Skip to main content

Mamba

Mamba is a new state space model architecture showing promising performance on information-dense data such as language modeling, where previous subquadratic models fall short of Transformers.

This backend:

  • provides support for Mamba models
  • requires CUDA runtime
note

This is an experimental backend and it may change in the future.

Example

warning

Please make sure to change syntax to #syntax=ghcr.io/sozercan/aikit:latest in the examples below.

https://github.com/sozercan/aikit/blob/main/test/aikitfile-mamba.yaml