Search papers, labs, and topics across Lattice.
TokenDial introduces a method for continuous attribute control in text-to-video generation by applying learned, attribute-specific offsets in the spatiotemporal visual patch-token space of a pretrained model. These offsets, trained using semantic direction matching and motion-magnitude scaling, allow for fine-grained adjustments to appearance and motion dynamics without retraining the entire backbone. The approach achieves superior controllability and edit quality compared to existing methods, as validated through quantitative metrics and human evaluations.
Unlock slider-style control over video attributes like motion magnitude and effect intensity without retraining your text-to-video model – just nudge the spatiotemporal tokens.
We present TokenDial, a framework for continuous, slider-style attribute control in pretrained text-to-video generation models. While modern generators produce strong holistic videos, they offer limited control over how much an attribute changes (e.g., effect intensity or motion magnitude) without drifting identity, background, or temporal coherence. TokenDial is built on the observation: additive offsets in the intermediate spatiotemporal visual patch-token space form a semantic control direction, where adjusting the offset magnitude yields coherent, predictable edits for both appearance and motion dynamics. We learn attribute-specific token offsets without retraining the backbone, using pretrained understanding signals: semantic direction matching for appearance and motion-magnitude scaling for motion. We demonstrate TokenDial's effectiveness on diverse attributes and prompts, achieving stronger controllability and higher-quality edits than state-of-the-art baselines, supported by extensive quantitative evaluation and human studies.
