v1.0.0 - Initial Railway Deployment

Full Changelog: https://github.com/aypy01/nerv/commits/v1.0.0

NERV - Machine Learning & Django Integration

The Vision

NERV is an open-source bridge between complex Machine Learning documentation (like Google Colab) and a functional web interface.

This project was a personal "test of fire" to master the CS50 AI curriculum and deploy it for a non-technical audience. What took 30 days of research to build the first time, I can now replicate in less than a week.

---

Featured Model: Yapper (v0.0.1)

The core of NERV is Yapper, a sentiment analysis engine built with TensorFlow.

Technical Shift: Vectorization ➔ LSTM

During development, I found that standard vectorization struggled with long-form text. To solve this, I researched and implemented a Bidirectional LSTM (Long Short-Term Memory) network.

• Why LSTM? Better at retaining context in reviews over 50 words.
• Accuracy: Maintained a stable accuracy of $\pm 2%$ while significantly improving contextual reliability.
• Roadmap: Future versions will move beyond binary classification (Positive/Negative) to generative logic.

---

Tech Stack

Category	Technology
Backend	Django 4.2+
AI/ML	TensorFlow-CPU, NumPy
Deployment	Railway
Server	Gunicorn, WhiteNoise

---

Quick Start (Local Setup)

1. Clone & Install:

   git clone [https://github.com/YOUR_USERNAME/nerv.git](https://github.com/YOUR_USERNAME/nerv.git)
   pip install -r requirements.txt