https://github.com/guntas-13/nextcharacterpredictor

An MLP based next character prediction model trained on several famous corpora with varying embedding dimensions and model architecture size
https://github.com/guntas-13/nextcharacterpredictor

large-language-models mlp-networks token-generator

Last synced: about 2 months ago
JSON representation

An MLP based next character prediction model trained on several famous corpora with varying embedding dimensions and model architecture size

• Host: GitHub
• URL: https://github.com/guntas-13/nextcharacterpredictor
• Owner: guntas-13
• Created: 2024-03-12T22:18:56.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-05-05T18:05:49.000Z (about 1 year ago)
• Last Synced: 2025-01-30T17:39:08.636Z (4 months ago)
• Topics: large-language-models, mlp-networks, token-generator
• Language: Python
• Homepage: https://tensionflowcharpredict.streamlit.app/
• Size: 58.6 MB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Next Character Predictor

## Find the app hosted at [Link](https://tensionflowcharpredict.streamlit.app/)

**Datasets: [Link](https://cs.stanford.edu/people/karpathy/char-rnn/)**

## **App Homepage**

![](https://github.com/guntas-13/NextCharacterPredictor/blob/main/1.png)

## **Generated Text for Gulliver's Travels**

![](https://github.com/guntas-13/NextCharacterPredictor/blob/main/2.png)

## **Generated Text for Wikipedia Corpus**

![](https://github.com/guntas-13/NextCharacterPredictor/blob/main/3.png)

### **Embedding the characters as a Vector $\in \mathbb{R}^N$**

![](https://github.com/guntas-13/NextCharacterPredictor/blob/main/Embed.svg)

### **Input layer of the model and creating the training examples**

![](https://github.com/guntas-13/NextCharacterPredictor/blob/main/MLPToken.svg)

To make inference from the model trained on ```tolstoy.txt``` run the script:

```bash

python3 predictT.py

```

To make inference from the model trained on ```wonderland.txt``` run the script:

```bash

python3 predictW.py

```

To run the StreamLit App, run the script:

```bash

streamlit run app.py

```