reverting merge

Edmond Yoo

2018-09-16 721d5e410e5a2b47ec2de722cc02fd6b9ed6b2fd

 src/network_kernels.cu

@@ -22,7 +22,6 @@
#include "region_layer.h"
#include "convolutional_layer.h"
#include "activation_layer.h"
#include "deconvolutional_layer.h"
#include "maxpool_layer.h"
#include "reorg_layer.h"
#include "avgpool_layer.h"
@@ -37,6 +36,10 @@
#include "blas.h"
}

#ifdef OPENCV
#include "opencv2/highgui/highgui_c.h"
#endif

float * get_network_output_gpu_layer(network net, int i);
float * get_network_delta_gpu_layer(network net, int i);
float * get_network_output_gpu(network net);
@@ -51,50 +54,25 @@
        if(l.delta_gpu){
            fill_ongpu(l.outputs * l.batch, 0, l.delta_gpu, 1);
        }
        if(l.type == CONVOLUTIONAL){
            forward_convolutional_layer_gpu(l, state);
        } else if(l.type == DECONVOLUTIONAL){
            forward_deconvolutional_layer_gpu(l, state);
        } else if(l.type == ACTIVE){
            forward_activation_layer_gpu(l, state);
        } else if(l.type == LOCAL){
            forward_local_layer_gpu(l, state);
        } else if(l.type == DETECTION){
            forward_detection_layer_gpu(l, state);
        } else if(l.type == REGION){
            forward_region_layer_gpu(l, state);
        } else if(l.type == CONNECTED){
            forward_connected_layer_gpu(l, state);
        } else if(l.type == RNN){
            forward_rnn_layer_gpu(l, state);
        } else if(l.type == GRU){
            forward_gru_layer_gpu(l, state);
        } else if(l.type == CRNN){
            forward_crnn_layer_gpu(l, state);
        } else if(l.type == CROP){
            forward_crop_layer_gpu(l, state);
        } else if(l.type == COST){
            forward_cost_layer_gpu(l, state);
        } else if(l.type == SOFTMAX){
            forward_softmax_layer_gpu(l, state);
        } else if(l.type == NORMALIZATION){
            forward_normalization_layer_gpu(l, state);
        } else if(l.type == BATCHNORM){
            forward_batchnorm_layer_gpu(l, state);
        } else if(l.type == MAXPOOL){
            forward_maxpool_layer_gpu(l, state);
        } else if(l.type == REORG){
            forward_reorg_layer_gpu(l, state);
        } else if(l.type == AVGPOOL){
            forward_avgpool_layer_gpu(l, state);
        } else if(l.type == DROPOUT){
            forward_dropout_layer_gpu(l, state);
        } else if(l.type == ROUTE){
            forward_route_layer_gpu(l, net);
        } else if(l.type == SHORTCUT){
            forward_shortcut_layer_gpu(l, state);
        }
        l.forward_gpu(l, state);
        if(net.wait_stream)
            cudaStreamSynchronize(get_cuda_stream());
        state.input = l.output_gpu;
/*
        cuda_pull_array(l.output_gpu, l.output, l.batch*l.outputs);
        if (l.out_w >= 0 && l.out_h >= 1 && l.c >= 3) {
            int j;
            for (j = 0; j < l.out_c; ++j) {
                image img = make_image(l.out_w, l.out_h, 3);
                memcpy(img.data, l.output+ l.out_w*l.out_h*j, l.out_w*l.out_h * 1 * sizeof(float));
                char buff[256];
                sprintf(buff, "layer-%d slice-%d", i, j);
                show_image(img, buff);
            }
            cvWaitKey(0); // wait press-key in console
            cvDestroyAllWindows();
        }
*/
    }
}

@@ -107,6 +85,7 @@
    for(i = net.n-1; i >= 0; --i){
        state.index = i;
        layer l = net.layers[i];
        if (l.stopbackward) break;
        if(i == 0){
            state.input = original_input;
            state.delta = original_delta;
@@ -115,71 +94,21 @@
            state.input = prev.output_gpu;
            state.delta = prev.delta_gpu;
        }
        if(l.type == CONVOLUTIONAL){
            backward_convolutional_layer_gpu(l, state);
        } else if(l.type == DECONVOLUTIONAL){
            backward_deconvolutional_layer_gpu(l, state);
        } else if(l.type == ACTIVE){
            backward_activation_layer_gpu(l, state);
        } else if(l.type == LOCAL){
            backward_local_layer_gpu(l, state);
        } else if(l.type == MAXPOOL){
            if(i != 0) backward_maxpool_layer_gpu(l, state);
        } else if(l.type == REORG){
            backward_reorg_layer_gpu(l, state);
        } else if(l.type == AVGPOOL){
            if(i != 0) backward_avgpool_layer_gpu(l, state);
        } else if(l.type == DROPOUT){
            backward_dropout_layer_gpu(l, state);
        } else if(l.type == DETECTION){
            backward_detection_layer_gpu(l, state);
        } else if(l.type == REGION){
            backward_region_layer_gpu(l, state);
        } else if(l.type == NORMALIZATION){
            backward_normalization_layer_gpu(l, state);
        } else if(l.type == BATCHNORM){
            backward_batchnorm_layer_gpu(l, state);
        } else if(l.type == SOFTMAX){
            if(i != 0) backward_softmax_layer_gpu(l, state);
        } else if(l.type == CONNECTED){
            backward_connected_layer_gpu(l, state);
        } else if(l.type == RNN){
            backward_rnn_layer_gpu(l, state);
        } else if(l.type == GRU){
            backward_gru_layer_gpu(l, state);
        } else if(l.type == CRNN){
            backward_crnn_layer_gpu(l, state);
        } else if(l.type == COST){
            backward_cost_layer_gpu(l, state);
        } else if(l.type == ROUTE){
            backward_route_layer_gpu(l, net);
        } else if(l.type == SHORTCUT){
            backward_shortcut_layer_gpu(l, state);
        }
        l.backward_gpu(l, state);
    }
}

void update_network_gpu(network net)
{
    cuda_set_device(net.gpu_index);
    int i;
    int update_batch = net.batch*net.subdivisions;
    float rate = get_current_rate(net);
    for(i = 0; i < net.n; ++i){
        layer l = net.layers[i];
        if(l.type == CONVOLUTIONAL){
            update_convolutional_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        } else if(l.type == DECONVOLUTIONAL){
            update_deconvolutional_layer_gpu(l, rate, net.momentum, net.decay);
        } else if(l.type == CONNECTED){
            update_connected_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        } else if(l.type == GRU){
            update_gru_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        } else if(l.type == RNN){
            update_rnn_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        } else if(l.type == CRNN){
            update_crnn_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        } else if(l.type == LOCAL){
            update_local_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
        l.t = get_current_batch(net);
        if(l.update_gpu){
            l.update_gpu(l, update_batch, rate, net.momentum, net.decay);
        }
    }
}
@@ -203,7 +132,15 @@
    state.delta = 0;
    state.truth = *net.truth_gpu;
    state.train = 1;
#ifdef CUDNN_HALF
    int i;
    for (i = 0; i < net.n; ++i) {
        layer l = net.layers[i];
        cuda_convert_f32_to_f16(l.weights_gpu, l.c*l.n*l.size*l.size, l.weights_gpu16);
    }
#endif
    forward_network_gpu(net, state);
    //cudaStreamSynchronize(get_cuda_stream());
    backward_network_gpu(net, state);
}

@@ -271,20 +208,9 @@
{
    int update_batch = net.batch*net.subdivisions;
    float rate = get_current_rate(net);
    if(l.type == CONVOLUTIONAL){
        update_convolutional_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    } else if(l.type == DECONVOLUTIONAL){
        update_deconvolutional_layer_gpu(l, rate, net.momentum, net.decay);
    } else if(l.type == CONNECTED){
        update_connected_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    } else if(l.type == RNN){
        update_rnn_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    } else if(l.type == GRU){
        update_gru_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    } else if(l.type == CRNN){
        update_crnn_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    } else if(l.type == LOCAL){
        update_local_layer_gpu(l, update_batch, rate, net.momentum, net.decay);
    l.t = get_current_batch(net);
    if(l.update_gpu){
        l.update_gpu(l, update_batch, rate, net.momentum, net.decay);
    }
}

@@ -463,14 +389,17 @@
    }
    for(i = 0; i < n; ++i){
        pthread_join(threads[i], 0);
        printf("%f\n", errors[i]);
        //printf("%f\n", errors[i]);
        sum += errors[i];
    }
    //cudaDeviceSynchronize();
    if (get_current_batch(nets[0]) % interval == 0) {
        printf("Syncing... ");
        fflush(stdout);
        sync_nets(nets, n, interval);
        printf("Done!\n");
    }
    //cudaDeviceSynchronize();
    free(threads);
    free(errors);
    return (float)sum/(n);
@@ -479,7 +408,7 @@
float *get_network_output_layer_gpu(network net, int i)
{
    layer l = net.layers[i];
    cuda_pull_array(l.output_gpu, l.output, l.outputs*l.batch);
    if(l.type != REGION) cuda_pull_array(l.output_gpu, l.output, l.outputs*l.batch);
    return l.output;
}

@@ -492,6 +421,8 @@

float *network_predict_gpu(network net, float *input)
{
    if (net.gpu_index != cuda_get_device())
        cuda_set_device(net.gpu_index);
    int size = get_network_input_size(net) * net.batch;
    network_state state;
    state.index = 0;