Newest papers

Training-free Acceleration for Blind Compressed Image Restoration via Activation-Calibrated Singular Value Decomposition

Suhang Yao, Xin Li, Xijun Wang, Yufei Wang, Yeying Jin, Zhibo Chen · 2026 · arXiv

DOI: 10.48550/arXiv.2610.29959 · open-access PDF

Portuguese

Modelos baseados em transformadores relatam um desempenho forte na restauração de imagem compressada cega (CIR) sob configurações de compressão desconhecidas, mas seus operadores de projeção densamente empilhados fazem a implantação caro. Nós estudamos a aceleração sem treinamento para tais modelos através de uma análise empírica, nível de operador. Especificamente, nós perfilamos onde a computação é concentrada, caracterizamos distribuições de ativação de modo de canal sob entradas comprimidas e examinamos a decomposição de valor único de pesos de projeção representativos. Estas medições identificam as projeções recorrentes na atenção, feed-forward, c

Machine translation (Meta M2M100); the English original follows.

Abstract

Transformer-based models achieve strong performance on blind compressed image restoration (CIR) under unknown compression settings, but their densely stacked projection operators make deployment costly. We study training-free acceleration for such models through an empirical, operator-level analysis. Specifically, we profile where computation is concentrated, characterize channel-wise activation distributions under compressed inputs, and examine the singular-value decay of representative projection weights. These measurements identify the recurring projections in the attention, feed-forward, c

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.29959, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange