AI
FrameMorrow: Future-guided Frame Selection with Prospective Tokens for Long-Horizon Video Generation
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 29/09/2026, 20:00:00 UTC
Briefing created: 05/10/2026, 03:24:10 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 05/10/2026, 03:18:24 UTC.
Briefings
FrameMorrow is a prospective frame selector that predicts future information needs to guide historical frame selection in long-horizon video generation. The system identifies relevant frames from history without generating the full future, enabling plug-and-play integration across diverse generative models. Experiments show consistent improvements in long-range consistency, visual quality, and action alignment across five benchmarks and eleven models.
Key points
- FrameMorrow predicts a small set of prospective tokens representing future information needs to guide historical frame selection.
- The method selects explicit historical frames rather than model-specific internal states, allowing integration with closed-source models.
Context and limitations
Evaluation spans five benchmarks and eleven generative models across long-video, interactive, and action-conditioned settings. Results are from experiments conducted at the time of publication.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 05/10/2026, 03:24:10 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeidvknnp4vbmitxav2h3ssr2r4hcaucmqwhjuuoguflac62npsdej4