kyegomez/zeta
A modular PyTorch framework providing reusable high-performance components for building AI models.
Not currently ranked — collecting fresh signals.
star history
Zeta is a library of modular building blocks for AI model development in PyTorch. It provides carefully optimized components including attention mechanisms (multi-query attention, flash attention, sigmoid attention), mixture of experts routing, feedforward networks, normalization layers, and other transformer architecture components. Developers can assemble state-of-the-art models by combining these pre-built, tested modules rather than implementing them from scratch.
Frequently asked
- What is kyegomez/zeta?
- A modular PyTorch framework providing reusable high-performance components for building AI models.
- Is zeta open source?
- Yes — kyegomez/zeta is open source, released under the Apache-2.0 license.
- What language is zeta written in?
- kyegomez/zeta is primarily written in Python.
- How popular is zeta?
- kyegomez/zeta has 597 stars on GitHub.
- Where can I find zeta?
- kyegomez/zeta is on GitHub at https://github.com/kyegomez/zeta.