update DoReFa readme

examples/DoReFa-Net/README.md

@@ -2,6 +2,22 @@ Code and model for the paper:
 
 [DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients](http://arxiv.org/abs/1606.06160), by Zhou et al.
 
+It also contains an implementation of the following papers:
++ [Binary Weight Network](https://arxiv.org/abs/1511.00363), with (W,A,G)=(1,32,32).
++ [Trained Ternary Quantization](https://arxiv.org/abs/1612.01064), with (W,A,G)=(t,32,32).
++ [Binarized Neural Networks](https://arxiv.org/abs/1602.02830), with (W,A,G)=(1,1,32).
+
+These different quantization techniques achieves the following accuracy in this implementation:
+
+| Model              | W,A,G       | Top 1 Error |
+|:-------------------|-------------|------------:|
+| Full Precision     | 32,32,32    |      41.4%  |
+| TTQ                | t,32,32     |      41.9%  |
+| BWN                | 1,32,32     |      44.3%  |
+| BNN                | 1,1,32      |      53.4%  |
+| DoReFa             | 1,2,6       |      47.6%  |
+| DoReFa             | 1,2,4       |      58.4%  |
+
 We hosted a demo at CVPR16 on behalf of Megvii, Inc, running a real-time 1/4-VGG size DoReFa-Net on ARM and half-VGG size DoReFa-Net on FPGA.
 We're not planning to release our C++ runtime for bit-operations.
 In this repo, bit operations are performed through `tf.float32`.
@@ -11,11 +27,6 @@ Pretrained model for (1,4,32)-ResNet18 and (1,2,6)-AlexNet are available at
 They're provided in the format of numpy dictionary, so it should be very easy to port into other applications.
 The __binary-weight 4-bit-activation ResNet-18__ model has 59.2% top-1 validation accuracy.
 
-Note that when (W,A,G) is set to (1,32,32), this code implements [Binary Weight Network](https://arxiv.org/abs/1511.00363).
-When (W,A,G) is set to (t,32,32), this code implements
-[Trained Ternary Quantization](https://arxiv.org/abs/1612.01064).
-But with (W,A,G) set to (1,1,32), it is not equivalent to [XNOR-Net](https://arxiv.org/abs/1603.05279), although it won't be hard to implement it.
-
 Alternative link to this page: [http://dorefa.net](http://dorefa.net)
 
 ## Preparation:

examples/DoReFa-Net/alexnet-dorefa.py

@@ -31,11 +31,10 @@ This is our attempt to reproduce it on tensorpack & TensorFlow.
 
 Accuracy:
     Trained with 4 GPUs and (W,A,G)=(1,2,6), it can reach top-1 single-crop validation error of 47.6%,
-    after 70 epochs. This number is better than what's in the paper
-    due to more sophisticated augmentations.
+    after 70 epochs. This number is better than what's in the paper due to more sophisticated augmentations.
 
     With (W,A,G)=(32,32,32) -- full precision baseline, 41.4% error.
-    With (W,A,G)=(t,32,32) -- TTQ, 42.3% error
+    With (W,A,G)=(t,32,32) -- TTQ, 41.9% error
     With (W,A,G)=(1,32,32) -- BWN, 44.3% error
     With (W,A,G)=(1,1,32) -- BNN, 53.4% error
     With (W,A,G)=(1,2,6), 47.6% error

setup.py

@@ -15,8 +15,10 @@ exec(open(libinfo_py, "rb").read())
 try:
     import pypandoc
     long_description = pypandoc.convert_file('README.md', 'rst')
+    description_type = 'text/x-rst'
 except ImportError:
     long_description = open('README.md').read()
+    description_type = 'text/markdown'
 
 # configure requirements
 reqfile = os.path.join(CURRENT_DIR, 'requirements.txt')
@@ -27,7 +29,7 @@ setup(
     version=__version__,
     description='Neural Network Toolbox on TensorFlow',
     long_description=long_description,
-
+    long_description_content_type=description_type,
     install_requires=req,
     tests_require=['flake8', 'scikit-image'],
     extras_require={