update readme

readme.md

@@ -1,6 +1,6 @@
 # GhostNet
 
-GhostNet: More Features from Cheap Operations
+GhostNet: More Features from Cheap Operations [[arXiv]](https://arxiv.org/abs/1911.11907)
 
 By Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, Chang Xu.
 
@@ -12,22 +12,25 @@ By Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, Chang Xu.
 
 - **Performance**
 
+We beat other SOTA lightweight CNNs such as **MobileNetV3** and **FBNet**.
+
 <div align="center">
    <img src="./fig/flops_latency.png" width="720">
 </div>
 
 ## Implementation
 
-The code provides the inference code and pretrained model of GhostNet on ImageNet.
+The code provides the TensorFlow code and pretrained model of GhostNet on ImageNet.
 
 `myconv2d.py` implemented `GhostModule` and `ghost_net.py` implemented `GhostNet`.
 
+### Requirements
+The code was verified on Python3.6, TensorFlow-1.13.1, Tensorpack-0.9.7. Not sure on other version.
+
 ### Usage
 Run `python test-ghostnet.py --eval --data_dir=/path/to/imagenet/dir/ --load=./models/ghostnet_checkpoint` to evaluate on `val` set.
-You'll get the accuracy: top-1 error=`0.26066`, top-5 error=`0.08614`.
 
-### Requirements
-The code was verified on Python3.6, TF-1.13.1, Tensorpack-0.9.7. Not sure on other version.
+You'll get the accuracy: top-1 error=`0.26066`, top-5 error=`0.08614` with only `141M` Flops (or say MAdds).
 
 ### Data Preparation
 ImageNet data dir should have the following structure, and `val` and `caffe_ilsvrc12` subdirs are essential:
@@ -50,7 +53,7 @@ caffe_ilsvrc12 data can be downloaded from http://dl.caffe.berkeleyvision.org/ca
 @article{ghostnet,
   title={GhostNet: More Features from Cheap Operations},
   author={Han, Kai and Wang, Yunhe and Tian, Qi and Guo, Jianyuan and Xu, Chunjing and Xu, Chang},
-  journal={arXiv preprint},
+  journal={arXiv},
   year={2019}
 }
 ```