📦 Ollama – v0.11.5
🎉 Hey AI Homebrew Enthusiasts! 🎉
Ollama version 0.11.5-rc2 is here with some fantastic improvements! Here’s what’s new:
- Faster Models: Performance boost for gpt-oss models! 🚀
- GPU Power: Better multi-GPU support and reduced VRAM usage. 💪
- Bug Fixes: Issues with Harmony tool calls & OpenAI API reasoning effort are now solved! ✅
- CPU Boost: `OLLAMA_FLASH_ATTENTION=1` now enables flash attention even on CPU! 🧠
- Smaller Footprint: Installation size reduced for Windows and Linux users. 💾
- Improved Memory Management: Thanks to @jessegross!
Plus, a big welcome to our new contributors: @vorburger, @dan-and, and @youzichuan! 🙌
You can find the full changelog here: https://github.com/jmorganca/ollama/compare/v0.11.4…v0.11.5-rc2
Happy Ollama-ing! 🦙✨