Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into macosx

Author: gangliao

.travis.yml

@@ -4,6 +4,7 @@ cache:
     - $HOME/third_party
     - $HOME/.ccache
     - $HOME/.cache/pip
+    - $HOME/Library/Caches/Homebrew
 sudo: required
 dist: trusty
 os:
@@ -54,7 +55,9 @@ before_install:
     fi
   - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then paddle/scripts/travis/before_install.osx.sh; fi
   - if [[ "$JOB" == "PRE_COMMIT" ]]; then sudo ln -s /usr/bin/clang-format-3.8 /usr/bin/clang-format; fi
-  - pip install numpy wheel protobuf sphinx recommonmark sphinx_rtd_theme virtualenv pre-commit requests==2.9.2 LinkChecker
+  # Paddle is using protobuf 3.1 currently. Protobuf 3.2 breaks the compatibility. So we specify the python 
+  # protobuf version.
+  - pip install numpy wheel 'protobuf==3.1' sphinx recommonmark sphinx_rtd_theme virtualenv pre-commit requests==2.9.2 LinkChecker
 script:
   - paddle/scripts/travis/main.sh
 notifications:

cmake/cblas.cmake

@@ -16,7 +16,8 @@
 set(CBLAS_FOUND OFF)
 
 ## Find MKL First.
-set(MKL_ROOT $ENV{MKLROOT} CACHE PATH "Folder contains MKL")
+set(INTEL_ROOT "/opt/intel" CACHE PATH "Folder contains intel libs")
+set(MKL_ROOT ${INTEL_ROOT}/mkl CACHE PATH "Folder contains MKL")
 
 find_path(MKL_INCLUDE_DIR mkl.h PATHS
   ${MKL_ROOT}/include)

demo/mnist/api_train.py

@@ -6,25 +6,15 @@
 
 The user api could be simpler and carefully designed.
 """
-import py_paddle.swig_paddle as api
-from py_paddle import DataProviderConverter
-import paddle.trainer.PyDataProvider2 as dp
-import numpy as np
 import random
-from mnist_util import read_from_mnist
-from paddle.trainer_config_helpers import *
-import paddle.v2
 
+import numpy as np
+import paddle.v2 as paddle_v2
+import py_paddle.swig_paddle as api
+from paddle.trainer_config_helpers import *
+from py_paddle import DataProviderConverter
 
-def network_config():
-    imgs = data_layer(name='pixel', size=784)
-    hidden1 = fc_layer(input=imgs, size=200)
-    hidden2 = fc_layer(input=hidden1, size=200)
-    inference = fc_layer(input=hidden2, size=10, act=SoftmaxActivation())
-    cost = classification_cost(
-        input=inference, label=data_layer(
-            name='label', size=10))
-    outputs(cost)
+from mnist_util import read_from_mnist
 
 
 def init_parameter(network):
@@ -67,7 +57,7 @@ def input_order_converter(generator):
 def main():
     api.initPaddle("-use_gpu=false", "-trainer_count=4")  # use 4 cpu cores
 
-    optimizer = paddle.v2.optimizer.Adam(
+    optimizer = paddle_v2.optimizer.Adam(
         learning_rate=1e-4,
         batch_size=1000,
         model_average=ModelAverage(average_window=0.5),
@@ -79,8 +69,20 @@ def main():
     updater = optimizer.create_local_updater()
     assert isinstance(updater, api.ParameterUpdater)
 
+    # define network
+    images = paddle_v2.layer.data(
+        name='pixel', type=paddle_v2.data_type.dense_vector(784))
+    label = paddle_v2.layer.data(
+        name='label', type=paddle_v2.data_type.integer_value(10))
+    hidden1 = paddle_v2.layer.fc(input=images, size=200)
+    hidden2 = paddle_v2.layer.fc(input=hidden1, size=200)
+    inference = paddle_v2.layer.fc(input=hidden2,
+                                   size=10,
+                                   act=paddle_v2.activation.Softmax())
+    cost = paddle_v2.layer.classification_cost(input=inference, label=label)
+
     # Create Simple Gradient Machine.
-    model_config = parse_network_config(network_config)
+    model_config = paddle_v2.layer.parse_network(cost)
     m = api.GradientMachine.createFromConfigProto(model_config,
                                                   api.CREATE_MODE_NORMAL,
                                                   optimizer.enable_types())
@@ -97,8 +99,7 @@ def main():
 
     # DataProvider Converter is a utility convert Python Object to Paddle C++
     # Input. The input format is as same as Paddle's DataProvider.
-    converter = DataProviderConverter(
-        input_types=[dp.dense_vector(784), dp.integer_value(10)])
+    converter = DataProviderConverter(input_types=[images.type, label.type])
 
     train_file = './data/raw_data/train'
     test_file = './data/raw_data/t10k'

demo/mnist/api_train_v2.py

@@ -0,0 +1,61 @@
+import numpy
+import paddle.v2 as paddle
+
+import mnist_util
+
+
+def train_reader():
+    train_file = './data/raw_data/train'
+    generator = mnist_util.read_from_mnist(train_file)
+    for item in generator:
+        yield item
+
+
+def main():
+    paddle.init(use_gpu=False, trainer_count=1)
+
+    # define network topology
+    images = paddle.layer.data(
+        name='pixel', type=paddle.data_type.dense_vector(784))
+    label = paddle.layer.data(
+        name='label', type=paddle.data_type.integer_value(10))
+    hidden1 = paddle.layer.fc(input=images, size=200)
+    hidden2 = paddle.layer.fc(input=hidden1, size=200)
+    inference = paddle.layer.fc(input=hidden2,
+                                size=10,
+                                act=paddle.activation.Softmax())
+    cost = paddle.layer.classification_cost(input=inference, label=label)
+
+    parameters = paddle.parameters.create(cost)
+    for param_name in parameters.keys():
+        array = parameters.get(param_name)
+        array[:] = numpy.random.uniform(low=-1.0, high=1.0, size=array.shape)
+        parameters.set(parameter_name=param_name, value=array)
+
+    adam_optimizer = paddle.optimizer.Adam(learning_rate=0.01)
+
+    def event_handler(event):
+        if isinstance(event, paddle.event.EndIteration):
+            para = parameters.get('___fc_2__.w0')
+            print "Pass %d, Batch %d, Cost %f, Weight Mean Of Fc 2 is %f" % (
+                event.pass_id, event.batch_id, event.cost, para.mean())
+
+        else:
+            pass
+
+    trainer = paddle.trainer.SGD(update_equation=adam_optimizer)
+
+    trainer.train(train_data_reader=train_reader,
+                  topology=cost,
+                  parameters=parameters,
+                  event_handler=event_handler,
+                  batch_size=32,  # batch size should be refactor in Data reader
+                  data_types={  # data_types will be removed, It should be in
+                      # network topology
+                      'pixel': images.type,
+                      'label': label.type
+                  })
+
+
+if __name__ == '__main__':
+    main()

demo/sentiment/dataprovider.py

@@ -32,4 +32,6 @@ def process(settings, file_name):
             word_slot = [
                 settings.word_dict[w] for w in words if w in settings.word_dict
             ]
+            if not word_slot:
+                continue
             yield word_slot, label

demo/sentiment/predict.py

@@ -138,7 +138,11 @@ def main():
 
     batch = []
     for line in sys.stdin:
-        batch.append([predict.get_index(line)])
+        words = predict.get_index(line)
+        if words:
+            batch.append([words])
+        else:
+            print('All the words in [%s] are not in the dictionary.' % line)
         if len(batch) == batch_size:
             predict.batch_predict(batch)
             batch = []

doc/api/trainer_config_helpers/layers.rst

@@ -279,6 +279,12 @@ concat_layer
     :members: concat_layer
     :noindex:
 
+seq_concat_layer
+----------------
+..  automodule:: paddle.trainer_config_helpers.layers
+    :members: seq_concat_layer
+    :noindex:
+
 Reshaping Layers
 ================
 
@@ -302,6 +308,12 @@ repeat_layer
     :members: repeat_layer
     :noindex:
 
+seq_reshape_layer
+-----------------
+..  automodule:: paddle.trainer_config_helpers.layers
+    :members: seq_reshape_layer
+    :noindex:
+
 Math Layers
 ===========