[Doc] Update RKYOLO Docs (#1330)

* 更新docs * 修正docs错误 * 更新docs * 更新python example脚本和ppyoloe转换脚本 * 更新PaddleDetection文档 * 更新文档 * 更新文档 * 更新文档 * 更新文档 * 更新文档 * 更新RKYOLO系列模型文档 * 更新PaddleDetection python example
2025-10-08 10:00:29 +08:00 · 2023-02-15 17:55:58 +08:00
parent da94fc46cf
commit 32af870711
6 changed files with 128 additions and 14 deletions
--- a/examples/vision/detection/paddledetection/rknpu2/README_CN.md
+++ b/examples/vision/detection/paddledetection/rknpu2/README_CN.md
@@ -50,7 +50,7 @@ paddle2onnx --model_dir picodet_s_416_coco_lcnet \
 # 固定shape
 python -m paddle2onnx.optimize --input_model picodet_s_416_coco_lcnet/picodet_s_416_coco_lcnet.onnx \
                                --output_model picodet_s_416_coco_lcnet/picodet_s_416_coco_lcnet.onnx \
-                                --input_shape_dict "{'image':[1,3,416,416]}"
+                                --input_shape_dict "{'image':[1,3,416,416], 'scale_factor':[1,2]}"
 ```

 ### 编写yaml文件
@@ -73,11 +73,12 @@ std:
 ```

 **修改outputs参数**
+
 由于Paddle2ONNX版本的不同，转换模型的输出节点名称也有所不同，请使用[Netron](https://netron.app)对模型进行可视化，并找到以下蓝色方框标记的NonMaxSuppression节点，红色方框的节点名称即为目标名称。

 例如，使用Netron可视化后，得到以下图片:

-![](https://user-images.githubusercontent.com/58363586/212599781-e1952da7-6eae-4951-8ca7-bab7e6940692.png)
+![](https://ai-studio-static-online.cdn.bcebos.com/8bce6b904a6b479e8b30da9f7c719fad57517ffb2f234aeca3b8ace0761754d5)

 找到蓝色方框标记的NonMaxSuppression节点，可以看到红色方框标记的两个节点名称为p2o.Div.79和p2o.Concat.9,因此需要修改outputs参数，修改后如下:

@@ -96,6 +97,16 @@ python tools/rknpu2/export.py --config_path tools/rknpu2/config/picodet_s_416_co
                              --target_platform rk3588
 ```

+## RKNN模型列表
+
+为了方便大家测试，我们提供picodet和ppyoloe两个模型，解压后即可使用:
+
+| 模型名称                        | 下载地址                                                                              |
+|-----------------------------|-----------------------------------------------------------------------------------|
+| picodet_s_416_coco_lcnet    | https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/picodet_s_416_coco_lcnet.zip    |
+| ppyoloe_plus_crn_s_80e_coco | https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/ppyoloe_plus_crn_s_80e_coco.zip |
+
+

 ## 其他链接

--- a/examples/vision/detection/paddledetection/rknpu2/cpp/CMakeLists.txt
+++ b/examples/vision/detection/paddledetection/rknpu2/cpp/CMakeLists.txt
@@ -14,3 +14,6 @@ target_link_libraries(infer_picodet_demo ${FASTDEPLOY_LIBS})

 add_executable(infer_yolov8_demo ${PROJECT_SOURCE_DIR}/infer_yolov8_demo.cc)
 target_link_libraries(infer_yolov8_demo ${FASTDEPLOY_LIBS})
+
+add_executable(infer_ppyoloe_demo ${PROJECT_SOURCE_DIR}/infer_ppyoloe_demo.cc)
+target_link_libraries(infer_ppyoloe_demo ${FASTDEPLOY_LIBS})
--- a/examples/vision/detection/paddledetection/rknpu2/cpp/README_CN.md
+++ b/examples/vision/detection/paddledetection/rknpu2/cpp/README_CN.md
@@ -12,7 +12,7 @@
 以上步骤请参考[RK2代NPU部署库编译](../../../../../../docs/cn/build_and_install/rknpu2.md)实现

 ```bash
-以picodet为例进行推理部署
+# 以picodet为例进行推理部署

 mkdir build
 cd build
@@ -23,6 +23,8 @@ cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-x.x.x
 make -j

 # 下载PPYOLOE模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/picodet_s_416_coco_lcnet.zip
+unzip picodet_s_416_coco_lcnet.zip
 wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/000000014439.jpg

 # CPU推理
@@ -31,13 +33,6 @@ wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/0000000
 ./infer_picodet_demo ./picodet_s_416_coco_lcnet 000000014439.jpg 1
 ```

-## 运行例程
-
-```bash
-cd ./build/install
-./infer_picodet model/picodet_s_416_coco_lcnet images/000000014439.jpg
-```
-
 ## 文档导航

 - [模型介绍](../../)
--- a/examples/vision/detection/paddledetection/rknpu2/cpp/infer_ppyoloe_demo.cc
+++ b/examples/vision/detection/paddledetection/rknpu2/cpp/infer_ppyoloe_demo.cc
@@ -0,0 +1,95 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fastdeploy/vision.h"
+
+void ONNXInfer(const std::string& model_dir, const std::string& image_file) {
+  std::string model_file = model_dir + "/yolov8_n_500e_coco.onnx";
+  std::string params_file;
+  std::string config_file = model_dir + "/infer_cfg.yml";
+  auto option = fastdeploy::RuntimeOption();
+  option.UseCpu();
+  auto format = fastdeploy::ModelFormat::ONNX;
+
+  auto model = fastdeploy::vision::detection::PPYOLOE(
+      model_file, params_file, config_file, option, format);
+
+  fastdeploy::TimeCounter tc;
+  tc.Start();
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::DetectionResult res;
+  if (!model.Predict(im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return;
+  }
+  auto vis_im = fastdeploy::vision::VisDetection(im, res, 0.5);
+  tc.End();
+  tc.PrintInfo("PPDet in ONNX");
+
+  std::cout << res.Str() << std::endl;
+  cv::imwrite("infer_onnx.jpg", vis_im);
+  std::cout << "Visualized result saved in ./infer_onnx.jpg" << std::endl;
+}
+
+void RKNPU2Infer(const std::string& model_dir, const std::string& image_file) {
+  auto model_file = model_dir + "/ppyoloe_plus_crn_s_80e_coco_rk3588_quantized.rknn";
+  auto params_file = "";
+  auto config_file = model_dir + "/infer_cfg.yml";
+
+  auto option = fastdeploy::RuntimeOption();
+  option.UseRKNPU2();
+
+  auto format = fastdeploy::ModelFormat::RKNN;
+
+  auto model = fastdeploy::vision::detection::PPYOLOE(
+      model_file, params_file, config_file, option, format);
+
+  model.GetPreprocessor().DisablePermute();
+  model.GetPreprocessor().DisableNormalize();
+  model.GetPostprocessor().ApplyDecodeAndNMS();
+
+  auto im = cv::imread(image_file);
+
+  fastdeploy::vision::DetectionResult res;
+  fastdeploy::TimeCounter tc;
+  tc.Start();
+  if (!model.Predict(&im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return;
+  }
+  tc.End();
+  tc.PrintInfo("PPDet in RKNPU2");
+
+  std::cout << res.Str() << std::endl;
+  auto vis_im = fastdeploy::vision::VisDetection(im, res, 0.5);
+  cv::imwrite("infer_rknpu2.jpg", vis_im);
+  std::cout << "Visualized result saved in ./infer_rknpu2.jpg" << std::endl;
+}
+
+int main(int argc, char* argv[]) {
+  if (argc < 4) {
+    std::cout
+        << "Usage: infer_demo path/to/model_dir path/to/image run_option, "
+           "e.g ./infer_model ./picodet_model_dir ./test.jpeg"
+        << std::endl;
+    return -1;
+  }
+
+  if (std::atoi(argv[3]) == 0) {
+    ONNXInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 1) {
+    RKNPU2Infer(argv[1], argv[2]);
+  }
+  return 0;
+}
--- a/examples/vision/detection/paddledetection/rknpu2/python/infer.py
+++ b/examples/vision/detection/paddledetection/rknpu2/python/infer.py
@@ -22,11 +22,11 @@ def parse_arguments():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--model_file",
-        default="./picodet_s_416_coco_lcnet_non_postprocess/picodet_xs_416_coco_lcnet.onnx",
+        default="./picodet_s_416_coco_lcnet/picodet_s_416_coco_lcnet_rk3588_unquantized.rknn",
        help="Path of rknn model.")
    parser.add_argument(
        "--config_file",
-        default="./picodet_s_416_coco_lcnet_non_postprocess/infer_cfg.yml",
+        default="./picodet_s_416_coco_lcnet/infer_cfg.yml",
        help="Path of config.")
    parser.add_argument(
        "--image",
--- a/examples/vision/detection/rkyolo/README_CN.md
+++ b/examples/vision/detection/rkyolo/README_CN.md
@@ -6,11 +6,21 @@ RKYOLO参考[rknn_model_zoo](https://github.com/airockchip/rknn_model_zoo/tree/m

 ## 支持模型列表

+FastDeploy目前支持以下三个模型的部署:
+
 * RKYOLOV5
+* RKYOLOX
+* RKYOLOv7

-## 模型转换example
+为了方便大家测试，我们提供了三个转换过后的模型，大家可以直接下载使用。
+如果你有转换模型的需求，请参考[RKNN_model_convert](https://github.com/airockchip/rknn_model_zoo/tree/main/models/CV/object_detection/yolo/RKNN_model_convert)
+
+| 模型名称               | 下载地址                                                                |
+|--------------------|---------------------------------------------------------------------|
+| yolov5-s-relu-int8 | https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/yolov5-s-relu.zip |
+| yolov7-tiny-int8   | https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/yolov7-tiny.zip   |
+| yolox-s-int8       | https://bj.bcebos.com/paddlehub/fastdeploy/rknpu2/yolox-s.zip       |

-请参考[RKNN_model_convert](https://github.com/airockchip/rknn_model_zoo/tree/main/models/CV/object_detection/yolo/RKNN_model_convert)


 ## 其他链接