diff --git a/configs/pedestrian/README.md b/configs/pedestrian/README.md index f140360ffd00e12f66542ca6cc676954d34ff9cc..f9ba42a1985cb3dcad00d6a3b621d24f37e338ac 100644 --- a/configs/pedestrian/README.md +++ b/configs/pedestrian/README.md @@ -45,6 +45,6 @@ python -u tools/infer.py -c configs/pedestrian/pedestrian_yolov3_darknet.yml \ Some inference results are visualized below: -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/PedestrianDetection_001.png) +![](../../docs/images/PedestrianDetection_001.png) -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/PedestrianDetection_004.png) +![](../../docs/images/PedestrianDetection_004.png) diff --git a/configs/pedestrian/README_cn.md b/configs/pedestrian/README_cn.md index 6439e1f1d9ae70167e8d50158bae305f1c9cc6d3..a1d8b86dbf941427ec4a56e2b99b6fb7cc6a2004 100644 --- a/configs/pedestrian/README_cn.md +++ b/configs/pedestrian/README_cn.md @@ -46,6 +46,6 @@ python -u tools/infer.py -c configs/pedestrian/pedestrian_yolov3_darknet.yml \ 预测结果示例： -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/PedestrianDetection_001.png) +![](../../docs/images/PedestrianDetection_001.png) -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/PedestrianDetection_004.png) +![](../../docs/images/PedestrianDetection_004.png) diff --git a/configs/ppyolo/README.md b/configs/ppyolo/README.md index 1dec65f93dd768e91f433d8233526495079defa8..fe811c009ec5463adfe29aebf8909b643a153a26 100644 --- a/configs/ppyolo/README.md +++ b/configs/ppyolo/README.md @@ -11,12 +11,12 @@ English | [简体中文](README_cn.md) ## Introduction -[PP-YOLO](https://arxiv.org/abs/2007.12099) is a optimized model based on YOLOv3 in PaddleDetection，whose performance(mAP on COCO) and inference spped are better than [YOLOv4](https://arxiv.org/abs/2004.10934)，PaddlePaddle 2.0.0rc1(available on pip now) or [Daily Version](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/Tables.html#whl-release) is required to run this PP-YOLO。 +[PP-YOLO](https://arxiv.org/abs/2007.12099) is a optimized model based on YOLOv3 in PaddleDetection，whose performance(mAP on COCO) and inference spped are better than [YOLOv4](https://arxiv.org/abs/2004.10934)，PaddlePaddle 2.0.2(available on pip now) or [Daily Version](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/Tables.html#whl-develop) is required to run this PP-YOLO。 PP-YOLO reached mmAP(IoU=0.5:0.95) as 45.9% on COCO test-dev2017 dataset, and inference speed of FP32 on single V100 is 72.9 FPS, inference speed of FP16 with TensorRT on single V100 is 155.6 FPS.
- +
PP-YOLO improved performance and speed of YOLOv3 with following methods: @@ -213,7 +213,7 @@ 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 [Model Zoo](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/docs/MODEL_ZOO.md) for details. +- [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. ## Citation diff --git a/configs/ppyolo/README_cn.md b/configs/ppyolo/README_cn.md index 704b45a77d34dc709eca6f639e3db2832bafb16e..4a17db7a20bf3b0d27dbc9f41cda282fadae6fcc 100644 --- a/configs/ppyolo/README_cn.md +++ b/configs/ppyolo/README_cn.md @@ -11,12 +11,12 @@ ## 简介 -[PP-YOLO](https://arxiv.org/abs/2007.12099)是PaddleDetection优化和改进的YOLOv3的模型，其精度(COCO数据集mAP)和推理速度均优于[YOLOv4](https://arxiv.org/abs/2004.10934)模型，要求使用PaddlePaddle 2.0.0rc1(可使用pip安装) 或适当的[develop版本](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/Tables.html#whl-release)。 +[PP-YOLO](https://arxiv.org/abs/2007.12099)是PaddleDetection优化和改进的YOLOv3的模型，其精度(COCO数据集mAP)和推理速度均优于[YOLOv4](https://arxiv.org/abs/2004.10934)模型，要求使用PaddlePaddle 2.0.2(可使用pip安装) 或适当的[develop版本](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/install/Tables.html#whl-develop)。 PP-YOLO在[COCO](http://cocodataset.org) test-dev2017数据集上精度达到45.9%，在单卡V100上FP32推理速度为72.9 FPS, V100上开启TensorRT下FP16推理速度为155.6 FPS。
- +
PP-YOLO从如下方面优化和提升YOLOv3模型的精度和速度： @@ -207,7 +207,7 @@ PP-YOLO模型相对于YOLOv3模型优化项消融实验数据如下表所示。 - 精度与推理速度数据均为使用输入图像尺寸为608的测试结果 - Box AP为在COCO train2017数据集训练，val2017和test-dev2017数据集上评估`mAP(IoU=0.5:0.95)`数据 - 推理速度为单卡V100上，batch size=1, 使用上述benchmark测试方法的测试结果，测试环境配置为CUDA 10.2，CUDNN 7.5.1 -- [YOLOv3-DarkNet53](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolov3/yolov3_darknet53_270e_coco.yml)精度38.9为PaddleDetection优化后的YOLOv3模型，可参见[模型库](https://github.com/PaddlePaddle/PaddleDetection/blob/master/docs/MODEL_ZOO.md) +- [YOLOv3-DarkNet53](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolov3/yolov3_darknet53_270e_coco.yml)精度38.9为PaddleDetection优化后的YOLOv3模型，可参见[YOLOv3](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/yolov3/README.md) ## 引用 diff --git a/configs/vehicle/README.md b/configs/vehicle/README.md index 28a524eb17b9a416641d709c99016667bf3cf6e0..5e20c6ffa86dcbc6e7d9fa8fbf57a9ae98eccb74 100644 --- a/configs/vehicle/README.md +++ b/configs/vehicle/README.md @@ -48,6 +48,6 @@ python -u tools/infer.py -c configs/vehicle/vehicle_yolov3_darknet.yml \ Some inference results are visualized below: -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/VehicleDetection_001.jpeg) +![](../../docs/images/VehicleDetection_001.jpeg) -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/VehicleDetection_005.png) +![](../../docs/images/VehicleDetection_005.png) diff --git a/configs/vehicle/README_cn.md b/configs/vehicle/README_cn.md index 275075d69fe9787eff57643e269c707aa1a844aa..2bd09bb10bb4ab6e56f15fb4411ecd012249b677 100644 --- a/configs/vehicle/README_cn.md +++ b/configs/vehicle/README_cn.md @@ -49,6 +49,6 @@ python -u tools/infer.py -c configs/vehicle/vehicle_yolov3_darknet.yml \ 预测结果示例： -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/VehicleDetection_001.jpeg) +![](../../docs/images/VehicleDetection_001.jpeg) -![](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/docs/images/VehicleDetection_005.png) +![](../../docs/images/VehicleDetection_005.png) diff --git a/docs/advanced_tutorials/MODEL_TECHNICAL.md b/docs/advanced_tutorials/MODEL_TECHNICAL.md index 614f0985e11ecce1e4cb187818a0534c76a7a7eb..1d3e58d909c9e5ae48028c4bc0beb71ad08bf363 100644 --- a/docs/advanced_tutorials/MODEL_TECHNICAL.md +++ b/docs/advanced_tutorials/MODEL_TECHNICAL.md @@ -398,7 +398,7 @@ OptimizerBuilder: type: L2 ``` **几点说明：** -- 可以通过OptimizerBuilder.optimizer指定优化器的类型及参数，目前支持的优化可以参考[PaddlePaddle官方文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/api/paddle/optimizer/Overview_cn.html) +- 可以通过OptimizerBuilder.optimizer指定优化器的类型及参数，目前支持的优化器可以参考[PaddlePaddle官方文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/api/paddle/optimizer/Overview_cn.html) - 可以设置LearningRate.schedulers设置不同学习率调整策略的组合，PaddlePaddle目前支持多种学习率调整策略，具体也可参考[PaddlePaddle官方文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/api/paddle/optimizer/Overview_cn.html)。需要注意的是，你需要对于PaddlePaddle中的学习率调整策略进行简单的封装，具体可参考源码`ppdet/optimizer.py`。 ##### 2.2.3Reader配置文件 diff --git a/docs/advanced_tutorials/READER.md b/docs/advanced_tutorials/READER.md index 32df651c77dea9c8becc787fc3896c4947238b1d..bc087f8959f008b8a9ac6b5613c1aca710f239a1 100644 --- a/docs/advanced_tutorials/READER.md +++ b/docs/advanced_tutorials/READER.md @@ -65,7 +65,7 @@ PaddleDetection的数据处理模块的所有代码逻辑在`ppdet/data/`中， } ``` -xxx_rec中的内容也可以通过`DetDataSet`的data_fields参数来控制，即可以过滤掉一些不需要的字段，但大多数情况下不需要修改，按照`configs/dataset`中的默认配置即可。 +xxx_rec中的内容也可以通过`DetDataSet`的data_fields参数来控制，即可以过滤掉一些不需要的字段，但大多数情况下不需要修改，按照`configs/datasets`中的默认配置即可。 此外，在parse_dataset函数中，保存了类别名到id的映射的一个字典`cname2cid`。在coco数据集中，会利用[COCO API](https://github.com/cocodataset/cocoapi)从标注文件中加载数据集的类别名，并设置此字典。在voc数据集中，如果设置`use_default_label=False`，将从`label_list.txt`中读取类别列表，反之将使用voc默认的类别列表。 @@ -153,7 +153,7 @@ COCO数据集目前分为COCO2014和COCO2017，主要由json文件和image文件 from . import xxx from .xxx import * ``` -完成以上两步就将新的数据源`XXXDataSet`添加好了，你可以参考[配置及运行](#配置及运行)实现自定义数据集的使用。 +完成以上两步就将新的数据源`XXXDataSet`添加好了，你可以参考[配置及运行](#5.配置及运行)实现自定义数据集的使用。 ### 3.数据预处理 diff --git a/docs/images/PedestrianDetection_001.png b/docs/images/PedestrianDetection_001.png new file mode 100644 index 0000000000000000000000000000000000000000..5194d6ff891b9507fedfc53f36de4f00219c7f30 Binary files /dev/null and b/docs/images/PedestrianDetection_001.png differ diff --git a/docs/images/PedestrianDetection_004.png b/docs/images/PedestrianDetection_004.png new file mode 100644 index 0000000000000000000000000000000000000000..7c62be5051f9a47c5f5e98ccd9f45c3fa5f30257 Binary files /dev/null and b/docs/images/PedestrianDetection_004.png differ diff --git a/docs/images/VehicleDetection_001.jpeg b/docs/images/VehicleDetection_001.jpeg new file mode 100644 index 0000000000000000000000000000000000000000..aa2b679d4d2a73487edd5f9c67323ab18df93893 Binary files /dev/null and b/docs/images/VehicleDetection_001.jpeg differ diff --git a/docs/images/VehicleDetection_005.png b/docs/images/VehicleDetection_005.png new file mode 100644 index 0000000000000000000000000000000000000000..57f918a30fcc5bf7bda284c1a1a0304e8822d325 Binary files /dev/null and b/docs/images/VehicleDetection_005.png differ diff --git a/docs/images/ppyolo_map_fps.png b/docs/images/ppyolo_map_fps.png new file mode 100644 index 0000000000000000000000000000000000000000..c66ad2fb490d661fa9a773aa382ea5911957994e Binary files /dev/null and b/docs/images/ppyolo_map_fps.png differ