add new information concerning latest changes

Author: Coderx7

Cifar/README.md

@@ -0,0 +1,65 @@
+# SimpleNet in Pytorch
+SimpleNetV1 architecture implementation in Pytorch 
+
+## Lets Keep it simple, Using simple architectures to outperform deeper and more complex architectures (2016).
+
+![GitHub Logo](https://github.com/Coderx7/SimpleNet/raw/master/SimpNet_V1/images(plots)/SimpleNet_Arch_Larged.jpg)
+
+
+This is the pytorch implementation of our architecture [SimpleNetV1(2016)](https://arxiv.org/abs/1608.06037) .   
+Pytorch is different from caffe in several sections, and this made it a bit harder to have the architecture properly ported especially since I'm a complete newbie in Pytorch. However, thanks to [this great work](https://github.com/D-X-Y/ResNeXt-DenseNet), I could easily focus on the model and port the architecture and hopefully achieve my reported results in Caffe and also exceed them as well! 
+
+The pytorch implementation is also very effieicent and the whole model takes only **1239MB** with the batch size of 64! (compare this to other architectures such as ResNet,WRN, DenseNet which a 800K model takes more than 6G of vram!)   
+
+The original Caffe implementation can be found here : [Original Caffe implementation - 2016](https://github.com/Coderx7/SimpleNet)     
+  
+#### CIFAR10/100 Results achieved using this implementation :
+
+| Dataset | Accuracy |
+|------------|----------|
+| CIFAR10    | **95.51**    |
+| CIFAR100   | **78.37**   |
+
+### CIFAR10/100 top results (2016): 
+
+| **Method**                   | **\#Params** |  **CIFAR10**  | **CIFAR100** |
+| :--------------------------- | :----------: | :-----------: | :----------: |
+| VGGNet(16L) /Enhanced        |     138m     | 91.4 / 92.45  |      \-      |
+| ResNet-110L / 1202L  \*      |  1.7/10.2m   | 93.57 / 92.07 | 74.84/72.18  |
+| SD-110L / 1202L              |  1.7/10.2m   | 94.77 / 95.09 |  75.42 / -   |
+| WRN-(16/8)/(28/10)           |    11/36m    | 95.19 / 95.83 |  77.11/79.5  |
+| Highway Network              |     N/A      |     92.40     |    67.76     |
+| FitNet                       |      1M      |     91.61     |    64.96     |
+| FMP\* (1 tests)              |     12M      |     95.50     |    73.61     |
+| Max-out(k=2)                 |      6M      |     90.62     |    65.46     |
+| Network in Network           |      1M      |     91.19     |    64.32     |
+| DSN                          |      1M      |     92.03     |    65.43     |
+| Max-out NIN                  |      \-      |     93.25     |    71.14     |
+| LSUV                         |     N/A      |     94.16     |     N/A      |
+| SimpleNet                    |    5.48M     |   **95.51**   | **78.37**    |
+
+
+
+#### Models and logs  
+-- Models and training logs can be found in [snapshot folder](https://github.com/Coderx7/SimpleNet_Pytorch/tree/master/snapshots).
+
+
+
+#### How to run ? 
+Simply initiate the training like :  
+`python3 main.py ./data/cifar.python --dataset cifar10 --arch simplenet_cifar_5m --save_path ./snapshots/simplenet --epochs 540 --batch_size 100 --workers 2`
+
+
+Note 1: the initial learning rate, and optimization policy is hard coded just like caffe.  
+Note 2: for testing the cifar10/100 weights located in the repository, use the `simplenet_cifar_5m_extra_pool` model instead. see [issue #5](https://github.com/Coderx7/SimpleNet_Pytorch/issues/5) for more information. 
+
+
+## Citation
+If you find SimpleNet useful in your research, please consider citing:
+
+    @article{hasanpour2016lets,
+      title={Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures},
+      author={Hasanpour, Seyyed Hossein and Rouhani, Mohammad and Fayyaz, Mohsen and Sabokrou, Mohammad},
+      journal={arXiv preprint arXiv:1608.06037},
+      year={2016}
+    }