REVIEW 2 cited by
ToSA: Token Merging with Spatial Awareness
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Token merging has emerged as an effective strategy to accelerate Vision Transformers (ViT) by reducing computational costs. However, existing methods primarily rely on the visual token's feature similarity for token merging, overlooking the potential of integrating spatial information, which can serve as a reliable criterion for token merging in the early layers of ViT, where the visual tokens only possess weak visual information. In this paper, we propose ToSA, a novel token merging method that combines both semantic and spatial awareness to guide the token merging process. ToSA leverages the depth image as input to generate pseudo spatial tokens, which serve as auxiliary spatial information for the visual token merging process. With the introduced spatial awareness, ToSA achieves a more informed merging strategy that better preserves critical scene structure. Experimental results demonstrate that ToSA outperforms previous token merging methods across multiple benchmarks on visual and embodied question answering while largely reducing the runtime of the ViT, making it an efficient solution for ViT acceleration. The code will be available at: https://github.com/hsiangwei0903/ToSA
Forward citations
Cited by 2 Pith papers
-
WaveZip: Wavelet-Driven Space-Time Decoupling for Video Token Condensation
WaveZip, a training-free wavelet method for video token condensation, reports about 10x token reduction while retaining roughly 98-99% accuracy on multi-benchmark LVLM evaluation.
-
WaveZip: Wavelet-Driven Space-Time Decoupling for Video Token Condensation
Wavelet-based temporal and spatial token condensation preserves 99.6% of full-token video question-answering accuracy at 10x compression without retraining.
Discussion (0). Sign in to comment.