improve the doc of paddle.fluid.memory_optimize, test=develop (#17473)

* improve the doc of paddle.fluid.memory_optimize, test=develop

* fix typo, test=develop

paddle/fluid/API.spec

@@ -27,7 +27,7 @@ paddle.fluid.DistributeTranspiler.get_pserver_programs (ArgSpec(args=['self', 'e
 paddle.fluid.DistributeTranspiler.get_startup_program (ArgSpec(args=['self', 'endpoint', 'pserver_program', 'startup_program'], varargs=None, keywords=None, defaults=(None, None)), ('document', 'd796fc0c8d51503b556fcf6dc15c4f0c'))
 paddle.fluid.DistributeTranspiler.get_trainer_program (ArgSpec(args=['self', 'wait_port'], varargs=None, keywords=None, defaults=(True,)), ('document', '736330e31a7a54abccc0c7fd9119d9ff'))
 paddle.fluid.DistributeTranspiler.transpile (ArgSpec(args=['self', 'trainer_id', 'program', 'pservers', 'trainers', 'sync_mode', 'startup_program', 'current_endpoint'], varargs=None, keywords=None, defaults=(None, '127.0.0.1:6174', 1, True, None, '127.0.0.1:6174')), ('document', '06ce55338dfe96311ad1078235ab3bf4'))
-paddle.fluid.memory_optimize (ArgSpec(args=['input_program', 'skip_opt_set', 'print_log', 'level', 'skip_grads'], varargs=None, keywords=None, defaults=(None, False, 0, False)), ('document', '3f11d536c8039c7b5fd57970078c344b'))
+paddle.fluid.memory_optimize (ArgSpec(args=['input_program', 'skip_opt_set', 'print_log', 'level', 'skip_grads'], varargs=None, keywords=None, defaults=(None, False, 0, False)), ('document', '97fdf61cf1ed4fb8f6a4b58aebce26a2'))
 paddle.fluid.release_memory (ArgSpec(args=['input_program', 'skip_opt_set'], varargs=None, keywords=None, defaults=(None,)), ('document', 'b524b73e10f9ccdfa6d189e2e535fc17'))
 paddle.fluid.DistributeTranspilerConfig.__init__ 
 paddle.fluid.ParallelExecutor.__init__ (ArgSpec(args=['self', 'use_cuda', 'loss_name', 'main_program', 'share_vars_from', 'exec_strategy', 'build_strategy', 'num_trainers', 'trainer_id', 'scope'], varargs=None, keywords=None, defaults=(None, None, None, None, None, 1, 0, None)), ('document', '6adf97f83acf6453d4a6a4b1070f3754'))
@@ -428,7 +428,7 @@ paddle.fluid.transpiler.DistributeTranspiler.get_pserver_programs (ArgSpec(args=
 paddle.fluid.transpiler.DistributeTranspiler.get_startup_program (ArgSpec(args=['self', 'endpoint', 'pserver_program', 'startup_program'], varargs=None, keywords=None, defaults=(None, None)), ('document', 'd796fc0c8d51503b556fcf6dc15c4f0c'))
 paddle.fluid.transpiler.DistributeTranspiler.get_trainer_program (ArgSpec(args=['self', 'wait_port'], varargs=None, keywords=None, defaults=(True,)), ('document', '736330e31a7a54abccc0c7fd9119d9ff'))
 paddle.fluid.transpiler.DistributeTranspiler.transpile (ArgSpec(args=['self', 'trainer_id', 'program', 'pservers', 'trainers', 'sync_mode', 'startup_program', 'current_endpoint'], varargs=None, keywords=None, defaults=(None, '127.0.0.1:6174', 1, True, None, '127.0.0.1:6174')), ('document', '06ce55338dfe96311ad1078235ab3bf4'))
-paddle.fluid.transpiler.memory_optimize (ArgSpec(args=['input_program', 'skip_opt_set', 'print_log', 'level', 'skip_grads'], varargs=None, keywords=None, defaults=(None, False, 0, False)), ('document', '3f11d536c8039c7b5fd57970078c344b'))
+paddle.fluid.transpiler.memory_optimize (ArgSpec(args=['input_program', 'skip_opt_set', 'print_log', 'level', 'skip_grads'], varargs=None, keywords=None, defaults=(None, False, 0, False)), ('document', '97fdf61cf1ed4fb8f6a4b58aebce26a2'))
 paddle.fluid.transpiler.release_memory (ArgSpec(args=['input_program', 'skip_opt_set'], varargs=None, keywords=None, defaults=(None,)), ('document', 'b524b73e10f9ccdfa6d189e2e535fc17'))
 paddle.fluid.transpiler.HashName.__init__ (ArgSpec(args=['self', 'pserver_endpoints'], varargs=None, keywords=None, defaults=None), ('document', '6adf97f83acf6453d4a6a4b1070f3754'))
 paddle.fluid.transpiler.HashName.dispatch (ArgSpec(args=['self', 'varlist'], varargs=None, keywords=None, defaults=None), ('document', '6adf97f83acf6453d4a6a4b1070f3754'))

python/paddle/fluid/transpiler/memory_optimization_transpiler.py

@@ -498,28 +498,54 @@ def memory_optimize(input_program,
                     print_log=False,
                     level=0,
                     skip_grads=False):
-    """Optimize memory by reusing var memory.
-
-      Note: it doesn't not support subblock nested in subblock.
+    """
+    | Legacy memory optimization strategy, reduce total memory consumption by reuse variable memory between different operators.
+    | Simple sample to explain the algorithm:
+    
+        ..  code-block:: python
+        
+            c = a + b  # assume this is the last time a is used
+            d = b * c
+         
+    | since **a** will not be used anymore after **"c = a + b"**, and the size of **a** and **d** are the same, 
+      we can use variable **a** to replace variable **d**, so actually we can optimize the above code to below:
 
+        ..  code-block:: python
+        
+            c = a + b
+            a = b * c 
+          
+    
+    | Please notice that, in this legacy design, we are using variable **a** to replace **d** directly, which means 
+      after you call this API, some variables may disappear, and some variables may hold unexpected values, like 
+      the above case, actually **a** holds the value of **d** after execution. 
+    
+    | So to protect important variables from being reused/removed in the optimization, we provide skip_opt_set 
+      to allow you specify a variable whitelist. 
+      The variables in the skip_opt_set will not be affected by memory_optimize API.
+    
+    Note: 
+        | **This API is deprecated, please avoid to use it in your new code.**
+        | Does not support operators which will create sub-block like While, IfElse etc.
+    
     Args:
         input_program(str): Input Program
         skip_opt_set(set): vars wil be skipped in memory optimze
         print_log(bool): whether to print debug log.
-        level(int): If level=0, reuse if the shape is completely equal, o
+        level(int): 0 or 1, 0 means we replace a with b only when a.size == b.size, 1 means we can replace a with b if a.size <= b.size
     Returns:
         None
     Examples:
         ..  code-block:: python
             
             import paddle.fluid as fluid
-
+    
             # build network
             # ...
             
             # deprecated API
             fluid.release_memory(fluid.default_main_program())
-
+    
     """
     sys.stderr.write('memory_optimize is deprecated. '
                      'Use CompiledProgram and Executor\n')