Skip to content
Branch: master
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

SSD: Single Shot MultiBox Object Detector

SSD is an unified framework for object detection with a single network.

You can use the code to train/evaluate/test for object detection task.

Gluon Implementation

You can find a Gluon implementation on gluon-cv.


This is a re-implementation of original SSD which is based on caffe. The official repository is available here. The arXiv paper is available here.

This example is intended for reproducing the nice detector while fully utilize the remarkable traits of MXNet.

  • Model converter from caffe is available now!
  • The result is almost identical to the original version. However, due to different implementation details, the results might differ slightly.

Due to the permission issue, this example is maintained in this repository separately. You can use the link regarding specific per example issues.

What's new

  • Support uint8 inference on CPU with MKL-DNN backend. Uint8 inference achieves 0.8364 mAP, which is a comparable accuracy to FP32 (0.8366 mAP).
  • Added live camera capture and detection display (run with --camera flag). Example: ./ --camera --cpu --frame-resize 0.5
  • Added multiple trained models.
  • Added a much simpler way to compose network from mainstream classification networks (resnet, inception...) and Guide.
  • Update to the latest version according to caffe version, with 5% mAP increase.
  • Use C++ record iterator based on back-end multi-thread engine to achieve huge speed up on multi-gpu environments.
  • Monitor validation mAP during training.
  • More network symbols under development and test.
  • Extra operators are now in mxnet/src/operator/contrib.
  • Old models are incompatible, use e06c55d or e4f73f1 for backward compatibility. Or, you can modify the json file to update the symbols if you are familiar with it, because only names have changed while weights and bias should still be good.

Demo results

demo1 demo2 demo3


Model Training data Test data mAP Note
VGG16_reduced 300x300 VOC07+12 trainval VOC07 test 77.8 fast
VGG16_reduced 512x512 VOC07+12 trainval VOC07 test 79.9 slow
Inception-v3 512x512 VOC07+12 trainval VOC07 test 78.9 fastest
Resnet-50 512x512 VOC07+12 trainval VOC07 test 78.9 fast


Model GPU CUDNN Batch-size FPS*
VGG16_reduced 300x300 TITAN X(Maxwell) v5.1 16 95
VGG16_reduced 300x300 TITAN X(Maxwell) v5.1 8 95
VGG16_reduced 300x300 TITAN X(Maxwell) v5.1 1 64
VGG16_reduced 300x300 TITAN X(Maxwell) N/A 8 36
VGG16_reduced 300x300 TITAN X(Maxwell) N/A 1 28
Forward time only, data loading and drawing excluded.

Getting started

  • You will need python modules: cv2, matplotlib and numpy. If you use mxnet-python api, you probably have already got them. You can install them via pip or package managers, such as apt-get:
sudo apt-get install python-opencv python-matplotlib python-numpy
  • Build MXNet: Follow the official instructions
# for Ubuntu/Debian
cp make/ ./
# enable cuda, cudnn if applicable

Remember to enable CUDA if you want to be able to train, since CPU training is insanely slow. Using CUDNN is optional, but highly recommended.

Try the demo

# cd /path/to/incubator-mxnet/example/ssd
# download the test images
python data/demo/
# run the demo
python --gpu 0
# play with examples:
python --epoch 0 --images ./data/demo/dog.jpg --thresh 0.5
python --cpu --network resnet50 --data-shape 512
# wait for library to load for the first time
  • Check python --help for more options.

Live Camera detection

Use to download the trained model. You can use ./ --camera to use a video capture device with opencv such as a webcam. This will open a window that will display the camera output together with the detections. You can play with the detection threshold to get more or less detections.

Train the model

  • Note that we recommend to use gluon-cv to train the model, please refer to gluon-cv ssd. This example only covers training on Pascal VOC dataset. Other datasets should be easily supported by adding subclass derived from class Imdb in dataset/ See example of dataset/ for details.
  • Download the converted pretrained vgg16_reduced model here, unzip .param and .json files into model/ directory by default.
  • Download the PASCAL VOC dataset, skip this step if you already have one.
cd /path/to/where_you_store_datasets/
# Extract the data.
tar -xvf VOCtrainval_11-May-2012.tar
tar -xvf VOCtrainval_06-Nov-2007.tar
tar -xvf VOCtest_06-Nov-2007.tar
  • We are going to use trainval set in VOC2007/2012 as a common strategy. The suggested directory structure is to store VOC2007 and VOC2012 directories in the same VOCdevkit folder.
  • Then link VOCdevkit folder to data/VOCdevkit by default:
ln -s /path/to/VOCdevkit /path/to/incubator-mxnet/example/ssd/data/VOCdevkit

Use hard link instead of copy could save us a bit disk space.

  • Create packed binary file for faster training:
# cd /path/to/incubator-mxnet/example/ssd
bash tools/
# or if you are using windows
python tools/ --dataset pascal --year 2007,2012 --set trainval --target ./data/train.lst
python tools/ --dataset pascal --year 2007 --set test --target ./data/val.lst --no-shuffle
  • Start training:
# cd /path/to/incubator-mxnet/example/ssd
  • By default, this example will use batch-size=32 and learning_rate=0.002. You might need to change the parameters a bit if you have different configurations. Check python --help for more training options. For example, if you have 4 GPUs, use:
# note that a perfect training parameter set is yet to be discovered for multi-GPUs
python --gpus 0,1,2,3 --batch-size 32

Evalute trained model

Make sure you have val.rec as validation dataset. It's the same one as used in training. Use:

# cd /path/to/incubator-mxnet/example/ssd
python --gpus 0,1 --batch-size 128 --epoch 0

Quantize model

Follow the Train instructions to train a FP32 SSD-VGG16_reduced_300x300 model based on Pascal VOC dataset. You can also download our SSD-VGG16 pre-trained model and packed binary data. Create model and data directories if they're not exist, extract the zip files, then rename the uncompressed files as follows (eg, rename ssd-val-fc19a535.idx to val.idx, ssd-val-fc19a535.lst to val.lst, ssd-val-fc19a535.rec to val.rec, ssd_vgg16_reduced_300-dd479559.params to ssd_vgg16_reduced_300-0000.params, ssd_vgg16_reduced_300-symbol-dd479559.json to ssd_vgg16_reduced_300-symbol.json.)


Then, use the following command for quantization. By default, this script uses 5 batches (32 samples per batch) for naive calibration:


After quantization, INT8 models will be saved in model/ dictionary. Use the following command to launch inference.


# Launch FP32 Inference
python --cpu --num-batch 10 --batch-size 224 --deploy --prefix=./model/ssd_

# Launch INT8 Inference
python --cpu --num-batch 10 --batch-size 224 --deploy --prefix=./model/cqssd_

# Launch dummy data Inference
python --deploy --prefix=./model/ssd_
python --deploy --prefix=./model/cqssd_

Convert model to deploy mode

This simply removes all loss layers, and attach a layer for merging results and non-maximum suppression. Useful when loading python symbol is not available.

# cd /path/to/incubator-mxnet/example/ssd
python --num-class 20

Convert caffe model

Converter from caffe is available at /path/to/incubator-mxnet/example/ssd/tools/caffe_converter

This is specifically modified to handle custom layer in caffe-ssd. Usage:

cd /path/to/incubator-mxnet/example/ssd/tools/caffe_converter
python deploy.prototxt name_of_pretrained_caffe_model.caffemodel ssd_converted
# you will use this model in deploy mode without loading from python symbol(layer names inconsistent)
python --prefix ssd_converted --epoch 1 --deploy

There is no guarantee that conversion will always work, but at least it's good for now.

Legacy models

Since the new interface for composing network is introduced, the old models have inconsistent names for weights. You can still load the previous model by rename the symbol to and call with python train/ --network legacy_xxx For example:

python --network '' --prefix model/ssd_300 --epoch 0
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.