change clip1 to clip

PaddleNLP/language_model/train.py

@@ -137,7 +137,7 @@ def main():
                 res_vars = res_vars[:-1]
             loss, last_hidden, last_cell, feed_order = res_vars
 
-            clip1 = fluid.clip.GradientClipByGlobalNorm(
+            clip = fluid.clip.GradientClipByGlobalNorm(
                 clip_norm=config.max_grad_norm)
 
             learning_rate = fluid.layers.create_global_var(
@@ -148,7 +148,7 @@ def main():
                 persistable=True)
 
             optimizer = fluid.optimizer.SGD(learning_rate=learning_rate,
-                                            grad_clip=clip1)
+                                            grad_clip=clip)
             optimizer.minimize(loss)
 
     # define inference program

PaddleNLP/pretrain_language_models/BERT/optimization.py

@@ -102,10 +102,9 @@ def optimization(loss,
         raise ValueError("Unkown learning rate scheduler, should be "
                          "'noam_decay' or 'linear_warmup_decay'")
 
-    clip1 = fluid.clip.GradientClipByGlobalNorm(clip_norm=1.0)
+    clip = fluid.clip.GradientClipByGlobalNorm(clip_norm=1.0)
 
-    optimizer = fluid.optimizer.Adam(
-        learning_rate=scheduled_lr, grad_clip=clip1)
+    optimizer = fluid.optimizer.Adam(learning_rate=scheduled_lr, grad_clip=clip)
 
     def exclude_from_weight_decay(param):
         name = param.name.rstrip(".master")

PaddleNLP/pretrain_language_models/XLNet/optimization.py

@@ -110,10 +110,9 @@ def optimization(loss,
                 return True
         return False
 
-    clip1 = fluid.clip.GradientClipByGlobalNorm(clip_norm=1.0)
+    clip = fluid.clip.GradientClipByGlobalNorm(clip_norm=1.0)
 
-    optimizer = fluid.optimizer.Adam(
-        learning_rate=scheduled_lr, grad_clip=clip1)
+    optimizer = fluid.optimizer.Adam(learning_rate=scheduled_lr, grad_clip=clip)
 
     param_list = dict()