float *in)
 {
-    int i;
+    int out_h = convolutional_out_height(layer);
+    int out_w = convolutional_out_width(layer);
+
+    int m = layer.n;
+    int k = layer.size*layer.size*layer.c;
+    int n = out_h*out_w*layer.batch;
+
+    float *a = layer.filters;
+    float *b = layer.col_image;
+    float *c = layer.output;
+    im2col_cpu(in,layer.batch, layer.c, layer.h, layer.w, 
+        layer.size, layer.stride, b);
+    bias_output(layer);
+    gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
+    activate_array(layer.output, m*n, layer.activation, 0.);
+}
+
+#ifdef GPU
+void forward_convolutional_layer_gpu(convolutional_layer layer, cl_mem in)
+{
     int m = layer.n;
     int k = layer.size*layer.size*layer.c;
     int n = convolutional_out_height(layer)*
         convolutional_out_width(layer)*
         layer.batch;
 
-    float *a = layer.filters;
-    float *b = layer.col_image;
-    float *c = layer.output;
-    for(i = 0; i < layer.batch; ++i){
-        im2col_gpu(in+i*(n/layer.batch),  layer.c,  layer.h,  layer.w,  layer.size,  layer.stride, b+i*(n/layer.batch));
-    }
-    gemm(0,0,m,n,k,1,a,k,b,n,0,c,n);
-    activate_array(layer.output, m*n, layer.activation);
+    cl_write_array(layer.filters_cl, layer.filters, m*k);
+    cl_mem a = layer.filters_cl;
+    cl_mem b = layer.col_image_cl;
+    cl_mem c = layer.output_cl;
+    im2col_ongpu(in, layer.batch, layer.c,  layer.h,  layer.w,  layer.size,  layer.stride, b);
+    gemm_ongpu(0,0,m,n,k,1,a,k,b,n,0,c,n);
+    activate_array_ongpu(layer.output_cl, m*n, layer.activation, 0.);
+    cl_read_array(layer.output_cl, layer.output, m*n);
 }
+#endif
 
 void learn_bias_convolutional_layer(convolutional_layer layer)
 {
-    int i,j,b;
+    int i,b;
     int size = convolutional_out_height(layer)
         *convolutional_out_width(layer);
     for(b = 0; b < layer.batch; ++b){
         for(i = 0; i < layer.n; ++i){
-            float sum = 0;
-            for(j = 0; j < size; ++j){
-                sum += layer.delta[j+size*(i+b*layer.n)];
-            }
-            layer.bias_updates[i] += sum/size;
+            layer.bias_updates[i] += mean_array(layer.delta+size*(i+b*layer.n), size);
         }
     }
 }
 
-void learn_convolutional_layer(convolutional_layer layer)
+void backward_convolutional_layer(convolutional_layer layer, float *delta)
 {
     int m = layer.n;
     int n = layer.size*layer.size*layer.c;
@@ -137,20 +167,18 @@ void learn_convolutional_layer(convolutional_layer layer)
     float *c = layer.filter_updates;
 
     gemm(0,1,m,n,k,1,a,k,b,k,1,c,n);
-}
 
-void backward_convolutional_layer(convolutional_layer layer, float *delta)
-{
+    if(delta){
         int i;
-    int m = layer.size*layer.size*layer.c;
-    int k = layer.n;
-    int n = convolutional_out_height(layer)*
+        m = layer.size*layer.size*layer.c;
+        k = layer.n;
+        n = convolutional_out_height(layer)*
             convolutional_out_width(layer)*
             layer.batch;
 
-    float *a = layer.filters;
-    float *b = layer.delta;
-    float *c = layer.col_image;
+        a = layer.filters;
+        b = layer.delta;
+        c = layer.col_image;
 
         gemm(1,0,m,n,k,1,a,m,b,n,0,c,n);
 
@@ -159,6 +187,7 @@ void backward_convolutional_layer(convolutional_layer layer, float *delta)
             col2im_cpu(c+i*n/layer.batch,  layer.c,  layer.h,  layer.w,  layer.size,  layer.stride, delta+i*n/layer.batch);
         }
     }
+}
 
 void update_convolutional_layer(convolutional_layer layer, float step, float momentum, float decay)
 {
@@ -171,32 +200,6 @@ void update_convolutional_layer(convolutional_layer layer, float step, float mom
     scal_cpu(size, momentum, layer.filter_updates, 1);
 }
 
-void test_convolutional_layer()
-{
-    convolutional_layer l = *make_convolutional_layer(1,4,4,1,1,3,1,LINEAR);
-    float input[] =    {1,2,3,4,
-                        5,6,7,8,
-                        9,10,11,12,
-                        13,14,15,16};
-    float filter[] =   {.5, 0, .3,
-                        0  , 1,  0,
-                        .2 , 0,  1};
-    float delta[] =    {1, 2,
-                        3,  4};
-    float in_delta[] = {.5,1,.3,.6,
-                        5,6,7,8,
-                        9,10,11,12,
-                        13,14,15,16};
-    l.filters = filter;
-    forward_convolutional_layer(l, input);
-    l.delta = delta;
-    learn_convolutional_layer(l);
-    image filter_updates = float_to_image(3,3,1,l.filter_updates);
-    print_image(filter_updates);
-    printf("Delta:\n");
-    backward_convolutional_layer(l, in_delta);
-    pm(4,4,in_delta);
-}
 
 image get_convolutional_filter(convolutional_layer layer, int i)
 {

src/convolutional_layer.h

@@ -1,6 +1,10 @@
 #ifndef CONVOLUTIONAL_LAYER_H
 #define CONVOLUTIONAL_LAYER_H
 
+#ifdef GPU
+#include "opencl.h"
+#endif
+
 #include "image.h"
 #include "activations.h"
 
@@ -22,13 +26,30 @@ typedef struct {
     float *delta;
     float *output;
 
+    #ifdef GPU
+    cl_mem filters_cl;
+    cl_mem filter_updates_cl;
+    cl_mem filter_momentum_cl;
+
+    cl_mem biases_cl;
+    cl_mem bias_updates_cl;
+    cl_mem bias_momentum_cl;
+
+    cl_mem col_image_cl;
+    cl_mem delta_cl;
+    cl_mem output_cl;
+    #endif
+
     ACTIVATION activation;
 } convolutional_layer;
 
+#ifdef GPU
+void forward_convolutional_layer_gpu(convolutional_layer layer, cl_mem in);
+#endif
+
 convolutional_layer *make_convolutional_layer(int batch, int h, int w, int c, int n, int size, int stride, ACTIVATION activation);
 void resize_convolutional_layer(convolutional_layer *layer, int h, int w, int c);
 void forward_convolutional_layer(const convolutional_layer layer, float *in);
-void learn_convolutional_layer(convolutional_layer layer);
 void update_convolutional_layer(convolutional_layer layer, float step, float momentum, float decay);
 image *visualize_convolutional_layer(convolutional_layer layer, char *window, image *prev_filters);
 

src/gemm.c

@@ -0,0 +1,283 @@
+#include "mini_blas.h"
+
+void gemm(int TA, int TB, int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float BETA,
+        float *C, int ldc)
+{
+#ifdef GPU
+    gemm_gpu( TA,  TB,  M, N, K, ALPHA,A,lda, B, ldb,BETA,C,ldc);
+#else
+    gemm_cpu( TA,  TB,  M, N, K, ALPHA,A,lda, B, ldb,BETA,C,ldc);
+#endif
+}
+
+void gemm_nn(int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float *C, int ldc)
+{
+    int i,j,k;
+    for(i = 0; i < M; ++i){
+        for(k = 0; k < K; ++k){
+            register float A_PART = ALPHA*A[i*lda+k];
+            for(j = 0; j < N; ++j){
+                C[i*ldc+j] += A_PART*B[k*ldb+j];
+            }
+        }
+    }
+}
+
+void gemm_nt(int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float *C, int ldc)
+{
+    int i,j,k;
+    for(i = 0; i < M; ++i){
+        for(j = 0; j < N; ++j){
+            register float sum = 0;
+            for(k = 0; k < K; ++k){
+                sum += ALPHA*A[i*lda+k]*B[k+j*ldb];
+            }
+            C[i*ldc+j] += sum;
+        }
+    }
+}
+
+void gemm_tn(int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float *C, int ldc)
+{
+    int i,j,k;
+    for(i = 0; i < M; ++i){
+        for(k = 0; k < K; ++k){
+            register float A_PART = ALPHA*A[k*lda+i];
+            for(j = 0; j < N; ++j){
+                C[i*ldc+j] += A_PART*B[k*ldb+j];
+            }
+        }
+    }
+}
+void gemm_tt(int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float *C, int ldc)
+{
+    int i,j,k;
+    for(i = 0; i < M; ++i){
+        for(j = 0; j < N; ++j){
+            for(k = 0; k < K; ++k){
+                C[i*ldc+j] += ALPHA*A[i+k*lda]*B[k+j*ldb];
+            }
+        }
+    }
+}
+
+
+void gemm_cpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float BETA,
+        float *C, int ldc)
+{
+    int i, j;
+    for(i = 0; i < M; ++i){
+        for(j = 0; j < N; ++j){
+            C[i*ldc + j] *= BETA;
+        }
+    }
+    if(!TA && !TB)
+        gemm_nn(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
+    else if(TA && !TB)
+        gemm_tn(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
+    else if(!TA && TB)
+        gemm_nt(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
+    else
+        gemm_tt(M, N, K, ALPHA,A,lda, B, ldb,C,ldc);
+}
+
+#ifdef GPU
+
+#include "opencl.h"
+#include <math.h>
+
+#define STR_HELPER(x) #x
+#define STR(x) STR_HELPER(x)
+
+#define BLOCK 8
+
+cl_kernel get_gemm_kernel()
+{
+    static int init = 0;
+    static cl_kernel gemm_kernel;
+    if(!init){
+        gemm_kernel = get_kernel("src/gemm.cl", "gemm", "-D BLOCK=" STR(BLOCK) );
+        init = 1;
+    }
+    return gemm_kernel;
+}
+
+void gemm_ongpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        cl_mem A_gpu, int lda, 
+        cl_mem B_gpu, int ldb,
+        float BETA,
+        cl_mem C_gpu, int ldc)
+{
+    cl_setup();
+    cl_kernel gemm_kernel = get_gemm_kernel();
+    cl_command_queue queue = cl.queue;
+
+    cl_uint i = 0;
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(TA), (void*) &TA);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(TB), (void*) &TB);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(M), (void*) &M);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(N), (void*) &N);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(K), (void*) &K);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(ALPHA), (void*) &ALPHA);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(A_gpu), (void*) &A_gpu);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(lda), (void*) &lda);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(B_gpu), (void*) &B_gpu);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(ldb), (void*) &ldb);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(BETA), (void*) &BETA);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(C_gpu), (void*) &C_gpu);
+    cl.error = clSetKernelArg(gemm_kernel, i++, sizeof(ldc), (void*) &ldc);
+    check_error(cl);
+
+    const size_t global_size[] = {ceil((float)M/BLOCK)*BLOCK, ceil((float)N/BLOCK)*BLOCK};
+    const size_t local_size[] = {BLOCK, BLOCK};
+
+    clEnqueueNDRangeKernel(queue, gemm_kernel, 2, 0, global_size, local_size, 0, 0, 0);
+    check_error(cl);
+}
+
+
+void gemm_gpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        float *A, int lda, 
+        float *B, int ldb,
+        float BETA,
+        float *C, int ldc)
+{
+    cl_setup();
+    cl_context context = cl.context;
+    cl_command_queue queue = cl.queue;
+
+    size_t size = sizeof(float)*(TA ? lda*K:lda*M);
+    cl_mem A_gpu = clCreateBuffer(context,
+            CL_MEM_READ_ONLY|CL_MEM_COPY_HOST_PTR,
+            size, A, &cl.error);
+    check_error(cl);
+
+    size = sizeof(float)*(TB ? ldb*N:ldb*K);
+    cl_mem B_gpu = clCreateBuffer(context,
+            CL_MEM_READ_ONLY|CL_MEM_COPY_HOST_PTR,
+            size, B, &cl.error);
+    check_error(cl);
+
+    size = sizeof(float)*(ldc*M);
+    cl_mem C_gpu = clCreateBuffer(context,
+            CL_MEM_READ_WRITE|CL_MEM_COPY_HOST_PTR,
+            size, C, &cl.error);
+    check_error(cl);
+
+    gemm_ongpu(TA, TB, M, N, K, ALPHA, A_gpu, lda, B_gpu, ldb, BETA, C_gpu, ldc);
+
+    clEnqueueReadBuffer(queue, C_gpu, CL_TRUE, 0, size, C, 0, 0, 0);
+    check_error(cl);
+
+    clReleaseMemObject(A_gpu);
+    clReleaseMemObject(B_gpu);
+    clReleaseMemObject(C_gpu);
+}
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+
+void time_gpu_random_matrix(int TA, int TB, int m, int k, int n)
+{
+    float *a;
+    if(!TA) a = random_matrix(m,k);
+    else a = random_matrix(k,m);
+    int lda = (!TA)?k:m;
+    float *b;
+    if(!TB) b = random_matrix(k,n);
+    else b = random_matrix(n,k);
+    int ldb = (!TB)?n:k;
+
+    float *c = random_matrix(m,n);
+    int i;
+    clock_t start = clock(), end;
+    for(i = 0; i<1000; ++i){
+        gemm_gpu(TA,TB,m,n,k,1,a,lda,b,ldb,1,c,n);
+    }
+    end = clock();
+    printf("Matrix Multiplication %dx%d * %dx%d, TA=%d, TB=%d: %lf ms\n",m,k,k,n, TA, TB, (float)(end-start)/CLOCKS_PER_SEC);
+    free(a);
+    free(b);
+    free(c);
+}
+
+void test_gpu_accuracy(int TA, int TB, int m, int k, int n)
+{
+    srand(0);
+    float *a;
+    if(!TA) a = random_matrix(m,k);
+    else a = random_matrix(k,m);
+    int lda = (!TA)?k:m;
+    float *b;
+    if(!TB) b = random_matrix(k,n);
+    else b = random_matrix(n,k);
+    int ldb = (!TB)?n:k;
+
+    float *c = random_matrix(m,n);
+    float *c_gpu = random_matrix(m,n);
+    memset(c, 0, m*n*sizeof(float));
+    memset(c_gpu, 0, m*n*sizeof(float));
+    int i;
+    //pm(m,k,b);
+    gemm_gpu(TA,TB,m,n,k,1,a,lda,b,ldb,1,c_gpu,n);
+    //pm(m, n, c_gpu);
+    gemm_cpu(TA,TB,m,n,k,1,a,lda,b,ldb,1,c,n);
+    //pm(m, n, c);
+    double sse = 0;
+    for(i = 0; i < m*n; ++i) {
+        //printf("%f %f\n", c[i], c_gpu[i]);
+        sse += pow(c[i]-c_gpu[i], 2);
+    }
+    printf("Matrix Multiplication %dx%d * %dx%d, TA=%d, TB=%d: %g MSE\n",m,k,k,n, TA, TB, sse/(m*n));
+    free(a);
+    free(b);
+    free(c);
+}
+
+void test_gpu_blas()
+{
+    test_gpu_accuracy(0,0,17,10,10); 
+    test_gpu_accuracy(1,0,17,10,10); 
+    test_gpu_accuracy(0,1,17,10,10); 
+    test_gpu_accuracy(1,1,17,10,10); 
+
+    test_gpu_accuracy(0,0,1000,10,100); 
+    test_gpu_accuracy(1,0,1000,10,100); 
+    test_gpu_accuracy(0,1,1000,10,100); 
+    test_gpu_accuracy(1,1,1000,10,100); 
+
+    time_gpu_random_matrix(0,0,1000,1000,100); 
+    time_random_matrix(0,0,1000,1000,100); 
+
+    time_gpu_random_matrix(0,1,1000,1000,100); 
+    time_random_matrix(0,1,1000,1000,100); 
+
+    time_gpu_random_matrix(1,0,1000,1000,100); 
+    time_random_matrix(1,0,1000,1000,100); 
+
+    time_gpu_random_matrix(1,1,1000,1000,100); 
+    time_random_matrix(1,1,1000,1000,100); 
+
+}
+#endif
+

src/im2col.c

@@ -0,0 +1,121 @@
+#include "mini_blas.h"
+
+//From Berkeley Vision's Caffe!
+//https://github.com/BVLC/caffe/blob/master/LICENSE
+void im2col_cpu(float* data_im,
+    const int batch, const int channels, const int height, const int width,
+    const int ksize, const int stride, float* data_col) 
+{
+    int c,h,w,b;
+    int height_col = (height - ksize) / stride + 1;
+    int width_col = (width - ksize) / stride + 1;
+    int channels_col = channels * ksize * ksize;
+    int im_size = height*width*channels;
+    int col_size = height_col*width_col*channels_col;
+    for(b = 0; b < batch; ++b){
+        for ( c = 0; c < channels_col; ++c) {
+            int w_offset = c % ksize;
+            int h_offset = (c / ksize) % ksize;
+            int c_im = c / ksize / ksize;
+            for ( h = 0; h < height_col; ++h) {
+                for ( w = 0; w < width_col; ++w) {
+                    data_col[(c * height_col + h) * width_col + w] =
+                    data_im[(c_im * height + h * stride + h_offset) * width
+                        + w * stride + w_offset];
+                }
+            }
+        }
+        data_im += im_size;
+        data_col+= col_size;
+    }
+}
+
+
+#ifdef GPU
+
+#include "opencl.h"
+#include <math.h>
+
+cl_kernel get_im2col_kernel()
+{
+    static int init = 0;
+    static cl_kernel im2col_kernel;
+    if(!init){
+        im2col_kernel = get_kernel("src/im2col.cl", "im2col", 0);
+        init = 1;
+    }
+    return im2col_kernel;
+}
+
+
+void im2col_ongpu(cl_mem data_im, const int batch,
+        const int channels, const int height, const int width,
+        const int ksize, const int stride, cl_mem data_col) 
+{
+    cl_setup();
+    cl_kernel im2col_kernel = get_im2col_kernel();
+    cl_command_queue queue = cl.queue;
+
+    cl_uint i = 0;
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(data_im), (void*) &data_im);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(batch), (void*) &batch);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(channels), (void*) &channels);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(height), (void*) &height);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(width), (void*) &width);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(ksize), (void*) &ksize);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(stride), (void*) &stride);
+    cl.error = clSetKernelArg(im2col_kernel, i++, sizeof(data_col), (void*) &data_col);
+    check_error(cl);
+
+    int height_col = (height - ksize) / stride + 1;
+    int width_col = (width - ksize) / stride + 1;
+    int channels_col = channels * ksize * ksize;
+
+    size_t global_size[2];
+    size_t local_size[2];
+    global_size[0] = batch;
+    global_size[1] = channels_col;
+    local_size[0] = height_col;
+    local_size[1] = width_col;
+
+    clEnqueueNDRangeKernel(queue, im2col_kernel, 2, 0,
+            global_size, local_size, 0, 0, 0);
+    check_error(cl);
+}
+
+void im2col_gpu(float *data_im,
+        const int batch, const int channels, const int height, const int width,
+        const int ksize, const int stride,
+        float *data_col) 
+{
+    cl_setup();
+    cl_context context = cl.context;
+    cl_command_queue queue = cl.queue;
+
+    size_t size = sizeof(float)*(channels*height*width*batch);
+    cl_mem im_gpu = clCreateBuffer(context,
+            CL_MEM_READ_ONLY|CL_MEM_COPY_HOST_PTR,
+            size, data_im, &cl.error);
+    check_error(cl);
+
+    int height_col = (height - ksize) / stride + 1;
+    int width_col = (width - ksize) / stride + 1;
+    int channels_col = channels * ksize * ksize;
+
+    size = sizeof(float)*(height_col*width_col*channels_col*batch);
+    cl_mem col_gpu = clCreateBuffer(context,
+            CL_MEM_WRITE_ONLY|CL_MEM_COPY_HOST_PTR,
+            size, data_col, &cl.error);
+    check_error(cl);
+
+    im2col_ongpu(im_gpu, batch, channels, height, width,
+            ksize, stride, col_gpu);
+
+    clEnqueueReadBuffer(queue, col_gpu, CL_TRUE, 0, size, data_col, 0, 0, 0);
+    check_error(cl);
+
+    clReleaseMemObject(col_gpu);
+    clReleaseMemObject(im_gpu);
+}
+
+#endif

src/im2col.cl

@@ -0,0 +1,26 @@
+
+__kernel void im2col(__global float *data_im,
+    const int batch, const int channels, const int height, const int width,
+    const int ksize, const int stride, __global float *data_col) 
+{
+    int b = get_global_id(0);
+    int c = get_global_id(1);
+
+    int h = get_local_id(0);
+    int w = get_local_id(1);
+
+    int height_col = (height - ksize) / stride + 1;
+    int width_col = (width - ksize) / stride + 1;
+    int channels_col = channels * ksize * ksize;
+
+    int im_offset = height*width*channels*b;
+    int col_offset = height_col*width_col*channels_col*b;
+
+    int w_offset = c % ksize;
+    int h_offset = (c / ksize) % ksize;
+    int c_im = c / ksize / ksize;
+
+    data_col[(c * height_col + h) * width_col + w + col_offset] =
+        data_im[(c_im * height + h * stride + h_offset) * width
+        + w * stride + w_offset + im_offset];
+}

src/mini_blas.h

@@ -1,3 +1,5 @@
+#include "opencl.h"
+
 void pm(int M, int N, float *A);
 void gemm(int TA, int TB, int M, int N, int K, float ALPHA, 
                     float *A, int lda, 
@@ -6,12 +8,27 @@ void gemm(int TA, int TB, int M, int N, int K, float ALPHA,
                     float *C, int ldc);
 float *random_matrix(int rows, int cols);
 void time_random_matrix(int TA, int TB, int m, int k, int n);
-void im2col_gpu(float* data_im, const int channels,
-        const int height, const int width, const int ksize, const int stride,
-        float* data_col);
-void im2col_cpu(float* data_im, const int channels,
-        const int height, const int width, const int ksize, const int stride,
-        float* data_col);
+
+#ifdef GPU
+void im2col_ongpu(cl_mem data_im, const int batch,
+        const int channels, const int height, const int width,
+        const int ksize, const int stride, cl_mem data_col);
+
+void im2col_gpu(float *data_im,
+    const int batch, const int channels, const int height, const int width,
+    const int ksize, const int stride, float *data_col);
+
+void gemm_ongpu(int TA, int TB, int M, int N, int K, float ALPHA, 
+        cl_mem A_gpu, int lda, 
+        cl_mem B_gpu, int ldb,
+        float BETA,
+        cl_mem C_gpu, int ldc);
+#endif
+
+void im2col_cpu(float* data_im,
+    const int batch, const int channels, const int height, const int width,
+    const int ksize, const int stride, float* data_col);
+
 void col2im_cpu(float* data_col, const int channels,
     const int height, const int width, const int ksize, const int stride,
     float* data_im);

src/network.c

@@ -19,6 +19,9 @@ network make_network(int n, int batch)
     net.types = calloc(net.n, sizeof(LAYER_TYPE));
     net.outputs = 0;
     net.output = 0;
+    #ifdef GPU
+    net.input_cl = 0;
+    #endif
     return net;
 }
 
@@ -40,17 +43,6 @@ void print_convolutional_cfg(FILE *fp, convolutional_layer *l, int first)
     fprintf(fp, "data=");
     for(i = 0; i < l->n; ++i) fprintf(fp, "%g,", l->biases[i]);
     for(i = 0; i < l->n*l->c*l->size*l->size; ++i) fprintf(fp, "%g,", l->filters[i]);
-    /*
-    int j,k;
-    for(i = 0; i < l->n; ++i) fprintf(fp, "%g,", l->biases[i]);
-    for(i = 0; i < l->n; ++i){
-        for(j = l->c-1; j >= 0; --j){
-            for(k = 0; k < l->size*l->size; ++k){
-                fprintf(fp, "%g,", l->filters[i*(l->c*l->size*l->size)+j*l->size*l->size+k]);
-            }
-        }
-    }
-    */
     fprintf(fp, "\n\n");
 }
 void print_connected_cfg(FILE *fp, connected_layer *l, int first)
@@ -121,18 +113,34 @@ void save_network(network net, char *filename)
     fclose(fp);
 }
 
-void forward_network(network net, float *input)
+void forward_network(network net, float *input, int train)
 {
     int i;
+    #ifdef GPU
+    cl_setup();
+    size_t size = get_network_input_size(net);
+    if(!net.input_cl){
+        net.input_cl = clCreateBuffer(cl.context,
+            CL_MEM_READ_WRITE, size*sizeof(float), 0, &cl.error);
+        check_error(cl);
+    }
+    cl_write_array(net.input_cl, input, size);
+    cl_mem input_cl = net.input_cl;
+    #endif
     for(i = 0; i < net.n; ++i){
         if(net.types[i] == CONVOLUTIONAL){
             convolutional_layer layer = *(convolutional_layer *)net.layers[i];
+            #ifdef GPU
+            forward_convolutional_layer_gpu(layer, input_cl);
+            input_cl = layer.output_cl;
+            #else
             forward_convolutional_layer(layer, input);
+            #endif
             input = layer.output;
         }
         else if(net.types[i] == CONNECTED){
             connected_layer layer = *(connected_layer *)net.layers[i];
-            forward_connected_layer(layer, input);
+            forward_connected_layer(layer, input, train);
             input = layer.output;
         }
         else if(net.types[i] == SOFTMAX){
@@ -263,9 +271,7 @@ float backward_network(network net, float *input, float *truth)
         }
         if(net.types[i] == CONVOLUTIONAL){
             convolutional_layer layer = *(convolutional_layer *)net.layers[i];
-            learn_convolutional_layer(layer);
-            //learn_convolutional_layer(layer);
-            if(i != 0) backward_convolutional_layer(layer, prev_delta);
+            backward_convolutional_layer(layer, prev_delta);
         }
         else if(net.types[i] == MAXPOOL){
             maxpool_layer layer = *(maxpool_layer *)net.layers[i];
@@ -281,8 +287,7 @@ float backward_network(network net, float *input, float *truth)
         }
         else if(net.types[i] == CONNECTED){
             connected_layer layer = *(connected_layer *)net.layers[i];
-            learn_connected_layer(layer, prev_input);
-            if(i != 0) backward_connected_layer(layer, prev_input, prev_delta);
+            backward_connected_layer(layer, prev_input, prev_delta);
         }
     }
     return error;
@@ -290,7 +295,7 @@ float backward_network(network net, float *input, float *truth)
 
 float train_network_datum(network net, float *x, float *y, float step, float momentum, float decay)
 {
-    forward_network(net, x);
+    forward_network(net, x, 1);
     //int class = get_predicted_class_network(net);
     float error = backward_network(net, x, y);
     update_network(net, step, momentum, decay);
@@ -332,7 +337,7 @@ float train_network_batch(network net, data d, int n, float step, float momentum
         int index = rand()%d.X.rows;
         float *x = d.X.vals[index];
         float *y = d.y.vals[index];
-        forward_network(net, x);
+        forward_network(net, x, 1);
         int class = get_predicted_class_network(net);
         backward_network(net, x, y);
         correct += (y[class]?1:0);
@@ -359,6 +364,27 @@ void train_network(network net, data d, float step, float momentum, float decay)
     fprintf(stderr, "Accuracy: %f\n", (float)correct/d.X.rows);
 }
 
+int get_network_input_size_layer(network net, int i)
+{
+    if(net.types[i] == CONVOLUTIONAL){
+        convolutional_layer layer = *(convolutional_layer *)net.layers[i];
+        return layer.h*layer.w*layer.c;
+    }
+    else if(net.types[i] == MAXPOOL){
+        maxpool_layer layer = *(maxpool_layer *)net.layers[i];
+        return layer.h*layer.w*layer.c;
+    }
+    else if(net.types[i] == CONNECTED){
+        connected_layer layer = *(connected_layer *)net.layers[i];
+        return layer.inputs;
+    }
+    else if(net.types[i] == SOFTMAX){
+        softmax_layer layer = *(softmax_layer *)net.layers[i];
+        return layer.inputs;
+    }
+    return 0;
+}
+
 int get_network_output_size_layer(network net, int i)
 {
     if(net.types[i] == CONVOLUTIONAL){
@@ -382,36 +408,6 @@ int get_network_output_size_layer(network net, int i)
     return 0;
 }
 
-/*
-   int resize_network(network net, int h, int w, int c)
-   {
-   int i;
-   for (i = 0; i < net.n; ++i){
-   if(net.types[i] == CONVOLUTIONAL){
-   convolutional_layer *layer = (convolutional_layer *)net.layers[i];
-   layer->h = h;
-   layer->w = w;
-   layer->c = c;
-   image output = get_convolutional_image(*layer);
-   h = output.h;
-   w = output.w;
-   c = output.c;
-   }
-   else if(net.types[i] == MAXPOOL){
-   maxpool_layer *layer = (maxpool_layer *)net.layers[i];
-   layer->h = h;
-   layer->w = w;
-   layer->c = c;
-   image output = get_maxpool_image(*layer);
-   h = output.h;
-   w = output.w;
-   c = output.c;
-   }
-   }
-   return 0;
-   }
- */
-
 int resize_network(network net, int h, int w, int c)
 {
     int i;
@@ -450,6 +446,11 @@ int get_network_output_size(network net)
     return get_network_output_size_layer(net, i);
 }
 
+int get_network_input_size(network net)
+{
+    return get_network_output_size_layer(net, 0);
+}
+
 image get_network_image_layer(network net, int i)
 {
     if(net.types[i] == CONVOLUTIONAL){
@@ -497,7 +498,7 @@ void visualize_network(network net)
 
 float *network_predict(network net, float *input)
 {
-    forward_network(net, input);
+    forward_network(net, input, 0);
     float *out = get_network_output(net);
     return out;
 }

src/network.h

@@ -2,6 +2,7 @@
 #ifndef NETWORK_H
 #define NETWORK_H
 
+#include "opencl.h"
 #include "image.h"
 #include "data.h"
 
@@ -20,10 +21,15 @@ typedef struct {
     LAYER_TYPE *types;
     int outputs;
     float *output;
+
+    #ifdef GPU
+    cl_mem input_cl;
+    cl_mem output_cl;
+    #endif
 } network;
 
 network make_network(int n, int batch);
-void forward_network(network net, float *input);
+void forward_network(network net, float *input, int train);
 float backward_network(network net, float *input, float *truth);
 void update_network(network net, float step, float momentum, float decay);
 float train_network_sgd(network net, data d, int n, float step, float momentum,float decay);
@@ -44,6 +50,7 @@ void print_network(network net);
 void visualize_network(network net);
 void save_network(network net, char *filename);
 int resize_network(network net, int h, int w, int c);
+int get_network_input_size(network net);
 
 #endif
 

src/opencl.c

@@ -88,4 +88,18 @@ cl_kernel get_kernel(char *filename, char *kernelname, char *options)
 	return kernel;
 }
 
+void cl_read_array(cl_mem mem, float *x, int n)
+{
+    cl_setup();
+    clEnqueueReadBuffer(cl.queue, mem, CL_TRUE, 0, sizeof(float)*n,x,0,0,0);
+    check_error(cl);
+}
+
+void cl_write_array(cl_mem mem, float *x, int n)
+{
+    cl_setup();
+    clEnqueueWriteBuffer(cl.queue, mem, CL_TRUE, 0,sizeof(float)*n,x,0,0,0);
+    check_error(cl);
+}
+
 #endif

src/opencl.h

@@ -1,3 +1,6 @@
+#ifdef GPU
+#ifndef OPENCL_H
+#define OPENCL_H
 #ifdef __APPLE__
 #include <OpenCL/opencl.h>
 #else
@@ -18,4 +21,7 @@ extern cl_info cl;
 void cl_setup();
 void check_error(cl_info info);
 cl_kernel get_kernel(char *filename, char *kernelname, char *options);
-
+void cl_read_array(cl_mem mem, float *x, int n);
+void cl_write_array(cl_mem mem, float *x, int n);
+#endif
+#endif

src/parser.c

@@ -89,6 +89,7 @@ connected_layer *parse_connected(list *options, network net, int count)
     int i;
     int input;
     int output = option_find_int(options, "output",1);
+    float dropout = option_find_float(options, "dropout", 0.);
     char *activation_s = option_find_str(options, "activation", "sigmoid");
     ACTIVATION activation = get_activation(activation_s);
     if(count == 0){
@@ -97,7 +98,7 @@ connected_layer *parse_connected(list *options, network net, int count)
     }else{
         input =  get_network_output_size_layer(net, count-1);
     }
-    connected_layer *layer = make_connected_layer(net.batch, input, output, activation);
+    connected_layer *layer = make_connected_layer(net.batch, input, output, dropout, activation);
     char *data = option_find_str(options, "data", 0);
     if(data){
         char *curr = data;

src/tests.c

@@ -52,7 +52,7 @@ void test_convolve_matrix()
 	int i;
 	clock_t start = clock(), end;
 	for(i = 0; i < 1000; ++i){
-		im2col_cpu(dog.data,  dog.c,  dog.h,  dog.w,  size,  stride, matrix);
+		im2col_cpu(dog.data,  1, dog.c,  dog.h,  dog.w,  size,  stride, matrix);
 		gemm(0,0,n,mw,mh,1,filters,mh,matrix,mw,1,edge.data,mw);
 	}
 	end = clock();
@@ -168,7 +168,7 @@ void test_parser()
 		float v = ((float)rand()/RAND_MAX);
 		float truth = v*v;
 		input[0] = v;
-		forward_network(net, input);
+		forward_network(net, input, 1);
 		float *out = get_network_output(net);
 		float *delta = get_network_delta(net);
 		float err = pow((out[0]-truth),2.);
@@ -245,7 +245,7 @@ void test_full()
 		normalize_data_rows(test);
 		for(j = 0; j < test.X.rows; ++j){
 			float *x = test.X.vals[j];
-			forward_network(net, x);
+			forward_network(net, x, 0);
 			int class = get_predicted_class_network(net);
 			fprintf(fp, "%d\n", class);
 		}
@@ -317,22 +317,14 @@ void test_nist()
 	int batch = 10000;
 	while(++count <= 10000){
 		float loss = train_network_sgd(net, train, batch, lr, momentum, decay);
-		printf("%5f %5f\n",(double)count*batch/train.X.rows, loss);
+		float test_acc = network_accuracy(net, test);
+		printf("%3d %5f %5f\n",count, loss, test_acc);
 		//printf("%5d Training Loss: %lf, Params: %f %f %f, ",count*1000, loss, lr, momentum, decay);
 		//end = clock();
 		//printf("Time: %lf seconds\n", (float)(end-start)/CLOCKS_PER_SEC);
 		//start=end;
-		/*
-		   if(count%5 == 0){
-		   float train_acc = network_accuracy(net, train);
-		   fprintf(stderr, "\nTRAIN: %f\n", train_acc);
-		   float test_acc = network_accuracy(net, test);
-		   fprintf(stderr, "TEST: %f\n\n", test_acc);
-		   printf("%d, %f, %f\n", count, train_acc, test_acc);
 		//lr *= .5;
 	}
-		*/
-	}
 }
 
 void test_ensemble()
@@ -387,7 +379,7 @@ void test_random_classify()
 			int index = rand()%m.rows;
 			//image p = float_to_image(1690,1,1,m.vals[index]);
 			//normalize_image(p);
-			forward_network(net, m.vals[index]);
+			forward_network(net, m.vals[index], 1);
 			float *out = get_network_output(net);
 			float *delta = get_network_delta(net);
 			//printf("%f\n", out[0]);
@@ -408,7 +400,7 @@ void test_random_classify()
 	matrix test = csv_to_matrix("test.csv");
 	truth = pop_column(&test, 0);
 	for(i = 0; i < test.rows; ++i){
-		forward_network(net, test.vals[i]);
+		forward_network(net, test.vals[i], 0);
 		float *out = get_network_output(net);
 		if(fabs(out[0]) < .5) fprintf(fp, "0\n");
 		else fprintf(fp, "1\n");
@@ -439,7 +431,7 @@ void test_im2row()
 	float *matrix = calloc(msize, sizeof(float));
 	int i;
 	for(i = 0; i < 1000; ++i){
-		im2col_cpu(test.data,  c,  h,  w,  size,  stride, matrix);
+		im2col_cpu(test.data, 1, c,  h,  w,  size,  stride, matrix);
 		//image render = float_to_image(mh, mw, mc, matrix);
 	}
 }
@@ -506,7 +498,7 @@ image features_output_size(network net, IplImage *src, int outh, int outw)
 	//normalize_array(im.data, im.h*im.w*im.c);
 	translate_image(im, -144);
 	resize_network(net, im.h, im.w, im.c);
-	forward_network(net, im.data);
+	forward_network(net, im.data, 0);
 	image out = get_network_image(net);
 	free_image(im);
 	cvReleaseImage(&sized);
@@ -558,7 +550,7 @@ void visualize_imagenet_topk(char *filename)
 		resize_network(net, im.h, im.w, im.c);
 		//scale_image(im, 1./255);
 		translate_image(im, -144);
-		forward_network(net, im.data);
+		forward_network(net, im.data, 0);
 		image out = get_network_image(net);
 
 		int dh = (im.h - h)/(out.h-1);
@@ -620,7 +612,7 @@ void visualize_imagenet_features(char *filename)
 		image im = load_image(image_path, 0, 0);
 		printf("Processing %dx%d image\n", im.h, im.w);
 		resize_network(net, im.h, im.w, im.c);
-		forward_network(net, im.data);
+		forward_network(net, im.data, 0);
 		image out = get_network_image(net);
 
 		int dh = (im.h - h)/h;
@@ -653,7 +645,7 @@ void visualize_cat()
 	image im = load_image("data/cat.png", 0, 0);
 	printf("Processing %dx%d image\n", im.h, im.w);
 	resize_network(net, im.h, im.w, im.c);
-	forward_network(net, im.data);
+	forward_network(net, im.data, 0);
 
 	visualize_network(net);
 	cvWaitKey(0);