Merge pull request #8543 from reyoung/feature/enhance_layer_generator

Enhance layer_function_generator

Author: reyoung

.gitignore

@@ -27,7 +27,6 @@ third_party/
 cmake-build-*
 
 # generated while compiling
-python/paddle/v2/fluid/core.so
 paddle/pybind/pybind.h
 CMakeFiles
 cmake_install.cmake

python/paddle/fluid/.gitignore

@@ -1 +1,2 @@
 proto
+core.so

python/paddle/fluid/layers/layer_function_generator.py

@@ -130,7 +130,7 @@ def generate_layer_fn(op_type):
     o_name = not_intermediate_outputs[0].name
     intermediate_output_names = [output.name for output in intermediate_outputs]
 
-    def infer_and_check_dtype(op_proto, **kwargs):
+    def infer_and_check_dtype(op_proto, *args, **kwargs):
         """
         This function performs the sanity check for dtype and
         instance type.
@@ -141,6 +141,10 @@ def infer_and_check_dtype(op_proto, **kwargs):
             val = kwargs.pop(name, [])
             if not isinstance(val, list) and not isinstance(val, tuple):
                 val = [val]
+            if len(val) == 0:
+                val = [args[0]]
+                args = args[1:]
+
             for each in val:
                 if not isinstance(each, Variable):
                     raise ValueError("input of {0} must be variable".format(
@@ -155,17 +159,20 @@ def infer_and_check_dtype(op_proto, **kwargs):
 
         return dtype
 
-    def func(**kwargs):
+    def func(*args, **kwargs):
         helper = LayerHelper(op_type, **kwargs)
 
-        dtype = infer_and_check_dtype(op_proto, **kwargs)
+        dtype = infer_and_check_dtype(op_proto, *args, **kwargs)
 
         inputs = dict()
         for ipt in op_proto.inputs:
             name = _convert_(ipt.name)
             val = kwargs.pop(name, [])
             if not isinstance(val, list) and not isinstance(val, tuple):
                 val = [val]
+            if len(val) == 0 and len(args) != 0:
+                val = args[0]
+                args = args[1:]
             inputs[ipt.name] = val
 
         outputs = dict()

python/paddle/fluid/layers/tensor.py

@@ -160,8 +160,8 @@ def sums(input, out=None):
           a0 = layers.array_read(array=tmp, i=i)
           i = layers.increment(x=i)
           a1 = layers.array_read(array=tmp, i=i)
-          mean_a0 = layers.mean(x=a0)
-          mean_a1 = layers.mean(x=a1)
+          mean_a0 = layers.mean(a0)
+          mean_a1 = layers.mean(a1)
           a_sum = layers.sums(input=[mean_a0, mean_a1])
     """
     helper = LayerHelper('sum', **locals())

python/paddle/fluid/tests/book/notest_rnn_encoder_decoer.py

@@ -147,7 +147,7 @@ def seq_to_seq_net():
     label = fluid.layers.data(
         name='label_sequence', shape=[1], dtype='int64', lod_level=1)
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     return avg_cost, prediction
 

python/paddle/fluid/tests/book/test_fit_a_line.py

@@ -29,7 +29,7 @@ def train(use_cuda, save_dirname):
     y = fluid.layers.data(name='y', shape=[1], dtype='float32')
 
     cost = fluid.layers.square_error_cost(input=y_predict, label=y)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
     sgd_optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book/test_image_classification.py

@@ -110,7 +110,7 @@ def train(net_type, use_cuda, save_dirname):
 
     predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
     cost = fluid.layers.cross_entropy(input=predict, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     acc = fluid.layers.accuracy(input=predict, label=label)
 
     # Test program 

python/paddle/fluid/tests/book/test_label_semantic_roles.py

@@ -164,7 +164,7 @@ def train(use_cuda, save_dirname=None):
         label=target,
         param_attr=fluid.ParamAttr(
             name='crfw', learning_rate=mix_hidden_lr))
-    avg_cost = fluid.layers.mean(x=crf_cost)
+    avg_cost = fluid.layers.mean(crf_cost)
 
     # TODO(qiao)
     # check other optimizers and check why out will be NAN

python/paddle/fluid/tests/book/test_machine_translation.py

@@ -178,7 +178,7 @@ def train_main(use_cuda, is_sparse):
     label = pd.data(
         name="target_language_next_word", shape=[1], dtype='int64', lod_level=1)
     cost = pd.cross_entropy(input=rnn_out, label=label)
-    avg_cost = pd.mean(x=cost)
+    avg_cost = pd.mean(cost)
 
     optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
     optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book/test_recognize_digits.py

@@ -48,7 +48,7 @@ def parse_arg():
 def loss_net(hidden, label):
     prediction = fluid.layers.fc(input=hidden, size=10, act='softmax')
     loss = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_loss = fluid.layers.mean(x=loss)
+    avg_loss = fluid.layers.mean(loss)
     acc = fluid.layers.accuracy(input=prediction, label=label)
     return prediction, avg_loss, acc
 
@@ -101,8 +101,8 @@ def train(nn_type, use_cuda, parallel, save_dirname, save_param_filename):
 
         avg_loss, acc = pd()
         # get mean loss and acc through every devices.
-        avg_loss = fluid.layers.mean(x=avg_loss)
-        acc = fluid.layers.mean(x=acc)
+        avg_loss = fluid.layers.mean(avg_loss)
+        acc = fluid.layers.mean(acc)
     else:
         prediction, avg_loss, acc = net_conf(img, label)