# MatFormer architecture

> Matryoshka Transformer. Matryoshka Representation Learning 개념을 Transformer architecture 전체에 적용하는 방식.

Matryoshka [Transformer](https://wiki.g15e.com/pages/Transformer%20architecture.txt). [Matryoshka Representation Learning](https://wiki.g15e.com/pages/Matryoshka%20Representation%20Learning.txt) 개념을 [Transformer architecture](https://wiki.g15e.com/pages/Transformer%20architecture.txt) 전체에 적용하는 방식.

## Papers

- [MatFormer: Nested Transformer for Elastic Inference](https://wiki.g15e.com/pages/MatFormer%20-%20Nested%20Transformer%20for%20Elastic%20Inference.txt)

## Models

- [Gemma 3n](https://wiki.g15e.com/pages/Gemma%203n.txt)