Skip to content

[Feature]: Alternating local-global attention layers #9464

Closed as not planned
Closed as not planned
@griff4692

Description

@griff4692

🚀 The feature, motivation and pitch

Gemma-2 and new Ministral models use alternating sliding window and full attention layers to reduce the size of the KV cache.

The KV cache is a huge inference bottleneck and this technique could be fine-tuned into other models to make them much more memory efficient, especially for large batch sizes.

Alternatives

No response

Additional context

No response

Before submitting a new issue...

  • Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

Metadata

Metadata

Assignees

No one assigned

    Labels

    feature requestNew feature or requeststaleOver 90 days of inactivity

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions