Neural Network Testing App
Maaz Ahmad
A neural network for the MNIST dataset, built and deployed using Streamlit. This app allows users to experiment with different network architectures, train models, and visualize the results.
Key Features:
Customizable Architecture: Adjust the number of hidden layers and units per layer.
Training & Evaluation: Train the model on the fly and evaluate its performance on training and validation sets.
Insights: Get feedback on whether the model is overfitting, underfitting, or well-balanced.
This project was a great opportunity to deepen my understanding of neural networks and deployment using Streamlit. It also provided hands-on experience in visualizing model performance and making data-driven decisions.
Check out my code here: https://lnkd.in/dGBG5C7B
Like this project
Posted Aug 3, 2024
Streamlit web-app to to visualize effect of various neural net architectures on model accuracy
