Still under development. 71 mAP on VOC2007 achieved so far.
This is a pre-released version.
This is a re-implementation of original yolo v2 which is based on darknet. The arXiv paper is available here.
- Build from source, this is required because this example is not merged, some custom operators are not presented in official MXNet. Instructions
- Install required packages:
cv2,matplotlib
- Download the pretrained model, and extract to
model/directory. - Run
# cd /paht/to/mxnet-yolo
python demo.py --cpu
# available options
python demo.py -h
- Grab a pretrained model, e.g.
darknet19 - Download PASCAL VOC dataset.
cd /path/to/where_you_store_datasets/
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
# Extract the data.
tar -xvf VOCtrainval_11-May-2012.tar
tar -xvf VOCtrainval_06-Nov-2007.tar
tar -xvf VOCtest_06-Nov-2007.tar
ln -s /path/to/VOCdevkit /path/to/mxnet-yolo/data/VOCdevkit
- Create packed binary file for faster training
# cd /path/to/mxnet-ssd
bash tools/prepare_pascal.sh
# or if you are using windows
python tools/prepare_dataset.py --dataset pascal --year 2007,2012 --set trainval --target ./data/train.lst
python tools/prepare_dataset.py --dataset pascal --year 2007 --set test --target ./data/val.lst --shuffle False
- Start training
python train.py --gpus 0,1,2,3 --epoch 0
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent