?

Makefile

@@ -25,7 +25,7 @@ CFLAGS+=-DGPU
 LDFLAGS+= -L/usr/local/cuda/lib64 -lcuda -lcudart -lcublas -lcurand
 endif
 
-OBJ=gemm.o utils.o cuda.o deconvolutional_layer.o convolutional_layer.o list.o image.o activations.o im2col.o col2im.o blas.o crop_layer.o dropout_layer.o maxpool_layer.o softmax_layer.o data.o matrix.o network.o connected_layer.o cost_layer.o parser.o option_list.o darknet.o detection_layer.o imagenet.o captcha.o detection.o route_layer.o
+OBJ=gemm.o utils.o cuda.o deconvolutional_layer.o convolutional_layer.o list.o image.o activations.o im2col.o col2im.o blas.o crop_layer.o dropout_layer.o maxpool_layer.o softmax_layer.o data.o matrix.o network.o connected_layer.o cost_layer.o parser.o option_list.o darknet.o detection_layer.o imagenet.o captcha.o detection.o route_layer.o writing.o
 ifeq ($(GPU), 1) 
 OBJ+=convolutional_kernels.o deconvolutional_kernels.o activation_kernels.o im2col_kernels.o col2im_kernels.o blas_kernels.o crop_layer_kernels.o dropout_layer_kernels.o maxpool_layer_kernels.o softmax_layer_kernels.o network_kernels.o
 endif

src/darknet.c

@@ -11,6 +11,7 @@
 
 extern void run_imagenet(int argc, char **argv);
 extern void run_detection(int argc, char **argv);
+extern void run_writing(int argc, char **argv);
 extern void run_captcha(int argc, char **argv);
 
 void del_arg(int argc, char **argv, int index)
@@ -105,6 +106,8 @@ int main(int argc, char **argv)
         run_imagenet(argc, argv);
     } else if (0 == strcmp(argv[1], "detection")){
         run_detection(argc, argv);
+    } else if (0 == strcmp(argv[1], "writing")){
+        run_writing(argc, argv);
     } else if (0 == strcmp(argv[1], "test")){
         test_resize(argv[2]);
     } else if (0 == strcmp(argv[1], "captcha")){

src/data.c

@@ -49,6 +49,33 @@ char **get_random_paths(char **paths, int n, int m)
     return random_paths;
 }
 
+char **find_replace_paths(char **paths, int n, char *find, char *replace)
+{
+    char **replace_paths = calloc(n, sizeof(char*));
+    int i;
+    for(i = 0; i < n; ++i){
+        char *replaced = find_replace(paths[i], find, replace);
+        replace_paths[i] = copy_string(replaced);
+    }
+    return replace_paths;
+}
+
+matrix load_image_paths_gray(char **paths, int n, int w, int h)
+{
+    int i;
+    matrix X;
+    X.rows = n;
+    X.vals = calloc(X.rows, sizeof(float*));
+    X.cols = 0;
+
+    for(i = 0; i < n; ++i){
+        image im = load_image(paths[i], w, h);
+        X.vals[i] = im.data;
+        X.cols = im.h*im.w*im.c;
+    }
+    return X;
+}
+
 matrix load_image_paths(char **paths, int n, int w, int h)
 {
     int i;
@@ -497,6 +524,21 @@ pthread_t load_data_detection_thread(int n, char **paths, int m, int classes, in
     return thread;
 }
 
+data load_data_writing(char **paths, int n, int m, int w, int h)
+{
+    if(m) paths = get_random_paths(paths, n, m);
+    char **replace_paths = find_replace_paths(paths, n, ".png", "label.png");
+    data d;
+    d.shallow = 0;
+    d.X = load_image_paths(paths, n, w, h);
+    d.y = load_image_paths_gray(replace_paths, n, w/4, h/4);
+    if(m) free(paths);
+    int i;
+    for(i = 0; i < n; ++i) free(replace_paths[i]);
+    free(replace_paths);
+    return d;
+}
+
 data load_data(char **paths, int n, int m, char **labels, int k, int w, int h)
 {
     if(m) paths = get_random_paths(paths, n, m);

src/data.h

@@ -41,6 +41,8 @@ pthread_t load_data_localization_thread(int n, char **paths, int m, int classes,
 data load_cifar10_data(char *filename);
 data load_all_cifar10();
 
+data load_data_writing(char **paths, int n, int m, int w, int h);
+
 list *get_paths(char *filename);
 char **get_labels(char *filename);
 void get_random_batch(data d, int n, float *X, float *y);

src/detection_layer.c

@@ -368,18 +368,11 @@ void forward_detection_layer(const detection_layer l, network_state state)
 
 
             *(l.cost) += pow((1-iou), 2);
-            if(0){
-                l.delta[j+0] = (state.truth[j+0] - l.output[j+0]);
-                l.delta[j+1] = (state.truth[j+1] - l.output[j+1]);
-                l.delta[j+2] = (state.truth[j+2] - l.output[j+2]);
-                l.delta[j+3] = (state.truth[j+3] - l.output[j+3]);
-            }else{
-                l.delta[j+0] = 4 * (state.truth[j+0] - l.output[j+0]) / 7;
-                l.delta[j+1] = 4 * (state.truth[j+1] - l.output[j+1]) / 7;
-                l.delta[j+2] = 4 * (state.truth[j+2] - l.output[j+2]);
-                l.delta[j+3] = 4 * (state.truth[j+3] - l.output[j+3]);
-            }
-            if(0){
+            l.delta[j+0] = 4 * (state.truth[j+0] - l.output[j+0]);
+            l.delta[j+1] = 4 * (state.truth[j+1] - l.output[j+1]);
+            l.delta[j+2] = 4 * (state.truth[j+2] - l.output[j+2]);
+            l.delta[j+3] = 4 * (state.truth[j+3] - l.output[j+3]);
+            if(1){
                 for (j = offset; j < offset+classes; ++j) {
                     if(state.truth[j]) state.truth[j] = iou;
                     l.delta[j] =  state.truth[j] - l.output[j];

src/imagenet.c

@@ -47,7 +47,7 @@ void train_imagenet(char *cfgfile, char *weightfile)
         avg_loss = avg_loss*.9 + loss*.1;
         printf("%d: %f, %f avg, %lf seconds, %d images\n", i, loss, avg_loss, sec(clock()-time), net.seen);
         free_data(train);
-        if((i % 20000) == 0) net.learning_rate *= .1;
+        if((i % 15000) == 0) net.learning_rate *= .1;
         //if(i%100 == 0 && net.learning_rate > .00001) net.learning_rate *= .97;
         if(i%1000==0){
             char buff[256];