- We use a multi-scale evaluation strategy to get the mAP in `Easy/Medium/Hard Set`. Please refer to the [evaluation on the WIDER FACE dataset](#Evaluated-on-the-WIDER-FACE-Dataset) for details.
We provide standalone deploy of PaddleInference(Server-GPU)、PaddleLite(mobile、ARM)、Third-Engine(MNN、OpenVino), which is independent of training codes。For detail, please click [Deploy-docs](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/README_en.md)。
We provide standalone deploy of PaddleInference(Server-GPU)、PaddleLite(mobile、ARM)、Third-Engine(MNN、OpenVino), which is independent of training codes。For detail, please click [Deploy-docs](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/README_en.md)。
## Train with custom data
...
...
@@ -225,7 +225,7 @@ For more configs, please refer to [KeyPointConfigGuide](../../docs/tutorials/Key
## BenchMark
We provide benchmarks in different runtime environments for your reference when choosing models. See [Keypoint Inference Benchmark](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/keypoint/KeypointBenchmark.md) for details.
We provide benchmarks in different runtime environments for your reference when choosing models. See [Keypoint Inference Benchmark](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/keypoint/KeypointBenchmark.md) for details.
@@ -17,7 +17,7 @@ The network for detecting vehicles is YOLOv3, the backbone of which is Dacknet53
### 2. Configuration for training
PaddleDetection provides users with a configuration file [yolov3_darknet53_270e_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/yolov3/yolov3_darknet53_270e_coco.yml) to train YOLOv3 on the COCO dataset, compared with this file, we modify some parameters as followed to conduct the training for pedestrian detection:
PaddleDetection provides users with a configuration file [yolov3_darknet53_270e_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/yolov3/yolov3_darknet53_270e_coco.yml) to train YOLOv3 on the COCO dataset, compared with this file, we modify some parameters as followed to conduct the training for pedestrian detection:
-<aname="latency">Latency:</a> All our models test on `Intel core i7 10750H` CPU with MKLDNN by 12 threads and `Qualcomm Snapdragon 865(4xA77+4xA55)` with 4 threads by arm8 and with FP16. In the above table, test CPU latency on Paddle-Inference and testing Mobile latency with `Lite`->[Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite).
- PicoDet is trained on COCO train2017 dataset and evaluated on COCO val2017. And PicoDet used 4 GPUs for training and all checkpoints are trained with default settings and hyperparameters.
- Benchmark test: When testing the speed benchmark, the post-processing is not included in the exported model, you need to set `-o export.benchmark=True` or manually modify [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/runtime.yml#L12).
- Benchmark test: When testing the speed benchmark, the post-processing is not included in the exported model, you need to set `-o export.benchmark=True` or manually modify [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/runtime.yml#L12).
</details>
...
...
@@ -82,8 +82,8 @@ We developed a series of lightweight models, named `PP-PicoDet`. Because of the
- If no post processing is required, please specify: `-o export.benchmark=True` (if -o has already appeared, delete -o here) or manually modify corresponding fields in [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/runtime.yml).
- If no NMS is required, please specify: `-o export.nms=True` or manually modify corresponding fields in [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/runtime.yml). Many scenes exported to ONNX only support single input and fixed shape output, so if exporting to ONNX, it is recommended not to export NMS.
- If no post processing is required, please specify: `-o export.benchmark=True` (if -o has already appeared, delete -o here) or manually modify corresponding fields in [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/runtime.yml).
- If no NMS is required, please specify: `-o export.nms=True` or manually modify corresponding fields in [runtime.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/runtime.yml). Many scenes exported to ONNX only support single input and fixed shape output, so if exporting to ONNX, it is recommended not to export NMS.
Please refer this [documentation](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/picodet/legacy_model/pruner/README.md) for details such as requirements, training and deployment.
Please refer this [documentation](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/picodet/legacy_model/pruner/README.md) for details such as requirements, training and deployment.
</details>
## Application
-**Pedestrian detection:** model zoo of `PicoDet-S-Pedestrian` please refer to [PP-TinyPose](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/keypoint/tiny_pose#%E8%A1%8C%E4%BA%BA%E6%A3%80%E6%B5%8B%E6%A8%A1%E5%9E%8B)
-**Pedestrian detection:** model zoo of `PicoDet-S-Pedestrian` please refer to [PP-TinyPose](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose#%E8%A1%8C%E4%BA%BA%E6%A3%80%E6%B5%8B%E6%A8%A1%E5%9E%8B)
-**Mainbody detection:** model zoo of `PicoDet-L-Mainbody` please refer to [mainbody detection](./legacy_model/application/mainbody_detection/README.md)
...
...
@@ -315,7 +315,7 @@ Please use `PicoDet-LCNet` model, which has fewer `transpose` operators.
<details>
<summary>How to count model parameters.</summary>
You can insert below code at [here](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/ppdet/engine/trainer.py#L141) to count learnable parameters.
You can insert below code at [here](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/ppdet/engine/trainer.py#L141) to count learnable parameters.
- PP-YOLO is trained on COCO train2017 dataset and evaluated on val2017 & test-dev2017 dataset，Box AP<sup>test</sup> is evaluation results of `mAP(IoU=0.5:0.95)`.
- PP-YOLO used 8 GPUs for training and mini-batch size as 24 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/docs/tutorials/FAQ).
- PP-YOLO used 8 GPUs for training and mini-batch size as 24 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/docs/tutorials/FAQ).
- PP-YOLO inference speed is tesed on single Tesla V100 with batch size as 1, CUDA 10.2, CUDNN 7.5.1, TensorRT 5.1.2.2 in TensorRT mode.
- PP-YOLO FP32 inference speed testing uses inference model exported by `tools/export_model.py` and benchmarked by running `depoly/python/infer.py` with `--run_benchmark`. All testing results do not contains the time cost of data reading and post-processing(NMS), which is same as [YOLOv4(AlexyAB)](https://github.com/AlexeyAB/darknet) in testing method.
- TensorRT FP16 inference speed testing exclude the time cost of bounding-box decoding(`yolo_box`) part comparing with FP32 testing above, which means that data reading, bounding-box decoding and post-processing(NMS) is excluded(test method same as [YOLOv4(AlexyAB)](https://github.com/AlexeyAB/darknet) too)
...
...
@@ -69,26 +69,26 @@ PP-YOLO and PP-YOLOv2 improved performance and speed of YOLOv3 with following me
| Model | GPU number | images/GPU | Model Size | input shape | Box AP<sup>val</sup> | Box AP50<sup>val</sup> | Kirin 990 1xCore(FPS) | download | config |
- PP-YOLO_MobileNetV3 is trained on COCO train2017 datast and evaluated on val2017 dataset，Box AP<sup>val</sup> is evaluation results of `mAP(IoU=0.5:0.95)`, Box AP<sup>val</sup> is evaluation results of `mAP(IoU=0.5)`.
- PP-YOLO_MobileNetV3 used 4 GPUs for training and mini-batch size as 32 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/docs/tutorials/FAQ).
- PP-YOLO_MobileNetV3 used 4 GPUs for training and mini-batch size as 32 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/docs/tutorials/FAQ).
- PP-YOLO_MobileNetV3 inference speed is tested on Kirin 990 with 1 thread.
### PP-YOLO tiny
| Model | GPU number | images/GPU | Model Size | Post Quant Model Size | input shape | Box AP<sup>val</sup> | Kirin 990 4xCore(FPS) | download | config | post quant model |
- PP-YOLO-tiny is trained on COCO train2017 datast and evaluated on val2017 dataset，Box AP<sup>val</sup> is evaluation results of `mAP(IoU=0.5:0.95)`, Box AP<sup>val</sup> is evaluation results of `mAP(IoU=0.5)`.
- PP-YOLO-tiny used 8 GPUs for training and mini-batch size as 32 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/docs/tutorials/FAQ/README.md).
- PP-YOLO-tiny used 8 GPUs for training and mini-batch size as 32 on each GPU, if GPU number and mini-batch size is changed, learning rate and iteration times should be adjusted according [FAQ](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/docs/tutorials/FAQ/README.md).
- PP-YOLO-tiny inference speed is tested on Kirin 990 with 4 threads by arm8
- we alse provide PP-YOLO-tiny post quant inference model, which can compress model to **1.3MB** with nearly no inference on inference speed and performance
...
...
@@ -98,9 +98,9 @@ PP-YOLO trained on Pascal VOC dataset as follows:
| Model | GPU number | images/GPU | backbone | input shape | Box AP50<sup>val</sup> | download | config |
@@ -212,7 +212,7 @@ Optimizing method and ablation experiments of PP-YOLO compared with YOLOv3.
- Performance and inference spedd are measure with input shape as 608
- All models are trained on COCO train2017 datast and evaluated on val2017 & test-dev2017 dataset，`Box AP` is evaluation results as `mAP(IoU=0.5:0.95)`.
- Inference speed is tested on single Tesla V100 with batch size as 1 following test method and environment configuration in benchmark above.
-[YOLOv3-DarkNet53](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolov3/yolov3_darknet53_270e_coco.yml) with mAP as 39.0 is optimized YOLOv3 model in PaddleDetection，see [YOLOv3](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/yolov3/README.md) for details.
-[YOLOv3-DarkNet53](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/yolov3/yolov3_darknet53_270e_coco.yml) with mAP as 39.0 is optimized YOLOv3 model in PaddleDetection，see [YOLOv3](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/yolov3/README.md) for details.
| Backbone | Network type | Number of images per GPU | Learning rate strategy | Inferring time(fps) | Box AP | Mask AP | Download | Configuration File |
- The above V100 prediction delay non-quantified model is tested by TensorRT FP32, and the quantified model is tested by TensorRT INT8, and both of them include NMS time.
...
...
@@ -153,8 +153,8 @@ Description:
| Model | Compression Strategy | Input Size | Box AP | Download | Model Configuration File | Compression Strategy Configuration File |
COCO数据集作为目标检测任务的训练目标难度更大，意味着teacher网络会预测出更多的背景bbox，如果直接用teacher的预测输出作为student学习的`soft label`会有严重的类别不均衡问题。解决这个问题需要引入新的方法，详细背景请参考论文:[Object detection at 200 Frames Per Second](https://arxiv.org/abs/1805.06361)。
**Attention:** Due to the overall upgrade of the dynamic graph framework, the weighting model published by PaddleDetection of PAF Net needs to be evaluated with a --bias field, for example
@@ -17,7 +17,7 @@ The network for detecting vehicles is YOLOv3, the backbone of which is Dacknet53
### 2. Configuration for training
PaddleDetection provides users with a configuration file [yolov3_darknet53_270e_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/yolov3/yolov3_darknet53_270e_coco.yml) to train YOLOv3 on the COCO dataset, compared with this file, we modify some parameters as followed to conduct the training for vehicle detection:
PaddleDetection provides users with a configuration file [yolov3_darknet53_270e_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/configs/yolov3/yolov3_darknet53_270e_coco.yml) to train YOLOv3 on the COCO dataset, compared with this file, we modify some parameters as followed to conduct the training for vehicle detection:
* (3) 下载`JetPack`，请参考[NVIDIA Jetson Linux Developer Guide](https://docs.nvidia.com/jetson/l4t/index.html) 中的`Preparing a Jetson Developer Kit for Use`章节内容进行刷写系统镜像。
本文档使用`Paddle`在`JetPack4.3`上预先编译好的预测库，请根据硬件在[安装与编译 Linux 预测库](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/05_inference_deployment/inference/build_and_install_lib_cn.html) 中选择对应版本的`Paddle`预测库。
@@ -14,7 +14,7 @@ PP-Human supports various inputs such as images, single-camera, and multi-camera
## 📣 Updates
- 🔥 **2022.7.13：PP-Human v2 launched with a full upgrade of four industrial features: behavior analysis, attributes recognition, visitor traffic statistics and ReID. It provides a strong core algorithm for pedestrian detection, tracking and attribute analysis with a simple and detailed development process and model optimization strategy.**
- 🔥 **2022.7.13：PP-Human v2 launched with a full upgrade of four industrial features: behavior analysis, attributes recognition, visitor traffic statistics and ReID. It provides a strong core algorithm for pedestrian detection, tracking and attribute analysis with a simple and detailed release/2.5ment process and model optimization strategy.**
- 2022.4.18: Add PP-Human practical tutorials, including training, deployment, and action expansion. Details for AIStudio project please see [Link](https://aistudio.baidu.com/aistudio/projectdetail/3842982)
- 2022.4.10: Add PP-Human examples; empower refined management of intelligent community management. A quick start for AIStudio [Link](https://aistudio.baidu.com/aistudio/projectdetail/3679564)
...
...
@@ -79,7 +79,7 @@ Click to download the model, then unzip and save it in the `. /output_inference`
@@ -97,7 +97,7 @@ SKELETON_ACTION: # Config for skeleton-based action recognition model
1: Others
```
- The falling action recognition model uses [ST-GCN](https://arxiv.org/abs/1801.07455), and employ the [PaddleVideo](https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/model_zoo/recognition/stgcn.md) toolkit to complete model training.
- The falling action recognition model uses [ST-GCN](https://arxiv.org/abs/1801.07455), and employ the [PaddleVideo](https://github.com/PaddlePaddle/PaddleVideo/blob/release/2.5/docs/zh-CN/model_zoo/recognition/stgcn.md) toolkit to complete model training.
@@ -138,7 +138,7 @@ ID_BASED_CLSACTION: # config for classfication-based action recognition model
1. Get the pedestrian detection box and the tracking ID number of the video input through object detection and multi-object tracking. The adopted model is PP-YOLOE, and for details, please refer to [PP-YOLOE](../../../configs/ppyoloe).
2. Capture every pedestrian in frames of the input video accordingly by using the coordinate of the detection box.
3. With image classification through pedestrian images at the frame level, when the category to which the image belongs is the corresponding behavior, it is considered that the character is in the behavior state for a certain period of time. This task is implemented with [PP-HGNet](https://github.com/PaddlePaddle/PaddleClas/blob/develop/docs/zh_CN/models/PP-HGNet.md). In current version, the behavior of calling is supported and the relationship between the action type and `class id` is:
3. With image classification through pedestrian images at the frame level, when the category to which the image belongs is the corresponding behavior, it is considered that the character is in the behavior state for a certain period of time. This task is implemented with [PP-HGNet](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.5/docs/zh_CN/models/PP-HGNet.md). In current version, the behavior of calling is supported and the relationship between the action type and `class id` is:
```
0: Calling
...
...
@@ -247,7 +247,7 @@ The result is shown as follow:
Data source and copyright owner: Surveillance Camera Fight Dataset.
### Introduction to the Solution
The current fight recognition model is using [PP-TSM](https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/model_zoo/recognition/pp-tsm.md), and adaptated to complete the model training. For the input video or video stream, we extraction frame at a certain interval. When the video frame accumulates to the specified number, it is input into the video classification model to determine whether there is fighting.
The current fight recognition model is using [PP-TSM](https://github.com/PaddlePaddle/PaddleVideo/blob/release/2.5/docs/zh-CN/model_zoo/recognition/pp-tsm.md), and adaptated to complete the model training. For the input video or video stream, we extraction frame at a certain interval. When the video frame accumulates to the specified number, it is input into the video classification model to determine whether there is fighting.
@@ -37,7 +37,7 @@ PP-Tracking supports GUI predict and deployment. Please refer to this [doc](http
PP-Tracking supports two paradigms: single camera tracking (MOT) and multi-camera tracking (MTMCT).
- Single camera tracking supports **FairMOT** and **DeepSORT** two MOT models, multi-camera tracking only support **DeepSORT**.
- The applications of single camera tracking include pedestrian tracking, vehicle tracking, multi-class tracking, small object tracking and traffic statistics. The models are mainly optimized based on FairMOT to achieve the effect of real-time tracking. At the same time, PP-Tracking provides pre-training models based on different application scenarios.
- In DeepSORT (including DeepSORT used in multi-camera tracking), the selected detectors are PaddeDetection's self-developed high-performance detector [PP-YOLOv2](../../configs/ppyolo/) and lightweight detector [PP-PicoDet](../../configs/picodet/), and the selected ReID model is PaddleClas's self-developed ultra lightweight backbone [PP-LCNet](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.3/docs/zh_CN/models/PP-LCNet.md)
- In DeepSORT (including DeepSORT used in multi-camera tracking), the selected detectors are PaddeDetection's self-release/2.5ed high-performance detector [PP-YOLOv2](../../configs/ppyolo/) and lightweight detector [PP-PicoDet](../../configs/picodet/), and the selected ReID model is PaddleClas's self-release/2.5ed ultra lightweight backbone [PP-LCNet](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.3/docs/zh_CN/models/PP-LCNet.md)
PP-Tracking provids multi-scenario pre-training models and the exported models for deployment:
-[Fixed Virtual Platform (FVP) based on Arm(R) Corstone(TM)-300 software](https://developer.arm.com/tools-and-software/open-source-software/arm-platforms-software/arm-ecosystem-fvps)
-[Fixed Virtual Platform (FVP) based on Arm(R) Corstone(TM)-300 software](https://release/2.5er.arm.com/tools-and-software/open-source-software/arm-platforms-software/arm-ecosystem-fvps)
-[GCC toolchain from Arm(R)](https://developer.arm.com/-/media/Files/downloads/gnu-rm/10-2020q4/gcc-arm-none-eabi-10-2020-q4-major-x86_64-linux.tar.bz2)
-[GCC toolchain from Arm(R)](https://release/2.5er.arm.com/-/media/Files/downloads/gnu-rm/10-2020q4/gcc-arm-none-eabi-10-2020-q4-major-x86_64-linux.tar.bz2)
[Intel's OpenVINO Toolkit](https://software.intel.com/content/www/us/en/develop/tools/openvino-toolkit.html). Most of the implements in this fold are same as *demo_ncnn*.
[Intel's OpenVINO Toolkit](https://software.intel.com/content/www/us/en/release/2.5/tools/openvino-toolkit.html). Most of the implements in this fold are same as *demo_ncnn*.
**Recommand** to use the xxx.tar.gz file to install instead of github method, [link](https://registrationcenter-download.intel.com/akdlm/irc_nas/18096/l_openvino_toolkit_p_2021.4.689.tgz).
## Install OpenVINO Toolkit
Go to [OpenVINO HomePage](https://software.intel.com/content/www/us/en/develop/tools/openvino-toolkit.html)
Go to [OpenVINO HomePage](https://software.intel.com/content/www/us/en/release/2.5/tools/openvino-toolkit.html)
[Intel's OpenVINO Toolkit](https://software.intel.com/content/www/us/en/develop/tools/openvino-toolkit.html). Most of the implements in this fold are same as *demo_ncnn*.
[Intel's OpenVINO Toolkit](https://software.intel.com/content/www/us/en/release/2.5/tools/openvino-toolkit.html). Most of the implements in this fold are same as *demo_ncnn*.
**Recommand** to use the xxx.tar.gz file to install instead of github method, [link](https://registrationcenter-download.intel.com/akdlm/irc_nas/18096/l_openvino_toolkit_p_2021.4.689.tgz).
## Install OpenVINO Toolkit
Go to [OpenVINO HomePage](https://software.intel.com/content/www/us/en/develop/tools/openvino-toolkit.html)
Go to [OpenVINO HomePage](https://software.intel.com/content/www/us/en/release/2.5/tools/openvino-toolkit.html)
Download a suitable version and install.
...
...
@@ -104,7 +104,7 @@ make
Download PicoDet openvino model [PicoDet openvino model download link](https://paddledet.bj.bcebos.com/deploy/third_engine/picodet_m_416_openvino.zip).
Download TinyPose openvino model [TinyPose openvino model download link](https://paddledet.bj.bcebos.com/deploy/third_engine/tinypose_256_openvino.zip).
move picodet and tinypose openvino model files to the demo's weight folder.
move picodet and tinypose openvino model files to the demo's weight folder.
# Secondary Development for Action Recognition Task
In the process of industrial implementation, the application of action recognition algorithms will inevitably lead to the need for customized types of action, or the optimization of existing action recognition models to improve the performance of the model in specific scenarios. In view of the diversity of behaviors, PP-Human supports the identification of five abnormal behavioras of smoking, making phone calls, falling, fighting, and people intrusion. At the same time, according to the different behaviors, PP-Human integrates five action recognition technology solutions based on video classification, detection-based, image-based classification, tracking-based and skeleton-based, which can cover 90%+ action type recognition and meet various development needs. In this document, we use a case to introduce how to select a action recognition solution according to the expected behavior, and use PaddleDetection to carry out the secondary development of the action recognition algorithm, including: solution selection, data preparation, model optimization and development process for adding new actions.
In the process of industrial implementation, the application of action recognition algorithms will inevitably lead to the need for customized types of action, or the optimization of existing action recognition models to improve the performance of the model in specific scenarios. In view of the diversity of behaviors, PP-Human supports the identification of five abnormal behavioras of smoking, making phone calls, falling, fighting, and people intrusion. At the same time, according to the different behaviors, PP-Human integrates five action recognition technology solutions based on video classification, detection-based, image-based classification, tracking-based and skeleton-based, which can cover 90%+ action type recognition and meet various release/2.5ment needs. In this document, we use a case to introduce how to select a action recognition solution according to the expected behavior, and use PaddleDetection to carry out the secondary release/2.5ment of the action recognition algorithm, including: solution selection, data preparation, model optimization and release/2.5ment process for adding new actions.
@@ -182,7 +182,7 @@ At this point, this model can be used in PP-Human.
### Custom Action Output
In the model of action recognition based on classification with human id, the task is defined as a picture-level classification task of corresponding person. The type of the corresponding classification is finally regarded as the action type of the current stage. Therefore, on the basis of completing the training and deployment of the custom model, it is also necessary to convert the classification model results to the final action recognition results as output, and the displayed result of the visualization should be modified.
Please modify the [postprocessing function](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/pphuman/action_infer.py#L509).
Please modify the [postprocessing function](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/pphuman/action_infer.py#L509).
The core code are:
```python
...
...
@@ -221,4 +221,4 @@ else:
```
#### Modify Visual Output
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
@@ -170,7 +170,7 @@ At this point, this model can be used in PP-Human.
In the model of action recognition based on detection with human id, the task is defined to detect target objects in images of corresponding person. When the target object is detected, the behavior type of the character in a certain period of time. The type of the corresponding classification is regarded as the action of the current period. Therefore, on the basis of completing the training and deployment of the custom model, it is also necessary to convert the detection model results to the final action recognition results as output, and the displayed result of the visualization should be modified.
#### Convert to Action Recognition Result
Please modify the [postprocessing function](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/pphuman/action_infer.py#L338).
Please modify the [postprocessing function](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/pphuman/action_infer.py#L338).
The core code are:
```python
...
...
@@ -196,4 +196,4 @@ else:
```
#### Modify Visual Output
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
In the skeleton-based action recognition, the classification result of the model represents the behavior type of the character in a certain period of time. The type of the corresponding classification is regarded as the action of the current period. Therefore, on the basis of completing the training and deployment of the custom model, the model output is directly used as the final result, and the displayed result of the visualization should be modified.
#### Modify Visual Output
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
At present, ID-based action recognition is displayed based on the results of action recognition and predefined category names. For the detail, please refer to [here](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/pipeline.py#L1024-L1043). If the custom action needs to be modified to another display name, please modify it accordingly to output the corresponding result.
In the practical application of object detection algorithms in a specific industry, additional training is often required for practical use. The project iteration will also need to modify categories. This document details how to use PaddleDetection for a customized object detection algorithm. The process includes data preparation, model optimization roadmap, and modifying the category development process.
In the practical application of object detection algorithms in a specific industry, additional training is often required for practical use. The project iteration will also need to modify categories. This document details how to use PaddleDetection for a customized object detection algorithm. The process includes data preparation, model optimization roadmap, and modifying the category release/2.5ment process.
@@ -60,7 +60,7 @@ Augmentation of covered data in keypoint model training to improve model perform
The keypoint model is trained and predicted on the basis of image, and video input is also predicted by splitting the video into frames. Although the content is mostly similar between frames, small differences may still lead to large changes in the output of the model. As a result of that, although the predicted coordinates are roughly correct, there may be jitters in the visual effect.
By adding a smoothing filter process, the performance of the video output can be effectively improved by combining the predicted results of each frame and the historical results. For this part, please see [Filter Smoothing](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/python/det_keypoint_unite_infer.py#L206).
By adding a smoothing filter process, the performance of the video output can be effectively improved by combining the predicted results of each frame and the historical results. For this part, please see [Filter Smoothing](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/python/det_keypoint_unite_infer.py#L206).
You need to modify the following configuration in the configuration file `PaddleClas/blob/develop/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`
You need to modify the following configuration in the configuration file `PaddleClas/blob/release/2.5/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`
```
DataLoader:
...
...
@@ -196,7 +196,7 @@ If the attributes need to be added or deleted, you need to
2) Modify the number and name of attributes used in train.txt corresponding to the training.
3) Modify the training configuration, for example, the number of attributes in the ``PaddleClas/blob/develop/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`` file, for details, please see the ``Modify configuration to start training`` section above.
3) Modify the training configuration, for example, the number of attributes in the ``PaddleClas/blob/release/2.5/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`` file, for details, please see the ``Modify configuration to start training`` section above.
Example of adding attributes.
...
...
@@ -209,7 +209,7 @@ Example of adding attributes.
2) Modify the number and name of attributes used in train.txt corresponding to the training.
3) Modify the training configuration, for example, the number of attributes in the ``PaddleClas/blob/develop/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`` file, for details, please see the ``Modify configuration to start training`` section above.
3) Modify the training configuration, for example, the number of attributes in the ``PaddleClas/blob/release/2.5/ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`` file, for details, please see the ``Modify configuration to start training`` section above.
When applying multi-object tracking algorithms in industrial applications, there will be inevitable demands for customized types of multi-object tracking or optimization of existing multi-object tracking models to improve the effectiveness of the models in specific scenarios. In this document, we present examples of how to choose a multi-object tracking solution based on the expected identified behavior, and how to use PaddleDetection for further development of multi-object tracking algorithms, including data preparation, model optimization ideas, and the development process of tracking category modification.
When applying multi-object tracking algorithms in industrial applications, there will be inevitable demands for customized types of multi-object tracking or optimization of existing multi-object tracking models to improve the effectiveness of the models in specific scenarios. In this document, we present examples of how to choose a multi-object tracking solution based on the expected identified behavior, and how to use PaddleDetection for further release/2.5ment of multi-object tracking algorithms, including data preparation, model optimization ideas, and the release/2.5ment process of tracking category modification.
You need to change the following configuration items in the configuration file [softmax_triplet_with_center.yaml](https://github.com/PaddlePaddle/PaddleClas/blob/develop/ppcls/configs/reid/strong_ baseline/softmax_triplet_with_center.yaml):
You need to change the following configuration items in the configuration file [softmax_triplet_with_center.yaml](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.5/ppcls/configs/reid/strong_ baseline/softmax_triplet_with_center.yaml):
@@ -223,7 +223,7 @@ The same applies to the deletion of attributes.
After modifying the attribute definition, the post-processing part of the pipeline also needs to be modified accordingly, which mainly affects the display results when the results are visualized.
The code is at [file](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/deploy/pipeline/ppvehicle/vehicle_attr.py#L108), that is, the `postprocess` function.
The code is at [file](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/deploy/pipeline/ppvehicle/vehicle_attr.py#L108), that is, the `postprocess` function.
The function implementation is described as follows:
PaddleDetection has been a vibrant open-source project and has a large amout of contributors and maintainers around it. It is an AI framework which enables developers to quickly integrate AI capacities into their own projects and applications.
PaddleDetection has been a vibrant open-source project and has a large amout of contributors and maintainers around it. It is an AI framework which enables release/2.5ers to quickly integrate AI capacities into their own projects and applications.
Intel OpenVINO is a widely used free toolkit. It facilitates the optimization of a deep learning model from a framework and deployment using an inference engine onto Intel hardware.
Apparently, the upstream(Paddle) and the downstream(Intel OpenVINO) can work together to streamline and simplify the process of developing an AI model and deploying the model onto hardware, which, in turn, makes our lives easier.
Apparently, the upstream(Paddle) and the downstream(Intel OpenVINO) can work together to streamline and simplify the process of release/2.5ing an AI model and deploying the model onto hardware, which, in turn, makes our lives easier.
This article will show you how to use a PaddleDetection model [FairMOT](../../../configs/mot/fairmot/README.md) from the Model Zoo in PaddleDetection and use it with OpenVINO to do the inference.
This document covers how to install PaddleDetection and its dependencies
(including PaddlePaddle), together with COCO and Pascal VOC dataset.
For general information about PaddleDetection, please see [README.md](https://github.com/PaddlePaddle/PaddleDetection/tree/develop).
For general information about PaddleDetection, please see [README.md](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5).
## Requirements:
...
...
@@ -116,7 +116,7 @@ OK
We provide docker images containing the latest PaddleDetection code, and all environment and package dependencies are pre-installed. All you have to do is to **pull and run the docker image**. Then you can enjoy PaddleDetection without any extra steps.
Get these images and guidance in [docker hub](https://hub.docker.com/repository/docker/paddlecloud/paddledetection), including CPU, GPU, ROCm environment versions.
Get these images and guidance in [docker hub](https://hub.docker.com/repository/docker/paddlecloud/paddledetection), including CPU, GPU, ROCm environment versions.
If you have some customized requirements about automatic building docker images, you can get it in github repo [PaddlePaddle/PaddleCloud](https://github.com/PaddlePaddle/PaddleCloud/tree/main/tekton).
@@ -147,7 +147,7 @@ Then, it will generate following contents:
#### Annotation file(json)—>COCO Dataset
Convert the data annotated by LabelMe to COCO dataset by the script [x2coco.py](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/tools/x2coco.py) provided by PaddleDetection.
Convert the data annotated by LabelMe to COCO dataset by the script [x2coco.py](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/tools/x2coco.py) provided by PaddleDetection.
**PaddleDetection supports the format of VOC or COCO.** The annotation file generated by LabelImg needs to be converted by VOC or COCO. You can refer to [PrepareDataSet](./PrepareDataSet.md#%E5%87%86%E5%A4%87%E8%AE%AD%E7%BB%83%E6%95%B0%E6%8D%AE).
Convert the data annotated by LabelMe to COCO dataset by this script [x2coco.py](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/tools/x2coco.py).
Convert the data annotated by LabelMe to COCO dataset by this script [x2coco.py](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.5/tools/x2coco.py).
If new data in the dataset needs to be added to paddedetection, you can refer to the [add new data source] (../advanced_tutorials/READER.md#2.3_Customizing_Dataset) document section in the data processing document to develop corresponding code to complete the new data source support. At the same time, you can read the [data processing document] (../advanced_tutorials/READER.md) for specific code analysis of data processing
If new data in the dataset needs to be added to paddedetection, you can refer to the [add new data source] (../advanced_tutorials/READER.md#2.3_Customizing_Dataset) document section in the data processing document to release/2.5 corresponding code to complete the new data source support. At the same time, you can read the [data processing document] (../advanced_tutorials/READER.md) for specific code analysis of data processing
The configuration file for the Dataset exists in the `configs/datasets` folder. For example, the COCO dataset configuration file is as follows:
