| src/im2col.cl | ●●●●● patch | view | raw | blame | history | |
| src/network.c | ●●●●● patch | view | raw | blame | history |
src/im2col.cl
@@ -28,7 +28,7 @@ int im_row = h_offset + h * stride - pad; int im_col = w_offset + w * stride - pad; int im_index = im_col + width*(im_row + height*(im_channel+batch*channels)); int im_index = im_col + width*(im_row + height*(im_channel+b*channels)); float val = (im_row < 0 || im_col < 0 || im_row >= height || im_col >= width) ? 0 : im[im_index]; data_col[col_index] = val; @@ -61,7 +61,7 @@ int im_row = h_offset + h * stride; int im_col = w_offset + w * stride; int im_index = im_col + width*(im_row + height*(im_channel+batch*channels)); int im_index = im_col + width*(im_row + height*(im_channel+b*channels)); float val = (im_row < 0 || im_col < 0 || im_row >= height || im_col >= width) ? 0 : im[im_index]; data_col[col_index] = val; src/network.c
@@ -38,7 +38,7 @@ //printf("start\n"); int i; for(i = 0; i < net.n; ++i){ clock_t time = clock(); //clock_t time = clock(); if(net.types[i] == CONVOLUTIONAL){ convolutional_layer layer = *(convolutional_layer *)net.layers[i]; forward_convolutional_layer_gpu(layer, input); @@ -63,7 +63,7 @@ forward_softmax_layer_gpu(layer, input); input = layer.output_cl; } printf("%d %f\n", i, sec(clock()-time)); //printf("%d %f\n", i, sec(clock()-time)); /* else if(net.types[i] == CROP){ crop_layer layer = *(crop_layer *)net.layers[i]; @@ -85,7 +85,7 @@ cl_mem prev_input; cl_mem prev_delta; for(i = net.n-1; i >= 0; --i){ clock_t time = clock(); //clock_t time = clock(); if(i == 0){ prev_input = input; prev_delta = 0; @@ -113,7 +113,7 @@ softmax_layer layer = *(softmax_layer *)net.layers[i]; backward_softmax_layer_gpu(layer, prev_delta); } printf("back: %d %f\n", i, sec(clock()-time)); //printf("back: %d %f\n", i, sec(clock()-time)); } }