Update README and requirements for consistency and accuracy

README.md

@@ -1,411 +1,11 @@
----# Next-Token Predictor# Next-Token Predictor# Next-Token Predictor---
-
-title: Next-Token Predictor
-
+---
+title: Next Token Predictor
 emoji: 🔮
-
 colorFrom: indigo
-
-colorTo: purpleA clean, modern web app that demonstrates how AI language models predict the next word in a sequence. Built with Gradio for easy deployment and secure token management.
-
+colorTo: purple
 sdk: gradio
-
-sdk_version: 4.44.0
-
+sdk_version: 4.44.1
 app_file: app.py
-
-pinned: false![Demo](https://img.shields.io/badge/Demo-Live-brightgreen) ![Python](https://img.shields.io/badge/Python-3.8+-blue) ![Gradio](https://img.shields.io/badge/Gradio-4.44+-orange)A clean, modern web app that demonstrates how AI language models predict the next word in a sequence. Built with Gradio for easy deployment and secure token management.title:# Next-Token Predictor
-
+pinned: false
 short_description: Demonstrate how AI language models predict the next word in a sequence
-
----
-
-
-
-# Next-Token Predictor## ✨ Features
-
-
-
-A clean, modern web app that demonstrates how AI language models predict the next word in a sequence. Built with Gradio for easy deployment and secure token management.
-
-
-
-![Demo](https://img.shields.io/badge/Demo-Live-brightgreen) ![Python](https://img.shields.io/badge/Python-3.8+-blue) ![Gradio](https://img.shields.io/badge/Gradio-4.44+-orange)- **🔮 Real-time Predictions**: Updates automatically as you type![Demo](https://img.shields.io/badge/Demo-Live-brightgreen) ![Python](https://img.shields.io/badge/Python-3.8+-blue) ![Gradio](https://img.shields.io/badge/Gradio-4.44+-orange)A clean, modern web app that demonstrates how AI language models predict the next word in a sequence. Built with Gradio for easy deployment and secure token management.
-
-
-
-## ✨ Features- **🔒 Secure**: Uses environment variables for API tokens  
-
-
-
-- **🔮 Real-time Predictions**: Updates automatically as you type- **⚡ Fast**: Serverless API calls with instant results
-
-- **🔒 Secure**: Uses environment variables for API tokens  
-
-- **⚡ Fast**: Serverless API calls with instant results- **📱 Responsive**: Clean, mobile-friendly interface
-
-- **📱 Responsive**: Clean, mobile-friendly interface
-
-- **🚀 Easy Deploy**: One-click deployment to HF Spaces- **🚀 Easy Deploy**: One-click deployment to HF Spaces## ✨ Features🎯 **Note**: This project has been converted to a Gradio app! See `README-gradio.md` for the latest version.
-
-
-
-## 🚀 Quick Start
-
-
-
-1. **Install dependencies:**## 🚀 Quick Start
-
-   ```bash
-
-   pip install -r requirements.txt
-
-   ```
-
-1. **Install dependencies:**- **🔮 Real-time Predictions**: Updates automatically as you type![Demo](https://img.shields.io/badge/Demo-Live-brightgreen) ![Python](https://img.shields.io/badge/Python-3.8+-blue) ![Gradio](https://img.shields.io/badge/Gradio-4.44+-orange)
-
-2. **Set your HF token:**
-
-      ```bash
-
-   **Option A: Using .env file (Recommended for local development)**
-
-   ```bash   pip install -r requirements.txt- **🔒 Secure**: Uses environment variables for API tokens  
-
-   # Copy the template
-
-   cp .env.example .env   ```
-
-   
-
-   # Edit .env and add your token:- **⚡ Fast**: Serverless API calls with instant resultsA web application that demonstrates how AI language models predict the next word in a sequence. Originally built with HTML/JavaScript, now available as a streamlined Gradio app for easier deployment and better security.
-
-   # HF_NEXT_TOKEN_PREDICTOR_TOKEN=hf_your_actual_token_here
-
-   ```2. **Set your HF token:**
-
-   
-
-   **Option B: Using environment variable**   - **📱 Responsive**: Clean, mobile-friendly interface
-
-   ```bash
-
-   export HF_NEXT_TOKEN_PREDICTOR_TOKEN="hf_your_token_here"  # Linux/Mac   **Option A: Using .env file (Recommended for local development)**
-
-   $env:HF_NEXT_TOKEN_PREDICTOR_TOKEN="hf_your_token_here"    # Windows PowerShell
-
-   ```   ```bash- **🚀 Easy Deploy**: One-click deployment to HF Spaces## ✨ Features
-
-
-
-3. **Run the app:**   # Copy the template
-
-   ```bash
-
-   python app.py   cp .env.example .env
-
-   ```
-
-   
-
-4. **Open browser:** Navigate to `http://127.0.0.1:7860`
-
-   # Edit .env and add your token:## 🚀 Quick Start## 🚀 Quick Start
-
-## 🔧 Deploy to HF Spaces
-
-   # HF_NEXT_TOKEN_PREDICTOR_TOKEN=hf_your_actual_token_here
-
-1. Create a new **Gradio Space** on Hugging Face
-
-2. Upload `app.py` and `requirements.txt`   ```
-
-3. Set `HF_NEXT_TOKEN_PREDICTOR_TOKEN` as a **repository secret** (not as .env file)
-
-4. Your demo is live! ✨   
-
-
-
-## 💰 Cost   **Option B: Using environment variable**1. **Install dependencies:**- **🔮 Real-time Predictions**: Updates automatically as you type
-
-
-
-With HF Pro account ($20/month):   ```bash
-
-- **$2.00/month** in free API calls
-
-- **~$0.0001** per prediction     export HF_NEXT_TOKEN_PREDICTOR_TOKEN="hf_your_token_here"  # Linux/Mac   ```bash
-
-- **2,000-20,000** free predictions/month
-
-   $env:HF_NEXT_TOKEN_PREDICTOR_TOKEN="hf_your_token_here"    # Windows PowerShell
-
-## 🎯 Try It
-
-   ```   pip install -r requirements.txt- **🔒 Secure**: Uses environment variables for API tokens  **For the latest Gradio version**, see `README-gradio.md` and run:
-
-Type "Twinkle, twinkle, little " and watch the AI predict "star"! 
-
-
-
-The app demonstrates how language models work by showing real-time probability distributions for the next token in any sequence.
-
-3. **Run the app:**   ```
-
-## 🔒 Security Notes
-
-   ```bash
-
-- **✅ .env file**: Safe for local development (ignored by git)
-
-- **✅ Environment variables**: Work everywhere   python app.py- **⚡ Fast**: Serverless API calls with instant results```bash
-
-- **✅ HF Spaces secrets**: Secure for production deployment
-
-- **❌ Never commit**: Your actual token to version control   ```
-
-
-
-## 📁 Project Structure2. **Set your HF token:**
-
-
-
-```4. **Open browser:** Navigate to `http://127.0.0.1:7860`
-
-├── app.py              # Main Gradio application
-
-├── requirements.txt    # Python dependencies   - **📱 Responsive**: Clean, mobile-friendly interfacepython app.py
-
-├── .env.example        # Template for environment variables
-
-├── .env               # Your actual tokens (git ignored)## 🔧 Deploy to HF Spaces
-
-├── .gitignore         # Protects secrets from being committed
-
-└── README.md          # This file   **Option A: Using .env file (Recommended for local development)**
-
-```
-
-1. Create a new **Gradio Space** on Hugging Face
-
-## 🤝 Contributing
-
-2. Upload `app.py` and `requirements.txt`   ```bash- **🚀 Easy Deploy**: One-click deployment to HF Spaces```
-
-Feel free to improve the API parsing, add features, or enhance the design!
-3. Set `HF_NEXT_TOKEN_PREDICTOR_TOKEN` as a **repository secret** (not as .env file)
-
-4. Your demo is live! ✨   # Copy the template
-
-
-
-## 💰 Cost   cp .env.example .env
-
-
-
-With HF Pro account ($20/month):   
-
-- **$2.00/month** in free API calls
-
-- **~$0.0001** per prediction     # Edit .env and add your token:## 🚀 Quick Start**For the original HTML version**, see below:
-
-- **2,000-20,000** free predictions/month
-
-   # HF_TOKEN=hf_your_actual_token_here
-
-## 🎯 Try It
-
-   ```
-
-Type "Twinkle, twinkle, little " and watch the AI predict "star"! 
-
-   
-
-The app demonstrates how language models work by showing real-time probability distributions for the next token in any sequence.
-
-   **Option B: Using environment variable**1. **Install dependencies:**---
-
-## 🔒 Security Notes
-
-   ```bash
-
-- **✅ .env file**: Safe for local development (ignored by git)
-
-- **✅ Environment variables**: Work everywhere   export HF_TOKEN="hf_your_token_here"  # Linux/Mac   ```bash
-
-- **✅ HF Spaces secrets**: Secure for production deployment
-
-- **❌ Never commit**: Your actual token to version control   $env:HF_TOKEN="hf_your_token_here"    # Windows PowerShell
-
-
-
-## 📁 Project Structure   ```   pip install -r requirements.txt## Original HTML VersionNext Token Predictor
-
-
-
-```
-
-├── app.py              # Main Gradio application
-
-├── requirements.txt    # Python dependencies3. **Run the app:**   ```emoji: 🔮
-
-├── .env.example        # Template for environment variables
-
-├── .env               # Your actual tokens (git ignored)   ```bash
-
-├── .gitignore         # Protects secrets from being committed
-
-└── README.md          # This file   python app.pycolorFrom: gray
-
-```
-
-   ```
-
-## 🤝 Contributing
-
-2. **Set your HF token:**colorTo: blue
-
-Feel free to improve the API parsing, add features, or enhance the design!
-4. **Open browser:** Navigate to `http://127.0.0.1:7860`
-
-   ```bashsdk: static
-
-## 🔧 Deploy to HF Spaces
-
-   export HF_TOKEN="hf_your_token_here"pinned: false
-
-1. Create a new **Gradio Space** on Hugging Face
-
-2. Upload `app.py` and `requirements.txt`   ```license: mit
-
-3. Set `HF_TOKEN` as a **repository secret** (not as .env file)
-
-4. Your demo is live! ✨short_description: See how AI predicts the next token and explore semantic relationships
-
-
-
-## 💰 Cost3. **Run the app:**---
-
-
-
-With HF Pro account ($20/month):   ```bash
-
-- **$2.00/month** in free API calls
-
-- **~$0.0001** per prediction     python app.py# Next Token Predictor
-
-- **2,000-20,000** free predictions/month
-
-   ```
-
-## 🎯 Try It
-
-Explore how AI language models think! This interactive demo shows how a large language model predicts the next token in a sentence, and visualizes token relationships in a semantic space that mimics neural connections in the brain.
-
-Type "Twinkle, twinkle, little " and watch the AI predict "star"! 
-
-4. **Open browser:** Navigate to `http://127.0.0.1:7860`
-
-The app demonstrates how language models work by showing real-time probability distributions for the next token in any sequence.
-
-## Features
-
-## 🔒 Security Notes
-
-## 🔧 Deploy to HF Spaces
-
-- **✅ .env file**: Safe for local development (ignored by git)
-
-- **✅ Environment variables**: Work everywhere- **Real-time Token Prediction**: Type any text and see the model's top predictions for the next token
-
-- **✅ HF Spaces secrets**: Secure for production deployment
-
-- **❌ Never commit**: Your actual token to version control1. Create a new **Gradio Space** on Hugging Face- **Semantic Neighborhood Map**: Interactive 2D visualization of token embeddings, showing how AI associates words
-
-
-
-## 📁 Project Structure2. Upload `app.py` and `requirements.txt`- **Educational Tool**: Designed to teach non-technical users about AI language processing
-
-
-
-```3. Set `HF_TOKEN` as a **repository secret**- **Browser-Based**: Runs entirely in your browser using Hugging Face Serverless Inference API
-
-├── app.py              # Main Gradio application
-
-├── requirements.txt    # Python dependencies4. Your demo is live! ✨- **Instant Predictions**: No model downloads - predictions happen instantly via API calls
-
-├── .env.example        # Template for environment variables
-
-├── .env               # Your actual tokens (git ignored)
-
-├── .gitignore         # Protects secrets from being committed
-
-└── README.md          # This file## 💰 Cost## How It Works
-
-```
-
-
-
-## 🤝 Contributing
-
-With HF Pro account ($20/month):1. **Token Prediction**: The model analyzes your input and predicts the most likely next words/tokens based on its training data
-
-Feel free to improve the API parsing, add features, or enhance the design!
-- **$2.00/month** in free API calls2. **Semantic Map**: Tokens are positioned in 2D space based on their embeddings. Closer tokens are more semantically similar
-
-- **~$0.0001** per prediction  3. **Neural Connections**: Lines show relationships between tokens, illustrating how AI "thinks" through associations
-
-- **2,000-20,000** free predictions/month
-
-## Usage
-
-## 🎯 Try It
-
-- Enter your Hugging Face API token (Pro account required) in the token field
-
-Type "Twinkle, twinkle, little " and watch the AI predict "star"! - Type a sentence in the input box
-
-- View the top predicted tokens on the right
-
-The app demonstrates how language models work by showing real-time probability distributions for the next token in any sequence.- Click on predictions to append them to your text
-
-- Hover over tokens in the list to see them highlighted in the semantic map
-
-## 📁 Project Structure- Use the "Run Twinkle demo" to see a classic example
-
-
-
-```## Technical Details
-
-├── app.py              # Main Gradio application
-
-├── requirements.txt    # Python dependencies- **Model**: Qwen3-0.6B via Hugging Face Serverless Inference API
-
-└── README.md          # This file- **API**: Hugging Face Inference API (requires Pro account token)
-
-```- **Performance**: Instant predictions with no model downloads
-
-- **Embeddings**: Pre-computed PCA projections of Qwen token embeddings for visualization
-
-## 🤝 Contributing
-
-## Deployment on Hugging Face Spaces
-
-Feel free to improve the API parsing, add features, or enhance the design!
-This project is designed to run as a static space on Hugging Face:
-
-1. Create a new Space with "Static" SDK
-2. Upload all files from this repository
-3. The `index.html` serves as the main page
-4. Ensure `assets/` folder is included for embeddings and vendor libraries
-5. The model is fetched directly from the Hugging Face Hub (no local model files needed)
-
-## Educational Goals
-
-This tool helps users understand:
-- How AI processes language at the token level
-- The concept of embeddings and semantic similarity
-- How predictions are made based on statistical patterns
-- The "neural network" metaphor through visual connections
-
-Enjoy exploring AI's "mind"!
+---

requirements.txt

@@ -1,3 +1,3 @@
-gradio==4.44.0
+gradio==4.44.1
 requests==2.31.0
 python-dotenv==1.0.0