TabbyML Newsletter #3: GitHub 10K 🌟, M1/M2 metal inference, Tabby in HuggingFace Spaces 🤗, and more

👋 Hello Tabby lovers,

 

Welcome to another issue of TabbyML Newsletter! This time we have some super exciting product features and community milestones to share. Here we go 🥳

 

After over 6-months of active development, Tabby has just reached over 🌟10K stars🌟 on GitHub! (11K as of Oct 9) 🚀🚀🚀 A massive thank you to every community member for your invaluable contributions and for making TabbyML a thriving open-source ecosystem. Cheers to many more achievements ahead! 🍻🥂

 

     https://star-history.com/#TabbyML/tabby&Date

 

🍎 Metal inference on Apple M1/M2 locally for StarCoder model series, with contribution of Tabby team to the amazing llama.cpp project (release notes)

✍️  Secret sauce to ensure Tabby's swift performance with code completion cancellation - Stream Laziness in Tabby

🆕 Documentation redesign 🎈

🔬Model directory release, with the latest integration of WizardCoder 🔮

❇️ Alternative model registry support in modelscope.cn to speed up model download for developers from mainland China regions

 

🤗 HuggingFace Space just integrated Tabby in its official doc! Come and deploy your first Tabby instance on GPU via HuggingFace Space!

🎹 AlphaSignal posed a great question about whether Tabby is the opensource GitHub copilot. We have a definitive answer in mind. What do you think? 💬

📺 Amazing video from Tabby community member Fahd Mirza about a step-by-step tutorial to install Tabby locally!

📌 Tabby is mentioned in Top 5 Trending Developer Tools: Discover the Latest in Web Development.

⚙️ Tabby is also highlighted as one of the five Awesome Rust Projects.

 

 

Speaking of Rust... We have a new Rust Engineer job opening, in addition to our eager search for Full-Stack and Machine Learning engineers to roll with us! Check out TabbyML Careers to see more details (🤐spoiler: 100% remote 🏖, open source ☕️, and transparent compensation💰).