v1.0.0: Initial Release - Pure NumPy Deep Learning Framework

🚀 First Official Release of miniFlow

We are excited to announce the first stable release of miniFlow, a lightweight deep learning framework built entirely from scratch using NumPy.

🚀 Key Highlights

• Pure NumPy Implementation: No reliance on Autograd engines like PyTorch or TensorFlow.
• High Performance: Achieved ~94% accuracy on the MNIST dataset using a simple MLP.
• Modular Design: Includes Dense, Flatten, Dropout, and BatchNorm layers.
• Advanced Optimization: Implemented Adam Optimizer and Softmax Cross Entropy with numerical stability.

📦 What's Included?

• Full implementation of Forward and Backward propagation.
• main.py script to train and visualize results on MNIST.
• Professional visualization tools for Loss and Predictions.
• He and Xavier weight initialization support.

🛠 Installation & Usage

Simply clone the repository and run the main script:

git clone https://github.com/alitkbbl/miniFlow.git
python main.py

Happy coding!