Update README.md

01_mysteries_of_neural_networks/06_numpy_convolutional_neural_net/README.md

@@ -21,6 +21,26 @@ source .env/bin/activate
 pip install -r requirements.txt
 ```
 
+## Demo
+
+[Check][5] how to use the created library to create a Convolutional Neural Network capable of classifying MNIST images, with 90% accuracy, using only NumPy.
+
+<p align="center"> 
+    <img width="600" src="./viz/cnn_acc.png" alt="Accuracy plot">
+</p>
+
+<p align="center"> 
+    <b>Figure 1.</b> Validation set accuracy / epoch 
+</p>
+
+<p align="center"> 
+    <img width="600" src="./viz/cnn_loss.png" alt="Accuracy plot">
+</p>
+
+<p align="center"> 
+    <b>Figure 2.</b> Validation set loss / epoch 
+</p>
+
 ## References
 
 1. CS231n: Convolutional Neural Networks for Visual Recognition - [link][1]
@@ -32,3 +52,4 @@ pip install -r requirements.txt
 [2]: https://github.com/lpraat/numpyCNN
 [3]: https://hal.inria.fr/inria-00112631/document
 [4]: https://www.youtube.com/watch?v=agYZcJNmbsU
+[5]: https://github.com/SkalskiP/ILearnDeepLearning.py/blob/master/01_mysteries_of_neural_networks/06_numpy_convolutional_neural_net/notebooks/CompleteCNN.ipynb