update release 2.6 doc, test=document_fix (#7754)

docs/CHANGELOG.md

@@ -4,6 +4,38 @@
 
 ## 最新版本信息
 
+### 2.6(02.15/2023)
+
+- 特色模型
+  - 发布旋转框检测模型PP-YOLOE-R：Anchor-free旋转框检测SOTA模型，精度速度双高、云边一体，s/m/l/x四个模型适配不用算力硬件、部署友好，避免使用特殊算子，能够轻松使用TensorRT加速；
+  - 发布小目标检测模型PP-YOLOE-SOD：基于切图的端到端检测方案、基于原图的检测模型，精度达VisDrone开源最优；
+  - 发布密集检测模型：基于PP-YOLOE+的密集检测算法，SKU数据集检测精度60.3，达到开源最优
+- 前沿算法
+  - YOLO家族新增前沿算法YOLOv8，更新YOLOv6-v3.0
+  - 新增目标检测算法DINO，YOLOF
+  - 新增ViTDet系列检测模型，PP-YOLOE+ViT_base, Mask RCNN + ViT_base, Mask RCNN + ViT_large
+  - 新增多目标跟踪算法CenterTrack
+  - 新增旋转框检测算法FCOSR
+  - 新增实例分割算法QueryInst
+  - 新增3D关键点检测算法Metro3d
+  - 新增模型蒸馏算法FGD，LD，CWD，新增PP-YOLOE+模型蒸馏，精度提升1.1 mAP
+  - 新增半监督检测算法 DenseTeacher，并适配PP-YOLOE+
+  - 新增少样本迁移学习方案，包含Co-tuning，Contrastive learning两类算法
+- 场景能力
+  - PP-Human v2开源边缘端实时检测模型，精度45.7，Jetson AGX速度80FPS
+  - PP-Vehicle开源边缘端实时检测模型，精度53.5，Jetson AGX速度80FPS
+  - PP-Human v2，PP-Vehicle支持多路视频流部署能力，实现Jetson AGX 4路视频流端到端20FPS实时部署
+  - PP-Vehicle新增车辆压线检测和车辆逆行检测能力
+- 框架能力
+  - 功能新增
+    - 新增检测热力图可视化能力，适配FasterRCNN/MaskRCNN系列, PP-YOLOE系列, BlazeFace, SSD, RetinaNet
+  - 功能完善/Bug修复
+    - 支持python3.10版本
+    - EMA支持过滤不更新参数
+    - 简化PP-YOLOE architecture架构代码
+    - AdamW适配paddle2.4.1版本
+
+
 ### 2.5(08.26/2022)
 
 - 特色模型

docs/CHANGELOG_en.md

@@ -4,6 +4,44 @@ English | [简体中文](./CHANGELOG.md)
 
 ## Last Version Information
 
+### 2.6(02.15/2023)
+
+- Featured model
+
+  - Release rotated object detector PP-YOLOE-R：SOTA Anchor-free rotated object detection model with high accuracy and efficiency. It has a series of models, named s/m/l/x, for cloud and edge devices and avoids using special operators to be deployed friendly with TensorRT.
+    - Release small object detector PP-YOLOE-SOD: End-to-end detection pipeline based on sliced images and SOTA model on VisDrone based on original images.
+    - Release crowded object detector: Crowded object detection model with top accuracy on SKU dataset.
+
+- Functions in different scenarios
+
+  - Release real-time object detection model on edge device in PP-Human v2. The model reaches 45.7mAP and 80FPS on Jetson AGX
+  - Release real-time object detection model on edge device in PP-Vehicle. The model reaches 53.5mAP and 80FPS on Jetson AGX
+  - Support multi-stream deployment in PP-Human v2 and PP-Vehicle. Achieved 20FPS in 4-stream deployment on Jetson AGX
+  - Support retrograde and press line detection in PP-Vehicle
+
+- Cutting-edge algorithms
+
+  - Release YOLOv8 and YOLOv6 3.0 in YOLO Family
+  - Release object detection algorithm DINO, YOLOF
+  - Rich ViTDet series including PP-YOLOE+ViT_base, Mask RCNN + ViT_base, Mask RCNN + ViT_large
+  - Release MOT algorithm CenterTrack
+  - Release oriented object detection algorithm FCOSR
+  - Release instance segmentation algorithm QueryInst
+  - Release 3D keypoint detection algorithm Metro3d
+  - Release distillation algorithm FGD，LD，CWD and PP-YOLOE+ distillation with improvement of 1.1+ mAP
+  - Release SSOD algorithm DenseTeacher and adapt for PP-YOLOE+
+  - Release few shot finetuning algorithm, including Co-tuning and Contrastive learning
+
+- Framework capabilities
+
+  - New functions
+    - Release Grad-CAM for heatmap visualization. Support Faster RCNN, Mask RCNN, PP-YOLOE, BlazeFace, SSD, RetinaNet.
+  - Improvement and fixes
+    - Support python 3.10
+    - Fix EMA for no-grad parameters
+    - Simplify PP-YOLOE architecture
+    - Support AdamW for Paddle 2.4.1
+
 ### 2.5(08.26/2022)
 
 - Featured model

docs/MODEL_ZOO_cn.md

@@ -139,6 +139,10 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型
 
 请参考[Vision Transformer](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/vitdet)
 
+### DINO
+
+请参考[DINO](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/dino)
+
 ### YOLOX
 
 请参考[YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolox)
@@ -162,6 +166,10 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型
 
 请参考[SOLOv2](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/solov2/)
 
+### QueryInst
+
+请参考[QueryInst](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/queryinst)
+
 
 ## [PaddleYOLO](https://github.com/PaddlePaddle/PaddleYOLO)
 

docs/MODEL_ZOO_en.md

@@ -138,6 +138,10 @@ Please refer to [Sparse R-CNN](https://github.com/PaddlePaddle/PaddleDetection/t
 
 Please refer to [Vision Transformer](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/vitdet)
 
+### DINO
+
+Please refer to [DINO](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/dino)
+
 ### YOLOX
 
 Please refer to [YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolox)
@@ -161,6 +165,10 @@ Please refer to [Cascade R-CNN](https://github.com/PaddlePaddle/PaddleDetection/
 
 Please refer to [SOLOv2](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/solov2/)
 
+### QueryInst
+
+Please refer to [QueryInst](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/queryinst)
+
 
 ## [PaddleYOLO](https://github.com/PaddlePaddle/PaddleYOLO)
 

docs/tutorials/INSTALL.md

@@ -12,7 +12,7 @@ For general information about PaddleDetection, please see [README.md](https://gi
 
 - PaddlePaddle 2.2
 - OS 64 bit
-- Python 3(3.5.1+/3.6/3.7/3.8/3.9)，64 bit
+- Python 3(3.5.1+/3.6/3.7/3.8/3.9/3.10)，64 bit
 - pip/pip3(9.0.1+), 64 bit
 - CUDA >= 10.2
 - cuDNN >= 7.6
@@ -22,7 +22,8 @@ Dependency of PaddleDetection and PaddlePaddle:
 
 | PaddleDetection version | PaddlePaddle version  |    tips    |
 | :----------------: | :---------------: | :-------: |
-|    develop           |       develop   |     Dygraph mode is set as default    |
+|    develop           |       >= 2.3.2   |     Dygraph mode is set as default    |
+|    release/2.6       |       >= 2.3.2   |     Dygraph mode is set as default    |
 |    release/2.5       |       >= 2.2.2   |     Dygraph mode is set as default    |
 |    release/2.4       |       >= 2.2.2   |     Dygraph mode is set as default    |
 |    release/2.3       |       >= 2.2.0rc |     Dygraph mode is set as default    |
@@ -42,10 +43,10 @@ Dependency of PaddleDetection and PaddlePaddle:
 ```
 
 # CUDA10.2
-python -m pip install paddlepaddle-gpu==2.2.2 -i https://mirror.baidu.com/pypi/simple
+python -m pip install paddlepaddle-gpu==2.3.2 -i https://mirror.baidu.com/pypi/simple
 
 # CPU
-python -m pip install paddlepaddle==2.2.2 -i https://mirror.baidu.com/pypi/simple
+python -m pip install paddlepaddle==2.3.2 -i https://mirror.baidu.com/pypi/simple
 ```
 
 - For more CUDA version or environment to quick install, please refer to the [PaddlePaddle Quick Installation document](https://www.paddlepaddle.org.cn/install/quick)

docs/tutorials/INSTALL_cn.md

@@ -7,9 +7,9 @@
 
 ## 环境要求
 
-- PaddlePaddle 2.2
+- PaddlePaddle 2.3.2
 - OS 64位操作系统
-- Python 3(3.5.1+/3.6/3.7/3.8/3.9)，64位版本
+- Python 3(3.5.1+/3.6/3.7/3.8/3.9/3.10)，64位版本
 - pip/pip3(9.0.1+)，64位版本
 - CUDA >= 10.2
 - cuDNN >= 7.6
@@ -18,7 +18,8 @@ PaddleDetection 依赖 PaddlePaddle 版本关系：
 
 |  PaddleDetection版本  | PaddlePaddle版本  |    备注    |
 | :------------------: | :---------------: | :-------: |
-|    develop           |       develop    |     默认使用动态图模式    |
+|    develop           |       >=2.3.2     |     默认使用动态图模式    |
+|    release/2.6       |       >=2.3.2     |     默认使用动态图模式    |
 |    release/2.5       |       >= 2.2.2    |     默认使用动态图模式    |
 |    release/2.4       |       >= 2.2.2    |     默认使用动态图模式    |
 |    release/2.3       |       >= 2.2.0rc  |     默认使用动态图模式    |
@@ -36,10 +37,10 @@ PaddleDetection 依赖 PaddlePaddle 版本关系：
 
 ```
 # CUDA10.2
-python -m pip install paddlepaddle-gpu==2.2.2 -i https://mirror.baidu.com/pypi/simple
+python -m pip install paddlepaddle-gpu==2.3.2 -i https://mirror.baidu.com/pypi/simple
 
 # CPU
-python -m pip install paddlepaddle==2.2.2 -i https://mirror.baidu.com/pypi/simple
+python -m pip install paddlepaddle==2.3.2 -i https://mirror.baidu.com/pypi/simple
 ```
 - 更多CUDA版本或环境快速安装，请参考[PaddlePaddle快速安装文档](https://www.paddlepaddle.org.cn/install/quick)
 - 更多安装方式例如conda或源码编译安装方法，请参考[PaddlePaddle安装文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/index_cn.html)