https://github.com/voyager466920/analog_gpt
💫 AnalogGPT is a Large Language Model that runs fully on the paper. No GPU, no computer required. Just a pen, paper and patience required. I added a Youtube to demonstrate how it works!
https://github.com/voyager466920/analog_gpt
ai analog analoggpt artificial gpt inteligence paper transformer-architecture
Last synced: 10 months ago
JSON representation
💫 AnalogGPT is a Large Language Model that runs fully on the paper. No GPU, no computer required. Just a pen, paper and patience required. I added a Youtube to demonstrate how it works!
- Host: GitHub
- URL: https://github.com/voyager466920/analog_gpt
- Owner: Voyager466920
- Created: 2025-06-13T05:30:01.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-06-26T12:50:19.000Z (about 1 year ago)
- Last Synced: 2025-07-01T04:29:45.003Z (about 1 year ago)
- Topics: ai, analog, analoggpt, artificial, gpt, inteligence, paper, transformer-architecture
- Language: Python
- Homepage: https://www.youtube.com/watch?v=6gqxgyYj3Hs&feature=youtu.be
- Size: 10.6 MB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
[![Contributors][contributors-shield]][contributors-url]
[![Forks][forks-shield]][forks-url]
[![Stargazers][stars-shield]][stars-url]
[![Issues][issues-shield]][issues-url]
AnalogGPT
AnalogGPT is a Large Language Model that runs fully on the paper.
Explore the docs »
View Demo
·
Report Bug
·
Request Feature
This is the Youtube Video about AnalogGPT!
Table of Contents
## About The Project
🚨 To download the PDF file, please hit a star on this repo, then click `AnalogGPT.pdf`, press `download raw file` button on the upper right corner.
🥫 These days, a lot of companies and even countries are trying to get their hands on GPUs. Modern AI needs a ton of GPU power for calculation. But imagine this — what if there’s a large language model that doesn’t need a GPU at all? Or even run *without* a computer?
I created **AnalogGPT** which is a fully paper-based version of GPT built from scratch.
🚀 Via **AnalogGPT** you can fully understand how Generative Pretrained Transformer works by calculating with your bare hands. You are becoming a GPU that calculates these numbers. As the model size is 15K, it's representation capacity is limited. I'll notify the hyper parameters below. Feel free to modify my code and test various things.
## Key Features
🖥️ **[No GPU]** You don't need a GPU or even computer to run this model.
💡 **[Learn how Transformer works]** With calculating with your bare hands, you can understand what attention is.
🔧 **[Next Token Prediction]** This is dedicated for next token prediction task.
### Built With
* [![Pytorch][Pytorch]][Pytorch-url]
* [![Pycharm][Pycharm]][Pycharm-url]
### Prerequisites
💫 Before using, please hit star on this repo!
📜 You can download the PDF from github
If you want to modify the code or train more, follow the following steps.
### Installation
1. Clone the repo
```sh
git clone https://github.com/DutchVandaline/Analog_GPT.git
```
2. Excecute on PyCharm
3. Use your code to modify the model.
## Contributing
If you liked this project, please buy me a coffee! It helps me do creative stuffs.
Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are **greatly appreciated**.
If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement".
Don't forget to give the project a star! Thanks again!
1. Fork the Project
2. Create your Feature Branch (`git checkout -b feature/AmazingFeature`)
3. Commit your Changes (`git commit -m 'Add some AmazingFeature'`)
4. Push to the Branch (`git push origin feature/AmazingFeature`)
5. Open a Pull Request
## Contact
Dutch Vanaline - developerha0013@gmail.com
Project Link: [https://github.com/DutchVandaline?tab=repositories](https://github.com/DutchVandaline?tab=repositories)
[contributors-shield]: https://img.shields.io/github/contributors/DutchVandaline/Analog_GPT.svg?style=for-the-badge
[contributors-url]: https://github.com/DutchVandaline/Wait_However/graphs/contributors
[forks-shield]: https://img.shields.io/github/forks/DutchVandaline/Analog_GPT.svg?style=for-the-badge
[forks-url]: https://github.com/DutchVandaline/Wait_However/network/members
[stars-shield]: https://img.shields.io/github/stars/DutchVandaline/Analog_GPT.svg?style=for-the-badge
[stars-url]: https://github.com/DutchVandaline/Wait_However/stargazers
[issues-shield]: https://img.shields.io/github/issues/DutchVandaline/Analog_GPT.svg?style=for-the-badge
[issues-url]: https://github.com/DutchVandaline/Wait_However/issues
[Pytorch]: https://img.shields.io/badge/PyTorch-EE4C2C?style=for-the-badge&logo=pytorch&logoColor=white
[Pytorch-url]: https://pytorch.org/
[Pycharm]: https://img.shields.io/badge/PyCharm-000000?style=for-the-badge&logo=pycharm&logoColor=white
[Pycharm-url]: https://www.jetbrains.com/pycharm/