Follow comments

Author: reyoung

demo/mnist/api_train_v2.py

@@ -50,10 +50,10 @@ def event_handler(event):
         output=inference,
         parameters=parameters,
         reader=paddle.reader.batched(
-            paddle.reader.limited(
+            paddle.reader.firstn(
                 paddle.reader.map_readers(lambda item: (item[0], ),
                                           paddle.dataset.mnist.test()),
-                limit=100),
+                n=100),
             batch_size=32))
     print probs.shape
 

python/paddle/v2/inferencer.py

@@ -5,10 +5,10 @@
 import itertools
 import numpy
 
-__all__ = ['InferenceEngine', 'infer']
+__all__ = ['Inference', 'infer']
 
 
-class InferenceEngine(object):
+class Inference(object):
     def __init__(self, output, parameters):
         topo = topology.Topology(output)
         gm = api.GradientMachine.createFromConfigProto(
@@ -55,5 +55,5 @@ def default_reader_dict(self):
 
 
 def infer(output, parameters, reader, reader_dict=None, field='value'):
-    inferer = InferenceEngine(output=output, parameters=parameters)
+    inferer = Inference(output=output, parameters=parameters)
     return inferer.infer(field=field, reader=reader, reader_dict=reader_dict)

python/paddle/v2/reader/decorator.py

@@ -14,7 +14,7 @@
 
 __all__ = [
     'map_readers', 'buffered', 'compose', 'chain', 'shuffle',
-    'ComposeNotAligned', 'batched', 'limited'
+    'ComposeNotAligned', 'batched', 'firstn'
 ]
 
 import itertools
@@ -215,15 +215,18 @@ def batched_reader():
     return batched_reader
 
 
-def limited(reader, limit):
+def firstn(reader, n):
     """
     Limit the max number of samples that reader could return.
     """
 
-    def limited_reader():
+    # TODO(yuyang18): Check if just drop the reader, could clean the opened
+    # resource or not?
+
+    def firstn_reader():
         for i, item in enumerate(reader()):
-            if i == limit:
+            if i == n:
                 break
             yield item
 
-    return limited_reader
+    return firstn_reader