You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Our project builds a Convolutional Neural Network (CNN) model to accurately classify handwritten digits from the MNIST dataset. We preprocess the data and design a CNN architecture. Additionally, we create a user-friendly web interface using Flask for easy digit classification.
Leveraging the mapreduce paradigm we propose a solution to parallelize the feedforward operation of neural networks in order to speed it up for sufficiently large NN architectures and for sufficiently large datasets. Tested Using the MNIST dataset results can be found in the results.html and results.ipynb files.
Made with the ONNX.js framework from Microsoft. Similar to tensorflow.js, ONNX.js is another framework to provide the capability of running machine learning models on the web with JavaScript