Trim the features

README.md

@@ -16,26 +16,18 @@ Contents
 
 Features
 --------
-* Lightweight: small but sharp knife
-  - mxnet contains concise implementation of state-of-art deep learning models
-  - The project maintains a minimum dependency that makes it portable and easy to build
-* Scalable and beyond
-  - The package scales to multiple GPUs already with an easy to use kvstore.
-  - The same code can be ported to distributed version when the distributed kvstore is ready.
-* Multi-GPU NDArray/Tensor API with auto parallelization
-  - The package supports a flexible ndarray interface that runs on both CPU and GPU, more importantly
-    automatically parallelize the computation for you.
+* To Mix and Maximize
+  - Mix all flavors of programming models to maximize flexiblity and efficiency.
+* Lightweight and scalable
+  - Minimum build dependency, scales to multi-GPU and ready toward distributed.
+* Auto parallelization
+  - Write serial ndarray GPU programs, and let the engine parallelize it for you.
 * Language agnostic
-  - The package currently support C++ and python, with a clean C API.
-  - This makes the package being easily portable to other languages and platforms.
+  - With support for python, c++, more to come.
 * Cloud friendly
-  - MXNet is ready to work with cloud storages including S3, HDFS, AZure for data source and model saving.
-  - This means you do can put data on S3 directly using it to train your deep model.
-* Easy extensibility with no requirement on GPU programming
-  - The package can be extended in several scopes, including python, c++.
-  - In all these levels, developers can write numpy style expressions, either via python
-    or [mshadow expression template](https://github.com/dmlc/mshadow).
-  - It brings concise and readable code, with performance matching hand crafted kernels
+  - Directly load/save from S3, HDFS, AZure
+* Easy extensibility
+  - Extending no requirement on GPU programming.
 
 Bug Reporting
 -------------

doc/developer-guide/index.md

@@ -6,7 +6,7 @@ Overview of the Design
 ----------------------
 * [Execution Engine](engine.md)
 
-List of Resources
------------------
-* [Doxygen Version of C++ API](https://mxnet.readthedocs.org/en/latest/doxygen)
-* [Contributor Guide](../contribute.md)
+List of Other Resources
+-----------------------
+* [Doxygen Version of C++ API](https://mxnet.readthedocs.org/en/latest/doxygen) gives a comprehensive document of C++ API.
+* [Contributor Guide](../contribute.md) gives guidelines on how to push changes to the project.

example/README.md

@@ -0,0 +1,17 @@
+MXNet Examples
+==============
+This folder contains examples of MXNet.
+
+Contents
+--------
+* [mnist](mnist) gives examples on training mnist.
+* [cifar10](cifar10) gives examples on CIFAR10 dataset.
+
+
+Python Howto
+------------
+[Python Howto](python-howto) is a folder containing short snippet of code
+introducing a certain feature of mxnet.
+
+***List of Examples***
+* [Configuring Net to get Multiple Ouputs](python-howto/multiple_outputs.py)

example/python-howto/multiple_outputs.py

@@ -0,0 +1,21 @@
+"""Create a Multiple output configuration.
+
+This example shows how to create a multiple output configuration.
+"""
+import mxnet as mx
+
+net = mx.symbol.Variable('data')
+fc1 = mx.symbol.FullyConnected(data=net, name='fc1', num_hidden=128)
+net = mx.symbol.Activation(data=fc1, name='relu1', act_type="relu")
+net = mx.symbol.FullyConnected(data=net, name='fc2', num_hidden=64)
+out = mx.symbol.Softmax(data=net, name='softmax')
+# group fc1 and out together
+group = mx.symbol.Group([fc1, out])
+print group.list_outputs()
+
+# You can go ahead and bind on the group
+# executor = group.simple_bind(data=data_shape)
+# executor.forward()
+# executor.output[0] will be value of fc1
+# executor.output[1] will be value of softmax
+

python/mxnet/symbol.py

@@ -1,6 +1,10 @@
 # coding: utf-8
 # pylint: disable=invalid-name, protected-access, fixme, too-many-arguments
-"""Symbol support of mxnet"""
+"""Symbolic support of mxnet.
+
+Symbolic API of MXNet
+
+"""
 from __future__ import absolute_import
 
 import ctypes
@@ -341,10 +345,10 @@ def _get_ndarray_handle(arg_key, args, arg_names, allow_missing):
         arg_key : str
             The name of argument, used for error message.
 
-        args : list of NDArray or dict of str->NDArray
+        args : list of NDArray or dict of str to NDArray
             Input arguments to the symbols.
             If type is list of NDArray, the position is in the same order of arg_names.
-            If type is dict of str->NDArray, then it maps the name of arguments
+            If type is dict of str to NDArray, then it maps the name of arguments
             to the corresponding NDArray,
 
         args_names : list of string
@@ -366,22 +370,22 @@ def _get_ndarray_handle(arg_key, args, arg_names, allow_missing):
                 raise ValueError('Length of %s do not match number of arguments' % arg_key)
             for narr in args:
                 if not isinstance(narr, NDArray):
-                    raise TypeError('Only Accept list of NDArrays or dict of str->NDArray')
+                    raise TypeError('Only Accept list of NDArrays or dict of str to NDArray')
                 arg_handles.append(narr.handle)
         elif isinstance(args, dict):
             for name in arg_names:
                 if name in arg_names:
                     narr = args[name]
                     if not isinstance(narr, NDArray):
-                        raise TypeError('Only Accept list of NDArrays or dict of str->NDArray')
+                        raise TypeError('Only Accept list of NDArrays or dict of str to NDArray')
                     arg_handles.append(narr.handle)
                 else:
                     if allow_missing:
                         arg_handles.append(None)
                     else:
                         raise ValueError('Must specify all the arguments in %s' % arg_key)
         else:
-            raise TypeError('Only Accept list of NDArrays or dict of str->NDArray')
+            raise TypeError('Only Accept list of NDArrays or dict of str to NDArray')
         return c_array(NDArrayHandle, arg_handles)
 
     def simple_bind(self, ctx, grad_req='write', **kwargs):
@@ -396,12 +400,12 @@ def simple_bind(self, ctx, grad_req='write', **kwargs):
         ctx : Context
             The device context the generated executor to run on.
         grad_req: string
-            {'write', 'add', 'null'}, or list of str or dict of str->str, optional
+            {'write', 'add', 'null'}, or list of str or dict of str to str, optional
             Specifies how we should update the gradient to the args_grad.
             - 'write' means everytime gradient is write to specified args_grad NDArray.
             - 'add' means everytime gradient is add to the specified NDArray.
             - 'null' means no action is taken, the gradient may not be calculated.
-        kwargs : dict of str->NDArray
+        kwargs : dict of str to NDArray
 
         Returns
         -------
@@ -436,34 +440,38 @@ def bind(self, ctx, args, args_grad=None, grad_req='write', aux_states=None):
         ctx : Context
             The device context the generated executor to run on.
 
-        args : list of NDArray or dict of str->NDArray
+        args : list of NDArray or dict of str to NDArray
             Input arguments to the symbol.
+
             - If type is list of NDArray, the position is in the same order of list_arguments.
-            - If type is dict of str->NDArray, then it maps the name of arguments
-              to the corresponding NDArray,
+            - If type is dict of str to NDArray, then it maps the name of arguments
+              to the corresponding NDArray.
             - In either case, all the arguments must be provided.
 
-        args_grad : list of NDArray or dict of str->NDArray, optional
+        args_grad : list of NDArray or dict of str to NDArray, optional
             When specified, args_grad provide NDArrays to hold
             the result of gradient value in backward.
+
             - If type is list of NDArray, the position is in the same order of list_arguments.
-            - If type is dict of str->NDArray, then it maps the name of arguments
+            - If type is dict of str to NDArray, then it maps the name of arguments
               to the corresponding NDArray.
-            - When the type is dict of str->NDArray, users only need to provide the dict
+            - When the type is dict of str to NDArray, users only need to provide the dict
               for needed argument gradient.
               Only the specified argument gradient will be calculated.
 
-        grad_req : {'write', 'add', 'null'}, or list of str or dict of str->str, optional
+        grad_req : {'write', 'add', 'null'}, or list of str or dict of str to str, optional
             Specifies how we should update the gradient to the args_grad.
+
             - 'write' means everytime gradient is write to specified args_grad NDArray.
             - 'add' means everytime gradient is add to the specified NDArray.
             - 'null' means no action is taken, the gradient may not be calculated.
 
-        aux_states : list of NDArray, or dict of str->NDArray, optional
+        aux_states : list of NDArray, or dict of str to NDArray, optional
             Input auxiliary states to the symbol, only need to specify when
             list_auxiliary_states is not empty.
+
             - If type is list of NDArray, the position is in the same order of list_auxiliary_states
-            - If type is dict of str->NDArray, then it maps the name of auxiliary_states
+            - If type is dict of str to NDArray, then it maps the name of auxiliary_states
               to the corresponding NDArray,
             - In either case, all the auxiliary_states need to be provided.
 
@@ -579,7 +587,7 @@ def Variable(name):
 
 
 def Group(symbols):
-    """Create a symbolic variable that groups several symbols together.
+    """Create a symbol that groups symbols together.
 
     Parameters
     ----------
@@ -613,10 +621,11 @@ def load(fname):
     Parameters
     ----------
     fname : str
-        The name of the file
-        - s3://my-bucket/path/my-s3-symbol
-        - hdfs://my-bucket/path/my-hdfs-symbol
-        - /path-to/my-local-symbol
+        The name of the file, examples:
+
+        - `s3://my-bucket/path/my-s3-symbol`
+        - `hdfs://my-bucket/path/my-hdfs-symbol`
+        - `/path-to/my-local-symbol`
 
     Returns
     -------