Pith. sign in

REVIEW 1 cited by

Vision Transformer with Attention Map Hallucination and FFN Compaction

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

arxiv 2306.10875 v1 pith:HCDDGH7T submitted 2023-06-19 cs.CV

Vision Transformer with Attention Map Hallucination and FFN Compaction

classification cs.CV
keywords attentionvisioncompactionhallucinationmapsmhsaoperationspropose
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
abstract

Vision Transformer(ViT) is now dominating many vision tasks. The drawback of quadratic complexity of its token-wise multi-head self-attention (MHSA), is extensively addressed via either token sparsification or dimension reduction (in spatial or channel). However, the therein redundancy of MHSA is usually overlooked and so is the feed-forward network (FFN). To this end, we propose attention map hallucination and FFN compaction to fill in the blank. Specifically, we observe similar attention maps exist in vanilla ViT and propose to hallucinate half of the attention maps from the rest with much cheaper operations, which is called hallucinated-MHSA (hMHSA). As for FFN, we factorize its hidden-to-output projection matrix and leverage the re-parameterization technique to strengthen its capability, making it compact-FFN (cFFN). With our proposed modules, a 10$\%$-20$\%$ reduction of floating point operations (FLOPs) and parameters (Params) is achieved for various ViT-based backbones, including straight (DeiT), hybrid (NextViT) and hierarchical (PVT) structures, meanwhile, the performances are quite competitive.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Explainable AI in Speaker Recognition -- Attention Map Visualisation and Evaluation

    eess.AS 2026-06 unverdicted novelty 5.0

    The paper introduces Modified RISE-eval to evaluate GradCAM and LayerCAM attention maps on speaker recognition networks and reports distinct advantages for each method under different conditions.