yolo踩过的坑

输入./darknet 那个测试命令后，加载不出来网络，报Float *****(core dumped)

折腾了半天，结果就差了个sudo....OTZ

应该就是cuda的权限问题，之前在openpose里也出过这个睿智错误，但那是上半年的事了，我现在给忘了，结果白白折腾了好几天。

++++++++++++++++++++++++++++++++++++++++++++++++

yolo的python借口在v3更新了，结果报错找不到make_boxes，这就导致读不了nparray格式的图更新代码如下：

from ctypes import *
import math
import random
import sys, os
sys.path.append(os.path.join(os.getcwd(),'python/'))#包含了darknet包的路径，原来还有这种操作= =
import darknet as dn
import cv2
import numpy as np
from numpy import *
def sample(probs):
    s = sum(probs)
    probs = [a/s for a in probs]
    r = random.uniform(0, 1)
    for i in range(len(probs)):
        r = r - probs[i]
        if r <= 0:
            return i
    return len(probs)-1

def c_array(ctype, values):
    arr = (ctype*len(values))()
    arr[:] = values
    return arr

class BOX(Structure):
    _fields_ = [("x", c_float),
                ("y", c_float),
                ("w", c_float),
                ("h", c_float)]

class DETECTION(Structure):
    _fields_ = [("bbox", BOX),
                ("classes", c_int),
                ("prob", POINTER(c_float)),
                ("mask", POINTER(c_float)),
                ("objectness", c_float),
                ("sort_class", c_int)]

class IMAGE(Structure):
    _fields_ = [("w", c_int),
                ("h", c_int),
                ("c", c_int),
                ("data", POINTER(c_float))]

class METADATA(Structure):
    _fields_ = [("classes", c_int),
                ("names", POINTER(c_char_p))]

    

#lib = CDLL("/home/pjreddie/documents/darknet/libdarknet.so", RTLD_GLOBAL)
lib = CDLL("/home/a421/zjs_test/darknet-master/libdarknet.so", RTLD_GLOBAL)
lib.network_width.argtypes = [c_void_p]
lib.network_width.restype = c_int
lib.network_height.argtypes = [c_void_p]
lib.network_height.restype = c_int

predict = lib.network_predict
predict.argtypes = [c_void_p, POINTER(c_float)]
predict.restype = POINTER(c_float)

set_gpu = lib.cuda_set_device
set_gpu.argtypes = [c_int]

make_image = lib.make_image
make_image.argtypes = [c_int, c_int, c_int]
make_image.restype = IMAGE

get_network_boxes = lib.get_network_boxes
get_network_boxes.argtypes = [c_void_p, c_int, c_int, c_float, c_float, POINTER(c_int), c_int, POINTER(c_int)]
get_network_boxes.restype = POINTER(DETECTION)

make_network_boxes = lib.make_network_boxes
make_network_boxes.argtypes = [c_void_p]
make_network_boxes.restype = POINTER(DETECTION)

free_detections = lib.free_detections
free_detections.argtypes = [POINTER(DETECTION), c_int]

free_ptrs = lib.free_ptrs
free_ptrs.argtypes = [POINTER(c_void_p), c_int]

network_predict = lib.network_predict
network_predict.argtypes = [c_void_p, POINTER(c_float)]

reset_rnn = lib.reset_rnn
reset_rnn.argtypes = [c_void_p]

load_net = lib.load_network
load_net.argtypes = [c_char_p, c_char_p, c_int]
load_net.restype = c_void_p

do_nms_obj = lib.do_nms_obj
do_nms_obj.argtypes = [POINTER(DETECTION), c_int, c_int, c_float]

do_nms_sort = lib.do_nms_sort
do_nms_sort.argtypes = [POINTER(DETECTION), c_int, c_int, c_float]

free_image = lib.free_image
free_image.argtypes = [IMAGE]

letterbox_image = lib.letterbox_image
letterbox_image.argtypes = [IMAGE, c_int, c_int]
letterbox_image.restype = IMAGE

load_meta = lib.get_metadata
lib.get_metadata.argtypes = [c_char_p]
lib.get_metadata.restype = METADATA

load_image = lib.load_image_color
load_image.argtypes = [c_char_p, c_int, c_int]
load_image.restype = IMAGE

rgbgr_image = lib.rgbgr_image
rgbgr_image.argtypes = [IMAGE]

predict_image = lib.network_predict_image
predict_image.argtypes = [c_void_p, IMAGE]
predict_image.restype = POINTER(c_float)

def classify(net, meta, im):
    out = predict_image(net, im)
    res = []
    for i in range(meta.classes):
        res.append((meta.names[i], out[i]))
    res = sorted(res, key=lambda x: -x[1])
    return res
def array_to_image(arr):
    # need to return old values to avoid python freeing memory
    arr = arr.transpose(2,0,1)
    c, h, w = arr.shape[0:3]
    arr = np.ascontiguousarray(arr.flat, dtype=np.float32) / 255.0
    data = arr.ctypes.data_as(POINTER(c_float))
    im = IMAGE(w,h,c,data)
    return im, arr

def detect(net, meta, image, thresh=.5, hier_thresh=.5, nms=.45):
    """if isinstance(image, bytes):  
        # image is a filename
        # i.e. image = b'/darknet/data/dog.jpg'
        im = load_image(image, 0, 0)
    else:  
        # image is an nparray
        # i.e. image = cv2.imread('/darknet/data/dog.jpg')
        im, image = array_to_image(image)
        rgbgr_image(im)
    """
    im, image = array_to_image(image)
    rgbgr_image(im)
    num = c_int(0)
    pnum = pointer(num)
    predict_image(net, im)
    dets = get_network_boxes(net, im.w, im.h, thresh,
                             hier_thresh, None, 0, pnum)
    num = pnum[0]
    if nms: do_nms_obj(dets, num, meta.classes, nms)

    res = []
    for j in range(num):
        a = dets[j].prob[0:meta.classes]
        if any(a):
            ai = np.array(a).nonzero()[0]
            for i in ai:
                b = dets[j].bbox
                res.append((meta.names[i], dets[j].prob[i],
                           (b.x, b.y, b.w, b.h)))

    res = sorted(res, key=lambda x: -x[1])
    if isinstance(image, bytes): free_image(im)
    free_detections(dets, num)
    return res
'''
def detect(net, meta, image, thresh=.5, hier_thresh=.5, nms=.45):
    im = load_image(image, 0, 0)
    num = c_int(0)
    pnum = pointer(num)
    predict_image(net, im)
    dets = get_network_boxes(net, im.w, im.h, thresh, hier_thresh, None, 0, pnum)
    num = pnum[0]
    if (nms): do_nms_obj(dets, num, meta.classes, nms);

    res = []
    for j in range(num):
        for i in range(meta.classes):
            if dets[j].prob[i] > 0:
                b = dets[j].bbox
                res.append((meta.names[i], dets[j].prob[i], (b.x, b.y, b.w, b.h)))
    res = sorted(res, key=lambda x: -x[1])
    free_image(im)
    free_detections(dets, num)
    return res
'''  
if __name__ == "__main__":
    #net = load_net("cfg/densenet201.cfg", "/home/pjreddie/trained/densenet201.weights", 0)
    #im = load_image("data/wolf.jpg", 0, 0)
    #meta = load_meta("cfg/imagenet1k.data")
    #r = classify(net, meta, im)
    #print r[:10]
    net = load_net("/home/a421/zjs_test/darknet-master/cfg/yolov3.cfg", "/home/a421/zjs_test/darknet-master/yolov3.weights", 0)
    im = cv2.imread('/home/a421/zjs_test/darknet-master/data/dog.jpg')
    cv2.imshow('test',im)
    cv2.waitKey(30)
    meta = load_meta("/home/a421/zjs_test/darknet-master/cfg/coco.data")
    #im = array_to_image(arr)
    #dn.rgbgr_image(im)
    r = detect(net, meta, im)
    
    
    #r = detect(net, meta, "/home/a421/zjs_test/darknet-master/data/dog.jpg")
    print r