Skip to content

Speculative execution during user speech? #479

Description

@bloodraven66

Any thoughts on supporting speculative generation while user is still active? goal is to reduce response latency as low as possible. This is possible and I've explored it in my recent paper - https://arxiv.org/abs/2606.13450.

This would rely on another turn-taking signal to continuously predict an upcoming end-of-turn. This would lead to much lower response latency than the current speculative framework. The audio will be played only when an actual end-of-turn is predicted.

open-source options for end-of-turn anticipation signal -

  1. Voice activity projection (https://github.com/inokoj/VAP-Realtime)
  2. Endpoint anticipation (https://huggingface.co/viks66/endpoint-anticipation)
  3. Kyutai/DSM VAD head 0/1 can signal an upcoming end of turn.

Happy to add a PR for it if there is interest!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions