multilayer [shortcut] in progress

build/darknet/x64/partial.cmd

@@ -6,6 +6,10 @@ rem Download Yolo9000: http://pjreddie.com/media/files/yolo9000.weights
 rem darknet.exe partial cfg/tiny-yolo-voc.cfg tiny-yolo-voc.weights tiny-yolo-voc.conv.13 13
 
 
+
+darknet.exe partial cfg/csresnext50.cfg csresnext50.weights csresnext50.conv.75 75
+
+
 darknet.exe partial cfg/darknet53_448.cfg darknet53_448.weights darknet53.conv.74 74
 
 

include/darknet.h

@@ -328,6 +328,8 @@ struct layer {
     int   * indexes;
     int   * input_layers;
     int   * input_sizes;
+    float **layers_output;
+    float **layers_delta;
     int   * map;
     int   * counts;
     float ** sums;
@@ -575,6 +577,10 @@ struct layer {
 
     float *gt_gpu;
     float *a_avg_gpu;
+
+    int *input_sizes_gpu;
+    float **layers_output_gpu;
+    float **layers_delta_gpu;
 #ifdef CUDNN
     cudnnTensorDescriptor_t srcTensorDesc, dstTensorDesc;
     cudnnTensorDescriptor_t srcTensorDesc16, dstTensorDesc16;

src/blas.c

@@ -68,6 +68,60 @@ void weighted_delta_cpu(float *a, float *b, float *s, float *da, float *db, floa
     }
 }
 
+void shortcut_multilayer_cpu(int size, int src_outputs, int batch, int n, int *outputs_of_layers, float **layers_output, float *out, float *in)
+{
+    int id;
+    #pragma omp parallel for
+    for (id = 0; id < size; ++id) {
+
+        int src_id = id;
+        int src_i = src_id % src_outputs;
+        src_id /= src_outputs;
+        int src_b = src_id;
+
+        out[id] = in[id];
+
+        // layers
+        for (int i = 0; i < n; ++i) {
+            int add_outputs = outputs_of_layers[i];
+            if (src_i < add_outputs) {
+                int add_index = add_outputs*src_b + src_i;
+                int out_index = id;
+
+                float *add = layers_output[i];
+                out[out_index] += add[add_index];
+            }
+        }
+    }
+}
+
+void backward_shortcut_multilayer_cpu(int size, int src_outputs, int batch, int n, int *outputs_of_layers,
+    float **layers_delta, float *delta_out, float *delta_in)
+{
+    int id;
+    #pragma omp parallel for
+    for (id = 0; id < size; ++id) {
+        int src_id = id;
+        int src_i = src_id % src_outputs;
+        src_id /= src_outputs;
+        int src_b = src_id;
+
+        delta_out[id] += delta_in[id];
+
+        // layers
+        for (int i = 0; i < n; ++i) {
+            int add_outputs = outputs_of_layers[i];
+            if (src_i < add_outputs) {
+                int add_index = add_outputs*src_b + src_i;
+                int out_index = id;
+
+                float *layer_delta = layers_delta[i];
+                layer_delta[add_index] += delta_in[id];
+            }
+        }
+    }
+}
+
 void shortcut_cpu(int batch, int w1, int h1, int c1, float *add, int w2, int h2, int c2, float *out)
 {
     int stride = w1/w2;

src/blas.h

@@ -32,6 +32,9 @@ void fill_cpu(int N, float ALPHA, float * X, int INCX);
 float dot_cpu(int N, float *X, int INCX, float *Y, int INCY);
 void test_gpu_blas();
 void shortcut_cpu(int batch, int w1, int h1, int c1, float *add, int w2, int h2, int c2, float *out);
+void shortcut_multilayer_cpu(int size, int src_outputs, int batch, int n, int *outputs_of_layers, float **layers_output, float *out, float *in);
+void backward_shortcut_multilayer_cpu(int size, int src_outputs, int batch, int n, int *outputs_of_layers,
+    float **layers_delta, float *delta_out, float *delta_in);
 
 void mean_cpu(float *x, int batch, int filters, int spatial, float *mean);
 void variance_cpu(float *x, float *mean, int batch, int filters, int spatial, float *variance);
@@ -84,6 +87,8 @@ void fast_variance_delta_gpu(float *x, float *delta, float *mean, float *varianc
 void fast_variance_gpu(float *x, float *mean, int batch, int filters, int spatial, float *variance);
 void fast_mean_gpu(float *x, int batch, int filters, int spatial, float *mean);
 void shortcut_gpu(int batch, int w1, int h1, int c1, float *add, int w2, int h2, int c2, float *out);
+void shortcut_multilayer_gpu(int src_outputs, int batch, int n, int *outputs_of_layers_gpu, float **layers_output_gpu, float *out, float *in);
+void backward_shortcut_multilayer_gpu(int src_outputs, int batch, int n, int *outputs_of_layers_gpu, float **layers_delta_gpu, float *delta_out, float *delta_in);
 void input_shortcut_gpu(float *in, int batch, int w1, int h1, int c1, float *add, int w2, int h2, int c2, float *out);
 void scale_bias_gpu(float *output, float *biases, int batch, int n, int size);
 void backward_scale_gpu(float *x_norm, float *delta, int batch, int n, int size, float *scale_updates);

src/blas_kernels.cu

@@ -668,6 +668,75 @@ extern "C" void fill_ongpu(int N, float ALPHA, float * X, int INCX)
     CHECK_CUDA(cudaPeekAtLastError());
 }
 
+__global__ void shortcut_multilayer_kernel(int size, int src_outputs, int batch, int n, int *outputs_of_layers_gpu, float **layers_output_gpu, float *out, float *in)
+{
+    const int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
+    if (id >= size) return;
+
+    int src_id = id;
+    int src_i = src_id % src_outputs;
+    src_id /= src_outputs;
+    int src_b = src_id;
+
+    out[id] = in[id];
+
+    // layers
+    for (int i = 0; i < n; ++i) {
+        int add_outputs = outputs_of_layers_gpu[i];
+        if (src_i < add_outputs) {
+            int add_index = add_outputs*src_b + src_i;
+            int out_index = id;
+
+            float *add = layers_output_gpu[i];
+            out[out_index] += add[add_index];
+        }
+    }
+}
+
+extern "C" void shortcut_multilayer_gpu(int src_outputs, int batch, int n, int *outputs_of_layers_gpu, float **layers_output_gpu, float *out, float *in)
+{
+    //printf(" src_outputs = %d, batch = %d, n = %d \n", src_outputs, batch, n);
+    int size = batch * src_outputs;
+    shortcut_multilayer_kernel << <cuda_gridsize(size), BLOCK, 0, get_cuda_stream() >> > (size, src_outputs, batch, n, outputs_of_layers_gpu, layers_output_gpu, out, in);
+    CHECK_CUDA(cudaPeekAtLastError());
+}
+
+
+
+__global__ void backward_shortcut_multilayer_kernel(int size, int src_outputs, int batch, int n, int *outputs_of_layers_gpu,
+    float **layers_delta_gpu, float *delta_out, float *delta_in)
+{
+    const int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
+    if (id >= size) return;
+
+    int src_id = id;
+    int src_i = src_id % src_outputs;
+    src_id /= src_outputs;
+    int src_b = src_id;
+
+    delta_out[id] += delta_in[id];
+
+    // layers
+    for (int i = 0; i < n; ++i) {
+        int add_outputs = outputs_of_layers_gpu[i];
+        if (src_i < add_outputs) {
+            int add_index = add_outputs*src_b + src_i;
+            int out_index = id;
+
+            float *layer_delta = layers_delta_gpu[i];
+            layer_delta[add_index] += delta_in[id];
+        }
+    }
+}
+
+extern "C" void backward_shortcut_multilayer_gpu(int src_outputs, int batch, int n, int *outputs_of_layers_gpu, float **layers_delta_gpu, float *delta_out, float *delta_in)
+{
+    //printf(" src_outputs = %d, batch = %d, n = %d \n", src_outputs, batch, n);
+    int size = batch * src_outputs;
+    backward_shortcut_multilayer_kernel << <cuda_gridsize(size), BLOCK, 0, get_cuda_stream() >> > (size, src_outputs, batch, n, outputs_of_layers_gpu, layers_delta_gpu, delta_out, delta_in);
+    CHECK_CUDA(cudaPeekAtLastError());
+}
+
 __global__ void shortcut_kernel(int size, int minw, int minh, int minc, int stride, int sample, int batch, int w1, int h1, int c1, float *add, int w2, int h2, int c2, float *out)
 {
     int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;

src/dark_cuda.c

@@ -369,6 +369,21 @@ float *cuda_make_array(float *x, size_t n)
     return x_gpu;
 }
 
+void **cuda_make_array_pointers(void **x, size_t n)
+{
+    void **x_gpu;
+    size_t size = sizeof(void*) * n;
+    cudaError_t status = cudaMalloc((void **)&x_gpu, size);
+    if (status != cudaSuccess) fprintf(stderr, " Try to set subdivisions=64 in your cfg-file. \n");
+    CHECK_CUDA(status);
+    if (x) {
+        status = cudaMemcpyAsync(x_gpu, x, size, cudaMemcpyDefault, get_cuda_stream());
+        CHECK_CUDA(status);
+    }
+    if (!x_gpu) error("Cuda malloc failed\n");
+    return x_gpu;
+}
+
 void cuda_random(float *x_gpu, size_t n)
 {
     static curandGenerator_t gen[16];

src/dark_cuda.h

@@ -62,6 +62,7 @@ extern "C" {
     float *cuda_make_array_pinned_preallocated(float *x, size_t n);
     float *cuda_make_array_pinned(float *x, size_t n);
     float *cuda_make_array(float *x, size_t n);
+    void **cuda_make_array_pointers(void **x, size_t n);
     int *cuda_make_int_array(size_t n);
 	int *cuda_make_int_array_new_api(int *x, size_t n);
     void cuda_push_array(float *x_gpu, float *x, size_t n);

src/layer.c

@@ -65,6 +65,8 @@ void free_layer_custom(layer l, int keep_cudnn_desc)
     if (l.indexes)            free(l.indexes);
     if (l.input_layers)       free(l.input_layers);
     if (l.input_sizes)        free(l.input_sizes);
+    if (l.layers_output)      free(l.layers_output);
+    if (l.layers_delta)       free(l.layers_delta);
     if (l.map)                free(l.map);
     if (l.rand)               free(l.rand);
     if (l.cost)               free(l.cost);
@@ -190,6 +192,9 @@ void free_layer_custom(layer l, int keep_cudnn_desc)
     if (l.rand_gpu)                cuda_free(l.rand_gpu);
     if (l.squared_gpu)             cuda_free(l.squared_gpu);
     if (l.norms_gpu)               cuda_free(l.norms_gpu);
+    if (l.input_sizes_gpu)         cuda_free(l.input_sizes_gpu);
+    if (l.layers_output_gpu)       cuda_free(l.layers_output_gpu);
+    if (l.layers_delta_gpu)        cuda_free(l.layers_delta_gpu);
 
     // CONV-LSTM
     if (l.f_gpu)                   cuda_free(l.f_gpu);

src/network.c

@@ -538,7 +538,7 @@ int resize_network(network *net, int w, int h)
         }else if(l.type == ROUTE){
             resize_route_layer(&l, net);
         }else if (l.type == SHORTCUT) {
-            resize_shortcut_layer(&l, w, h);
+            resize_shortcut_layer(&l, w, h, net);
         }else if (l.type == SCALE_CHANNELS) {
             resize_scale_channels_layer(&l, net);
         }else if (l.type == DROPOUT) {

src/network_kernels.cu

@@ -130,8 +130,8 @@ void forward_network_gpu(network net, network_state state)
         printf("\n\nSorted by time:\n");
         qsort(sorted_avg_time_per_layer, net.n, sizeof(time_benchmark_layers), time_comparator);
         for (i = 0; i < net.n; ++i) {
-            //printf("\layer %d - type: %d - avg_time %lf ms \n", avg_time_per_layer[i].layer_id, avg_time_per_layer[i].layer_type, avg_time_per_layer[i].time);
-            printf("\%d - layer %d - type: %d - avg_time %lf ms \n", i, sorted_avg_time_per_layer[i].layer_id, sorted_avg_time_per_layer[i].layer_type, sorted_avg_time_per_layer[i].time);
+            //printf("layer %d - type: %d - avg_time %lf ms \n", avg_time_per_layer[i].layer_id, avg_time_per_layer[i].layer_type, avg_time_per_layer[i].time);
+            printf("%d - layer %d - type: %d - avg_time %lf ms \n", i, sorted_avg_time_per_layer[i].layer_id, sorted_avg_time_per_layer[i].layer_type, sorted_avg_time_per_layer[i].time);
         }
     }
 

src/parser.c

@@ -817,15 +817,54 @@ layer parse_shortcut(list *options, size_params params, network net)
     ACTIVATION activation = get_activation(activation_s);
 
     int assisted_excitation = option_find_float_quiet(options, "assisted_excitation", 0);
+    //char *l = option_find(options, "from");
+    //int index = atoi(l);
+    //if(index < 0) index = params.index + index;
+
     char *l = option_find(options, "from");
-    int index = atoi(l);
-    if(index < 0) index = params.index + index;
+    int len = strlen(l);
+    if (!l) error("Route Layer must specify input layers: from = ...");
+    int n = 1;
+    int i;
+    for (i = 0; i < len; ++i) {
+        if (l[i] == ',') ++n;
+    }
 
-    int batch = params.batch;
-    layer from = net.layers[index];
-    if (from.antialiasing) from = *from.input_layer;
+    int* layers = (int*)calloc(n, sizeof(int));
+    int* sizes = (int*)calloc(n, sizeof(int));
+    float **layers_output = (float **)calloc(n, sizeof(float *));
+    float **layers_delta = (float **)calloc(n, sizeof(float *));
+    float **layers_output_gpu = (float **)calloc(n, sizeof(float *));
+    float **layers_delta_gpu = (float **)calloc(n, sizeof(float *));
+
+    for (i = 0; i < n; ++i) {
+        int index = atoi(l);
+        l = strchr(l, ',') + 1;
+        if (index < 0) index = params.index + index;
+        layers[i] = index;
+        sizes[i] = params.net.layers[index].outputs;
+        layers_output[i] = params.net.layers[index].output;
+        layers_delta[i] = params.net.layers[index].delta;
+
+    }
+
+#ifdef GPU
+    for (i = 0; i < n; ++i) {
+        layers_output_gpu[i] = params.net.layers[layers[i]].output_gpu;
+        layers_delta_gpu[i] = params.net.layers[layers[i]].delta_gpu;
+    }
+#endif// GPU
+
+    layer s = make_shortcut_layer(params.batch, n, layers, sizes, params.w, params.h, params.c, layers_output, layers_delta, layers_output_gpu, layers_delta_gpu, activation, params.train);
 
-    layer s = make_shortcut_layer(batch, index, params.w, params.h, params.c, from.out_w, from.out_h, from.out_c, assisted_excitation, activation, params.train);
+    for (i = 0; i < n; ++i) {
+        int index = layers[i];
+        assert(params.w == net.layers[index].out_w && params.h == net.layers[index].out_h);
+
+        if (params.w != net.layers[index].out_w || params.h != net.layers[index].out_h || params.c != net.layers[index].out_c)
+            fprintf(stderr, " w = %d, w2 = %d, h = %d, h2 = %d, c = %d, c2 = %d \n",
+                params.w, net.layers[index].out_w, params.h, net.layers[index].out_h, params.c, params.net.layers[index].out_c);
+    }
 
     return s;
 }

src/shortcut_layer.c

@@ -6,29 +6,36 @@
 #include <stdio.h>
 #include <assert.h>
 
-layer make_shortcut_layer(int batch, int index, int w, int h, int c, int w2, int h2, int c2, int assisted_excitation, ACTIVATION activation, int train)
+layer make_shortcut_layer(int batch, int n, int *input_layers, int* input_sizes, int w, int h, int c,
+    float **layers_output, float **layers_delta, float **layers_output_gpu, float **layers_delta_gpu, ACTIVATION activation, int train)
 {
-    if(assisted_excitation) fprintf(stderr, "Shortcut Layer - AE: %d\n", index);
-    else fprintf(stderr,"Shortcut Layer: %d\n", index);
+    fprintf(stderr, "Shortcut Layer: ");
+    int i;
+    for(i = 0; i < n; ++i) fprintf(stderr, "%d, ", input_layers[i]);
+
     layer l = { (LAYER_TYPE)0 };
     l.train = train;
     l.type = SHORTCUT;
     l.batch = batch;
     l.activation = activation;
-    l.w = w2;
-    l.h = h2;
-    l.c = c2;
-    l.out_w = w;
-    l.out_h = h;
-    l.out_c = c;
+    l.n = n;
+    l.input_layers = input_layers;
+    l.input_sizes = input_sizes;
+    l.layers_output = layers_output;
+    l.layers_delta = layers_delta;
+
+    //l.w = w2;
+    //l.h = h2;
+    //l.c = c2;
+    l.w = l.out_w = w;
+    l.h = l.out_h = h;
+    l.c = l.out_c = c;
     l.outputs = w*h*c;
     l.inputs = l.outputs;
 
-    l.assisted_excitation = assisted_excitation;
-
-    if(w != w2 || h != h2 || c != c2) fprintf(stderr, " w = %d, w2 = %d, h = %d, h2 = %d, c = %d, c2 = %d \n", w, w2, h, h2, c, c2);
+    //if(w != w2 || h != h2 || c != c2) fprintf(stderr, " w = %d, w2 = %d, h = %d, h2 = %d, c = %d, c2 = %d \n", w, w2, h, h2, c, c2);
 
-    l.index = index;
+    l.index = l.input_layers[0];
 
     if (train) l.delta = (float*)calloc(l.outputs * batch, sizeof(float));
     l.output = (float*)calloc(l.outputs * batch, sizeof(float));
@@ -47,17 +54,18 @@ layer make_shortcut_layer(int batch, int index, int w, int h, int c, int w2, int
 
     if (train) l.delta_gpu =  cuda_make_array(l.delta, l.outputs*batch);
     l.output_gpu = cuda_make_array(l.output, l.outputs*batch);
-    if (l.assisted_excitation)
-    {
-        const int size = l.out_w * l.out_h * l.batch;
-        l.gt_gpu = cuda_make_array(NULL, size);
-        l.a_avg_gpu = cuda_make_array(NULL, size);
-    }
+
+    l.input_sizes_gpu = cuda_make_array(input_sizes, l.n);
+    l.layers_output_gpu = cuda_make_array_pointers(layers_output_gpu, l.n);
+    l.layers_delta_gpu = cuda_make_array_pointers(layers_delta_gpu, l.n);
 #endif  // GPU
+
+    l.bflops = l.out_w * l.out_h * l.out_c * l.n / 1000000000.;
+    fprintf(stderr, " outputs:%4d x%4d x%4d %5.3f BF\n", l.out_w, l.out_h, l.out_c, l.bflops);
     return l;
 }
 
-void resize_shortcut_layer(layer *l, int w, int h)
+void resize_shortcut_layer(layer *l, int w, int h, network *net)
 {
     //assert(l->w == l->out_w);
     //assert(l->h == l->out_h);
@@ -68,6 +76,13 @@ void resize_shortcut_layer(layer *l, int w, int h)
     if (l->train) l->delta = (float*)realloc(l->delta, l->outputs * l->batch * sizeof(float));
     l->output = (float*)realloc(l->output, l->outputs * l->batch * sizeof(float));
 
+    int i;
+    for (i = 0; i < l->n; ++i) {
+        int index = l->input_layers[i];
+        l->input_sizes[i] = net->layers[index].outputs;
+        assert(l->w == net->layers[index].w && l->h == net->layers[index].h);
+    }
+
 #ifdef GPU
     cuda_free(l->output_gpu);
     l->output_gpu = cuda_make_array(l->output, l->outputs*l->batch);
@@ -82,7 +97,11 @@ void resize_shortcut_layer(layer *l, int w, int h)
 
 void forward_shortcut_layer(const layer l, network_state state)
 {
-    if (l.w == l.out_w && l.h == l.out_h && l.c == l.out_c) {
+    int from_w = state.net.layers[l.index].w;
+    int from_h = state.net.layers[l.index].h;
+    int from_c = state.net.layers[l.index].c;
+
+    if (l.n == 1 && from_w == l.w && from_h == l.h && from_c == l.c) {
         int size = l.batch * l.w * l.h * l.c;
         int i;
         #pragma omp parallel for
@@ -90,16 +109,16 @@ void forward_shortcut_layer(const layer l, network_state state)
             l.output[i] = state.input[i] + state.net.layers[l.index].output[i];
     }
     else {
-        copy_cpu(l.outputs*l.batch, state.input, 1, l.output, 1);
-        shortcut_cpu(l.batch, l.w, l.h, l.c, state.net.layers[l.index].output, l.out_w, l.out_h, l.out_c, l.output);
+        shortcut_multilayer_cpu(l.outputs * l.batch, l.outputs, l.batch, l.n, l.input_sizes, l.layers_output, l.output, state.input);
     }
 
+    //copy_cpu(l.outputs*l.batch, state.input, 1, l.output, 1);
+    //shortcut_cpu(l.batch, from_w, from_h, from_c, state.net.layers[l.index].output, l.out_w, l.out_h, l.out_c, l.output);
+
     //activate_array(l.output, l.outputs*l.batch, l.activation);
     if (l.activation == SWISH) activate_array_swish(l.output, l.outputs*l.batch, l.activation_input, l.output);
     else if (l.activation == MISH) activate_array_mish(l.output, l.outputs*l.batch, l.activation_input, l.output);
     else activate_array_cpu_custom(l.output, l.outputs*l.batch, l.activation);
-
-    if (l.assisted_excitation && state.train) assisted_excitation_forward(l, state);
 }
 
 void backward_shortcut_layer(const layer l, network_state state)
@@ -108,8 +127,11 @@ void backward_shortcut_layer(const layer l, network_state state)
     else if (l.activation == MISH) gradient_array_mish(l.outputs*l.batch, l.activation_input, l.delta);
     else gradient_array(l.output, l.outputs*l.batch, l.activation, l.delta);
 
-    axpy_cpu(l.outputs*l.batch, 1, l.delta, 1, state.delta, 1);
-    shortcut_cpu(l.batch, l.out_w, l.out_h, l.out_c, l.delta, l.w, l.h, l.c, state.net.layers[l.index].delta);
+    backward_shortcut_multilayer_cpu(l.outputs * l.batch, l.outputs, l.batch, l.n, l.input_sizes,
+        l.layers_delta, state.delta, l.delta);
+
+    //axpy_cpu(l.outputs*l.batch, 1, l.delta, 1, state.delta, 1);
+    //shortcut_cpu(l.batch, l.out_w, l.out_h, l.out_c, l.delta, l.w, l.h, l.c, state.net.layers[l.index].delta);
 }
 
 #ifdef GPU
@@ -118,13 +140,25 @@ void forward_shortcut_layer_gpu(const layer l, network_state state)
     //copy_ongpu(l.outputs*l.batch, state.input, 1, l.output_gpu, 1);
     //simple_copy_ongpu(l.outputs*l.batch, state.input, l.output_gpu);
     //shortcut_gpu(l.batch, l.w, l.h, l.c, state.net.layers[l.index].output_gpu, l.out_w, l.out_h, l.out_c, l.output_gpu);
-    input_shortcut_gpu(state.input, l.batch, l.w, l.h, l.c, state.net.layers[l.index].output_gpu, l.out_w, l.out_h, l.out_c, l.output_gpu);
+
+    //input_shortcut_gpu(state.input, l.batch, l.w, l.h, l.c, state.net.layers[l.index].output_gpu, l.out_w, l.out_h, l.out_c, l.output_gpu);
+
+    //-----------
+    //if (l.outputs == l.input_sizes[0])
+    //if(l.n == 1)
+    //{
+    //    input_shortcut_gpu(state.input, l.batch, state.net.layers[l.index].w, state.net.layers[l.index].h, state.net.layers[l.index].c,
+    //        state.net.layers[l.index].output_gpu, l.out_w, l.out_h, l.out_c, l.output_gpu);
+    //}
+    //else
+    {
+        shortcut_multilayer_gpu(l.outputs, l.batch, l.n, l.input_sizes_gpu, l.layers_output_gpu, l.output_gpu, state.input);
+    }
 
     if (l.activation == SWISH) activate_array_swish_ongpu(l.output_gpu, l.outputs*l.batch, l.activation_input_gpu, l.output_gpu);
     else if (l.activation == MISH) activate_array_mish_ongpu(l.output_gpu, l.outputs*l.batch, l.activation_input_gpu, l.output_gpu);
     else activate_array_ongpu(l.output_gpu, l.outputs*l.batch, l.activation);
 
-    if (l.assisted_excitation && state.train) assisted_excitation_forward_gpu(l, state);
 }
 
 void backward_shortcut_layer_gpu(const layer l, network_state state)
@@ -133,7 +167,9 @@ void backward_shortcut_layer_gpu(const layer l, network_state state)
     else if (l.activation == MISH) gradient_array_mish_ongpu(l.outputs*l.batch, l.activation_input_gpu, l.delta_gpu);
     else gradient_array_ongpu(l.output_gpu, l.outputs*l.batch, l.activation, l.delta_gpu);
 
-    axpy_ongpu(l.outputs*l.batch, 1, l.delta_gpu, 1, state.delta, 1);
-    shortcut_gpu(l.batch, l.out_w, l.out_h, l.out_c, l.delta_gpu, l.w, l.h, l.c, state.net.layers[l.index].delta_gpu);
+    backward_shortcut_multilayer_gpu(l.outputs, l.batch, l.n, l.input_sizes_gpu, l.layers_delta_gpu, state.delta, l.delta_gpu);
+
+    //axpy_ongpu(l.outputs*l.batch, 1, l.delta_gpu, 1, state.delta, 1);
+    //shortcut_gpu(l.batch, l.out_w, l.out_h, l.out_c, l.delta_gpu, l.w, l.h, l.c, state.net.layers[l.index].delta_gpu);
 }
 #endif

src/shortcut_layer.h

@@ -7,10 +7,11 @@
 #ifdef __cplusplus
 extern "C" {
 #endif
-layer make_shortcut_layer(int batch, int index, int w, int h, int c, int w2, int h2, int c2, int assisted_excitation, ACTIVATION activation, int train);
+layer make_shortcut_layer(int batch, int n, int *input_layers, int* input_sizes, int w, int h, int c,
+    float **layers_output, float **layers_delta, float **layers_output_gpu, float **layers_delta_gpu, ACTIVATION activation, int train);
 void forward_shortcut_layer(const layer l, network_state state);
 void backward_shortcut_layer(const layer l, network_state state);
-void resize_shortcut_layer(layer *l, int w, int h);
+void resize_shortcut_layer(layer *l, int w, int h, network *net);
 
 #ifdef GPU
 void forward_shortcut_layer_gpu(const layer l, network_state state);