[Fix] fix np.arange() output not stable error

docs/en/basic_concepts/evaluation.md

@@ -60,8 +60,8 @@ In MMOCR, the calculation of `HmeanIOUMetric` can be summarized as the following
    It is worth noting that `pred_score_thrs` will **automatically search** for the **best threshold** within a certain range by default, and users can also customize the search range by manually modifying the configuration file:
 
    ```python
-   # By default, HmeanIOUMetric searches the best threshold within the range [0.3, 0.9] with a step size of 0.1
-   val_evaluator = dict(type='HmeanIOUMetric', pred_score_thrs=dict(start=0.3, stop=0.9, step=0.1))
+   # By default, HmeanIOUMetric searches the best threshold within the range [0.3, 1) with a step size of 0.1
+   val_evaluator = dict(type='HmeanIOUMetric', pred_score_thrs=dict(start=0.3, stop=1, step=0.1))
    ```
 
 2. Calculate the IoU matrix

docs/zh_cn/basic_concepts/evaluation.md

@@ -60,8 +60,8 @@ F_1=(1+\beta^2)\cdot\frac{PR}{\beta^2\cdot P+R} = \frac{2PR}{P+R}
    值得注意的是，`pred_score_thrs` 默认将**自动搜索**一定范围内的**最佳阈值**，用户也可以通过手动修改配置文件来自定义搜索范围：
 
    ```python
-   # HmeanIOUMetric 默认以 0.1 为步长搜索 [0.3, 0.9] 范围内的最佳得分阈值
-   val_evaluator = dict(type='HmeanIOUMetric', pred_score_thrs=dict(start=0.3, stop=0.9, step=0.1))
+   # HmeanIOUMetric 默认以 0.1 为步长搜索 [0.3, 1) 范围内的最佳得分阈值
+   val_evaluator = dict(type='HmeanIOUMetric', pred_score_thrs=dict(start=0.3, stop=1, step=0.1))
    ```
 
 2. 计算 IoU 矩阵

mmocr/evaluation/metrics/hmean_iou_metric.py

@@ -44,7 +44,7 @@ class HmeanIOUMetric(BaseMetric):
         ignore_precision_thr (float): Precision threshold when prediction and\
             gt ignored polygons are matched. Defaults to 0.5.
         pred_score_thrs (dict): Best prediction score threshold searching
-            space. Defaults to dict(start=0.3, stop=0.9, step=0.1).
+            space. Defaults to dict(start=0.3, stop=1 step=0.1).
         strategy (str): Polygon matching strategy. Options are 'max_matching'
             and 'vanilla'. 'max_matching' refers to the optimum strategy that
             maximizes the number of matches. Vanilla strategy matches gt and
@@ -63,14 +63,20 @@ class HmeanIOUMetric(BaseMetric):
     def __init__(self,
                  match_iou_thr: float = 0.5,
                  ignore_precision_thr: float = 0.5,
-                 pred_score_thrs: Dict = dict(start=0.3, stop=0.9, step=0.1),
+                 pred_score_thrs: Dict = dict(start=0.3, stop=1, step=0.1),
                  strategy: str = 'vanilla',
                  collect_device: str = 'cpu',
                  prefix: Optional[str] = None) -> None:
         super().__init__(collect_device=collect_device, prefix=prefix)
         self.match_iou_thr = match_iou_thr
         self.ignore_precision_thr = ignore_precision_thr
-        self.pred_score_thrs = np.arange(**pred_score_thrs)
+        self.pred_score_thrs = np.linspace(
+            pred_score_thrs['start'],
+            pred_score_thrs['stop'],
+            int(
+                np.round((pred_score_thrs['stop'] - pred_score_thrs['start']) /
+                         pred_score_thrs['step'])),
+            endpoint=False)
         assert strategy in ['max_matching', 'vanilla']
         self.strategy = strategy
 

projects/ABCNet/abcnet/metric/e2e_hmean_iou_metric.py

@@ -45,7 +45,7 @@ class E2EHmeanIOUMetric(BaseMetric):
         ignore_precision_thr (float): Precision threshold when prediction and\
             gt ignored polygons are matched. Defaults to 0.5.
         pred_score_thrs (dict): Best prediction score threshold searching
-            space. Defaults to dict(start=0.3, stop=0.9, step=0.1).
+            space. Defaults to dict(start=0.3, stop=1, step=0.1).
         strategy (str): Polygon matching strategy. Options are 'max_matching'
             and 'vanilla'. 'max_matching' refers to the optimum strategy that
             maximizes the number of matches. Vanilla strategy matches gt and
@@ -64,7 +64,7 @@ class E2EHmeanIOUMetric(BaseMetric):
     def __init__(self,
                  match_iou_thr: float = 0.5,
                  ignore_precision_thr: float = 0.5,
-                 pred_score_thrs: Dict = dict(start=0.3, stop=0.9, step=0.1),
+                 pred_score_thrs: Dict = dict(start=0.3, stop=1, step=0.1),
                  lexicon_path: Optional[str] = None,
                  word_spotting: bool = False,
                  min_length_case_word: int = 3,
@@ -75,7 +75,13 @@ def __init__(self,
         super().__init__(collect_device=collect_device, prefix=prefix)
         self.match_iou_thr = match_iou_thr
         self.ignore_precision_thr = ignore_precision_thr
-        self.pred_score_thrs = np.arange(**pred_score_thrs)
+        self.pred_score_thrs = np.linspace(
+            pred_score_thrs['start'],
+            pred_score_thrs['stop'],
+            int(
+                np.round((pred_score_thrs['stop'] - pred_score_thrs['start']) /
+                         pred_score_thrs['step'])),
+            endpoint=False)
         self.word_spotting = word_spotting
         self.min_length_case_word = min_length_case_word
         self.special_characters = special_characters

projects/SPTS/spts/metric/e2e_point_metric.py

@@ -34,7 +34,13 @@ def __init__(self,
                  collect_device: str = 'cpu',
                  prefix: Optional[str] = None) -> None:
         super().__init__(collect_device=collect_device, prefix=prefix)
-        self.text_score_thrs = np.arange(**text_score_thrs)
+        self.text_score_thrs = np.linspace(
+            text_score_thrs['start'],
+            text_score_thrs['stop'],
+            int(
+                np.round((text_score_thrs['stop'] - text_score_thrs['start']) /
+                         text_score_thrs['step'])),
+            endpoint=False)
         self.word_spotting = word_spotting
 
     def poly_center(self, poly_pts):