[cherry-pick] Fix some typos. (#5873)

8b0a2721 · Shuangchi He · GitHub · 3d1e2320 · 8b0a2721 · 8b0a2721
20 changed file
--- a/configs/picodet/README.md
+++ b/configs/picodet/README.md
@@ -232,9 +232,9 @@ paddle2onnx --model_dir output_inference/picodet_s_320_coco_lcnet/ \
 | Paddle Lite |  -    |  [C++](../../deploy/lite) | ✔︎ |
 | Android Demo |  -  |  [Paddle Lite](https://github.com/PaddlePaddle/Paddle-Lite-Demo/tree/develop/object_detection/android/app/cxx/picodet_detection_demo) | ✔︎ |
 | PaddleInference | [Python](../../deploy/python) |  [C++](../../deploy/cpp) | ✔︎ |
-| ONNXRuntime  | [Python](../../deploy/third_engine/demo_onnxruntime) | Comming soon | ✔︎ |
-| NCNN |  Comming soon  | [C++](../../deploy/third_engine/demo_ncnn) | ✘ |
-| MNN  | Comming soon | [C++](../../deploy/third_engine/demo_mnn) |  ✘ |
+| ONNXRuntime  | [Python](../../deploy/third_engine/demo_onnxruntime) | Coming soon | ✔︎ |
+| NCNN |  Coming soon  | [C++](../../deploy/third_engine/demo_ncnn) | ✘ |
+| MNN  | Coming soon | [C++](../../deploy/third_engine/demo_mnn) |  ✘ |




--- a/configs/picodet/README_en.md
+++ b/configs/picodet/README_en.md
@@ -224,13 +224,13 @@ paddle2onnx --model_dir output_inference/picodet_s_320_coco_lcnet/ \

 | Infer Engine     | Python | C++  | Predict With Postprocess |
 | :-------- | :--------: | :---------------------: | :----------------: |
-| OpenVINO | [Python](../../deploy/third_engine/demo_openvino/python) | [C++](../../deploy/third_engine/demo_openvino)（postprocess comming soon） |  ✔︎ |
+| OpenVINO | [Python](../../deploy/third_engine/demo_openvino/python) | [C++](../../deploy/third_engine/demo_openvino)（postprocess coming soon） |  ✔︎ |
 | Paddle Lite |  -    |  [C++](../../deploy/lite) | ✔︎ |
 | Android Demo |  -  |  [Paddle Lite](https://github.com/PaddlePaddle/Paddle-Lite-Demo/tree/develop/object_detection/android/app/cxx/picodet_detection_demo) | ✔︎ |
 | PaddleInference | [Python](../../deploy/python) |  [C++](../../deploy/cpp) | ✔︎ |
-| ONNXRuntime  | [Python](../../deploy/third_engine/demo_onnxruntime) | Comming soon | ✔︎ |
-| NCNN |  Comming soon  | [C++](../../deploy/third_engine/demo_ncnn) | ✘ |
-| MNN  | Comming soon | [C++](../../deploy/third_engine/demo_mnn) |  ✘ |
+| ONNXRuntime  | [Python](../../deploy/third_engine/demo_onnxruntime) | Coming soon | ✔︎ |
+| NCNN |  Coming soon  | [C++](../../deploy/third_engine/demo_ncnn) | ✘ |
+| MNN  | Coming soon | [C++](../../deploy/third_engine/demo_mnn) |  ✘ |


 Android demo visualization:
@@ -278,7 +278,7 @@ python tools/train.py -c configs/picodet/picodet_s_416_coco_lcnet.yml \
 ## Unstructured Pruning

 <details open>
-<summary>Toturial:</summary>
+<summary>Tutorial:</summary>

 Please refer this [documentation](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/picodet/legacy_model/pruner/README.md) for details such as requirements, training and deployment.


--- a/configs/rcnn_enhance/README.md
+++ b/configs/rcnn_enhance/README.md
@@ -2,7 +2,7 @@

 ### 简介

-* 近年来，学术界和工业界广泛关注图像中目标检测任务。基于[PaddleClas](https://github.com/PaddlePaddle/PaddleClas)中SSLD蒸馏方案训练得到的ResNet50_vd预训练模型(ImageNet1k验证集上Top1 Acc为82.39%)，结合PaddleDetection中的丰富算子，飞桨提供了一种面向服务器端实用的目标检测方案PSS-DET(Practical Server Side Detection)。基于COCO2017目标检测数据集，V100单卡预测速度为为61FPS时，COCO mAP可达41.2%。
+* 近年来，学术界和工业界广泛关注图像中目标检测任务。基于[PaddleClas](https://github.com/PaddlePaddle/PaddleClas)中SSLD蒸馏方案训练得到的ResNet50_vd预训练模型(ImageNet1k验证集上Top1 Acc为82.39%)，结合PaddleDetection中的丰富算子，飞桨提供了一种面向服务器端实用的目标检测方案PSS-DET(Practical Server Side Detection)。基于COCO2017目标检测数据集，V100单卡预测速度为61FPS时，COCO mAP可达41.2%。


 ### 模型库

--- a/deploy/pptracking/python/mot/tracker/jde_tracker.py
+++ b/deploy/pptracking/python/mot/tracker/jde_tracker.py
@@ -116,7 +116,7 @@ class JDETracker(object):

        Return:
            output_stracks_dict (dict(list)): The list contains information
-                regarding the online_tracklets for the recieved image tensor.
+                regarding the online_tracklets for the received image tensor.
        """
        self.frame_id += 1
        if self.frame_id == 1:

--- a/deploy/python/keypoint_postprocess.py
+++ b/deploy/python/keypoint_postprocess.py
@@ -35,7 +35,7 @@ class HrHRNetPostProcess(object):
        heat_thresh (float): value of topk below this threshhold will be ignored
        tag_thresh (float): coord's value sampled in tagmap below this threshold belong to same people for init

-        inputs(list[heatmap]): the output list of modle, [heatmap, heatmap_maxpool, tagmap], heatmap_maxpool used to get topk
+        inputs(list[heatmap]): the output list of model, [heatmap, heatmap_maxpool, tagmap], heatmap_maxpool used to get topk
        original_height, original_width (float): the original image size
    """


--- a/docs/tutorials/FAQ/FAQ第零期.md
+++ b/docs/tutorials/FAQ/FAQ第零期.md
@@ -59,7 +59,7 @@ TrainReader:
    - Gt2YoloTarget: {anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]], anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]], downsample_ratios: [32, 16, 8]}
  # 训练时batch_size
  batch_size: 24
-  # 读取数据是是否乱序
+  # 读取数据是否乱序
  shuffle: true
  # 是否丢弃最后不能完整组成batch的数据
  drop_last: true

--- a/docs/tutorials/config_annotation/faster_rcnn_r50_fpn_1x_coco_annotation.md
+++ b/docs/tutorials/config_annotation/faster_rcnn_r50_fpn_1x_coco_annotation.md
@@ -90,7 +90,7 @@ TrainReader:
  - PadBatch: {pad_to_stride: 32}
  # 训练时batch_size
  batch_size: 1
-  # 读取数据是是否乱序
+  # 读取数据是否乱序
  shuffle: true
  # 是否丢弃最后不能完整组成batch的数据
  drop_last: true
@@ -110,7 +110,7 @@ EvalReader:
  - PadBatch: {pad_to_stride: 32}
  # 评估时batch_size
  batch_size: 1
-  # 读取数据是是否乱序
+  # 读取数据是否乱序
  shuffle: false
  # 是否丢弃最后不能完整组成batch的数据
  drop_last: false
@@ -130,7 +130,7 @@ TestReader:
  - PadBatch: {pad_to_stride: 32}
  # 测试时batch_size
  batch_size: 1
-  # 读取数据是是否乱序
+  # 读取数据是否乱序
  shuffle: false
  # 是否丢弃最后不能完整组成batch的数据
  drop_last: false

--- a/docs/tutorials/config_annotation/ppyolo_r50vd_dcn_1x_coco_annotation.md
+++ b/docs/tutorials/config_annotation/ppyolo_r50vd_dcn_1x_coco_annotation.md
@@ -102,7 +102,7 @@ TrainReader:
    - Gt2YoloTarget: {anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]], anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]], downsample_ratios: [32, 16, 8]}
  # 训练时batch_size
  batch_size: 24
-  # 读取数据是是否乱序
+  # 读取数据是否乱序
  shuffle: true
  # 是否丢弃最后不能完整组成batch的数据
  drop_last: true

--- a/ppdet/data/transform/mot_operators.py
+++ b/ppdet/data/transform/mot_operators.py
@@ -529,7 +529,7 @@ class Gt2FairMOTTarget(Gt2TTFTarget):
    Generate FairMOT targets by ground truth data.
    Difference between Gt2FairMOTTarget and Gt2TTFTarget are:
        1. the gaussian kernal radius to generate a heatmap.
-        2. the targets needed during traing.
+        2. the targets needed during training.
    
    Args:
        num_classes(int): the number of classes.

--- a/ppdet/data/transform/operators.py
+++ b/ppdet/data/transform/operators.py
@@ -1054,7 +1054,7 @@ class CropWithSampling(BaseOperator):
           [max sample, max trial, min scale, max scale,
            min aspect ratio, max aspect ratio,
            min overlap, max overlap]
-            avoid_no_bbox (bool): whether to to avoid the
+            avoid_no_bbox (bool): whether to avoid the
                                  situation where the box does not appear.
        """
        super(CropWithSampling, self).__init__()
@@ -1145,7 +1145,7 @@ class CropWithDataAchorSampling(BaseOperator):
            das_anchor_scales (list[float]): a list of anchor scales in data
                anchor smapling.
            min_size (float): minimum size of sampled bbox.
-            avoid_no_bbox (bool): whether to to avoid the
+            avoid_no_bbox (bool): whether to avoid the
                                  situation where the box does not appear.
        """
        super(CropWithDataAchorSampling, self).__init__()

--- a/ppdet/metrics/mot_metrics.py
+++ b/ppdet/metrics/mot_metrics.py
@@ -557,7 +557,7 @@ class KITTIEvaluation(object):
                            "track ids are not unique for sequence %d: frame %d"
                            % (seq, t_data.frame))
                        logger.info(
-                            "track id %d occured at least twice for this frame"
+                            "track id %d occurred at least twice for this frame"
                            % t_data.track_id)
                        logger.info("Exiting...")
                        #continue # this allows to evaluate non-unique result files

--- a/ppdet/modeling/architectures/keypoint_hrhrnet.py
+++ b/ppdet/modeling/architectures/keypoint_hrhrnet.py
@@ -153,7 +153,7 @@ class HrHRNetPostProcess(object):
        heat_thresh (float): value of topk below this threshhold will be ignored
        tag_thresh (float): coord's value sampled in tagmap below this threshold belong to same people for init

-        inputs(list[heatmap]): the output list of modle, [heatmap, heatmap_maxpool, tagmap], heatmap_maxpool used to get topk
+        inputs(list[heatmap]): the output list of model, [heatmap, heatmap_maxpool, tagmap], heatmap_maxpool used to get topk
        original_height, original_width (float): the original image size
    '''


--- a/ppdet/modeling/losses/sparsercnn_loss.py
+++ b/ppdet/modeling/losses/sparsercnn_loss.py
@@ -198,7 +198,7 @@ class SparseRCNNLoss(nn.Layer):
        # Retrieve the matching between the outputs of the last layer and the targets
        indices = self.matcher(outputs_without_aux, targets)

-        # Compute the average number of target boxes accross all nodes, for normalization purposes
+        # Compute the average number of target boxes across all nodes, for normalization purposes
        num_boxes = sum(len(t["labels"]) for t in targets)
        num_boxes = paddle.to_tensor(
            [num_boxes],

--- a/ppdet/modeling/mot/tracker/jde_tracker.py
+++ b/ppdet/modeling/mot/tracker/jde_tracker.py
@@ -122,7 +122,7 @@ class JDETracker(object):

        Return:
            output_stracks_dict (dict(list)): The list contains information
-                regarding the online_tracklets for the recieved image tensor.
+                regarding the online_tracklets for the received image tensor.
        """
        self.frame_id += 1
        if self.frame_id == 1:

--- a/ppdet/modeling/ops.py
+++ b/ppdet/modeling/ops.py
@@ -281,7 +281,7 @@ def roi_align(input,
            rois_num = paddle.static.data(name='rois_num', shape=[None], dtype='int32')
            align_out = ops.roi_align(input=x,
                                               rois=rois,
-                                               ouput_size=(7, 7),
+                                               output_size=(7, 7),
                                               spatial_scale=0.5,
                                               sampling_ratio=-1,
                                               rois_num=rois_num)

--- a/static/configs/iou_loss/README_cn.md
+++ b/static/configs/iou_loss/README_cn.md
@@ -28,7 +28,7 @@ PaddleDetection也开源了基于faster rcnn的GIOU loss实现。使用GIOU loss

 GIOU loss解决了IOU loss中预测边框A与真值B的交并比为0时，模型无法给出优化方向的问题，但是仍然有2种情况难以解决，
 1. 当边框A和边框B处于包含关系的时候，GIOU loss退化为IOU loss，此时模型收敛较慢。
-2. 当A与B相交，若A和B的的x1与x2均相等或者y1与y2均相等，GIOU loss仍然会退化为IOU loss，收敛很慢。
+2. 当A与B相交，若A和B的x1与x2均相等或者y1与y2均相等，GIOU loss仍然会退化为IOU loss，收敛很慢。

 基于此，论文提出了DIOU loss与CIOU loss，解决收敛速度慢以及部分条件下无法收敛的问题。
 为加速收敛，论文在改进的loss中引入距离的概念，具体地，边框loss可以定义为如下形式：

--- a/static/configs/ppyolo/README.md
+++ b/static/configs/ppyolo/README.md
@@ -90,7 +90,7 @@ PP-YOLO and PP-YOLOv2 improved performance and speed of YOLOv3 with following me
 |:----------------------------:|:----------:|:----------:| :---------: | :-----------------------: | :--------: | :----------:| :------------------: | :-------------------: | :------: | :----------------------: | :-----: |
 | PP-YOLO_MobileNetV3_small    |     4      |      32    |     75%     | PP-YOLO_MobileNetV3_large |   4.2MB    |     320     |         16.2         |      39.8      | [model](https://paddlemodels.bj.bcebos.com/object_detection/ppyolo_mobilenet_v3_small_prune75_distillby_mobilenet_v3_large.pdparams) | [model](https://paddlemodels.bj.bcebos.com/object_detection/ppyolo_mobilenet_v3_small_prune75_distillby_mobilenet_v3_large.tar) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/static/configs/ppyolo/ppyolo_mobilenet_v3_small.yml)                   |

- Slim PP-YOLO is trained by slim traing method from [Distill pruned model](../../slim/extensions/distill_pruned_model/README.md)，distill training pruned PP-YOLO_MobileNetV3_small model with PP-YOLO_MobileNetV3_large model as the teacher model
+- Slim PP-YOLO is trained by slim training method from [Distill pruned model](../../slim/extensions/distill_pruned_model/README.md)，distill training pruned PP-YOLO_MobileNetV3_small model with PP-YOLO_MobileNetV3_large model as the teacher model
 - Pruning detectiom head of PP-YOLO model with ratio as 75%, while the arguments are `--pruned_params="yolo_block.0.2.conv.weights,yolo_block.0.tip.conv.weights,yolo_block.1.2.conv.weights,yolo_block.1.tip.conv.weights" --pruned_ratios="0.75,0.75,0.75,0.75"`
 - For Slim PP-YOLO training, evaluation, inference and model exporting, please see [Distill pruned model](../../slim/extensions/distill_pruned_model/README.md)


--- a/static/configs/rcnn_enhance/README.md
+++ b/static/configs/rcnn_enhance/README.md
@@ -2,7 +2,7 @@

 ### 简介

-* 近年来，学术界和工业界广泛关注图像中目标检测任务。基于[PaddleClas](https://github.com/PaddlePaddle/PaddleClas)中SSLD蒸馏方案训练得到的ResNet50_vd预训练模型(ImageNet1k验证集上Top1 Acc为82.39%)，结合PaddleDetection中的丰富算子，飞桨提供了一种面向服务器端实用的目标检测方案PSS-DET(Practical Server Side Detection)。基于COCO2017目标检测数据集，V100单卡预测速度为为61FPS时，COCO mAP可达41.6%；预测速度为20FPS时，COCO mAP可达47.8%。
+* 近年来，学术界和工业界广泛关注图像中目标检测任务。基于[PaddleClas](https://github.com/PaddlePaddle/PaddleClas)中SSLD蒸馏方案训练得到的ResNet50_vd预训练模型(ImageNet1k验证集上Top1 Acc为82.39%)，结合PaddleDetection中的丰富算子，飞桨提供了一种面向服务器端实用的目标检测方案PSS-DET(Practical Server Side Detection)。基于COCO2017目标检测数据集，V100单卡预测速度为61FPS时，COCO mAP可达41.6%；预测速度为20FPS时，COCO mAP可达47.8%。

 * 以标准的Faster RCNN ResNet50_vd FPN为例，下表给出了PSS-DET不同的模块的速度与精度收益。


--- a/static/deploy/android_demo/app/build.gradle
+++ b/static/deploy/android_demo/app/build.gradle
@@ -65,7 +65,7 @@ def archives = [
                'src' : 'https://paddlelite-demo.bj.bcebos.com/models/yolov3_mobilenet_v3_prune86_FPGM_320_fp32_for_hybrid_cpu_npu_v2_6_1.tar.gz',
                'dest' : 'src/main/assets/models/yolov3_mobilenet_v3_for_hybrid_cpu_npu'
        ],
-        // pp-yolo tiny comming soon
+        // pp-yolo tiny coming soon
        // ssd_mobilenet_v1 voc
        [
                'src' : 'https://paddlelite-demo.bj.bcebos.com/models/ssdlite_mobilenet_v3_large_for_cpu_nb.tar.gz',

--- a/static/ppdet/data/transform/operators.py
+++ b/static/ppdet/data/transform/operators.py
@@ -955,7 +955,7 @@ class CropImage(BaseOperator):
           [max sample, max trial, min scale, max scale,
            min aspect ratio, max aspect ratio,
            min overlap, max overlap]
-            avoid_no_bbox (bool): whether to to avoid the
+            avoid_no_bbox (bool): whether to avoid the
                                  situation where the box does not appear.
        """
        super(CropImage, self).__init__()
@@ -1047,7 +1047,7 @@ class CropImageWithDataAchorSampling(BaseOperator):
            das_anchor_scales (list[float]): a list of anchor scales in data
                anchor smapling.
            min_size (float): minimum size of sampled bbox.
-            avoid_no_bbox (bool): whether to to avoid the
+            avoid_no_bbox (bool): whether to avoid the
                                  situation where the box does not appear.
        """
        super(CropImageWithDataAchorSampling, self).__init__()