diff --git a/doc/fluid/api/transpiler.rst b/doc/fluid/api/transpiler.rst
index 943d39331d26c..d2ac04f1449c3 100644
--- a/doc/fluid/api/transpiler.rst
+++ b/doc/fluid/api/transpiler.rst
@@ -14,6 +14,15 @@ DistributeTranspiler
     :members:
     :noindex:
 
+.. _api_fluid_transpiler_InferenceTranspiler:
+
+InferenceTranspiler
+-------------------
+
+..  autoclass:: paddle.fluid.transpiler.InferenceTranspiler
+    :members:
+    :noindex:
+
 .. _api_fluid_transpiler_memory_optimize:
 
 memory_optimize
diff --git a/python/paddle/fluid/transpiler/inference_transpiler.py b/python/paddle/fluid/transpiler/inference_transpiler.py
index d32c69d148dfa..b8afeae5ebd6e 100644
--- a/python/paddle/fluid/transpiler/inference_transpiler.py
+++ b/python/paddle/fluid/transpiler/inference_transpiler.py
@@ -19,7 +19,7 @@
 from ..executor import global_scope
 
 
-class InferenceTranspiler:
+class InferenceTranspiler(object):
     '''
     Convert the fluid program to optimized inference program.