• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI

Husky is said to run one model on Mac and iPhone without Wi-Fi

A post describes Husky as a model-specific inference engine. A follow-up says its Metal kernels are tailored to the Woof model, with the goal of beating Apple's MLX.

5 Sources, 20d ago, first seen 20d ago

TLDR

The posts describe Husky as a runtime for one model, rather than many, and say it runs on Mac and iPhone even without Wi-Fi. The author says they wrote model-shaped Metal “megakernels” to beat Apple's MLX. Because Husky knows Woof's shapes ahead of time, they say each matrix uses one kernel to combine normalization, 4-bit unpacking and multiplication in a single pass.

Combined views

—

5 Sources, first seen 20d ago

— likes— comments— saves— reposts

Combined views

—

5 Sources, first seen 20d ago

— likes— comments— saves— reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

5 Sources

Sigil Wen@0xSigilTo beat Apple's MLX we wrote model-shaped Metal Megakernels. MLX is general, so each layer is a chain of small GPU calls with the CPU in between. Husky knows Woof's shapes ahead of time, so each matrix gets one kernel: norm, 4-bit unpack and multiply in one pass. The GPU never waits.20d
Elliot Arledge@elliotarledgeyour macbook is now basically a cerebras wafer due to better kernels and scheduling19d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    5 Sources

    Sigil Wen@0xSigilTo beat Apple's MLX we wrote model-shaped Metal Megakernels. MLX is general, so each layer is a chain of small GPU calls with the CPU in between. Husky knows Woof's shapes ahead of time, so each matrix gets one kernel: norm, 4-bit unpack and multiply in one pass. The GPU never waits.20d
    Elliot Arledge@elliotarledgeyour macbook is now basically a cerebras wafer due to better kernels and scheduling19d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet