Open Models Almanac
Methodsmodel architecture

specific method · filed under model architecture

Flow-matching action expert

A flow-matching action expert is a transformer-based action expert that generates action chunks from VLM features.1 It is trained as a conditional flow-matching transformer that outputs the vector field from the VLM features and noisy actions.2

Trade-off Interleaving cross and self-attention differs from prior VLA architectures that rely exclusively on one attention type.3

How this definition was written

Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.

  1. The action expert is trained to predict an action chunk from VLM features — SmolVLA: A vision-language-action model for affordable and efficient robotics
  2. is trained to output the vector field from the VLM features and the noisy actions — SmolVLA: A vision-language-action model for affordable and efficient robotics
  3. Differently from prior VLA architectures, we interleave cross and self-attention layers — SmolVLA: A vision-language-action model for affordable and efficient robotics

Also called Flow-matching continuous action expert.

sources
3
models
2
labs adopt it
2
strongest
core

Documented in

Allen Institute for AI
Hugging Face

Evidence

3 quotes from 3 sources, one card per model, each opening with its strongest.

How sources treat it
core 3
↑