baiyuetribe / ncnn-models Goto Github PK

D/vulkan: searching for layers in '/data/app/com.tencent.yolov7test-VQ5vdVed2_ZMl_KkC1Pr2w==/lib/arm64'
D/vulkan: searching for layers in '/system/fake-libs64'
D/vulkan: searching for layers in '/data/app/com.tencent.yolov7test-VQ5vdVed2_ZMl_KkC1Pr2w==/base.apk!/lib/arm64-v8a'
E/vulkan: invalid vkGetInstanceProcAddr(VK_NULL_HANDLE, "vkEnumerateInstanceVersion") call
I/cent.yolov7test: type=1400 audit(0.0:22277): avc: denied { open } for path="/dev/__properties__/u:object_r:tee_supplicant_prop:s0" dev="tmpfs" ino=15850 scontext=u:r:untrusted_app_25:s0:c512,c768 tcontext=u:object_r:tee_supplicant_prop:s0 tclass=file permissive=1
I/cent.yolov7test: type=1400 audit(0.0:22278): avc: denied { getattr } for path="/dev/__properties__/u:object_r:tee_supplicant_prop:s0" dev="tmpfs" ino=15850 scontext=u:r:untrusted_app_25:s0:c512,c768 tcontext=u:object_r:tee_supplicant_prop:s0 tclass=file permissive=1
I/cent.yolov7test: type=1400 audit(0.0:22279): avc: denied { open } for path="/dev/__properties__/u:object_r:firstboot_prop:s0" dev="tmpfs" ino=15862 scontext=u:r:untrusted_app_25:s0:c512,c768 tcontext=u:object_r:firstboot_prop:s0 tclass=file permissive=1
I/cent.yolov7test: type=1400 audit(0.0:22285): avc: denied { open } for path="/dev/__properties__/u:object_r:bluetooth_prop:s0" dev="tmpfs" ino=15894 scontext=u:r:untrusted_app_25:s0:c512,c768 tcontext=u:object_r:bluetooth_prop:s0 tclass=file permissive=1
I/cent.yolov7test: type=1400 audit(0.0:22286): avc: denied { getattr } for path="/dev/__properties__/u:object_r:bluetooth_prop:s0" dev="tmpfs" ino=15894 scontext=u:r:untrusted_app_25:s0:c512,c768 tcontext=u:object_r:bluetooth_prop:s0 tclass=file permissive=1
I/mali_so: [File] : hardware/arm/maliT760/driver/product/base/src/mali_base_kbase.c; [Line] : 876; [Func] : base_context_deal_with_version_affairs_rk_ext;
    arm_release_ver of this mali_so is 'r18p0-01rel0', rk_so_ver is '8@0'.
D/mali_so: [File] : hardware/arm/maliT760/driver/product/base/src/mali_base_kbase.c; [Line] : 881; [Func] : base_context_deal_with_version_affairs_rk_ext;
    current process is NOT sf, to bail out.
W/ncnn: [0 Mali-T860]  queueC=0[2]  queueG=0[2]  queueT=0[2]
W/ncnn: [0 Mali-T860]  bugsbn1=0  bugbilz=0  bugcopc=0  bugihfa=1
W/ncnn: [0 Mali-T860]  fp16-p/s/a=1/0/1  int8-p/s/a=1/0/0
W/ncnn: [0 Mali-T860]  subgroup=16  basic=0  vote=0  ballot=0  shuffle=0
I/mali_so: [File] : hardware/arm/maliT760/driver/product/base/src/mali_base_kbase.c; [Line] : 876; [Func] : base_context_deal_with_version_affairs_rk_ext;
    arm_release_ver of this mali_so is 'r18p0-01rel0', rk_so_ver is '8@0'.
D/mali_so: [File] : hardware/arm/maliT760/driver/product/base/src/mali_base_kbase.c; [Line] : 881; [Func] : base_context_deal_with_version_affairs_rk_ext;
    current process is NOT sf, to bail out.
I/zygote64: Do partial code cache collection, code=12KB, data=20KB
I/zygote64: After code cache collection, code=12KB, data=20KB
I/zygote64: Increasing code cache capacity to 128KB
I/zygote64: Do partial code cache collection, code=12KB, data=39KB
I/zygote64: After code cache collection, code=12KB, data=39KB
I/zygote64: Increasing code cache capacity to 256KB
I/zygote64: Compiler allocated 7MB to compile void android.widget.TextView.<init>(android.content.Context, android.util.AttributeSet, int, int)

static void generate_grids_and_stride(const int target_w, const int target_h, std::vector<int>& strides, std::vector<GridAndStride>& grid_strides)
{
    for (auto stride : strides)
    {
        int num_grid_w = target_w / stride;
        int num_grid_h = target_h / stride;
        for (int g1 = 0; g1 < num_grid_h; g1++)
        {
            for (int g0 = 0; g0 < num_grid_w; g0++)
            {
                GridAndStride gs;
                gs.grid0 = g0;
                gs.grid1 = g1;
                gs.stride = stride;
                grid_strides.push_back(gs);
            }
        }
    }
}

static void generate_proposals(std::vector<GridAndStride> grid_strides, const ncnn::Mat& feat_blob, float prob_threshold, std::vector<Object>& objects) {
    const int num_grid = feat_blob.h;
    fprintf(stderr, "output height: %d, width: %d, channels: %d, dims:%d\n", feat_blob.h, feat_blob.w, feat_blob.c, feat_blob.dims);

    const int num_anchors = grid_strides.size();

    const int num_class = feat_blob.c / num_anchors - 5;

    const float* feat_ptr = feat_blob.channel(0);

    for (int anchor_idx = 0; anchor_idx < num_anchors; anchor_idx++)
    {
        __android_log_print(ANDROID_LOG_DEBUG, "yolov7-custom", "anchor_idx %d", anchor_idx);

        const int grid0 = grid_strides[anchor_idx].grid0;

        __android_log_print(ANDROID_LOG_DEBUG, "yolov7-custom", "grid0 %d", grid0);
        const int grid1 = grid_strides[anchor_idx].grid1;

        __android_log_print(ANDROID_LOG_DEBUG, "yolov7-custom", "grid1 %d", grid1);
        const int stride = grid_strides[anchor_idx].stride;

        __android_log_print(ANDROID_LOG_DEBUG, "yolov7-custom", "stride %d", stride);

        // yolox/models/yolo_head.py decode logic
        //  outputs[..., :2] = (outputs[..., :2] + grids) * strides
        //  outputs[..., 2:4] = torch.exp(outputs[..., 2:4]) * strides
        float x_center = (feat_ptr[0] + grid0) * stride;
        float y_center = (feat_ptr[1] + grid1) * stride;
        float w = exp(feat_ptr[2]) * stride;
        float h = exp(feat_ptr[3]) * stride;
        float x0 = x_center - w * 0.5f;
        float y0 = y_center - h * 0.5f;

        float box_objectness = feat_ptr[4];
        for (int class_idx = 0; class_idx < num_class; class_idx++)
        {
            float box_cls_score = feat_ptr[5 + class_idx];
            float box_prob = box_objectness * box_cls_score;
            if (box_prob > prob_threshold)
            {
                Object obj;
                obj.rect.x = x0;
                obj.rect.y = y0;
                obj.rect.width = w;
                obj.rect.height = h;
                obj.label = class_idx;
                obj.prob = box_prob;

                objects.push_back(obj);
            }

        } // class loop
        feat_ptr += feat_blob.w;

    } // point anchor loop
}

int Yolo::detect(const cv::Mat &rgb, std::vector<Object> &objects, float prob_threshold,
                 float nms_threshold) {
    int img_w = rgb.cols;
    int img_h = rgb.rows;

    // letterbox pad to multiple of 32
    int w = img_w;
    int h = img_h;
    float scale = 1.f;
    if (w > h)
    {
        scale = (float)target_size / w;
        w = target_size;
        h = h * scale;
    }
    else
    {
        scale = (float)target_size / h;
        h = target_size;
        w = w * scale;
    }

    ncnn::Mat in = ncnn::Mat::from_pixels_resize(rgb.data, ncnn::Mat::PIXEL_RGB, img_w, img_h, w, h);

    // pad to target_size rectangle
    // yolov5/utils/datasets.py letterbox
    int wpad = (w + 31) / 32 * 32 - w;
    int hpad = (h + 31) / 32 * 32 - h;
    ncnn::Mat in_pad;
    ncnn::copy_make_border(in, in_pad, 0, hpad, 0, wpad, ncnn::BORDER_CONSTANT, 114.f);

    // so for 0-255 input image, rgb_mean should multiply 255 and norm should div by std.
    // new release of yolox has deleted this preprocess,if you are using new release please don't use this preprocess.
    in_pad.substract_mean_normalize(0, norm_vals);

    ncnn::Extractor ex = yolo.create_extractor();

    ex.input("images", in_pad);

    std::vector<Object> proposals;

    {
        ncnn::Mat out;
        ex.extract("output0", out);

        std::vector<int> strides = {8, 16, 32}; // might have stride=64
        std::vector<GridAndStride> grid_strides;
        generate_grids_and_stride(in_pad.w, in_pad.h, strides, grid_strides);
        generate_proposals(grid_strides, out, prob_threshold, proposals);
    }

    // sort all proposals by score from highest to lowest
    qsort_descent_inplace(proposals);

    // apply nms with nms_threshold
    std::vector<int> picked;
    nms_sorted_bboxes(proposals, picked, nms_threshold);

    int count = picked.size();

    objects.resize(count);
    for (int i = 0; i < count; i++)
    {
        objects[i] = proposals[picked[i]];

        // adjust offset to original unpadded
        float x0 = (objects[i].rect.x) / scale;
        float y0 = (objects[i].rect.y) / scale;
        float x1 = (objects[i].rect.x + objects[i].rect.width) / scale;
        float y1 = (objects[i].rect.y + objects[i].rect.height) / scale;

        // clip
        x0 = std::max(std::min(x0, (float)(img_w - 1)), 0.f);
        y0 = std::max(std::min(y0, (float)(img_h - 1)), 0.f);
        x1 = std::max(std::min(x1, (float)(img_w - 1)), 0.f);
        y1 = std::max(std::min(y1, (float)(img_h - 1)), 0.f);

        objects[i].rect.x = x0;
        objects[i].rect.y = y0;
        objects[i].rect.width = x1 - x0;
        objects[i].rect.height = y1 - y0;
    }

    return 0;
}