Summary

MosaiChunk is a memory system for long-horizon video generation that preserves previously established visual details. A lightweight learned router selects historical key-value entries across space and time for a frozen video generator under a fixed active-memory budget.

Key Features

Spatiotemporal KV selection.
Historical detail retrieval.
Frozen generator backbone.
Lightweight trainable router.
Fixed active-memory budget.
Text-to-video experiments.
Image-to-video experiments.
Consistency across scene revisits.

About MosaiChunk


Video-model developers can explore more consistent object and scene reappearances without retraining the entire generator. Core capabilities include Spatiotemporal KV selection, Historical detail retrieval, Frozen generator backbone, and Lightweight trainable router.


Public code or model weights are linked from the project, with setup instructions for running or adapting the system. Additional characteristics include Fixed active-memory budget, Text-to-video experiments, Image-to-video experiments, and Consistency across scene revisits.

Get more likes & reach the top of search results by adding this button on your site!

Embed button preview - Light theme
Embed button preview - Dark theme
TurboType Banner

Subscribe to the AI Search Newsletter

Get top updates in AI to your inbox every weekend. It's free!