SSD: Single Shot MultiBox Detector pytorch implementation focusing on simplicity

SSD: Single Shot MultiBox Detector

Introduction

Here is my pytorch implementation of 2 models: SSD-Resnet50 and SSDLite-MobilenetV2. These models are based on original model (SSD-VGG16) described in the paper SSD: Single Shot MultiBox Detector. This implementation supports mixed precision training.


An example of SSD Resnet50's output.

Motivation

Why this implementation exists while there are many ssd implementations already ?

I believe that many of you when seeing this implementation have this question in your mind. Indeed there are already many implementations for SSD and its variants in Pytorch. However most of them are either:

  • over-complicated
  • modularized
  • many improvements added
  • not evaluated/visualized

The above-mentioned points make learner hard to understand how original ssd looks like. Hence, I re-implement this well-known model, focusing on simplicity. I believe this implementation is suitable for ML/DL users from different levels, especially beginners. In compared to model described in the paper, there are some minor changes (e.g. backbone), but other parts follow paper strictly.

Datasets

Dataset Classes #Train images #Validation images
COCO2017 80 118k 5k
  • COCO: Download the coco images and annotations from coco website. Make sure to put the files as the following structure (The root folder names coco):
    coco
    ├── annotations
    │   ├── instances_train2017.json
    │   └── instances_val2017.json
    │── train2017
    └── val2017 
    

Docker

For being convenient, I provide Dockerfile which could be used for running training as well as test phases

Assume that docker image's name is ssd. You already created an empty folder name trained_models for storing trained weights. Then you clone this repository and cd into it.

Build:

docker build --network=host -t ssd .

Run:

docker run --rm -it -v path/to/your/coco:/coco -v path/to/trained_models:/trained_models --ipc=host --network=host ssd

How to use my code

Assume that at this step, you either already installed necessary libraries or you are inside docker container

Now, with my code, you can:

  • Train your model by running python -m torch.distributed.launch --nproc_per_node=NUM_GPUS_YOU_HAVE train.py --model [ssd|ssdlite] --batch-size [int] [--amp]. You could stop or resume your training process whenever you want. For example, if you stop your training process after 10 epochs, the next time you run the training script, your training process will continue from epoch 10. mAP evaluation, by default, will be run at the end of each epoch. Note: By specifying --amp flag, your model will be trained with mixed precision (FP32 and FP16) instead of full precision (FP32) by default. Mixed precision training reduces gpu usage and therefore allows you train your model with bigger batch size while sacrificing negligible accuracy. More infomation could be found at apex and pytorch.
  • Test your model for COCO dataset by running python test_dataset.py --pretrained_model path/to/trained_model
  • Test your model for image by running python test_image.py --pretrained_model path/to/trained_model --input path/to/input/file --output path/to/output/file
  • Test your model for video by running python test_video.py --pretrained_model path/to/trained_model --input path/to/input/file --output path/to/output/file

You could download my trained weight for SSD-Resnet50 at link

Experiments

I trained my models by using NVIDIA RTX 2080. Below is mAP evaluation for SSD-Resnet50 trained for 54 epochs on COCO val2017 dataset


SSD-Resnet50 evaluation.


SSD-Resnet50 tensorboard for training loss curve and validation mAP curve.

Results

Some predictions are shown below:

References

Owner
Viet Nguyen
M.Sc. in Computer Science, majoring in Artificial Intelligence and Robotics. Interest topics: Deep Learning in NLP and Computer Vision. Reinforcement Learning.
Viet Nguyen
Comments
  • training with a custom dataset

    training with a custom dataset

    I have two questions, how can I use this code to train the SSD with my own custom dataset? and after finishing the training process, how can I use the model for inference tasks on real-time camera applications? thank you for simplifying the SSD idea, I'm a beginner and I'm finding the existing implementations really complicated.

  • Thank you for the repo and some questions

    Thank you for the repo and some questions

    First of all, thanks for the repo. Was trying to find something like this. It will help me a lot. There are a few questions that I want to ask.

    1. The trained model that you are providing, is trained for 50 epochs, right?
    2. Also, any possibility to get trained weight for the SSDLite version anytime soon?

    Thanks again.

  • export the coco predictions in json file

    export the coco predictions in json file

    Hello Viet, thanks for this repo! I noticed that you used the coco eval API to evaluate the the trained model and the results are clear and impressive. I am wondering if the coco eval API also can be used in the test_dataset.py to evaluate the prediction results on coco dataset, such as coco eval2017 dataset? Any suggestion will be appreciated ! : )

  • Default Bounding Box Priors

    Default Bounding Box Priors

    Hi,

    Thanks for your implementation. This makes understanding the model much more easier. I did notice one thing though. In the paper, the authors mention about defining default bounding boxes as priors and define different aspect ratios and scales for the same as well. Then during training they calculate the offset values between the priors with the overlap greater than a threshold w.r.t the ground truth bounding boxes and use that to calculate the localization loss. Although, this implementation trains and works well, but I couldn't find the use of default bounding boxes anywhere. Am I missing something here?

    Thanks.

  • pytorch_version

    pytorch_version

    Hello there, Which pytorch version is used for this repository ? I am using pytorch 1.0, I am getting the error "No module named box_convert" from the ops/boxes.py . Could you please provide the solution ? Thank you !

    Best regards

A light and fast one class detection framework for edge devices. We provide face detector, head detector, pedestrian detector, vehicle detector......

A Light and Fast Face Detector for Edge Devices Big News: LFD, which is a big update of LFFD, now is released (2021.03.09). It is strongly recommended

Nov 20, 2022
A PyTorch Implementation of Single Shot Scale-invariant Face Detector.
A PyTorch Implementation of Single Shot Scale-invariant Face Detector.

S³FD: Single Shot Scale-invariant Face Detector A PyTorch Implementation of Single Shot Scale-invariant Face Detector. Eval python wider_eval_pytorch.

Nov 24, 2022
SSD-based Object Detection in PyTorch
SSD-based Object Detection in PyTorch

SSD-based Object Detection in PyTorch 서강대학교 현대모비스 SW 프로그램에서 진행한 인공지능 프로젝트입니다. Jetson nano를 이용해 pre-trained network를 fine tuning시켜 차량 및 신호등 인식을 구현하였습니다

Nov 16, 2021
SSD: A Unified Framework for Self-Supervised Outlier Detection [ICLR 2021]

SSD: A Unified Framework for Self-Supervised Outlier Detection [ICLR 2021] Pdf: https://openreview.net/forum?id=v5gjXpmR8J Code for our ICLR 2021 pape

Nov 27, 2022
A simple PyTorch Implementation of Generative Adversarial Networks, focusing on anime face drawing.
A simple PyTorch Implementation of Generative Adversarial Networks, focusing on anime face drawing.

AnimeGAN A simple PyTorch Implementation of Generative Adversarial Networks, focusing on anime face drawing. Randomly Generated Images The images are

Nov 29, 2022
An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicity.
An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicity.

Fast Face Classification (F²C) This is the code of our paper An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicit

Jun 27, 2021
A general-purpose programming language, focused on simplicity, safety and stability.
A general-purpose programming language, focused on simplicity, safety and stability.

The Rivet programming language A general-purpose programming language, focused on simplicity, safety and stability. Rivet's goal is to be a very power

Nov 11, 2022
Paaster is a secure by default end-to-end encrypted pastebin built with the objective of simplicity.
Paaster is a secure by default end-to-end encrypted pastebin built with the objective of simplicity.

Follow the development of our desktop client here Paaster Paaster is a secure by default end-to-end encrypted pastebin built with the objective of sim

Nov 3, 2022
A whale detector design for the Kaggle whale-detector challenge!
A whale detector design for the Kaggle whale-detector challenge!

CNN (InceptionV1) + STFT based Whale Detection Algorithm So, this repository is my PyTorch solution for the Kaggle whale-detection challenge. The obje

Sep 28, 2021
Lane follower: Lane-detector (OpenCV) + Object-detector (YOLO5) + CAN-bus
Lane follower: Lane-detector (OpenCV) + Object-detector (YOLO5) + CAN-bus

Lane Follower This code is for the lane follower, including perception and control, as shown below. Environment Hardware Industrial Camera Intel-NUC(1

Jul 7, 2022
HeartRate detector with ArduinoandPython - Use Arduino and Python create a heartrate detector.

Syllabus of Contents Syllabus of Contents Introduction Of Project Features Develop With Python code introduction Installation License Developer Contac

Jan 5, 2022
Video lie detector using xgboost - A video lie detector using OpenFace and xgboost

video_lie_detector_using_xgboost a video lie detector using OpenFace and xgboost

Jan 11, 2022
Imposter-detector-2022 - HackED 2022 Team 3IQ - 2022 Imposter Detector
Imposter-detector-2022 - HackED 2022 Team 3IQ - 2022 Imposter Detector

HackED 2022 Team 3IQ - 2022 Imposter Detector By Aneeljyot Alagh, Curtis Kan, Jo

Aug 20, 2022
A Kernel fuzzer focusing on race bugs

Razzer: Finding kernel race bugs through fuzzing Environment setup $ source scripts/envsetup.sh scripts/envsetup.sh sets up necessary environment var

Nov 27, 2022
Simulation of self-focusing of laser beams in condensed media
Simulation of self-focusing of laser beams in condensed media

What is it? Program for scientific research, which allows to simulate the phenomenon of self-focusing of different laser beams (including Gaussian, ri

Jun 2, 2022
Embracing Single Stride 3D Object Detector with Sparse Transformer

SST: Single-stride Sparse Transformer This is the official implementation of paper: Embracing Single Stride 3D Object Detector with Sparse Transformer

Nov 28, 2022
Official PyTorch Implementation of Mask-aware IoU and maYOLACT Detector [BMVC2021]
Official PyTorch Implementation of Mask-aware IoU and maYOLACT Detector [BMVC2021]

The official implementation of Mask-aware IoU and maYOLACT detector. Our implementation is based on mmdetection. Mask-aware IoU for Anchor Assignment

Oct 21, 2021
Pytorch implementation of 'Fingerprint Presentation Attack Detector Using Global-Local Model'
Pytorch implementation of 'Fingerprint Presentation Attack Detector Using Global-Local Model'

RTK-PAD This is an official pytorch implementation of 'Fingerprint Presentation Attack Detector Using Global-Local Model', which is accepted by IEEE T

Aug 1, 2022