fix bug in pipeline doc and update kpt training config (#6769) (#6783)

configs/pphuman/hrnet_w32_256x192.yml → configs/pphuman/dark_hrnet_w32_256x192.yml

@@ -101,7 +101,7 @@ TrainReader:
         flip_pairs: *flip_perm
     - TopDownAffine:
         trainsize: *trainsize
-    - ToHeatmapsTopDown:
+    - ToHeatmapsTopDown_DARK:
         hmsize: *hmsize
         sigma: 2
   batch_transforms:
@@ -125,6 +125,7 @@ EvalReader:
         is_scale: true
     - Permute: {}
   batch_size: 16
+  drop_empty: false
 
 TestReader:
   inputs_def:
@@ -139,4 +140,3 @@ TestReader:
         is_scale: true
     - Permute: {}
   batch_size: 1
-  fuse_normalize: false #whether to fuse nomalize layer into model while export model

deploy/pipeline/docs/tutorials/PPHuman_QUICK_STARTED.md

@@ -55,7 +55,7 @@ PP-Human提供了目标检测、属性识别、行为识别、ReID预训练模
 |  闯入识别  |   31.8ms | [多目标跟踪](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip) | 多目标跟踪：182M |
 |  打架识别  |   19.7ms | [视频分类](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip) | 90M |
 |  抽烟识别  |   单人15.1ms | [目标检测](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip)<br>[基于人体id的目标检测](https://bj.bcebos.com/v1/paddledet/models/pipeline/ppyoloe_crn_s_80e_smoking_visdrone.zip) | 目标检测：182M<br>基于人体id的目标检测：27M |
-|  打电话识别  |   单人ms | [目标检测](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip)<br>[基于人体id的图像分类](https://bj.bcebos.com/v1/paddledet/models/pipeline/PPHGNet_tiny_calling_halfbody.zip) | 目标检测：182M<br>基于人体id的图像分类：45M |
+|  打电话识别  |   单人6.0ms | [目标检测](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip)<br>[基于人体id的图像分类](https://bj.bcebos.com/v1/paddledet/models/pipeline/PPHGNet_tiny_calling_halfbody.zip) | 目标检测：182M<br>基于人体id的图像分类：45M |
 
 下载模型后，解压至`./output_inference`文件夹。
 

deploy/pipeline/docs/tutorials/PPVehicle_QUICK_STARTED.md

@@ -194,7 +194,7 @@ PP-Vehicle 整体方案如下图所示:
 
 ### 车牌识别
 - 使用PaddleOCR特色模型ch_PP-OCRv3_det+ch_PP-OCRv3_rec模型，识别车牌号码
-- 详细文档参考[属性识别](ppvehicle_plate.md)
+- 详细文档参考[车牌识别](ppvehicle_plate.md)
 
 ### 违法停车识别
 - 车辆跟踪模型使用高精度模型PP-YOLOE L，根据车辆的跟踪轨迹以及指定的违停区域判断是否违法停车，如果存在则展示违法停车车牌号。

deploy/pipeline/docs/tutorials/ppvehicle_attribute.md

@@ -6,12 +6,12 @@
 
 | 任务 | 算法 | 精度 | 预测速度 | 下载链接|
 |-----------|------|-----------|----------|---------------|
-| 车辆检测/跟踪 | PP-YOLOE | - | - | [预测部署模型](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_ppvehicle.zip) |
-| 车辆属性识别 | PPLCNet | 90.81 | 2.36 ms | [预测部署模型](https://bj.bcebos.com/v1/paddledet/models/pipeline/vehicle_attribute_model.zip) |
+| 车辆检测/跟踪 | PP-YOLOE | mAP 63.9 | 38.67ms | [预测部署模型](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_ppvehicle.zip) |
+| 车辆属性识别 | PPLCNet | 90.81 | 7.31 ms | [预测部署模型](https://bj.bcebos.com/v1/paddledet/models/pipeline/vehicle_attribute_model.zip) |
 
 
 注意：
-1. 属性模型预测速度是基于Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz 测试得到，开启 MKLDNN 加速策略，线程数为10。
+1. 属性模型预测速度是基于NVIDIA T4, 开启TensorRT FP16得到。模型预测速度包含数据预处理、模型预测、后处理部分。
 2. 关于PP-LCNet的介绍可以参考[PP-LCNet](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.4/docs/zh_CN/models/PP-LCNet.md)介绍，相关论文可以查阅[PP-LCNet paper](https://arxiv.org/abs/2109.15099)。
 3. 属性模型的训练和精度测试均基于[VeRi数据集](https://www.v7labs.com/open-datasets/veri-dataset)。
 

deploy/pipeline/docs/tutorials/ppvehicle_attribute_en.md

@@ -6,12 +6,12 @@ Vehicle attribute recognition is widely used in smart cities, smart transportati
 
 | Task | Algorithm | Precision | Inference Speed | Download |
 |-----------|------|-----------|----------|---------------------|
-| Vehicle Detection/Tracking | PP-YOLOE | - | - | [Inference and Deployment Model](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_ppvehicle.zip) |
-| Vehicle Attribute Recognition | PPLCNet | 90.81 | 2.36 ms | [Inference and Deployment Model](https://bj.bcebos.com/v1/paddledet/models/pipeline/vehicle_attribute_model.zip) |
+| Vehicle Detection/Tracking | PP-YOLOE | mAP 63.9 | 38.67ms | [Inference and Deployment Model](https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_ppvehicle.zip) |
+| Vehicle Attribute Recognition | PPLCNet | 90.81 | 7.31 ms | [Inference and Deployment Model](https://bj.bcebos.com/v1/paddledet/models/pipeline/vehicle_attribute_model.zip) |
 
 
 Note:
-1. The inference speed of the attribute model is obtained from the test on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz, with the MKLDNN acceleration strategy enabled, and 10 threads.
+1. The inference speed of the attribute model is obtained from the test on NVIDIA T4, with TensorRT FP16. The time includes data pre-process, model inference and post-process.
 2. For introductions, please refer to [PP-LCNet Series](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.4/docs/en/models/PP-LCNet_en.md). Related paper is available on PP-LCNet paper
 3. The training and test phase of vehicle attribute recognition model are both obtained from [VeRi dataset](https://www.v7labs.com/open-datasets/veri-dataset).