model/examples/nightmare.c

#include "darknet.h"

#include <math.h>

// ./darknet nightmare cfg/extractor.recon.cfg ~/trained/yolo-coco.conv frame6.png -reconstruct -iters 500 -i 3 -lambda .1 -rate .01 -smooth 2

float abs_mean(float *x, int n)
{
    int i;
    float sum = 0;
    for (i = 0; i < n; ++i){
        sum += fabs(x[i]);
    }
    return sum/n;
}

void calculate_loss(float *output, float *delta, int n, float thresh)
{
    int i;
    float mean = mean_array(output, n); 
    float var = variance_array(output, n);
    for(i = 0; i < n; ++i){
        if(delta[i] > mean + thresh*sqrt(var)) delta[i] = output[i];
        else delta[i] = 0;
    }
}

void optimize_picture(network *net, image orig, int max_layer, float scale, float rate, float thresh, int norm)
{
    //scale_image(orig, 2);
    //translate_image(orig, -1);
    net->n = max_layer + 1;

    int dx = rand()%16 - 8;
    int dy = rand()%16 - 8;
    int flip = rand()%2;

    image crop = crop_image(orig, dx, dy, orig.w, orig.h);
    image im = resize_image(crop, (int)(orig.w * scale), (int)(orig.h * scale));
    if(flip) flip_image(im);

    resize_network(net, im.w, im.h);
    layer last = net->layers[net->n-1];
    //net->layers[net->n - 1].activation = LINEAR;

    image delta = make_image(im.w, im.h, im.c);

#ifdef GPU
    net->delta_gpu = cuda_make_array(delta.data, im.w*im.h*im.c);
    copy_cpu(net->inputs, im.data, 1, net->input, 1);

    forward_network_gpu(net);
    copy_gpu(last.outputs, last.output_gpu, 1, last.delta_gpu, 1);

    cuda_pull_array(last.delta_gpu, last.delta, last.outputs);
    calculate_loss(last.delta, last.delta, last.outputs, thresh);
    cuda_push_array(last.delta_gpu, last.delta, last.outputs);

    backward_network_gpu(net);

    cuda_pull_array(net->delta_gpu, delta.data, im.w*im.h*im.c);
    cuda_free(net->delta_gpu);
    net->delta_gpu = 0;
#else
    printf("\nnet: %d %d %d im: %d %d %d\n", net->w, net->h, net->inputs, im.w, im.h, im.c);
    copy_cpu(net->inputs, im.data, 1, net->input, 1);
    net->delta = delta.data;
    forward_network(net);
    copy_cpu(last.outputs, last.output, 1, last.delta, 1);
    calculate_loss(last.output, last.delta, last.outputs, thresh);
    backward_network(net);
#endif

    if(flip) flip_image(delta);
    //normalize_array(delta.data, delta.w*delta.h*delta.c);
    image resized = resize_image(delta, orig.w, orig.h);
    image out = crop_image(resized, -dx, -dy, orig.w, orig.h);

    /*
       image g = grayscale_image(out);
       free_image(out);
       out = g;
     */

    //rate = rate / abs_mean(out.data, out.w*out.h*out.c);
    image gray = make_image(out.w, out.h, out.c);
    fill_image(gray, .5);
    axpy_cpu(orig.w*orig.h*orig.c, -1, orig.data, 1, gray.data, 1);
    axpy_cpu(orig.w*orig.h*orig.c, .1, gray.data, 1, out.data, 1);

    if(norm) normalize_array(out.data, out.w*out.h*out.c);
    axpy_cpu(orig.w*orig.h*orig.c, rate, out.data, 1, orig.data, 1);

    /*
       normalize_array(orig.data, orig.w*orig.h*orig.c);
       scale_image(orig, sqrt(var));
       translate_image(orig, mean);
     */

    //translate_image(orig, 1);
    //scale_image(orig, .5);
    //normalize_image(orig);

    constrain_image(orig);

    free_image(crop);
    free_image(im);
    free_image(delta);
    free_image(resized);
    free_image(out);

}

void smooth(image recon, image update, float lambda, int num)
{
    int i, j, k;
    int ii, jj;
    for(k = 0; k < recon.c; ++k){
        for(j = 0; j < recon.h; ++j){
            for(i = 0; i < recon.w; ++i){
                int out_index = i + recon.w*(j + recon.h*k);
                for(jj = j-num; jj <= j + num && jj < recon.h; ++jj){
                    if (jj < 0) continue;
                    for(ii = i-num; ii <= i + num && ii < recon.w; ++ii){
                        if (ii < 0) continue;
                        int in_index = ii + recon.w*(jj + recon.h*k);
                        update.data[out_index] += lambda * (recon.data[in_index] - recon.data[out_index]);
                    }
                }
            }
        }
    }
}

void reconstruct_picture(network *net, float *features, image recon, image update, float rate, float momentum, float lambda, int smooth_size, int iters)
{
    int iter = 0;
    for (iter = 0; iter < iters; ++iter) {
        image delta = make_image(recon.w, recon.h, recon.c);

#ifdef GPU
        layer l = get_network_output_layer(net);
        cuda_push_array(net->input_gpu, recon.data, recon.w*recon.h*recon.c);
        //cuda_push_array(net->truth_gpu, features, net->truths);
        net->delta_gpu = cuda_make_array(delta.data, delta.w*delta.h*delta.c);

        forward_network_gpu(net);
        cuda_push_array(l.delta_gpu, features, l.outputs);
        axpy_gpu(l.outputs, -1, l.output_gpu, 1, l.delta_gpu, 1);
        backward_network_gpu(net);

        cuda_pull_array(net->delta_gpu, delta.data, delta.w*delta.h*delta.c);

        cuda_free(net->delta_gpu);
#else
        net->input = recon.data;
        net->delta = delta.data;
        net->truth = features;

        forward_network(net);
        backward_network(net);
#endif

        //normalize_array(delta.data, delta.w*delta.h*delta.c);
        axpy_cpu(recon.w*recon.h*recon.c, 1, delta.data, 1, update.data, 1);
        //smooth(recon, update, lambda, smooth_size);

        axpy_cpu(recon.w*recon.h*recon.c, rate, update.data, 1, recon.data, 1);
        scal_cpu(recon.w*recon.h*recon.c, momentum, update.data, 1);

        float mag = mag_array(delta.data, recon.w*recon.h*recon.c);
        printf("mag: %f\n", mag);
        //scal_cpu(recon.w*recon.h*recon.c, 600/mag, recon.data, 1);

        constrain_image(recon);
        free_image(delta);
    }
}

/*
void run_lsd(int argc, char **argv)
{
    srand(0);
    if(argc < 3){
        fprintf(stderr, "usage: %s %s [cfg] [weights] [image] [options! (optional)]\n", argv[0], argv[1]);
        return;
    }

    char *cfg = argv[2];
    char *weights = argv[3];
    char *input = argv[4];

    int norm = find_int_arg(argc, argv, "-norm", 1);
    int rounds = find_int_arg(argc, argv, "-rounds", 1);
    int iters = find_int_arg(argc, argv, "-iters", 10);
    float rate = find_float_arg(argc, argv, "-rate", .04);
    float momentum = find_float_arg(argc, argv, "-momentum", .9);
    float lambda = find_float_arg(argc, argv, "-lambda", .01);
    char *prefix = find_char_arg(argc, argv, "-prefix", 0);
    int reconstruct = find_arg(argc, argv, "-reconstruct");
    int smooth_size = find_int_arg(argc, argv, "-smooth", 1);

    network net = parse_network_cfg(cfg);
    load_weights(&net, weights);
    char *cfgbase = basecfg(cfg);
    char *imbase = basecfg(input);

    set_batch_network(&net, 1);
    image im = load_image_color(input, 0, 0);

    float *features = 0;
    image update;
    if (reconstruct){
        im = letterbox_image(im, net->w, net->h);

        int zz = 0;
        network_predict(net, im.data);
        image out_im = get_network_image(net);
        image crop = crop_image(out_im, zz, zz, out_im.w-2*zz, out_im.h-2*zz);
        //flip_image(crop);
        image f_im = resize_image(crop, out_im.w, out_im.h);
        free_image(crop);
        printf("%d features\n", out_im.w*out_im.h*out_im.c);


        im = resize_image(im, im.w, im.h);
        f_im = resize_image(f_im, f_im.w, f_im.h);
        features = f_im.data;

        int i;
        for(i = 0; i < 14*14*512; ++i){
            features[i] += rand_uniform(-.19, .19);
        }

        free_image(im);
        im = make_random_image(im.w, im.h, im.c);
        update = make_image(im.w, im.h, im.c);

    }

    int e;
    int n;
    for(e = 0; e < rounds; ++e){
        fprintf(stderr, "Iteration: ");
        fflush(stderr);
        for(n = 0; n < iters; ++n){  
            fprintf(stderr, "%d, ", n);
            fflush(stderr);
            if(reconstruct){
                reconstruct_picture(net, features, im, update, rate, momentum, lambda, smooth_size, 1);
                //if ((n+1)%30 == 0) rate *= .5;
                show_image(im, "reconstruction");
#ifdef OPENCV
                cvWaitKey(10);
#endif
            }else{
                int layer = max_layer + rand()%range - range/2;
                int octave = rand()%octaves;
                optimize_picture(&net, im, layer, 1/pow(1.33333333, octave), rate, thresh, norm);
            }
        }
        fprintf(stderr, "done\n");
        char buff[256];
        if (prefix){
            sprintf(buff, "%s/%s_%s_%d_%06d",prefix, imbase, cfgbase, max_layer, e);
        }else{
            sprintf(buff, "%s_%s_%d_%06d",imbase, cfgbase, max_layer, e);
        }
        printf("%d %s\n", e, buff);
        save_image(im, buff);
        //show_image(im, buff);
        //cvWaitKey(0);

        if(rotate){
            image rot = rotate_image(im, rotate);
            free_image(im);
            im = rot;
        }
        image crop = crop_image(im, im.w * (1. - zoom)/2., im.h * (1.-zoom)/2., im.w*zoom, im.h*zoom);
        image resized = resize_image(crop, im.w, im.h);
        free_image(im);
        free_image(crop);
        im = resized;
    }
}
*/

void run_nightmare(int argc, char **argv)
{
    srand(0);
    if(argc < 4){
        fprintf(stderr, "usage: %s %s [cfg] [weights] [image] [layer] [options! (optional)]\n", argv[0], argv[1]);
        return;
    }

    char *cfg = argv[2];
    char *weights = argv[3];
    char *input = argv[4];
    int max_layer = atoi(argv[5]);

    int range = find_int_arg(argc, argv, "-range", 1);
    int norm = find_int_arg(argc, argv, "-norm", 1);
    int rounds = find_int_arg(argc, argv, "-rounds", 1);
    int iters = find_int_arg(argc, argv, "-iters", 10);
    int octaves = find_int_arg(argc, argv, "-octaves", 4);
    float zoom = find_float_arg(argc, argv, "-zoom", 1.);
    float rate = find_float_arg(argc, argv, "-rate", .04);
    float thresh = find_float_arg(argc, argv, "-thresh", 1.);
    float rotate = find_float_arg(argc, argv, "-rotate", 0);
    float momentum = find_float_arg(argc, argv, "-momentum", .9);
    float lambda = find_float_arg(argc, argv, "-lambda", .01);
    char *prefix = find_char_arg(argc, argv, "-prefix", 0);
    int reconstruct = find_arg(argc, argv, "-reconstruct");
    int smooth_size = find_int_arg(argc, argv, "-smooth", 1);

    network *net = load_network(cfg, weights, 0);
    char *cfgbase = basecfg(cfg);
    char *imbase = basecfg(input);

    set_batch_network(net, 1);
    image im = load_image_color(input, 0, 0);
    if(0){
        float scale = 1;
        if(im.w > 512 || im.h > 512){
            if(im.w > im.h) scale = 512.0/im.w;
            else scale = 512.0/im.h;
        }
        image resized = resize_image(im, scale*im.w, scale*im.h);
        free_image(im);
        im = resized;
    }
    //im = letterbox_image(im, net->w, net->h);

    float *features = 0;
    image update;
    if (reconstruct){
        net->n = max_layer;
        im = letterbox_image(im, net->w, net->h);
        //resize_network(&net, im.w, im.h);

        network_predict(net, im.data);
        if(net->layers[net->n-1].type == REGION){
            printf("region!\n");
            zero_objectness(net->layers[net->n-1]);
        }
        image out_im = copy_image(get_network_image(net));
        /*
           image crop = crop_image(out_im, zz, zz, out_im.w-2*zz, out_im.h-2*zz);
        //flip_image(crop);
        image f_im = resize_image(crop, out_im.w, out_im.h);
        free_image(crop);
         */
        printf("%d features\n", out_im.w*out_im.h*out_im.c);

        features = out_im.data;

        /*
        int i;
           for(i = 0; i < 14*14*512; ++i){
        //features[i] += rand_uniform(-.19, .19);
        }
        free_image(im);
        im = make_random_image(im.w, im.h, im.c);
         */
        update = make_image(im.w, im.h, im.c);
    }

    int e;
    int n;
    for(e = 0; e < rounds; ++e){
        fprintf(stderr, "Iteration: ");
        fflush(stderr);
        for(n = 0; n < iters; ++n){  
            fprintf(stderr, "%d, ", n);
            fflush(stderr);
            if(reconstruct){
                reconstruct_picture(net, features, im, update, rate, momentum, lambda, smooth_size, 1);
                //if ((n+1)%30 == 0) rate *= .5;
                show_image(im, "reconstruction", 10);
            }else{
                int layer = max_layer + rand()%range - range/2;
                int octave = rand()%octaves;
                optimize_picture(net, im, layer, 1/pow(1.33333333, octave), rate, thresh, norm);
            }
        }
        fprintf(stderr, "done\n");
        if(0){
            image g = grayscale_image(im);
            free_image(im);
            im = g;
        }
        char buff[256];
        if (prefix){
            sprintf(buff, "%s/%s_%s_%d_%06d",prefix, imbase, cfgbase, max_layer, e);
        }else{
            sprintf(buff, "%s_%s_%d_%06d",imbase, cfgbase, max_layer, e);
        }
        printf("%d %s\n", e, buff);
        save_image(im, buff);
        //show_image(im, buff, 0);

        if(rotate){
            image rot = rotate_image(im, rotate);
            free_image(im);
            im = rot;
        }
        image crop = crop_image(im, im.w * (1. - zoom)/2., im.h * (1.-zoom)/2., im.w*zoom, im.h*zoom);
        image resized = resize_image(crop, im.w, im.h);
        free_image(im);
        free_image(crop);
        im = resized;
    }
}