PassionSR quantizes one-step diffusion super-resolution models to 6-8 bits via learnable quantizer boundaries, equivalent transformations, and a two-stage calibration, while keeping quality close to full precision.
Dual aggregation transformer for image super-resolution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution
PassionSR quantizes one-step diffusion super-resolution models to 6-8 bits via learnable quantizer boundaries, equivalent transformations, and a two-stage calibration, while keeping quality close to full precision.