VLA Foundry (Foundry-VLA-1.7B)
Toyota Research InstituteFoundry-VLM-1.3B vision-language backbone + (condition on last 4 layers) flow-matching diffusion action head (24 layers, 1024 hidden dim, 16 heads)
specific method · filed under model architecture
Uses a flow-matching prediction head for action generation; the supplied evidence does not settle whether the diffusion qualifier denotes a distinct mechanism.
The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See prediction head for the methods filed beside it.
Also called Flow-matching diffusion action head, Flow-matching prediction head, flow matching head.
3 quotes from 3 sources, one card per model, each opening with its strongest.
Foundry-VLM-1.3B vision-language backbone + (condition on last 4 layers) flow-matching diffusion action head (24 layers, 1024 hidden dim, 16 heads)
This model currently supports only the flow-matching action head for π₀.₅ training and inference.
we currently only support the flow matching head for both π₀.₅ training and inference