Merging to f6d8617 of upstream

cfg/alexnet.cfg

@@ -90,6 +90,3 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/cifar.cfg

@@ -119,6 +119,3 @@ activation=leaky
 
 [softmax]
 groups=1
-
-[cost]
-

cfg/cifar.test.cfg

@@ -115,5 +115,3 @@ activation=leaky
 groups=1
 temperature=3
 
-[cost]
-

cfg/darknet19.cfg

@@ -203,6 +203,3 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/darknet19_448.cfg

@@ -195,6 +195,3 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/densenet201.cfg

@@ -1949,6 +1949,3 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/extraction.cfg

@@ -201,6 +201,3 @@ activation=leaky
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/extraction22k.cfg

@@ -204,6 +204,3 @@ activation=leaky
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/go.cfg

@@ -130,6 +130,3 @@ stride=1
 
 [softmax]
 
-[cost]
-type=sse
-

cfg/go.test.cfg

@@ -129,6 +129,4 @@ stride=1
 
 [softmax]
 
-[cost]
-type=sse
 

cfg/gru.cfg

@@ -27,6 +27,4 @@ activation=linear
 
 [softmax]
 
-[cost]
-type=sse
 

cfg/resnet152.cfg

@@ -1458,6 +1458,3 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
-

cfg/resnet50.cfg

@@ -506,6 +506,4 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
 

cfg/rnn.cfg

@@ -35,6 +35,4 @@ activation=leaky
 
 [softmax]
 
-[cost]
-type=sse
 

cfg/rnn.train.cfg

@@ -35,6 +35,4 @@ activation=leaky
 
 [softmax]
 
-[cost]
-type=sse
 

cfg/strided.cfg

@@ -180,6 +180,3 @@ activation=ramp
 
 [softmax]
 
-[cost]
-type=sse
-

cfg/tiny.cfg

@@ -171,6 +171,4 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
 

cfg/vgg-16.cfg

@@ -148,6 +148,4 @@ activation=linear
 [softmax]
 groups=1
 
-[cost]
-type=sse
 

cfg/yolov3-tiny.cfg

@@ -0,0 +1,182 @@
+[net]
+# Testing
+batch=1
+subdivisions=1
+# Training
+# batch=64
+# subdivisions=2
+width=416
+height=416
+channels=3
+momentum=0.9
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.001
+burn_in=1000
+max_batches = 500200
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+[convolutional]
+batch_normalize=1
+filters=16
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=2
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[maxpool]
+size=2
+stride=1
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=1
+pad=1
+activation=leaky
+
+###########
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+
+
+[yolo]
+mask = 3,4,5
+anchors = 10,14,  23,27,  37,58,  81,82,  135,169,  344,319
+classes=80
+num=6
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+
+[route]
+layers = -4
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[upsample]
+stride=2
+
+[route]
+layers = -1, 8
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+[yolo]
+mask = 1,2,3
+anchors = 10,14,  23,27,  37,58,  81,82,  135,169,  344,319
+classes=80
+num=6
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1

cfg/yolov3.cfg

@@ -610,7 +610,7 @@ anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  37
 classes=80
 num=9
 jitter=.3
-ignore_thresh = .5
+ignore_thresh = .7
 truth_thresh = 1
 random=1
 
@@ -696,7 +696,7 @@ anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  37
 classes=80
 num=9
 jitter=.3
-ignore_thresh = .5
+ignore_thresh = .7
 truth_thresh = 1
 random=1
 
@@ -783,7 +783,7 @@ anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  37
 classes=80
 num=9
 jitter=.3
-ignore_thresh = .5
+ignore_thresh = .7
 truth_thresh = 1
 random=1
 

src/convolutional_kernels.cu

@@ -109,9 +109,13 @@ void forward_convolutional_layer_gpu(convolutional_layer l, network net)
             float *a = l.weights_gpu + j*l.nweights/l.groups;
             float *b = net.workspace;
             float *c = l.output_gpu + (i*l.groups + j)*n*m;
+            float *im = net.input_gpu + (i*l.groups + j)*l.c/l.groups*l.h*l.w;
 
-            im2col_gpu(net.input_gpu + (i*l.groups + j)*l.c/l.groups*l.h*l.w,
-                l.c/l.groups, l.h, l.w, l.size, l.stride, l.pad, b);
+            if (l.size == 1){
+                b = im;
+            } else {
+                im2col_gpu(im, l.c/l.groups, l.h, l.w, l.size, l.stride, l.pad, b);
+            }
             gemm_gpu(0,0,m,n,k,1,a,k,b,n,1,c,n);
         }
     }
@@ -234,22 +238,26 @@ void backward_convolutional_layer_gpu(convolutional_layer l, network net)
             float *b = net.workspace;
             float *c = l.weight_updates_gpu + j*l.nweights/l.groups;
 
-            float *im = net.input_gpu+(i*l.groups + j)*l.c/l.groups*l.h*l.w;
+            float *im  = net.input_gpu+(i*l.groups + j)*l.c/l.groups*l.h*l.w;
+            float *imd = net.delta_gpu+(i*l.groups + j)*l.c/l.groups*l.h*l.w;
 
-            im2col_gpu(im, l.c/l.groups, l.h, l.w,
-                    l.size, l.stride, l.pad, b);
+            im2col_gpu(im, l.c/l.groups, l.h, l.w, l.size, l.stride, l.pad, b);
             gemm_gpu(0,1,m,n,k,1,a,k,b,k,1,c,n);
 
-            if(net.delta_gpu){
-                if(l.binary || l.xnor) swap_binary(&l);
+            if (net.delta_gpu) {
+                if (l.binary || l.xnor) swap_binary(&l);
                 a = l.weights_gpu + j*l.nweights/l.groups;
                 b = l.delta_gpu + (i*l.groups + j)*m*k;
                 c = net.workspace;
+                if (l.size == 1) {
+                    c = imd;
+                }
 
                 gemm_gpu(1,0,n,k,m,1,a,n,b,k,0,c,k);
 
-                col2im_gpu(net.workspace, l.c/l.groups, l.h, l.w, l.size, l.stride, 
-                    l.pad, net.delta_gpu + (i*l.groups + j)*l.c/l.groups*l.h*l.w);
+                if (l.size != 1) {
+                    col2im_gpu(net.workspace, l.c/l.groups, l.h, l.w, l.size, l.stride, l.pad, imd);
+                }
                 if(l.binary || l.xnor) {
                     swap_binary(&l);
                 }