SSH fine-tunes large models by learning sparse Hartley-spectrum coefficients selected by energy of the pretrained weights, matching or beating LoRA and FourierFT with fewer parameters.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SSH: Sparse Spectrum Adaptation via Discrete Hartley Transformation
SSH fine-tunes large models by learning sparse Hartley-spectrum coefficients selected by energy of the pretrained weights, matching or beating LoRA and FourierFT with fewer parameters.