diff --git a/src/gemm.c b/src/gemm.c index 61f574bb1a7..0b06d987086 100644 --- a/src/gemm.c +++ b/src/gemm.c @@ -1031,7 +1031,7 @@ void convolution_2d(int w, int h, int ksize, int n, int c, int pad, int stride, } //for (i = 0; i < w*h*c; i += 8) { - //*((__m256*)&input[i]) = _mm256_and_ps(*((__m256*)&input[i]), _mm256_castsi256_ps(all256_sing1)); + //(*(__m256*)&input[i]) = _mm256_and_ps(*((__m256*)&input[i]), _mm256_castsi256_ps(all256_sing1)); //} @@ -1124,7 +1124,7 @@ void convolution_2d(int w, int h, int ksize, int n, int c, int pad, int stride, //__m256 out = *((__m256*)&output[output_index]); //out = _mm256_add_ps(out, sum256); - //*((__m256*)&output[output_index]) = out; + //(*(__m256*)&output[output_index]) = out; *((__m256*)&output[output_index]) = sum256; //_mm256_storeu_ps(&C[i*ldc + j], result256); diff --git a/src/im2col_kernels.cu b/src/im2col_kernels.cu index 54961d3a2d1..f924b3e5cd2 100644 --- a/src/im2col_kernels.cu +++ b/src/im2col_kernels.cu @@ -60,7 +60,7 @@ __global__ void im2col_gpu_kernel(const int n, const float* data_im, data_im_ptr[i * width + j] : 0; //data_im[(channel_in * height + h_in) * width + w_in + i * width + j]; - //*data_col_ptr = data_im_ptr[ii * width + jj]; + //(*data_col_ptr) = data_im_ptr[ii * width + jj]; data_col_ptr += height_col * width_col; } @@ -122,7 +122,7 @@ __global__ void im2col_align_gpu_kernel(const int n, const float* data_im, *data_col_ptr = val; //tmp_s[0] = val; - //*data_col_ptr = (h >= 0 && w >= 0 && h < height && w < width) ? + //(*data_col_ptr) = (h >= 0 && w >= 0 && h < height && w < width) ? // data_im_ptr[i * width + j] : 0; //float src_val = (h >= 0 && w >= 0 && h < height && w < width) ? data_im_ptr[i * width + j] : 0; @@ -178,11 +178,11 @@ __global__ void im2col_align_gpu_kernel(const int n, const float* data_im, int out_index = (channel_out + i*ksize + j) * bit_align + pre_out_index;// h_out * width_col + w_out; data_col[out_index] = val; - //*data_col_ptr = val; + //(*data_col_ptr) = val; //dst_s[threadIdx.x] = val; //tmp_s[0] = val; - //*data_col_ptr = (h >= 0 && w >= 0 && h < height && w < width) ? + //(*data_col_ptr) = (h >= 0 && w >= 0 && h < height && w < width) ? // data_im_ptr[i * width + j] : 0; //float src_val = (h >= 0 && w >= 0 && h < height && w < width) ? data_im_ptr[i * width + j] : 0; @@ -1175,7 +1175,7 @@ __global__ void gemm_nn_custom_bin_mean_transposed_gpu_kernel(int M, int N, int //if (i_cur < M && (index % N == 0 || threadIdx.x == 0)) { //for (int k = 0; k < K; k += 64) { // l.size*l.size*l.c - one filter size [27 - 9216] - //*((uint64_t *)(A_s + (local_i*lda + k) / 8)) = *((uint64_t *)(A + (i_cur*lda + k) / 8)); // weights + //(*(uint64_t *)(A_s + (local_i*lda + k) / 8)) = *((uint64_t *)(A + (i_cur*lda + k) / 8)); // weights // } //} diff --git a/src/yolo_layer.c b/src/yolo_layer.c index be56972c8bc..4e3f54e77aa 100644 --- a/src/yolo_layer.c +++ b/src/yolo_layer.c @@ -597,7 +597,7 @@ void forward_yolo_layer(const layer l, network_state state) if (count == 0) count = 1; if (class_count == 0) class_count = 1; - //*(l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2); + //(*l.cost) = pow(mag_array(l.delta, l.outputs * l.batch), 2); //printf("Region %d Avg IOU: %f, Class: %f, Obj: %f, No Obj: %f, .5R: %f, .75R: %f, count: %d\n", state.index, avg_iou / count, avg_cat / class_count, avg_obj / count, avg_anyobj / (l.w*l.h*l.n*l.batch), recall / count, recall75 / count, count); int stride = l.w*l.h;