Zzh-tju / DIoU

Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression (AAAI 2020)

310 stars 50 forks source link

求助：请问为什么用darknet yolov3训练时，log日志中显示的还是giou，而且训练loss比用giou的高，请问这是啥原因呢，谢谢！ #5

Closed 1343464520 closed 4 years ago

1343464520 commented 4 years ago

train_log_part

1343464520 commented 4 years ago

我的cfg文件如下： [net]

Testing

batch=1

subdivisions=1

Training

batch=64

subdivisions=16

batch=32 subdivisions=16

width=416 height=416

width=512

height=512

channels=3 momentum=0.9 decay=0.0005 angle=5 saturation = 1.5 exposure = 1.5 hue=.1

ddj add...

mixup=1

cutmix=1

mosaic=1

blur=1

################

learning_rate=0.001

learning_rate=0.01 burn_in=1000 max_batches = 50000 policy=steps steps=200,800,2000,4000,8000,15000,30000

steps=15000,16000

scales=.5,.5,.5,.5,.5,.5,.5

[convolutional] batch_normalize=1 filters=32 size=3 stride=1 pad=1 activation=leaky

Downsample

[convolutional] batch_normalize=1 filters=64 size=3 stride=2 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=32 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=64 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

Downsample

[convolutional] batch_normalize=1 filters=128 size=3 stride=2 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=64 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=64 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=128 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

Downsample

[convolutional] batch_normalize=1 filters=256 size=3 stride=2 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

Downsample

[convolutional] batch_normalize=1 filters=512 size=3 stride=2 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

Downsample

[convolutional] batch_normalize=1 filters=1024 size=3 stride=2 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=1024 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=1024 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=1024 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 filters=1024 size=3 stride=1 pad=1 activation=leaky

[shortcut] from=-3 activation=linear

######################

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky

[convolutional] batch_normalize=1 filters=512 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=1024 activation=leaky

[convolutional] size=1 stride=1 pad=1 filters=18 activation=linear

[yolo] mask = 6,7,8 anchors = 17,21, 40,21, 68,22, 94,23, 109,21, 122,23, 126,27, 160,22, 207,22 # sill new data 416x416 ubuntu shuffle classes=1 num=9 jitter=.3 ignore_thresh = .5 truth_thresh = 1 random=1

add giou loss

iou_normalizer=0.25

cls_normalizer=1.0

iou_loss=giou

add ciou loss

iou_normalizer=0.5 cls_normalizer=1.0 iou_loss=ciou nms_kind=greedynms beta1=0.6

[route] layers = -4

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[upsample] stride=2

[route] layers = -1, 61

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky

[convolutional] batch_normalize=1 filters=256 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=512 activation=leaky

[convolutional] size=1 stride=1 pad=1 filters=18 activation=linear

[yolo] mask = 3,4,5 anchors = 17,21, 40,21, 68,22, 94,23, 109,21, 122,23, 126,27, 160,22, 207,22 # sill new data 416x416 ubuntu shuffle classes=1 num=9 jitter=.3 ignore_thresh = .5 truth_thresh = 1 random=1

add giou loss

iou_normalizer=0.25

cls_normalizer=1.0

iou_loss=giou

add ciou loss

iou_normalizer=0.5 cls_normalizer=1.0 iou_loss=ciou nms_kind=greedynms beta1=0.6

[route] layers = -4

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[upsample] stride=2

[route] layers = -1, 36

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=256 activation=leaky

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=256 activation=leaky

[convolutional] batch_normalize=1 filters=128 size=1 stride=1 pad=1 activation=leaky

[convolutional] batch_normalize=1 size=3 stride=1 pad=1 filters=256 activation=leaky

[convolutional] size=1 stride=1 pad=1 filters=18 activation=linear

[yolo] mask = 0,1,2 anchors = 17,21, 40,21, 68,22, 94,23, 109,21, 122,23, 126,27, 160,22, 207,22 # sill new data 416x416 ubuntu shuffle

classes=1 num=9 jitter=.3 ignore_thresh = .5 truth_thresh = 1 random=1

add giou loss

iou_normalizer=0.25

cls_normalizer=1.0

iou_loss=giou

add ciou loss

iou_normalizer=0.5 cls_normalizer=1.0 iou_loss=ciou nms_kind=greedynms beta1=0.6

1343464520 commented 4 years ago

上面大号粗黑体在cfg中都是注释掉的，显示的问题。。。

Zzh-tju commented 4 years ago

显然这不是使用我的repo的训练输出。如果是https://github.com/Zzh-tju/DIoU-darknet 就没有问题。我刚刚测试了 ./darknet detector train cfg/voc-ciou.data cfg/voc-ciou.cfg darknet53.conv.74 --gpus 0,1

v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.160846, CIOU: 0.109033), Class: 0.455774, Obj: 0.497935, No Obj: 0.459067, .5R: 0.000000, .75R: 0.000000, count: 5
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.064099, CIOU: -0.048820), Class: 0.769042, Obj: 0.302707, No Obj: 0.513355, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.466621, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1137.725464
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.232801, CIOU: 0.211622), Class: 0.383240, Obj: 0.503172, No Obj: 0.462374, .5R: 0.000000, .75R: 0.000000, count: 2
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.215681, CIOU: 0.215456), Class: 0.342119, Obj: 0.121874, No Obj: 0.513232, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.469092, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1136.798950
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.210665, CIOU: 0.163074), Class: 0.469379, Obj: 0.561193, No Obj: 0.460269, .5R: 0.250000, .75R: 0.000000, count: 4
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.293309, CIOU: 0.263840), Class: 0.539059, Obj: 0.475476, No Obj: 0.512379, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: 0.551950, CIOU: 0.547578), Class: 0.125250, Obj: 0.204644, No Obj: 0.467130, .5R: 1.000000, .75R: 0.000000, count: 1
train_network_err: 1135.440063
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.119378, CIOU: 0.117855), Class: 0.689940, Obj: 0.460485, No Obj: 0.460324, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.275015, CIOU: 0.246272), Class: 0.296907, Obj: 0.615617, No Obj: 0.513529, .5R: 0.000000, .75R: 0.000000, count: 3
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.466939, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1132.094727
train_network sum: 4542.059082 / 8
7: 567.757385, 567.498413 avg, 0.000000 rate, 0.456348 seconds, 56 images
Loaded: 0.000046 seconds
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.211749, CIOU: 0.202135), Class: 0.658431, Obj: 0.322365, No Obj: 0.462490, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.262051, CIOU: 0.254771), Class: 0.426819, Obj: 0.782858, No Obj: 0.512206, .5R: 0.000000, .75R: 0.000000, count: 1
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.468074, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1128.728638
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.551529, CIOU: 0.526662), Class: 0.486327, Obj: 0.482804, No Obj: 0.460465, .5R: 0.500000, .75R: 0.500000, count: 2
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.512807, .5R: -nan, .75R: -nan, count: 0
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.468829, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1132.304077
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.122917, CIOU: -0.044101), Class: 0.570439, Obj: 0.484558, No Obj: 0.461125, .5R: 0.000000, .75R: 0.000000, count: 2
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.240738, CIOU: 0.204894), Class: 0.437907, Obj: 0.354213, No Obj: 0.513501, .5R: 0.000000, .75R: 0.000000, count: 3
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.467784, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1133.650391
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.255730, CIOU: 0.245869), Class: 0.405298, Obj: 0.524315, No Obj: 0.462522, .5R: 0.000000, .75R: 0.000000, count: 2
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.513409, .5R: -nan, .75R: -nan, count: 0
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.469859, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1134.588501
train_network sum: 4529.271484 / 8
8: 566.158936, 567.364441 avg, 0.000000 rate, 0.410363 seconds, 64 images
Loaded: 0.000040 seconds
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.062998, CIOU: -0.028079), Class: 0.722950, Obj: 0.497665, No Obj: 0.463482, .5R: 0.000000, .75R: 0.000000, count: 4
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.511521, .5R: -nan, .75R: -nan, count: 0
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.467789, .5R: -nan, .75R: -nan, count: 0
train_network_err: 1128.023071
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.158216, CIOU: 0.137621), Class: 0.589739, Obj: 0.456860, No Obj: 0.465134, .5R: 0.000000, .75R: 0.000000, count: 3
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.142218, CIOU: 0.011128), Class: 0.462262, Obj: 0.684095, No Obj: 0.513644, .5R: 0.000000, .75R: 0.000000, count: 2
v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: 0.190844, CIOU: 0.169876), Class: 0.442428, Obj: 0.458559, No Obj: 0.469251, .5R: 0.000000, .75R: 0.000000, count: 1
train_network_err: 1140.072632

1343464520 commented 4 years ago

谢谢大佬回复！是的，你的repo我简单改了下makefile文件在ubuntu上编译出了一些问题，我就改了下前面跟显卡型号相关的几行，没有深入去看，后面我了解yolov4作者用过你的loss，所以我直接在他的repo下跑的😂😂------------------ 原始邮件 ------------------ 发件人: "Zzh-tju"notifications@github.com 发送时间: 2020年5月8日(星期五) 晚上6:30 收件人: "Zzh-tju/DIoU"DIoU@noreply.github.com; 抄送: "1343464520"1343464520@qq.com;"Author"author@noreply.github.com; 主题: Re: [Zzh-tju/DIoU] 求助：请问为什么用darknet yolov3训练时，log日志中显示的还是giou，而且训练loss比用giou的高，请问这是啥原因呢，谢谢！ (#5)

显然这不是使用我的repo的训练输出。如果是https://github.com/Zzh-tju/DIoU-darknet 就没有问题。我刚刚测试了 ./darknet detector train cfg/voc-ciou.data cfg/voc-ciou.cfg darknet53.conv.74 --gpus 0,1 ` v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.160846, CIOU: 0.109033), Class: 0.455774, Obj: 0.497935, No Obj: 0.459067, .5R: 0.000000, .75R: 0.000000, count: 5 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.064099, CIOU: -0.048820), Class: 0.769042, Obj: 0.302707, No Obj: 0.513355, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.466621, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1137.725464 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.232801, CIOU: 0.211622), Class: 0.383240, Obj: 0.503172, No Obj: 0.462374, .5R: 0.000000, .75R: 0.000000, count: 2 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.215681, CIOU: 0.215456), Class: 0.342119, Obj: 0.121874, No Obj: 0.513232, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.469092, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1136.798950 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.210665, CIOU: 0.163074), Class: 0.469379, Obj: 0.561193, No Obj: 0.460269, .5R: 0.250000, .75R: 0.000000, count: 4 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.293309, CIOU: 0.263840), Class: 0.539059, Obj: 0.475476, No Obj: 0.512379, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: 0.551950, CIOU: 0.547578), Class: 0.125250, Obj: 0.204644, No Obj: 0.467130, .5R: 1.000000, .75R: 0.000000, count: 1 train_network_err: 1135.440063 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.119378, CIOU: 0.117855), Class: 0.689940, Obj: 0.460485, No Obj: 0.460324, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.275015, CIOU: 0.246272), Class: 0.296907, Obj: 0.615617, No Obj: 0.513529, .5R: 0.000000, .75R: 0.000000, count: 3 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.466939, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1132.094727 train_network sum: 4542.059082 / 8 7: 567.757385, 567.498413 avg, 0.000000 rate, 0.456348 seconds, 56 images Loaded: 0.000046 seconds v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.211749, CIOU: 0.202135), Class: 0.658431, Obj: 0.322365, No Obj: 0.462490, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.262051, CIOU: 0.254771), Class: 0.426819, Obj: 0.782858, No Obj: 0.512206, .5R: 0.000000, .75R: 0.000000, count: 1 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.468074, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1128.728638 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.551529, CIOU: 0.526662), Class: 0.486327, Obj: 0.482804, No Obj: 0.460465, .5R: 0.500000, .75R: 0.500000, count: 2 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.512807, .5R: -nan, .75R: -nan, count: 0 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.468829, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1132.304077 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.122917, CIOU: -0.044101), Class: 0.570439, Obj: 0.484558, No Obj: 0.461125, .5R: 0.000000, .75R: 0.000000, count: 2 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.240738, CIOU: 0.204894), Class: 0.437907, Obj: 0.354213, No Obj: 0.513501, .5R: 0.000000, .75R: 0.000000, count: 3 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.467784, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1133.650391 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.255730, CIOU: 0.245869), Class: 0.405298, Obj: 0.524315, No Obj: 0.462522, .5R: 0.000000, .75R: 0.000000, count: 2 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.513409, .5R: -nan, .75R: -nan, count: 0 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.469859, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1134.588501 train_network sum: 4529.271484 / 8 8: 566.158936, 567.364441 avg, 0.000000 rate, 0.410363 seconds, 64 images Loaded: 0.000040 seconds v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.062998, CIOU: -0.028079), Class: 0.722950, Obj: 0.497665, No Obj: 0.463482, .5R: 0.000000, .75R: 0.000000, count: 4 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.511521, .5R: -nan, .75R: -nan, count: 0 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: -nan, CIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.467789, .5R: -nan, .75R: -nan, count: 0 train_network_err: 1128.023071 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 82 Avg (IOU: 0.158216, CIOU: 0.137621), Class: 0.589739, Obj: 0.456860, No Obj: 0.465134, .5R: 0.000000, .75R: 0.000000, count: 3 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 94 Avg (IOU: 0.142218, CIOU: 0.011128), Class: 0.462262, Obj: 0.684095, No Obj: 0.513644, .5R: 0.000000, .75R: 0.000000, count: 2 v3 (ciou loss, Normalizer: (iou: 0.500000, cls: 1.000000) Region 106 Avg (IOU: 0.190844, CIOU: 0.169876), Class: 0.442428, Obj: 0.458559, No Obj: 0.469251, .5R: 0.000000, .75R: 0.000000, count: 1 train_network_err: 1140.072632

`

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

1343464520 commented 4 years ago

大佬。我在用你的repo编译时总是报错。。看了下跟opencv相关，应该是版本问题，我用的是3.4版本，不知道您的opencv用的哪个版本？

Zzh-tju commented 4 years ago

@1343464520 3.4.1