A temporal-spatial LSB mask over one shared weight buffer lets diffusion models use lower bit precision in less sensitive denoising stages, cutting compute by 25-50% on bit-serial hardware with no loss in image quality.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
TASQ: Temporal-Adaptive Bit Sparsification Quantization for Diffusion Models
A temporal-spatial LSB mask over one shared weight buffer lets diffusion models use lower bit precision in less sensitive denoising stages, cutting compute by 25-50% on bit-serial hardware with no loss in image quality.