shortcut layers, msr networks

Joseph Redmon

2015-12-14 db0397cfaaf488364e3d2e1669dfefae2ee6ea73

 src/softmax_layer_kernels.cu

@@ -1,11 +1,13 @@
#include "cuda_runtime.h"
#include "curand.h"
#include "cublas_v2.h"

extern "C" {
#include "softmax_layer.h"
#include "cuda.h"
#include "blas.h"
}

#define BLOCK 256

__global__ void forward_softmax_layer_kernel(int n, int batch, float *input, float *output)
{
    int b = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;