Connect home devices into a powerful cluster to accelerate LLM inference. More devices mean faster performance, leveraging HPipe parallelism and high-speed synchronization over Ethernet.
Our team(The gentlemen approach) is developing this project to realize HPipe and compare with other method.
Supports Linux, macOS, and Windows.
Special Thanks to original project developers, prof. LIM.
How to Run
- Sorry but this project is not finished so we will decribe it later.
News
- 27 Dec 2025 - Started to approach gentle.
- 28 Dec 2025 - Finished developing H-Pipe network (root, worker etc.)
This project is released under the MIT license.