README_en.md 8.0 KB
Newer Older
1 2
English | [简体中文](README.md)

Q
qingqing01 已提交
3 4
Documentation:[https://paddledetection.readthedocs.io](https://paddledetection.readthedocs.io)

5 6
# PaddleDetection

K
Kaipeng Deng 已提交
7 8 9 10 11 12 13
PaddleDetection is an end-to-end object detection development kit based on PaddlePaddle, which
aims to help developers in the whole development of training models, optimizing performance and
inference speed, and deploying models. PaddleDetection provides varied object detection architectures
in modular design, and wealthy data augmentation methods, network components, loss functions, etc.
PaddleDetection supported practical projects such as industrial quality inspection, remote sensing
image object detection, and automatic inspection with its practical features such as model compression
and multi-platform deployment.
14

W
wangguanzhong 已提交
15
**Now all models in PaddleDetection require PaddlePaddle version 1.7 or higher, or suitable develop version.**
16 17

<div align="center">
G
Guanghua Yu 已提交
18
  <img src="docs/images/000000570688.jpg" />
19 20 21 22 23 24 25
</div>


## Introduction

Features:

Q
qingqing01 已提交
26 27 28 29 30 31
- Rich models:

  PaddleDetection provides rich of models, including 100+ pre-trained models
such as object detection, instance segmentation, face detection etc. It covers
the champion models, the practical detection models for cloud and edge device.

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
- Production Ready:

  Key operations are implemented in C++ and CUDA, together with PaddlePaddle's
highly efficient inference engine, enables easy deployment in server environments.

- Highly Flexible:

  Components are designed to be modular. Model architectures, as well as data
preprocess pipelines, can be easily customized with simple configuration
changes.

- Performance Optimized:

  With the help of the underlying PaddlePaddle framework, faster training and
reduced GPU memory footprint is achieved. Notably, YOLOv3 training is
much faster compared to other frameworks. Another example is Mask-RCNN
(ResNet50), we managed to fit up to 4 images per GPU (Tesla V100 16GB) during
multi-GPU training.

Supported Architectures:

Q
qingqing01 已提交
53 54 55 56 57 58 59 60 61 62
|                     | ResNet | ResNet-vd <sup>[1](#vd)</sup> | ResNeXt-vd | SENet | MobileNet |  HRNet | Res2Net |
| ------------------- | :----: | ----------------------------: | :--------: | :---: | :-------: |:------:|:-----:  |
| Faster R-CNN        |   ✓    |                             ✓ |     x      |   ✓   |     ✗     |   ✗    |  ✗      |
| Faster R-CNN + FPN  |   ✓    |                             ✓ |     ✓      |   ✓   |     ✗     |   ✓    |  ✓      |
| Mask R-CNN          |   ✓    |                             ✓ |     x      |   ✓   |     ✗     |   ✗    |  ✗      |
| Mask R-CNN + FPN    |   ✓    |                             ✓ |     ✓      |   ✓   |     ✗     |   ✗    |  ✓      |
| Cascade Faster-RCNN |   ✓    |                             ✓ |     ✓      |   ✗   |     ✗     |   ✗    |  ✗      |
| Cascade Mask-RCNN   |   ✓    |                             ✗ |     ✗      |   ✓   |     ✗     |   ✗    |  ✗      |
| Libra R-CNN         |   ✗    |                             ✓ |     ✗      |   ✗   |     ✗     |   ✗    |  ✗      |
| RetinaNet           |   ✓    |                             ✗ |     ✗      |   ✗   |     ✗     |   ✗    |  ✗      |
63
| YOLOv3              |   ✓    |                             ✓ |     ✗      |   ✗   |     ✓     |   ✗    |  ✗      |
Q
qingqing01 已提交
64 65 66
| SSD                 |   ✗    |                             ✗ |     ✗      |   ✗   |     ✓     |   ✗    |  ✗      |
| BlazeFace           |   ✗    |                             ✗ |     ✗      |   ✗   |     ✗     |   ✗    |  ✗      |
| Faceboxes           |   ✗    |                             ✗ |     ✗      |   ✗   |     ✗     |   ✗    |  ✗      |
67 68 69

<a name="vd">[1]</a> [ResNet-vd](https://arxiv.org/pdf/1812.01187) models offer much improved accuracy with negligible performance cost.

70 71
**NOTE:** ✓ for config file and pretrain model provided in [Model Zoo](docs/MODEL_ZOO.md), ✗ for not provided but is supported generally.

Q
qingqing01 已提交
72 73 74 75 76 77 78
More models:

- EfficientDet
- FCOS
- CornerNet-Squeeze
- YOLOv4

Q
qingqing01 已提交
79 80 81 82 83 84 85
More Backbones:

- DarkNet
- VGG
- GCNet
- CBNet

86 87
Advanced Features:

Q
qingqing01 已提交
88
- [x] **Synchronized Batch Norm**
89 90 91
- [x] **Group Norm**
- [x] **Modulated Deformable Convolution**
- [x] **Deformable PSRoI Pooling**
Q
qingqing01 已提交
92
- [x] **Non-local and GCNet**
93 94 95

**NOTE:** Synchronized batch normalization can only be used on multiple GPU devices, can not be used on CPU devices or single GPU device.

K
Kaipeng Deng 已提交
96 97 98 99 100 101 102 103 104 105 106 107
The following is the relationship between COCO mAP and FPS on Tesla V100 of representative models of each architectures and backbones.

<div align="center">
  <img src="docs/images/map_fps.png" />
</div>

**NOTE:**
- `CBResNet` stands for `Cascade-Faster-RCNN-CBResNet200vd-FPN`, which has highest mAP on COCO as 53.3% in PaddleDetection models
- `Cascade-Faster-RCNN` stands for `Cascade-Faster-RCNN-ResNet50vd-DCN`, which has been optimized to 20 FPS inference speed when COCO mAP as 47.8%
- The enhanced `YOLOv3-ResNet50vd-DCN` is 10.6 absolute percentage points higher than paper on COCO mAP, and inference speed is nearly 70% faster than the darknet framework
- All these models can be get in [Model Zoo](#Model-Zoo)

G
Guanghua Yu 已提交
108
## Tutorials
109

G
Guanghua Yu 已提交
110 111 112 113 114

### Get Started

- [Installation guide](docs/tutorials/INSTALL.md)
- [Quick start on small dataset](docs/tutorials/QUICK_STARTED.md)
W
wangguanzhong 已提交
115
- [Train/Evaluation/Inference](docs/tutorials/GETTING_STARTED.md)
G
Guanghua Yu 已提交
116
- [FAQ](docs/FAQ.md)
G
Guanghua Yu 已提交
117 118 119 120 121 122

### Advanced Tutorial

- [Guide to preprocess pipeline and custom dataset](docs/advanced_tutorials/READER.md)
- [Models technical](docs/advanced_tutorials/MODEL_TECHNICAL.md)
- [Transfer learning document](docs/advanced_tutorials/TRANSFER_LEARNING.md)
G
Guanghua Yu 已提交
123 124 125
- [Parameter configuration](docs/advanced_tutorials/config_doc):
  - [Introduction to the configuration workflow](docs/advanced_tutorials/config_doc/CONFIG.md)
  - [Parameter configuration for RCNN model](docs/advanced_tutorials/config_doc/RCNN_PARAMS_DOC.md)
Q
qingqing01 已提交
126
- [IPython Notebook demo](demo/mask_rcnn_demo.ipynb)
G
Guanghua Yu 已提交
127
- [Model compression](slim)
G
Guanghua Yu 已提交
128 129 130 131 132
    - [Model compression benchmark](slim)
    - [Quantization](slim/quantization)
    - [Model pruning](slim/prune)
    - [Model distillation](slim/distillation)
    - [Neural Architecture Search](slim/nas)
G
Guanghua Yu 已提交
133 134 135 136
- [Deployment](deploy)
    - [Export model for inference](docs/advanced_tutorials/deploy/EXPORT_MODEL.md)
    - [Python inference](deploy/python)
    - [C++ inference](deploy/cpp)
G
Guanghua Yu 已提交
137
    - [Inference benchmark](docs/advanced_tutorials/inference/BENCHMARK_INFER_cn.md)
138 139 140 141

## Model Zoo

- Pretrained models are available in the [PaddleDetection model zoo](docs/MODEL_ZOO.md).
142
- [Mobile models](configs/mobile/README.md)
Q
qingqing01 已提交
143 144 145 146 147
- [Anchor free models](configs/anchor_free/README.md)
- [Face detection models](docs/featured_model/FACE_DETECTION_en.md)
- [Pretrained models for pedestrian detection](docs/featured_model/CONTRIB.md)
- [Pretrained models for vehicle detection](docs/featured_model/CONTRIB.md)
- [YOLOv3 enhanced model](docs/featured_model/YOLOv3_ENHANCEMENT.md): Compared to MAP of 33.0% in paper, enhanced YOLOv3 reaches the MAP of 43.6%, and inference speed is improved as well
148 149
- [Objects365 2019 Challenge champion model](docs/featured_model/champion_model/CACascadeRCNN.md)
- [Best single model of Open Images 2019-Object Detction](docs/featured_model/champion_model/OIDV5_BASELINE_MODEL.md)
150
- [Practical Server-side detection method](configs/rcnn_enhance/README_en.md): Inference speed on single V100 GPU can reach 20FPS when COCO mAP is 47.8%.
151 152


G
Guanghua Yu 已提交
153 154
## License
PaddleDetection is released under the [Apache 2.0 license](LICENSE).
155 156

## Updates
G
Guanghua Yu 已提交
157
v0.3.0 was released at `05/2020`, add anchor-free, EfficientDet, YOLOv4, etc. Launched mobile and server-side practical and efficient multiple models. For example, the YOLOv3-MobileNetv3 mobile side model is accelerated 3.5 times, the server side has optimized the two-stage model, and the speed and accuracy have high cost performance. We also refactored predictive deployment functions, and improved ease of use, fix many known bugs, etc.
G
Guanghua Yu 已提交
158
Please refer to [版本更新文档](docs/CHANGELOG.md) for details.
159 160 161 162

## Contributing

Contributions are highly welcomed and we would really appreciate your feedback!!