update docs

.gitignore

@@ -8,6 +8,7 @@ __pycache__/
 *.h5
 
 # test code
+notebook.ipynb
 perceptron/tmp/
 perceptron/utils/images/porn.jpeg
 

MANIFEST.in

@@ -0,0 +1 @@
+include perceptron/utils/images/*.png

README.md

@@ -11,6 +11,12 @@ Currently, you can use Perceptron either through its python API or its command l
 
 ## Getting Started
 
+### Installation 
+The `PyTorch` and `Tensorflow` packages are required if you want to test their models, we requir user to manually install them on demand. Otherwise, run the following command to install Perceptron Benchmark
+```
+pip install -e .
+```
+
 ### Running Examples via Command Lines
 In the docker shell, run the test through Perceptron command line interface
 ```bash

docs/source/conf.py

@@ -40,7 +40,9 @@
 # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
 # ones.
 extensions = [
+    'sphinx.ext.autosummary',
     'sphinx.ext.autodoc',
+    'numpydoc',
     'sphinx.ext.todo',
     'sphinx.ext.mathjax',
     'sphinx.ext.ifconfig',

docs/source/index.rst

@@ -4,7 +4,7 @@
    contain the root `toctree` directive.
 
 Welcome to Perceptron Robustness Benchmark's page!
-=======================================
+==================================================
 
 Perceptron is a benchmark to test safety and security properties of neural
 networks for perceptual tasks.
@@ -15,7 +15,7 @@ It comes with support for many frameworks to build models including
 * PyTorch
 * Keras
 * Cloud API
-* PaddlePaddle (In progress)
+* PaddlePaddle
 
 See currently supported evaluation metrics, models, adversarial criteria,
 and verification methods in :doc:`user/summary`.
@@ -49,78 +49,6 @@ toolbox such as :code:`cleverhans`, :code:`foolbox`, :code:`IBM ART`,
     - Gives :red:`verifiable` robustness bounds for security and safety
       properties.
 
-More specifically, we compare :code:`perceptron` with existing DNN benchmarks
-in the following table:
-
-.. list-table:: :shade:`DNN Benchmarks Comparison`
-   :widths: 50 24 24 24 24
-   :header-rows: 1
-
-   * - Properties
-     - Perceptron
-     - Cleverhans
-     - Foolbox
-     - IBM ART
-   * - Multi-platform support
-     - :math:`\checkmark`
-     - :math:`\checkmark`
-     - :math:`\checkmark`
-     - :math:`\checkmark`
-   * - Consistent API design
-     - :math:`\checkmark`
-     - :math:`\cdot`
-     - :math:`\checkmark`
-     - :math:`\cdot`
-   * - Custom adversarial criteria
-     - :math:`\checkmark`
-     - :math:`\cdot`
-     - :math:`\checkmark`
-     - :math:`\cdot`
-   * - Multiple perceptual tasks
-     - :math:`\checkmark`
-     - :math:`\cdot`
-     - :math:`\cdot`
-     - :math:`\cdot`
-   * - Standardized metrics
-     - :math:`\checkmark`
-     - :math:`\cdot`
-     - :math:`\checkmark`
-     - :math:`\cdot`
-   * - Verifiable robustness bounds
-     - :math:`\checkmark`
-     - :math:`\cdot`
-     - :math:`\cdot`
-     - :math:`\cdot`
-
-Explanation of compared properties:
-
-    - :blue:`Multi-platform support`: supports at least the
-      three deep learning frameworks, :code:`Tensoflow`,
-      :code:`PyTorch`, and :code:`Keras`.
-
-    - :blue:`Consistent API design`: implementations of evaluation methods
-      are platform-agnostic. More specifically, the same piece of code for
-      an evaluation method (e.g., a :code:`C&W` attack) can run against
-      models across all platforms (e.g., :code:`Tensorflow`,
-      :code:`PyTorch`, and :code:`cloud API`).
-
-    - :blue:`Custom adversarial criterion`: a criterion defines under what
-      circumstances an :code:`(input, label)` pair is considered an adversary.
-      Customized adversarial criteria other than :code:`misclassification`
-      should be supported.
-
-    - :blue:`Multiple perceptual tasks`: supports computer vision tasks other
-      than :code:`classification`, e.g., :code:`object detection` and
-      :code:`face recognition`.
-
-    - :blue:`Standardized metrics`: enables DNN models' robustness to be
-      comparable on all **security** and **safety** properties.
-
-    - :blue:`Verifiable robustness bounds`: supports verification of certain
-      safety properties. Returns either a verifiable bound, indicating that the
-      model is robust against perturbations within that bound, or return
-      counter-examples.
-
 Running benchmarks
 ==================
 

perceptron/models/classification/cloud.py

@@ -53,7 +53,7 @@ def __init__(
 
 
 class AipAntiPornModel(AipModel):
-    """Create a :class: `Model` instance from an `AipAntiPorn` model.
+    """Create a :class:`Model` instance from an `AipAntiPorn` model.
 
     Parameters
     ----------

perceptron/models/classification/keras.py

@@ -21,7 +21,7 @@
 
 
 class KerasModel(DifferentiableModel):
-    """Create a :class: `Model` instance from a `Keras` model.
+    """Create a :class:`Model` instance from a `Keras` model.
 
     Parameters
     ----------

perceptron/models/detection/keras_retina_resnet50.py

@@ -23,7 +23,7 @@
 
 
 class KerasResNet50RetinaNetModel(DifferentiableModel):
-    """Create a :class: `Model` instance from a `DifferentiableModel` model.
+    """Create a :class:`Model` instance from a `DifferentiableModel` model.
 
     Parameters
     ----------

perceptron/models/detection/keras_ssd300.py

@@ -33,7 +33,7 @@
 
 
 class KerasSSD300Model(DifferentiableModel):
-    """Create a :class: `Model` instance from a `DifferentiableModel` model.
+    """Create a :class:`Model` instance from a `DifferentiableModel` model.
 
     Parameters
     ----------

perceptron/models/detection/keras_yolov3.py

@@ -25,7 +25,7 @@
 
 
 class KerasYOLOv3Model(DifferentiableModel):
-    """Create a :class: `Model` instance from a `Keras` model.
+    """Create a :class:`Model` instance from a `Keras` model.
 
     Parameters
     ----------

setup.py

@@ -1,8 +1,23 @@
 from setuptools import setup
 from setuptools import find_packages
 
+requirements = [
+    'numpy',
+    'matplotlib',
+    'urllib3',
+    'tqdm',
+    'pillow',
+    'scipy',
+    'sphinx',
+    'sphinx_rtd_theme'
+]
+
 setup(
     name='perceptron',
-    version='0.1.0.dev',
-    packages=find_packages()
+    description='Robustness benchmark for deep learning models',
+    version='1.0.0.dev',
+    url="https://github.com/baidu-advbox/perceptron-benchmark",
+    packages=find_packages(),
+    include_package_data=True,
+    install_requires=requirements
 )