Skip to content

youngwanLEE/vovnet-detectron2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

VoVNet-v2 backbone networks in Detectron2

Efficient Backbone Network for Object Detection and Segmentation

[CenterMask(code)][CenterMask2(code)] [VoVNet-v1(arxiv)] [VoVNet-v2(arxiv)] [BibTeX]

In this project, we release code for VoVNet-v2 backbone network (introduced by CenterMask) in detectron2 as a extention form. VoVNet can extract diverse feature representation efficiently by using One-Shot Aggregation (OSA) module that concatenates subsequent layers at once. Since the OSA module can capture multi-scale receptive fields, the diversifed feature maps allow object detection and segmentation to address multi-scale objects and pixels well, especially robust on small objects. VoVNet-v2 improves VoVNet-v1 by adding identity mapping that eases the optimization problem and effective SE (Squeeze-and-Excitation) that enhances the diversified feature representation.

Highlight

Compared to ResNe(X)t backbone

  • Efficient : Faster speed
  • Accurate : Better performance, especially small object.

Update

  • Lightweight-VoVNet-19 has been released. (19/02/2020)
  • VoVNetV2-19-FPNLite has been released. (22/01/2020)
  • centermask2 has been released. (20/02/2020)

Results on MS-COCO in Detectron2

Note

We measure the inference time of all models with batch size 1 on the same V100 GPU machine.
We train all models using V100 8GPUs.

  • pytorch1.3.1
  • CUDA 10.1
  • cuDNN 7.3

Faster R-CNN

Lightweight-VoVNet with FPNLite

Backbone Param. lr sched inference time AP APs APm APl download
MobileNetV2 3.5M 3x 0.022 33.0 19.0 35.0 43.4 model | metrics
V2-19 11.2M 3x 0.034 38.9 24.8 41.7 49.3 model | metrics
V2-19-DW 6.5M 3x 0.027 36.7 22.7 40.0 46.0 model | metrics
V2-19-Slim 3.1M 3x 0.023 35.2 21.7 37.3 44.4 model | metrics
V2-19-Slim-DW 1.8M 3x 0.022 32.4 19.1 34.6 41.8 model | metrics
  • DW and Slim denote depthwise separable convolution and a thiner model with half the channel size, respectively.
Backbone Param. lr sched inference time AP APs APm APl download
V2-19-FPN 37.6M 3x 0.040 38.9 24.9 41.5 48.8 model | metrics
R-50-FPN 51.2M 3x 0.047 40.2 24.2 43.5 52.0 model | metrics
V2-39-FPN 52.6M 3x 0.047 42.7 27.1 45.6 54.0 model | metrics
R-101-FPN 70.1M 3x 0.063 42.0 25.2 45.6 54.6 model | metrics
V2-57-FPN 68.9M 3x 0.054 43.3 27.5 46.7 55.3 model | metrics
X-101-FPN 114.3M 3x 0.120 43.0 27.2 46.1 54.9 model | metrics
V2-99-FPN 96.9M 3x 0.073 44.1 28.1 47.0 56.4 model | metrics

Mask R-CNN

Backbone lr sched inference time box AP box APs box APm box APl mask AP mask APs mask APm mask APl download
V2-19-FPNLite 3x 0.036 39.7 25.1 42.6 50.8 36.4 19.9 38.8 50.8 model | metrics
V2-19-FPN 3x 0.044 40.1 25.4 43.0 51.0 36.6 19.7 38.7 51.2 model | metrics
R-50-FPN 3x 0.055 41.0 24.9 43.9 53.3 37.2 18.6 39.5 53.3 model | metrics
V2-39-FPN 3x 0.052 43.8 27.6 47.2 55.3 39.3 21.4 41.8 54.6 model | metrics
R-101-FPN 3x 0.070 42.9 26.4 46.6 56.1 38.6 19.5 41.3 55.3 model | metrics
V2-57-FPN 3x 0.058 44.2 28.2 47.2 56.8 39.7 21.6 42.2 55.6 model | metrics
X-101-FPN 3x 0.129 44.3 27.5 47.6 56.7 39.5 20.7 42.0 56.5 model | metrics
V2-99-FPN 3x 0.076 44.9 28.5 48.1 57.7 40.3 21.7 42.8 56.6 model | metrics

Panoptic-FPN on COCO

Name lr
sched
inference
time
(s/im)
box
AP
mask
AP
PQ download
R-50-FPN 3x 0.063 40.0 36.5 41.5 model | metrics
V2-39-FPN 3x 0.063 42.8 38.5 43.4 model | metrics
R-101-FPN 3x 0.078 42.4 38.5 43.0 model | metrics
V2-57-FPN 3x 0.070 43.4 39.2 44.3 model | metrics

Using this command with --num-gpus 1

python /path/to/vovnet-detectron2/train_net.py --config-file /path/to/vovnet-detectron2/configs/<config.yaml> --eval-only --num-gpus 1 MODEL.WEIGHTS <model.pth>

Installation

As this vovnet-detectron2 is implemented as a extension form (detectron2/projects) upon detectron2, you just install detectron2 following INSTALL.md.

Prepare for coco dataset following this instruction.

Training

ImageNet Pretrained Models

We provide backbone weights pretrained on ImageNet-1k dataset.

To train a model, run

python /path/to/vovnet-detectron2/train_net.py --config-file /path/to/vovnet-detectron2/configs/<config.yaml>

For example, to launch end-to-end Faster R-CNN training with VoVNetV2-39 backbone on 8 GPUs, one should execute:

python /path/to/vovnet-detectron2/train_net.py --config-file /path/to/vovnet-detectron2/configs/faster_rcnn_V_39_FPN_3x.yaml --num-gpus 8

Evaluation

Model evaluation can be done similarly:

python /path/to/vovnet-detectron2/train_net.py --config-file /path/to/vovnet-detectron2/configs/faster_rcnn_V_39_FPN_3x.yaml --eval-only MODEL.WEIGHTS <model.pth>

TODO

  • Adding Lightweight models
  • Applying VoVNet for other meta-architectures

Citing VoVNet

If you use VoVNet, please use the following BibTeX entry.

@inproceedings{lee2019energy,
  title = {An Energy and GPU-Computation Efficient Backbone Network for Real-Time Object Detection},
  author = {Lee, Youngwan and Hwang, Joong-won and Lee, Sangrok and Bae, Yuseok and Park, Jongyoul},
  booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops},
  year = {2019}
}

@article{lee2019centermask,
  title={CenterMask: Real-Time Anchor-Free Instance Segmentation},
  author={Lee, Youngwan and Park, Jongyoul},
  booktitle={CVPR},
  year={2020}
}