SJTU-IPADS / PowerInfer

High-speed Large Language Model Serving on PCs with Consumer-grade GPUs
MIT License
7.96k stars 412 forks source link

Meta: Implementing hybrid inference across key desktop platforms #92

Open hodlen opened 10 months ago

hodlen commented 10 months ago

As we embark on the initial phase of PowerInfer's development, our primary goal is to introduce the hybrid inference feature across all major desktop hardware and software platforms. Our current progress includes:

This milestone marks a significant step in our journey 🚀. We are diligently working to achieve this goal within the coming weeks. Stay connected for more updates and feel free to join us in this exciting endeavor by contributing code! 🤝💻