GPUStack 0.1.1 was released. This release brings support for embedding models / API and completions API. As a bit of background, OpenAI's API has emerged as the de facto standard for model APIs, particularly amongst open-source models. In the previous version, we introduced support for general text generation models and provided an OpenAI-compatible chat completions API. To support more scenarios, such as RAG and certain frameworks, this version includes support for embedding models / API and completions API.
For more details about GPUStack, visit:
GitHub repo: https://github.com/gpustack/gpustack
User guide: https://docs.gpustack.ai
What's New in GPUStack 0.1.1
Embedding Models / API and Completions API Support
Provide support and offer code examples for usage:
Docker and Docker Compose Support
We’ve introduced support for running GPUStack using Docker and Docker Compose on Ubuntu systems with NVIDIA GPUs. This feature streamlines the deployment process, making it easier to set up GPUStack and manage GPU resources.
PyPI Mirror Support for Faster Installation
Installation just got quicker! You can now specify a PyPI mirror during installation to speed up package downloads. This is particularly useful for users with limited bandwidth or those behind restrictive firewalls.
Upgrade Support Added
We’ve added support for upgrading GPUStack installations. This feature ensures that keeping your setup up-to-date is seamless and straightforward without reconfiguring.
Bug Fixes in GPUStack 0.1.1
Fix for 500 Error in Absence of Worker
Considering the possibility of separating the server and worker, the 500 error that occurred in the worker's absence has now been fixed.
Operating System Compatibility Fixes
We’ve addressed compatibility issues with certain operating systems, ensuring broader support and smoother operation across diverse setups.
Installation Script Fixes
Various issues related to the installation script have been resolved, improving the overall installation experience and reducing potential errors during setup.
Data Collection Fixes
We’ve also addressed issues with data collection in fastfetch, ensuring that your system information is reported accurately and reliably.
Join Our Community
You can see the release's full changelog here. As always, we'd love to hear your feedback via Discord or GitHub. Your feedback is invaluable in helping us make GPUStack better for everyone, and we welcome your contributions to the project.
