@TurrinaMartino@ivanfioravanti With oMLX, you can enjoy testing all possible forms of quantization, TurboCache, MTP, and dQuant. Additionally, you can generate oQ models and run a wide range of benchmark tests. It's the Swiss Army knife for MLX.
@ivanfioravanti LM Studio and GGUF are ready for production use. oMLX is a young product gaining traction, and within a few months it will become a benchmark on Mac for MLX π€
@silupescu@silupescu Corpus Domini di Campobasso?? πβ€οΈπ posso chiederti come fai a conoscere questa festa? Io sono di lΓ¬ e non credo lo sappiano in molti. Un π
Today I restarted my Mac G4 server that has been online 24/7 for about 24 years. Once a year.
Specs:β’ 450 MHz CPUβ’ 1 GB SDRAM .. And it still works. Once were warriors. #RetroComputing#Apple
@ivanfioravanti Currently, my two local AI productsβeditorial software with vectorized data stored on Qdrantβare performing great on oMLX, apart from some minor caching issues when handling heavy loads. But I'm looking forward to future developments/updates for LMStudio.
Ollama is now updated to run the fastest on Apple silicon, powered by MLX, Apple's machine learning framework.
This change unlocks much faster performance to accelerate demanding work on macOS:
- Personal assistants like OpenClaw
- Coding agents like Claude Code, OpenCode, or Codex