Skip to content

Commit c408e79

Browse files
jjykhjjykh
committed
updated README.md
1 parent e402cb7 commit c408e79

File tree

2 files changed

+12
-154
lines changed

2 files changed

+12
-154
lines changed

‎README.md

Lines changed: 2 additions & 20 deletions
Original file line numberDiff line numberDiff line change
@@ -9,28 +9,10 @@ Bagua is an ancient Chinese technique that can predict the future via looking ba
99
npm -g install node-gyp
1010
```
1111

12-
# bagua-caffe
13-
Caffe Backend
12+
# Supported Backends
13+
1. bagua-caffe
1414

15-
## How to build bagua-caffe
16-
Build the code. Please follow [Caffe instruction](http://caffe.berkeleyvision.org/installation.html) to install all necessary packages and build it.
1715

18-
```shell
19-
# Modify Makefile.config according to your Caffe installation.
20-
21-
make -j8
22-
23-
# Build Node.js Addons
24-
25-
node-gyp configure build
26-
```
27-
28-
## Demo
29-
Here we provided a SSD demo for object detection.
30-
31-
1. Download [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300.tar.gz), and only extract the **caffemodel** file into model-zoo/caffe/VOC0712/SSD_300x300.
32-
33-
2. cd demo/caffe-ssd && node demo.js
3416

3517

3618

‎bagua-caffe/README.md

Lines changed: 10 additions & 134 deletions
Original file line numberDiff line numberDiff line change
@@ -1,143 +1,19 @@
1-
# SSD: Single Shot MultiBox Detector
1+
# How to build bagua-caffe
2+
Build the code. Please follow [Caffe instruction](http://caffe.berkeleyvision.org/installation.html) to install all necessary packages and build it.
23

3-
[![Build Status](https://travis-ci.org/weiliu89/caffe.svg?branch=ssd)](https://travis-ci.org/weiliu89/caffe)
4-
[![License](https://img.shields.io/badge/license-BSD-blue.svg)](LICENSE)
5-
6-
By [Wei Liu](http://www.cs.unc.edu/~wliu/), [Dragomir Anguelov](https://www.linkedin.com/in/dragomiranguelov), [Dumitru Erhan](http://research.google.com/pubs/DumitruErhan.html), [Christian Szegedy](http://research.google.com/pubs/ChristianSzegedy.html), [Scott Reed](http://www-personal.umich.edu/~reedscot/), [Cheng-Yang Fu](http://www.cs.unc.edu/~cyfu/), [Alexander C. Berg](http://acberg.com).
7-
8-
### Introduction
9-
10-
SSD is an unified framework for object detection with a single network. You can use the code to train/evaluate a network for object detection task. For more details, please refer to our [arXiv paper](http://arxiv.org/abs/1512.02325) and our [slide](http://www.cs.unc.edu/~wliu/papers/ssd_eccv2016_slide.pdf).
11-
12-
<p align="center">
13-
<img src="http://www.cs.unc.edu/~wliu/papers/ssd.png" alt="SSD Framework" width="600px">
14-
</p>
15-
16-
| System | VOC2007 test *mAP* | **FPS** (Titan X) | Number of Boxes | Input resolution
17-
|:-------|:-----:|:-------:|:-------:|:-------:|
18-
| [Faster R-CNN (VGG16)](https://github.com/ShaoqingRen/faster_rcnn) | 73.2 | 7 | ~6000 | ~1000 x 600 |
19-
| [YOLO (customized)](http://pjreddie.com/darknet/yolo/) | 63.4 | 45 | 98 | 448 x 448 |
20-
| SSD300* (VGG16) | 77.2 | 46 | 8732 | 300 x 300 |
21-
| SSD512* (VGG16) | **79.8** | 19 | 24564 | 512 x 512 |
22-
23-
24-
<p align="left">
25-
<img src="http://www.cs.unc.edu/~wliu/papers/ssd_results.png" alt="SSD results on multiple datasets" width="800px">
26-
</p>
27-
28-
_Note: SSD300* and SSD512* are the latest models. Current code should reproduce these results._
29-
30-
### Citing SSD
31-
32-
Please cite SSD in your publications if it helps your research:
33-
34-
@inproceedings{liu2016ssd,
35-
title = {{SSD}: Single Shot MultiBox Detector},
36-
author = {Liu, Wei and Anguelov, Dragomir and Erhan, Dumitru and Szegedy, Christian and Reed, Scott and Fu, Cheng-Yang and Berg, Alexander C.},
37-
booktitle = {ECCV},
38-
year = {2016}
39-
}
40-
41-
### Contents
42-
1. [Installation](#installation)
43-
2. [Preparation](#preparation)
44-
3. [Train/Eval](#traineval)
45-
4. [Models](#models)
46-
47-
### Installation
48-
1. Get the code. We will call the directory that you cloned Caffe into `$CAFFE_ROOT`
49-
```Shell
50-
git clone https://github.com/weiliu89/caffe.git
51-
cd caffe
52-
git checkout ssd
53-
```
54-
55-
2. Build the code. Please follow [Caffe instruction](http://caffe.berkeleyvision.org/installation.html) to install all necessary packages and build it.
56-
```Shell
4+
```shell
575
# Modify Makefile.config according to your Caffe installation.
58-
cp Makefile.config.example Makefile.config
59-
make -j8
60-
# Make sure to include $CAFFE_ROOT/python to your PYTHONPATH.
61-
make py
62-
make test -j8
63-
# (Optional)
64-
make runtest -j8
65-
```
66-
67-
### Preparation
68-
1. Download [fully convolutional reduced (atrous) VGGNet](https://gist.github.com/weiliu89/2ed6e13bfd5b57cf81d6). By default, we assume the model is stored in `$CAFFE_ROOT/models/VGGNet/`
69-
70-
2. Download VOC2007 and VOC2012 dataset. By default, we assume the data is stored in `$HOME/data/`
71-
```Shell
72-
# Download the data.
73-
cd $HOME/data
74-
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar
75-
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
76-
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
77-
# Extract the data.
78-
tar -xvf VOCtrainval_11-May-2012.tar
79-
tar -xvf VOCtrainval_06-Nov-2007.tar
80-
tar -xvf VOCtest_06-Nov-2007.tar
81-
```
82-
83-
3. Create the LMDB file.
84-
```Shell
85-
cd $CAFFE_ROOT
86-
# Create the trainval.txt, test.txt, and test_name_size.txt in data/VOC0712/
87-
./data/VOC0712/create_list.sh
88-
# You can modify the parameters in create_data.sh if needed.
89-
# It will create lmdb files for trainval and test with encoded original image:
90-
# - $HOME/data/VOCdevkit/VOC0712/lmdb/VOC0712_trainval_lmdb
91-
# - $HOME/data/VOCdevkit/VOC0712/lmdb/VOC0712_test_lmdb
92-
# and make soft links at examples/VOC0712/
93-
./data/VOC0712/create_data.sh
94-
```
956

96-
### Train/Eval
97-
1. Train your model and evaluate the model on the fly.
98-
```Shell
99-
# It will create model definition files and save snapshot models in:
100-
# - $CAFFE_ROOT/models/VGGNet/VOC0712/SSD_300x300/
101-
# and job file, log file, and the python script in:
102-
# - $CAFFE_ROOT/jobs/VGGNet/VOC0712/SSD_300x300/
103-
# and save temporary evaluation results in:
104-
# - $HOME/data/VOCdevkit/results/VOC2007/SSD_300x300/
105-
# It should reach 77.* mAP at 120k iterations.
106-
python examples/ssd/ssd_pascal.py
107-
```
108-
If you don't have time to train your model, you can download a pre-trained model at [here](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300.tar.gz).
7+
make -j8
1098

110-
2. Evaluate the most recent snapshot.
111-
```Shell
112-
# If you would like to test a model you trained, you can do:
113-
python examples/ssd/score_ssd_pascal.py
114-
```
9+
# Build Node.js Addons
11510

116-
3. Test your model using a webcam. Note: press <kbd>esc</kbd> to stop.
117-
```Shell
118-
# If you would like to attach a webcam to a model you trained, you can do:
119-
python examples/ssd/ssd_pascal_webcam.py
11+
node-gyp configure build
12012
```
121-
[Here](https://drive.google.com/file/d/0BzKzrI_SkD1_R09NcjM1eElLcWc/view) is a demo video of running a SSD500 model trained on [MSCOCO](http://mscoco.org) dataset.
122-
123-
4. Check out [`examples/ssd_detect.ipynb`](https://github.com/weiliu89/caffe/blob/ssd/examples/ssd_detect.ipynb) or [`examples/ssd/ssd_detect.cpp`](https://github.com/weiliu89/caffe/blob/ssd/examples/ssd/ssd_detect.cpp) on how to detect objects using a SSD model. Check out [`examples/ssd/plot_detections.py`](https://github.com/weiliu89/caffe/blob/ssd/examples/ssd/plot_detections.py) on how to plot detection results output by ssd_detect.cpp.
124-
125-
5. To train on other dataset, please refer to data/OTHERDATASET for more details. We currently add support for COCO and ILSVRC2016. We recommend using [`examples/ssd.ipynb`](https://github.com/weiliu89/caffe/blob/ssd/examples/ssd_detect.ipynb) to check whether the new dataset is prepared correctly.
126-
127-
### Models
128-
We have provided the latest models that are trained from different datasets. To help reproduce the results in [Table 6](https://arxiv.org/pdf/1512.02325v4.pdf), most models contain a pretrained `.caffemodel` file, many `.prototxt` files, and python scripts.
129-
130-
1. PASCAL VOC models:
131-
* 07+12: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_512x512.tar.gz)
132-
* 07++12: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712Plus_SSD_300x300.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712Plus_SSD_512x512.tar.gz)
133-
* COCO<sup>[1]</sup>: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300_coco.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_512x512_coco.tar.gz)
134-
* 07+12+COCO: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300_ft.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_512x512_ft.tar.gz)
135-
* 07++12+COCO: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712Plus_SSD_300x300_ft.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712Plus_SSD_512x512_ft.tar.gz)
13613

137-
2. COCO models:
138-
* trainval35k: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_coco_SSD_300x300.tar.gz), [SSD512*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_coco_SSD_512x512.tar.gz)
14+
# Demo
15+
Here we provided a SSD demo for object detection.
13916

140-
3. ILSVRC models:
141-
* trainval1: [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_ILSVRC2016_SSD_300x300.tar.gz), [SSD500](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_ilsvrc15_SSD_500x500.tar.gz)
17+
1. Download [SSD300*](http://www.cs.unc.edu/~wliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300.tar.gz), and only extract the **caffemodel** file into model-zoo/caffe/VOC0712/SSD_300x300.
14218

143-
<sup>[1]</sup>We use [`examples/convert_model.ipynb`](https://github.com/weiliu89/caffe/blob/ssd/examples/convert_model.ipynb) to extract a VOC model from a pretrained COCO model.
19+
2. cd demo/caffe-ssd && node demo.js

0 commit comments

Comments
 (0)