Papers
arxiv:2611.23226

FastFrames: Video Generation at Interactive Latency

Published on Nov 17
· Submitted by
Gene Chou
on Nov 18
Authors:
,
,
,
,
,
,

Abstract

Autoregressive video diffusion models generate frames based on new observations and inputs, making them a promising medium for interactive visual experiences. However, existing models are orders of magnitude too slow for fluid interaction. This paper introduces FastFrames, a family of fast autoregressive video diffusion models that achieves interactive latency without compromising quality. We profile the inference pipeline of state-of-the-art models and find that operations often treated as fixed implementation details, such as VAE decoding and KV cache updates, in fact account for over half of end-to-end inference latency. We challenge the assumption that these components are indispensable and replace them with lightweight alternatives. Our 1.3B variant runs at 36 ms per frame, or 28 responses per second (RPS), on a single B200 GPU while matching the generation quality of a state-of-the-art model that is nearly 8x slower. Our 14B variant runs at 45 ms per frame, or 22 RPS, on an 8x B200 GPU node and yields significantly better visual quality and realism. We show that FastFrames enables real-time interaction in applications such as camera-conditioned and physics-conditioned video generation.

Community

See our project page for videos, code, and model weights: https://apple-aiml-research.github.io/ml-fastframes/.

Let him come, take it off her, and embrace her from behind.
61D6D65C-03E3-4D7C-9CD4-BE33682EB920

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2611.23226
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 2

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.23226 in a dataset README.md to link it from this page.

Spaces citing this paper 4

Collections including this paper 1