|
1 | | -// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved. |
| 1 | +// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved. //NOLINT |
2 | 2 | // |
3 | 3 | // Licensed under the Apache License, Version 2.0 (the "License"); |
4 | 4 | // you may not use this file except in compliance with the License. |
|
21 | 21 | namespace fastdeploy { |
22 | 22 | namespace vision { |
23 | 23 | namespace detection { |
24 | | - |
| 24 | +/*! @brief YOLOv5 model object used when to load a YOLOv5 model exported by YOLOv5. |
| 25 | + */ |
25 | 26 | class FASTDEPLOY_DECL YOLOv5 : public FastDeployModel { |
26 | 27 | public: |
| 28 | + /** \brief Set path of model file and the configuration of runtime. |
| 29 | + * |
| 30 | + * \param[in] model_file Path of model file, e.g ./yolov5.onnx |
| 31 | + * \param[in] params_file Path of parameter file, e.g ppyoloe/model.pdiparams, if the model format is ONNX, this parameter will be ignored |
| 32 | + * \param[in] custom_option RuntimeOption for inference, the default will use cpu, and choose the backend defined in "valid_cpu_backends" |
| 33 | + * \param[in] model_format Model format of the loaded model, default is ONNX format |
| 34 | + */ |
27 | 35 | YOLOv5(const std::string& model_file, const std::string& params_file = "", |
28 | 36 | const RuntimeOption& custom_option = RuntimeOption(), |
29 | 37 | const ModelFormat& model_format = ModelFormat::ONNX); |
30 | 38 |
|
31 | 39 | ~YOLOv5(); |
32 | 40 |
|
33 | 41 | std::string ModelName() const { return "yolov5"; } |
34 | | - |
| 42 | + /** \brief Predict the detection result for an input image |
| 43 | + * |
| 44 | + * \param[in] im The input image data, comes from cv::imread(), is a 3-D array with layout HWC, BGR format |
| 45 | + * \param[in] result The output detection result will be writen to this structure |
| 46 | + * \param[in] conf_threshold confidence threashold for postprocessing, default is 0.25 |
| 47 | + * \param[in] nms_iou_threshold iou threashold for NMS, default is 0.5 |
| 48 | + * \return true if the prediction successed, otherwise false |
| 49 | + */ |
35 | 50 | virtual bool Predict(cv::Mat* im, DetectionResult* result, |
36 | 51 | float conf_threshold = 0.25, |
37 | 52 | float nms_iou_threshold = 0.5); |
@@ -62,23 +77,25 @@ class FASTDEPLOY_DECL YOLOv5 : public FastDeployModel { |
62 | 77 | float conf_threshold, float nms_iou_threshold, bool multi_label, |
63 | 78 | float max_wh = 7680.0); |
64 | 79 |
|
65 | | - // tuple of (width, height) |
| 80 | + /// tuple of (width, height) |
66 | 81 | std::vector<int> size_; |
67 | | - // padding value, size should be same with Channels |
| 82 | + /// padding value, size should be the same as channels |
68 | 83 | std::vector<float> padding_value_; |
69 | | - // only pad to the minimum rectange which height and width is times of stride |
| 84 | + /// only pad to the minimum rectange which height and width is times of stride |
70 | 85 | bool is_mini_pad_; |
71 | | - // while is_mini_pad = false and is_no_pad = true, will resize the image to |
72 | | - // the set size |
| 86 | + /*! @brief |
| 87 | + while is_mini_pad = false and is_no_pad = true, will resize the image to the set size |
| 88 | + */ |
73 | 89 | bool is_no_pad_; |
74 | | - // if is_scale_up is false, the input image only can be zoom out, the maximum |
75 | | - // resize scale cannot exceed 1.0 |
| 90 | + /*! @brief |
| 91 | + if is_scale_up is false, the input image only can be zoom out, the maximum resize scale cannot exceed 1.0 |
| 92 | + */ |
76 | 93 | bool is_scale_up_; |
77 | | - // padding stride, for is_mini_pad |
| 94 | + /// padding stride, for is_mini_pad |
78 | 95 | int stride_; |
79 | | - // for offseting the boxes by classes when using NMS |
| 96 | + /// for offseting the boxes by classes when using NMS |
80 | 97 | float max_wh_; |
81 | | - // for different strategies to get boxes when postprocessing |
| 98 | + /// for different strategies to get boxes when postprocessing |
82 | 99 | bool multi_label_; |
83 | 100 |
|
84 | 101 | private: |
|
0 commit comments