cifar and go stuff

Makefile

@@ -1,5 +1,5 @@
-GPU=0
+GPU=1
-OPENCV=0
+OPENCV=1
 DEBUG=0
 
 ARCH= --gpu-architecture=compute_20 --gpu-code=compute_20 

cfg/cifar.cfg

@@ -0,0 +1,126 @@
+[net]
+batch=128
+subdivisions=1
+height=32
+width=32
+channels=3
+momentum=0.9
+decay=0.0005
+
+learning_rate=0.4
+policy=poly
+power=4
+max_batches = 50000
+
+[crop]
+crop_width=28
+crop_height=28
+flip=1
+angle=0
+saturation = 1
+exposure = 1
+noadjust=1
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[dropout]
+probability=.5
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[dropout]
+probability=.5
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[dropout]
+probability=.5
+
+[convolutional]
+filters=10
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[avgpool]
+
+[softmax]
+groups=1
+
+[cost]
+

cfg/cifar.test.cfg

@@ -0,0 +1,119 @@
+[net]
+batch=128
+subdivisions=1
+height=32
+width=32
+channels=3
+momentum=0.9
+decay=0.0005
+
+learning_rate=0.4
+policy=poly
+power=4
+max_batches = 50000
+
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[dropout]
+probability=.5
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[dropout]
+probability=.5
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[dropout]
+probability=.5
+
+[convolutional]
+filters=10
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[avgpool]
+
+[softmax]
+groups=1
+temperature=3
+
+[cost]
+

src/classifier.c

@@ -410,7 +410,7 @@ void validate_classifier_multi(char *datacfg, char *filename, char *weightfile)
 
     char **labels = get_labels(label_list);
     list *plist = get_paths(valid_list);
-    int scales[] = {192, 224, 288, 320, 352};
+    int scales[] = {160, 192, 224, 288, 320, 352, 384};
     int nscales = sizeof(scales)/sizeof(scales[0]);
 
     char **paths = (char **)list_to_array(plist);

src/data.c

@@ -917,8 +917,8 @@ data load_all_cifar10()
 data load_go(char *filename)
 {
     FILE *fp = fopen(filename, "rb");
-    matrix X = make_matrix(128, 361);
+    matrix X = make_matrix(3363059, 361);
-    matrix y = make_matrix(128, 361);
+    matrix y = make_matrix(3363059, 361);
     int row, col;
 
     if(!fp) file_error(filename);
@@ -943,6 +943,8 @@ data load_go(char *filename)
             X.vals[count][i] = val;
         }
         ++count;
+        free(label);
+        free(board);
     }
     X = resize_matrix(X, count);
     y = resize_matrix(y, count);

src/go.c

@@ -23,7 +23,11 @@ void train_go(char *cfgfile, char *weightfile)
 
     char *backup_directory = "/home/pjreddie/backup/";
 
-    data train = load_go("/home/pjreddie/backup/go.train");
+
+    char buff[256];
+    sprintf(buff, "/home/pjreddie/go.train.%02d", rand()%10);
+    data train = load_go(buff);
+
     int N = train.X.rows;
     int epoch = (*net.seen)/N;
     while(get_current_batch(net) < net.max_batches || net.max_batches == 0){
@@ -58,6 +62,10 @@ void train_go(char *cfgfile, char *weightfile)
             char buff[256];
             sprintf(buff, "%s/%s_%d.weights",backup_directory,base, epoch);
             save_weights(net, buff);
+
+            free_data(train);
+            sprintf(buff, "/home/pjreddie/go.train.%02d", epoch%10);
+            train = load_go(buff);
         }
         if(get_current_batch(net)%100 == 0){
             char buff[256];
@@ -65,7 +73,6 @@ void train_go(char *cfgfile, char *weightfile)
             save_weights(net, buff);
         }
     }
-    char buff[256];
     sprintf(buff, "%s/%s.weights", backup_directory, base);
     save_weights(net, buff);
 
@@ -122,9 +129,9 @@ void update_board(float *board)
     free(l);
 }
 
-void print_board(float *board, int swap)
+void print_board(float *board, int swap, int *indexes)
 {
-    int i,j;
+    int i,j,n;
     printf("\n\n");
     printf("   ");
     for(i = 0; i < 19; ++i){
@@ -135,9 +142,21 @@ void print_board(float *board, int swap)
         printf("%2d ", 19-j);
         for(i = 0; i < 19; ++i){
             int index = j*19 + i;
+            if(indexes){
+                int found = 0;
+                for(n = 0; n < 3; ++n){
+                    if(index == indexes[n]){
+                        found = 1;
+                        if(n == 0) printf("\uff11");
+                        else if(n == 1) printf("\uff12");
+                        else if(n == 2) printf("\uff13");
+                    }
+                }
+                if(found) continue;
+            }
             if(board[index]*-swap > 0) printf("\u25C9 ");
             else if(board[index]*-swap < 0) printf("\u25EF ");
-            else printf("  ");
+            else printf("\uFF0b");
         }
         printf("\n");
     }
@@ -166,7 +185,7 @@ void test_go(char *filename, char *weightfile)
         float *output = network_predict(net, board);
         copy_cpu(19*19, output, 1, move, 1);
         int i;
-        #ifdef GPU
+#ifdef GPU
         image bim = float_to_image(19, 19, 1, board);
         for(i = 1; i < 8; ++i){
             rotate_image_cw(bim, i);
@@ -184,7 +203,7 @@ void test_go(char *filename, char *weightfile)
             rotate_image_cw(bim, -i);
         }
         scal_cpu(19*19, 1./8., move, 1);
-        #endif
+#endif
         for(i = 0; i < 19*19; ++i){
             if(board[i]) move[i] = 0;
         }
@@ -192,7 +211,7 @@ void test_go(char *filename, char *weightfile)
         int indexes[3];
         int row, col;
         top_k(move, 19*19, 3, indexes);
-        print_board(board, color);
+        print_board(board, color, indexes);
         for(i = 0; i < 3; ++i){
             int index = indexes[i];
             row = index / 19;
@@ -217,6 +236,7 @@ void test_go(char *filename, char *weightfile)
             if (c == 'p'){
                 flip_board(board);
                 color = -color;
+                free(line);
                 continue;
             }else{
                 char g;
@@ -232,8 +252,10 @@ void test_go(char *filename, char *weightfile)
             if (col > 7) col -= 1;
             board[row*19 + col] = 1;
         }else{
+            free(line);
             continue;
         }
+        free(line);
         update_board(board);
         flip_board(board);
         color = -color;