code style and docs string improvement for interpreters

holyseven · holyseven · commit ed328df2d74c · 2022-04-25T12:45:29.000+08:00
diff --git a/interpretdl/interpreter/_normlime_base.py b/interpretdl/interpreter/_normlime_base.py
@@ -11,24 +11,24 @@
 
 class NormLIMECVInterpreter(LIMECVInterpreter):
     """
-    NormLIME Interpreter for CV tasks.
+    NormLIME Interpreter for CV tasks. 
+
+    (TODO) Some technical details will be complete soon.
 
     More details regarding the NormLIME method can be found in the original paper:
-    https://arxiv.org/abs/1909.04200
+    https://arxiv.org/abs/1909.04200.
+
+    Args:
+        paddle_model (_type_): 
+            A user-defined function that gives access to model predictions.
+            It takes the following arguments:
+            - data: Data inputs.
+            and outputs predictions.
+        device (str, optional): The device used for running `paddle_model`, options: ``cpu``, ``gpu:0``, ``gpu:1`` etc.
+        use_cuda (_type_, optional): Would be deprecated soon. Use ``device`` directly.    
     """
 
     def __init__(self, paddle_model, device='gpu:0', use_cuda=None):
-        """
-        
-        Args:
-            paddle_model (_type_): 
-                A user-defined function that gives access to model predictions.
-                It takes the following arguments:
-                - data: Data inputs.
-                and outputs predictions.
-            device (str, optional): The device used for running `paddle_model`, options: ``cpu``, ``gpu:0``, ``gpu:1`` etc.
-            use_cuda (_type_, optional): Would be deprecated soon. Use ``device`` directly.
-        """
 
         LIMECVInterpreter.__init__(self, paddle_model, use_cuda=use_cuda, device=device)
         self.lime_interpret = super().interpret
@@ -58,6 +58,8 @@ def interpret(self,
         """
         Main function of the interpreter.
 
+        (TODO) Some technical details will be complete soon.
+
         Args:
             image_paths (list of strs): A list of image filepaths.
             num_samples (int, optional): LIME sampling numbers. Larger number of samples usually gives more 
@@ -182,20 +184,18 @@ class NormLIMENLPInterpreter(LIMENLPInterpreter):
     NormLIME Interpreter for NLP tasks.
 
     More details regarding the NormLIME method can be found in the original paper:
-    https://arxiv.org/abs/1909.04200
-    """
+    https://arxiv.org/abs/1909.04200.
 
-    def __init__(self, paddle_model: callable, device: str = 'gpu:0', use_cuda=None):
-        """
+    Args:
+        paddle_model (callable): A user-defined function that gives access to model predictions.
+                It takes the following arguments:
 
-        Args:
-            paddle_model (callable): A user-defined function that gives access to model predictions.
-                    It takes the following arguments:
+                - data: Data inputs.
+                and outputs predictions. See the example at the end of ``interpret()``.
+        device (str): The device used for running `paddle_model`, options: ``cpu``, ``gpu:0``, ``gpu:1`` etc.    
+    """
 
-                    - data: Data inputs.
-                    and outputs predictions. See the example at the end of ``interpret()``.
-            device (str): The device used for running `paddle_model`, options: ``cpu``, ``gpu:0``, ``gpu:1`` etc.
-        """
+    def __init__(self, paddle_model: callable, device: str = 'gpu:0', use_cuda=None):
         LIMENLPInterpreter.__init__(self, paddle_model, device, use_cuda)
         self.lime_interpret = super().interpret
 
diff --git a/interpretdl/interpreter/gradient_cam.py b/interpretdl/interpreter/gradient_cam.py
@@ -52,13 +52,16 @@ def interpret(self,
         following by a ReLU activation to produce the final explanation.
 
         Args:
-            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy array of read images.
+            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy 
+                array of read images.
             target_layer_name (str): The target layer to calculate gradients.
             labels (list or tuple or numpy.ndarray, optional): The target labels to analyze. 
-                The number of labels should be equal to the number of images. If None, the most likely label for each image will be used. Default: None
-            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. Defaults to 224.
-            crop_to (int, optional): [description]. After resize, images will be center cropped to a square image with the size `crop_to`. 
-                If None, no crop will be performed. Defaults to None.
+                The number of labels should be equal to the number of images. If None, the most likely label for each
+                image will be used. Default: None
+            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. 
+                Defaults to 224.
+            crop_to (int, optional): [description]. After resize, images will be center cropped to a square image with 
+                the size `crop_to`. If None, no crop will be performed. Defaults to None.
             visual (bool, optional): Whether or not to visualize the processed image. Default: True
             save_path (str or list of strs or None, optional): The filepath(s) to save the processed image(s). 
                 If None, the image will not be saved. Default: None
diff --git a/interpretdl/interpreter/lime.py b/interpretdl/interpreter/lime.py
@@ -51,14 +51,18 @@ def interpret(self,
 
         Args:
             data (str): The input file path.
-            interpret_class (int, optional): The index of class to interpret. If None, the most likely label will be used. Default: None
-            num_samples (int, optional): LIME sampling numbers. Larger number of samples usually gives more accurate interpretation. Default: 1000
+            interpret_class (int, optional): The index of class to interpret. If None, the most likely label will be 
+                used. Default: None
+            num_samples (int, optional): LIME sampling numbers. Larger number of samples usually gives more accurate 
+                interpretation. Default: 1000
             batch_size (int, optional): Number of samples to forward each time. Default: 50
-            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. Defaults to 224.
-            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image with the size `crop_to`. 
-                If None, no crop will be performed. Defaults to None.
+            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. 
+                Defaults to 224.
+            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image 
+                with the size `crop_to`. If None, no crop will be performed. Defaults to None.
             visual (bool, optional): Whether or not to visualize the processed image. Default: True
-            save_path (str, optional): The path to save the processed image. If None, the image will not be saved. Default: None
+            save_path (str, optional): The path to save the processed image. If None, the image will not be saved. 
+                Default: None
 
         Returns:
             [dict]: LIME results: {interpret_label_i: weights on features}
diff --git a/interpretdl/interpreter/lime_prior.py b/interpretdl/interpreter/lime_prior.py
@@ -90,15 +90,19 @@ def interpret(self,
 
         Args:
             inputs (str): The input file path.
-            interpret_class (int, optional): The index of class to interpret. If None, the most likely label will be used. Default: None
+            interpret_class (int, optional): The index of class to interpret. If None, the most likely label will be 
+                used. Default: None
             prior_reg_force (float, optional): The regularization force to apply. Default: 1.0
-            num_samples (int, optional): LIME sampling numbers. Larger number of samples usually gives more accurate interpretation. Default: 1000
+            num_samples (int, optional): LIME sampling numbers. Larger number of samples usually gives more accurate 
+                interpretation. Default: 1000
             batch_size (int, optional): Number of samples to forward each time. Default: 50
-            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. Defaults to 224.
-            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image with the size `crop_to`. 
-                If None, no crop will be performed. Defaults to None.
+            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. 
+                Defaults to 224.
+            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image 
+                with the size `crop_to`. If None, no crop will be performed. Defaults to None.
             visual (bool, optional): Whether or not to visualize the processed image. Default: True
-            save_path (str, optional): The path to save the processed image. If None, the image will not be saved. Default: None
+            save_path (str, optional): The path to save the processed image. If None, the image will not be saved. 
+                Default: None
 
         Returns:
             [dict]: LIME results: {interpret_label_i: weights on features}
diff --git a/interpretdl/interpreter/score_cam.py b/interpretdl/interpreter/score_cam.py
@@ -41,15 +41,19 @@ def interpret(self,
         (TODO) The technical details will be described later.
 
         Args:
-            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy array of read images.
+            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy 
+                array of read images.
             target_layer_name (str): The target layer to calculate gradients.
-            labels (list or tuple or numpy.ndarray, optional): The target labels to analyze. The number of labels should be equal to the number of images. 
-                If None, the most likely label for each image will be used. Default: None
-            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. Defaults to 224.
-            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image with the size `crop_to`. 
-                If None, no crop will be performed. Defaults to None.
+            labels (list or tuple or numpy.ndarray, optional): The target labels to analyze. The number of labels 
+                should be equal to the number of images. If None, the most likely label for each image will be used. 
+                Default: None
+            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. 
+            Defaults to 224.
+            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image 
+                with the size `crop_to`. If None, no crop will be performed. Defaults to None.
             visual (bool, optional): Whether or not to visualize the processed image. Default: True
-            save_path (str or list of strs or None, optional): The filepath(s) to save the processed image(s). If None, the image will not be saved. Default: None
+            save_path (str or list of strs or None, optional): The filepath(s) to save the processed image(s). If None,
+                the image will not be saved. Default: None
 
         Returns:
             [numpy.ndarray]: interpretations/heatmap for images
diff --git a/interpretdl/interpreter/smooth_grad.py b/interpretdl/interpreter/smooth_grad.py
@@ -35,16 +35,22 @@ def __init__(self, paddle_model: callable, device: str = 'gpu:0', use_cuda=None)
         the gradients w.r.t. these noised inputs. The final explanation is averaged gradients.
 
         Args:
-            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy array of read images.
-            labels (list or tuple or numpy.ndarray, optional): The target labels to analyze. The number of labels should be equal to the number of images. If None, the most likely label for each image will be used. Default: None
+            inputs (str or list of strs or numpy.ndarray): The input image filepath or a list of filepaths or numpy 
+                array of read images.
+            labels (list or tuple or numpy.ndarray, optional): The target labels to analyze. The number of labels 
+                should be equal to the number of images. If None, the most likely label for each image will be used. 
+                Default: None
             noise_amount (float, optional): Noise level of added noise to the image.
-                                            The std of Guassian random noise is noise_amount * (x_max - x_min). Default: 0.1
+                                            The std of Guassian random noise is noise_amount * (x_max - x_min). 
+                                            Default: 0.1
             n_samples (int, optional): The number of new images generated by adding noise. Default: 50
-            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. Defaults to 224.
-            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image with the size `crop_to`. 
-                If None, no crop will be performed. Defaults to None.
+            resize_to (int, optional): [description]. Images will be rescaled with the shorter edge being `resize_to`. 
+                Defaults to 224.
+            crop_to ([type], optional): [description]. After resize, images will be center cropped to a square image 
+                with the size `crop_to`. If None, no crop will be performed. Defaults to None.
             visual (bool, optional): Whether or not to visualize the processed image. Default: True
-            save_path (str or list of strs or None, optional): The filepath(s) to save the processed image(s). If None, the image will not be saved. Default: None
+            save_path (str or list of strs or None, optional): The filepath(s) to save the processed image(s). If None,
+                the image will not be saved. Default: None
 
         :return: interpretations/gradients for each image
         :rtype: numpy.ndarray
