##Introduction

This repository implements SSD (Single Shot MultiBox Detector)and deploy it. The tools needed for this mini-project are:

• A server for inference: Cloud instances or simply your powerful laptop.
• Data: Pascal VOC or COCO
• PyTorch: To train the deep learning model.
• Docker: Make our life easier to create a container for our application.
• Flask: For API and user interface.

• The environment can be built by DOCKERFILE and the service starts by ./serveSSD.sh.
• Multi-GPU training and inference: We use DistributedDataParallel, you can train or test with arbitrary GPU(s), the training schema will change accordingly.
• Modular: Add your own modules without pain. We abstract backbone,Detector, BoxHead, BoxPredictor, etc. You can replace every component with your own code without change the code base. For example, You can add EfficientNet as backbone, just add efficient_net.py (ALREADY ADDED) and register it, specific it in the config file, It's done!
• Smooth and enjoyable training procedure: we save the state of model, optimizer, scheduler, training iter, you can stop your training and resume training exactly from the save point without change your training CMD.
• Batched inference: can perform inference using multiple images per batch per GPU.
• Evaluating during training: eval you model every eval_step to check performance improving or not.
• Metrics Visualization: visualize metrics details in tensorboard, like AP, APl, APm and APs for COCO dataset or mAP and 20 categories' AP for VOC dataset.
• Auto download: load pre-trained weights from URL and cache it.

1、Clone the repo 
   git clone https://github.com/kristine-li/SSD-Server.git
   cd SSD-Server

2、build the docker image
   docker build -t ssdServer .

3、run the container
   docker run -i -t -p 5000:5000 ssdServer:latest
   Note：This will run the application on 0.0.0.0:5000, only support for CPU
   TODO: support for GPU： nvidia-docker run -i -t -p 5000:5000 ssdServer:latest

Note：In this demo, before step 2，you should download model SSD(vgg_ssd300_voc0712) to model/

1、start 
   ./serveSSD.sh start

2、stop
  ./serveSSD.sh stop

Upload an image via URL or manually

pip3 install -i https://pypi.tuna.tsinghua.edu.cn/simple  -r requirements.txt

For Pascal VOC dataset, make the folder structure like this:

VOC_ROOT
|__ VOC2007
    |_ JPEGImages
    |_ Annotations
    |_ ImageSets
    |_ SegmentationClass
|__ VOC2012
    |_ JPEGImages
    |_ Annotations
    |_ ImageSets
    |_ SegmentationClass
|__ ...

Where VOC_ROOT default is datasets folder in current project, you can create symlinks to datasets or export VOC_ROOT="/path/to/voc_root".

If you train coco dataset, must install cocoapi.
cd ~/github
git clone https://github.com/cocodataset/cocoapi.git
cd cocoapi/PythonAPI
python setup.py build_ext install

For COCO dataset, make the folder structure like this:

COCO_ROOT
|__ annotations
    |_ instances_valminusminival2014.json
    |_ instances_minival2014.json
    |_ instances_train2014.json
    |_ instances_val2014.json
    |_ ...
|__ train2014
    |_ <im-1-name>.jpg
    |_ ...
    |_ <im-N-name>.jpg
|__ val2014
    |_ <im-1-name>.jpg
    |_ ...
    |_ <im-N-name>.jpg
|__ ...

Where COCO_ROOT default is datasets folder in current project, you can create symlinks to datasets or export COCO_ROOT="/path/to/coco_root".

# for example, train SSD300:
python train.py --config-file configs/vgg_ssd300_voc0712.yaml

# for example, train SSD300 with 4 GPUs:
export NGPUS=4
python -m torch.distributed.launch --nproc_per_node=$NGPUS train.py --config-file configs/vgg_ssd300_voc0712.yaml SOLVER.WARMUP_FACTOR 0.03333 SOLVER.WARMUP_ITERS 1000

The configuration files that I provide assume that we are running on single GPU. When changing number of GPUs, hyper-parameter (lr, max_iter, ...) will also changed according to this paper: Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour.

# for example, evaluate SSD300:
python test.py --config-file configs/vgg_ssd300_voc0712.yaml

# for example, evaluate SSD300 with 4 GPUs:
export NGPUS=4
python -m torch.distributed.launch --nproc_per_node=$NGPUS test.py --config-file configs/vgg_ssd300_voc0712.yaml

Predicting image in a folder is simple:

python demo.py --config-file configs/vgg_ssd300_voc0712.yaml --images_dir demo --ckpt model/vgg_ssd300_voc0712.pth

Then it will download and cache vgg_ssd300_voc0712.pth automatically and predicted images with boxes, scores and label names will saved to demo/result folder by default.

If you want to add your custom components, please see DEVELOP_GUIDE.md for more details.

If you have issues running or compiling this code, we have compiled a list of common issues in TROUBLESHOOTING.md. If your issue is not present there, please feel free to open a new issue.

The implementation is heavily influenced by the projects SSD,and ML-web-app.