update gradient clip english doc for new gradient clipping strategy

梯度裁剪的策略进行了升级，配合修复相应的 裁剪API、minimize、ParamAttr 的API英文文档。

对应API变动的文档： #23224 

对应中文文档PR：PaddlePaddle/FluidDoc#1942

python/paddle/fluid/optimizer.py

@@ -802,10 +802,11 @@ class Optimizer(object):
                 will be updated.
             no_grad_set (set, optional): Set of ``Variable``  or ``Variable.name`` that don't need
                 to be updated. The default value is None.   
-            grad_clip (GradClipBase, optional) : Gradient clipping strategy, static
-                graph mode does not need to use this argument. Currently, this argument
-                only supports gradient clipping in dygraph mode. In the future, this
-                argument my be adjusted. The default value is None.
+            grad_clip (GradientClipBase, optional): Gradient cliping strategy, it's an instance of 
+                some derived class of ``GradientClipBase`` . There are three cliping strategies 
+                ( :ref:`api_fluid_clip_GradientClipByGlobalNorm` , :ref:`api_fluid_clip_GradientClipByNorm` , 
+                :ref:`api_fluid_clip_GradientClipByValue` ). Default value: None, and there is no 
+                gradient clipping.
 
         Returns:
             tuple: tuple (optimize_ops, params_grads), A list of operators appended

python/paddle/fluid/param_attr.py

@@ -32,6 +32,12 @@ class ParamAttr(object):
     name, initializer, learning rate, regularizer, trainable, gradient clip,
     and model average.
     
+    Note:
+        ``gradient_clip`` of ``ParamAttr`` HAS BEEN DEPRECATED since 2.0. 
+        It is recommended to use ``minimize(loss, grad_clip=clip)`` to clip gradient. 
+        There are three clipping strategies: :ref:`api_fluid_clip_GradientClipByGlobalNorm` , 
+        :ref:`api_fluid_clip_GradientClipByNorm` , :ref:`api_fluid_clip_GradientClipByValue` .
+
     Parameters:
         name (str, optional): The parameter's name. Default None, meaning that the name
                 would be created automatically.
@@ -44,8 +50,6 @@ class ParamAttr(object):
         regularizer (WeightDecayRegularizer, optional): Regularization factor. Default None, meaning
                 there is no regularization.
         trainable (bool): Whether this parameter is trainable. Default True.
-        gradient_clip (BaseGradientClipAttr, optional): The method to clip this parameter's
-                gradient. Default None, meaning that there is no gradient clip.
         do_model_average (bool): Whether this parameter should do model average
                 when model average is enabled. Default False.
 
@@ -191,6 +195,12 @@ class WeightNormParamAttr(ParamAttr):
     Training of Deep Neural Networks
     <https://arxiv.org/pdf/1602.07868.pdf>`_.
       
+    Note:
+        ``gradient_clip`` of ``WeightNormParamAttr`` HAS BEEN DEPRECATED since 2.0. 
+        It is recommended to use ``minimize(loss, grad_clip=clip)`` to clip gradient. 
+        There are three clipping strategies: :ref:`api_fluid_clip_GradientClipByGlobalNorm` , 
+        :ref:`api_fluid_clip_GradientClipByNorm` , :ref:`api_fluid_clip_GradientClipByValue` .
+
     Args:
         dim(int): Dimension over which to compute the norm. Dim is a non-negative
             number which is less than the rank of weight Tensor. For Example, dim can
@@ -209,9 +219,6 @@ class WeightNormParamAttr(ParamAttr):
             ``regularizer = fluid.regularizer.L2DecayRegularizer(regularization_coeff=0.1)``.
             Default None, meaning that there is no regularization.
         trainable(bool, optional): Whether this parameter is trainable. Default True.
-        gradient_clip: The method to clip this parameter's gradient, such as
-            ``gradient_clip = fluid.clip.GradientClipByNorm(clip_norm=2.0))`` .
-            Default None, meaning that there is no gradient clip.
         do_model_average(bool, optional): Whether this parameter should do model average.
             Default False.
 
@@ -229,7 +236,6 @@ class WeightNormParamAttr(ParamAttr):
                                           learning_rate=1.0,
                                           regularizer=fluid.regularizer.L2DecayRegularizer(regularization_coeff=0.1),
                                           trainable=True,
-                                          gradient_clip=fluid.clip.GradientClipByNorm(clip_norm=2.0),
                                           do_model_average=False))
 
     """