jing-vision / lightnet

Lightweight turnkey solution for AI

lightnet

lightnet is a turnkey solution to real world problems accelerated with deep learning AI technology, including but not limited to object detection, image classification and human pose estimation.

• The source code
• Build instructions
• End2end Object Detection
• End2end Image Classification
• Human Pose Estimation
• FAQ

How to read the source code

This project is dependent on a few open-source projects:

• modules/darknet - the main engine for training & inferencing.
• modules/Yolo_mark - the toolkit to prepare training data for object detection.
• modules/yolo2_light - lightweighted inferencing engine [optional].
• modules/cvui - lightweighted GUI based purely on OpenCV.
• moudles/pytorch-caffe-darknet-convert - DL framework model converter
• modules/minitrace - library to generate tracing logs for Chrome "about:tracing"
• modules/readerwriterqueue - single-producer, single-consumer lock-free queue for C++
• modules/bhtsne - Barnes-Hut implementation of the t-SNE algorithm

How to build from Visual Studio 2015

Install NVIDIA SDK

• Download CUDA 11.x

• Download cuDNN v8.x

  • Extract to the same folder as CUDA SDK
  • e.g. c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7\

• Download zlib

Install OpenCV

• Latest OpenCV: e.g. https://github.com/opencv/opencv/releases/
• Extract to C:\opencv\

Build it

Execute the batch file

build.bat

Object Detection - inference w/ pre-trained weights

First you need to download the weights. You can read more details on darknet website.

cfg	weights	names
cfg/yolov2.cfg	https://pjreddie.com/media/files/yolov2.weights	data/coco.names
cfg/yolov2-tiny.cfg	https://pjreddie.com/media/files/yolov2-tiny.weights	coco.names
cfg/yolo9000.cfg	http://pjreddie.com/media/files/yolo9000.weights	cfg/9k.names
cfg/yolov3.cfg	https://pjreddie.com/media/files/yolov3.weights	cfg/coco.names
cfg/yolov3-openimages.cfg	https://pjreddie.com/media/files/yolov3-openimages.weights	data/openimages.names
cfg/yolov3-tiny.cfg	https://pjreddie.com/media/files/yolov3-tiny.weights	cfg/coco.names
cfg/yolov2_shoe.cfg	yolov2_shoe.weights	obj.names
cfg/yolov4.cfg	https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v3_optimal/yolov4.weights	cfg/coco.names

Syntax for object detection

darknet.exe detector demo <data> <cfg> <weights> -c <camera_idx> -i <gpu_idx>
darknet.exe detector demo <data> <cfg> <weights> <video_filename> -i <gpu_idx>
darknet.exe detector test <data> <cfg> <weights> <img_filename> -i <gpu_idx>

Default launch device combination is -i 0 -c 0.

Run yolov4 on camera #0

darknet.exe detector demo cfg/coco.data cfg/yolov4.cfg yolov4.weights

Run yolov3 on camera #0

darknet.exe detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights

Run yolo9000 on camera #0

darknet.exe detector demo cfg/combine9k.data cfg/yolo9000.cfg yolo9000.weights

Run yolo9000 on images

darknet.exe detector test cfg/combine9k.data cfg/yolo9000.cfg yolo9000.weights

Run yolo9000 CPU on camera #0

darknet_no_gpu.exe detector demo cfg/combine9k.data cfg/yolo9000.cfg yolo9000.weights

Object Detection - label images manually

• delete all files from directory my-yolo-net/img and put your .jpg-images in
• change numer of classes (objects for detection) in file my-yolo-net/obj.data: https://github.com/jing-vision/yolo-studio/blob/master/networks/yolo-template/obj.data#L1
• put names of objects, one for each line in file my-yolo-net/obj.names: https://github.com/jing-vision/yolo-studio/blob/master/networks/yolo-template/obj.names
• Run file: my-yolo-net/yolo_mark.cmd

Object Detection - train yolo v2 network

0. Fork __template-yolov2 to my-yolo-net

1. Download pre-trained weights for the convolutional layers: http://pjreddie.com/media/files/darknet19_448.conv.23 to bin/darknet19_448.conv.23

2. To training for your custom objects, you should change 2 lines in file obj.cfg:

• change classes in obj.data#L1
• set number of classes (objects) in obj.cfg#L230
• set filter-value equal to (classes + 5)*5 in obj.cfg#L224

3. Run my-yolo-net/train.cmd

Object Detection - train yolo v3 network

0. Fork __template-yolov3 to my-yolo-net

1. Download pre-trained weights for the convolutional layers: http://pjreddie.com/media/files/darknet53.conv.74 to bin/darknet53.conv.74

2. Create file obj.cfg with the same content as in yolov3.cfg (or copy yolov3.cfg to obj.cfg) and:

• change line batch to batch=64
• change line subdivisions to subdivisions=8
• change line classes=80 to your number of objects in each of 3 [yolo]-layers:
  • obj.cfg#L610
  • obj.cfg#L696
  • obj.cfg#L783
• change [filters=255] to filters=(classes + 5)x3 in the 3 [convolutional] before each [yolo] layer
  • obj.cfg#L603
  • obj.cfg#L689
  • obj.cfg#L776

So if classes=1 then should be filters=18. If classes=2 then write filters=21.

(Do not write in the cfg-file: filters=(classes + 5)x3)

(Generally filters depends on the classes, coords and number of masks, i.e. filters=(classes + coords + 1)*<number of mask>, where mask is indices of anchors. If mask is absence, then filters=(classes + coords + 1)*num)

So for example, for 2 objects, your file obj.cfg should differ from yolov3.cfg in such lines in each of 3 [yolo]-layers:

  [convolutional]
  filters=21

  [region]
  classes=2

Image Classification - inference w/ pre-trained weights

Again, you need download weights first. You can read more details on darknet website.

cfg	weights
cfg/alexnet.cfg	https://pjreddie.com/media/files/alexnet.weights
cfg/vgg-16.cfg	https://pjreddie.com/media/files/vgg-16.weights
cfg/extraction.cfg	https://pjreddie.com/media/files/extraction.weights
cfg/darknet.cfg	https://pjreddie.com/media/files/darknet.weights
cfg/darknet19.cfg	https://pjreddie.com/media/files/darknet19.weights
cfg/darknet19_448.cfg	https://pjreddie.com/media/files/darknet19_448.weights
cfg/darknet53.cfg	https://pjreddie.com/media/files/darknet53.weights
cfg/resnet50.cfg	https://pjreddie.com/media/files/resnet50.weights
cfg/resnet152.cfg	https://pjreddie.com/media/files/resnet152.weights
cfg/densenet201.cfg	https://pjreddie.com/media/files/densenet201.weights

Image Classification - train darknet19_448 network

0. Fork __template-darknet19_448 to my-darknet19-net

1. Download pre-trained weights for the convolutional layers: http://pjreddie.com/media/files/darknet19_448.conv.23 to bin/darknet19_448.conv.23

2. Create file obj.cfg with the same content as in darknet19_448.cfg (or copy darknet19_448.cfg to obj.cfg) and:

• set batch to 128 or 64 or 32 depends on your GPU memory in darknet19-classify.cfg#L4
• change line to subdivisions=4
• set filter-value equal to classes in darknet19-classify.cfg#L189

Human Pose Estimation - inference w/ pre-trained weights

This project lives in DancingGaga

For more details, please check the README there.

[Weight file] (darknet version openpose.weight)

https://drive.google.com/open?id=1BfY0Hx2d2nm3I4JFh0W1cK2aHD1FSGea

FAQ

blobFromImage() vs letterbox_image() vs resize_image()

AlexeyAB/darknet#232 (comment)

How to fix CUDA Error: no kernel image is available for execution on the device?

# Tesla V100
# ARCH= -gencode arch=compute_70,code=[sm_70,compute_70]

# GeForce RTX 2080 Ti, RTX 2080, RTX 2070, Quadro RTX 8000, Quadro RTX 6000, Quadro RTX 5000, Tesla T4, XNOR Tensor Cores
# ARCH= -gencode arch=compute_75,code=[sm_75,compute_75]

# Jetson XAVIER
# ARCH= -gencode arch=compute_72,code=[sm_72,compute_72]

# GTX 1080, GTX 1070, GTX 1060, GTX 1050, GTX 1030, Titan Xp, Tesla P40, Tesla P4
# ARCH= -gencode arch=compute_61,code=sm_61 -gencode arch=compute_61,code=compute_61

# GP100/Tesla P100 - DGX-1
# ARCH= -gencode arch=compute_60,code=sm_60

# For Jetson TX1, Tegra X1, DRIVE CX, DRIVE PX - uncomment:
# ARCH= -gencode arch=compute_53,code=[sm_53,compute_53]

# For Jetson Tx2 or Drive-PX2 uncomment:
# ARCH= -gencode arch=compute_62,code=[sm_62,compute_62]

How to fine tune a existing network?

https://github.com/pjreddie/darknet/wiki/YOLO:-Real-Time-Object-Detection

darknet.exe partial cfg/darknet.cfg darknet.weights darknet.conv.13 13 darknet.exe partial cfg/extraction.cfg extraction.weights extraction.conv.23 23 darknet.exe partial cfg/darknet19.cfg darknet19.weights darknet19.conv.23 23 darknet.exe partial cfg/darknet19_448.cfg darknet19_448.weights darknet19_448.conv.23 23 darknet.exe partial cfg/darknet53.cfg darknet53.weights darknet53.conv.74 74 darknet.exe partial cfg/resnet50.cfg resnet50.weights resnet50.conv.66 66

Explanation of yolo training output

https://github.com/rafaelpadilla/darknet#faq_yolo

CFG Parameters

• in net section
• in layer section

How to use scripts folder?

pip install -r scripts/requirements.txt

lightnet's Issues

where is get_network_layer() code

Hi,

i'm trying to build feature-viz, i included lightnet but when compiling get_network_layer is undefined.
where is the definition/source code to include?

Thanks

cuDNN Error: CUDNN_STATUS_BAD_PARAM

Hi,
With vs2015, CUDA 10.0, cudnn 7.6.2
After build, I execute :

darknet.exe detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights

and report the fllowing error

cuDNN status Error in: file: ..\modules\darknet\src\convolutional_layer.c : cudnn_convolutional_setup() : line: 236 : build time: Aug  7 2019 - 07:13:29
cuDNN Error: CUDNN_STATUS_BAD_PARAM

thx

feature-viz

您好:
請問執行feature-viz這專案之前需要事先建立那些環境呢?我目前已經照步驟建立了vs2015的資料夾，但執行sln檔的時候遇到許多error。我的方式是把ligntnet載下來，並在feature-viz資料夾下執行，謝謝!

OPENCV ERROR to use darknet_viz

Hi~
when I try to use darknet_viz by following command

opencv ERROR out

could you figure it out?

i dont see any thing in yolov3-tiny; i load a picture(it have dog,bike,car)

Getting a link error while trying to build feature-viz using VS2022 and VS2015

When trying to build feature-viz I am getting the following link-error

Severity	Code	Description	Project	File	Line	Suppression State
Error	LNK1104	cannot open file 'yolo_cpp_dll-d.lib'	feature-viz	C:\lightnet\feature-viz\vs2022\LINK	1

Performance and accuracy of yolov2_light

@vinjn hello thanks for the wonderful source code , i had few queries .

1. what is the weight size of the yolov2_light model
2. have you performed any quantization and pruning techniques for these models
3. what is the performance achieved with loss of accuracy

Jing-pose error

Hey again, you must love me asking all these questions lol.

I generated the vs files for jing pose and compiled with no errors.
I get the jing-pose.exe in bin folder, ive downloaded the openpose.weight file listed on the github main page and used the openpose.cfg file located in jing-pose irectory.

But I get the following issue:

C:\lightnet\bin>jing-pose.exe openpose.cfg openpose.weight person.jpg
Failed to open: openpose.cfg
warning: Error opening file (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:808)
warning: openpose.weight (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:809)
Failed to open: openpose.weight
Try to load cfg: openpose.cfg, weights: openpose.weight, clear = 0
batch: Using default '1'
learning_rate: Using default '0.001000'
momentum: Using default '0.900000'
decay: Using default '0.000100'
subdivisions: Using default '1'
policy: Using default 'constant'
max_batches: Using default '0'
layer filters size input output
0 conv 64 3 x 3 / 1 200 x 200 x 3 -> 200 x 200 x 64 0.138 BF
1 conv 64 3 x 3 / 1 200 x 200 x 64 -> 200 x 200 x 64 2.949 BF
2 max 2 x 2 / 2 200 x 200 x 64 -> 100 x 100 x 64 0.003 BF
3 conv 128 3 x 3 / 1 100 x 100 x 64 -> 100 x 100 x 128 1.475 BF
4 conv 128 3 x 3 / 1 100 x 100 x 128 -> 100 x 100 x 128 2.949 BF
5 max 2 x 2 / 2 100 x 100 x 128 -> 50 x 50 x 128 0.001 BF
6 conv 256 3 x 3 / 1 50 x 50 x 128 -> 50 x 50 x 256 1.475 BF
7 conv 256 3 x 3 / 1 50 x 50 x 256 -> 50 x 50 x 256 2.949 BF
8 conv 256 3 x 3 / 1 50 x 50 x 256 -> 50 x 50 x 256 2.949 BF
9 conv 256 3 x 3 / 1 50 x 50 x 256 -> 50 x 50 x 256 2.949 BF
10 max 2 x 2 / 2 50 x 50 x 256 -> 25 x 25 x 256 0.001 BF
11 conv 512 3 x 3 / 1 25 x 25 x 256 -> 25 x 25 x 512 1.475 BF
12 conv 512 3 x 3 / 1 25 x 25 x 512 -> 25 x 25 x 512 2.949 BF
13 conv 256 3 x 3 / 1 25 x 25 x 512 -> 25 x 25 x 256 1.475 BF
14 conv 128 3 x 3 / 1 25 x 25 x 256 -> 25 x 25 x 128 0.369 BF
15 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
16 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
17 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
18 conv 512 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 512 0.082 BF
19 conv 38 1 x 1 / 1 25 x 25 x 512 -> 25 x 25 x 38 0.024 BF
20 route 14
21 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
22 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
23 conv 128 3 x 3 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.184 BF
24 conv 512 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 512 0.082 BF
25 conv 19 1 x 1 / 1 25 x 25 x 512 -> 25 x 25 x 19 0.012 BF
26 route 19 25 14
27 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
28 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
29 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
30 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
31 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
32 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
33 conv 38 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 38 0.006 BF
34 route 26
35 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
36 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
37 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
38 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
39 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
40 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
41 conv 19 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 19 0.003 BF
42 route 33 41 14
43 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
44 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
45 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
46 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
47 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
48 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
49 conv 38 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 38 0.006 BF
50 route 42
51 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
52 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
53 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
54 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
55 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
56 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
57 conv 19 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 19 0.003 BF
58 route 49 57 14
59 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
60 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
61 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
62 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
63 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
64 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
65 conv 38 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 38 0.006 BF
66 route 58
67 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
68 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
69 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
70 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
71 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
72 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
73 conv 19 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 19 0.003 BF
74 route 65 73 14
75 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
76 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
77 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
78 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
79 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
80 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
81 conv 38 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 38 0.006 BF
82 route 74
83 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
84 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
85 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
86 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
87 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
88 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
89 conv 19 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 19 0.003 BF
90 route 81 89 14
91 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
92 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
93 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
94 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
95 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
96 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
97 conv 38 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 38 0.006 BF
98 route 90
99 conv 128 7 x 7 / 1 25 x 25 x 185 -> 25 x 25 x 128 1.450 BF
100 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
101 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
102 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
103 conv 128 7 x 7 / 1 25 x 25 x 128 -> 25 x 25 x 128 1.004 BF
104 conv 128 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 128 0.020 BF
105 conv 19 1 x 1 / 1 25 x 25 x 128 -> 25 x 25 x 19 0.003 BF
106 route 105 97
Total BFLOPS 80.306
Loading weights from openpose.weight...
seen 32
Done!
OpenCV Error: Assertion failed (!input.empty()) in create_netsize_im, file ..\src\post_process.cpp, line 495

Jing-pose has been moved to a standalone repository

https://github.com/jing-interactive/DancingGaga

activation_kernels.cu' does not exist?

Hi, Im trying to compile darknet but get the following error message in 2015:

Severity Code Description Project File Line Suppression State
Error The path specified for SourceFile at 'C:\lightnet\lightnet\modules\darknet\src\activation_kernels.cu' does not exist. darknet C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V140\BuildCustomizations\CUDA 9.1.targets 428

An help would be great, Cheers J

How to configure the visualization

How to configure the visualization file? Is similar to
feature-viz.exe -cfg=cfg\alexnet.cfg -weights=alexnet.weights
Feature-viz.exe -cfg=cfg\darknet.cfg -weights=darknet.weights
Feature-viz.exe -cfg=cfg\darknet19.cfg -weights=darknet19.weights
How to use it?

Is this project can be used on ubuntu?

where can i find __template-yolov3?

Hello,

sorry but where can i find __template-yolov3 mentioned by README ?

thx

Get Error while compiling Feature-viz

submodule checkout error

This is the error message:

fatal: reference is not a tree: 6bb99873873f0fde08e8ecde754d5cb0e0ff97b8
Unable to checkout '6bb99873873f0fde08e8ecde754d5cb0e0ff97b8' in submodule path 'modules/yolo2_light'

The repository yolo2_light is not up to date.

Feature visualization

Hi! Thank you for sharing your work. I want to use feature-viz with CPU, is that possible?

I have compiled yolo lib for CPU but there is a function in lightnet.cpp that seems to need the GPU: cuda_pull_array(l->output_gpu, l->output, l->outputs*l->batch))

VIDEOIO(cvCreateFileCapture_FFMPEG_proxy(filename)): raised unknown C++ exception!

@vinjn
当我运行时，出现如下错误，请问怎样解决啊？