AI RESEARCH

TMPDiff: Temporal Mixed-Precision for Diffusion Models

arXiv CS.LG

ArXi:2603.14062v1 Announce Type: cross Diffusion models are the go-to method for Text-to-Image generation, but their iterative denoising processes has high inference latency. Quantization reduces compute time by using lower bitwidths, but applies a fixed precision across all denoising timesteps, leaving an entire optimization axis unexplored. We propose TMPDiff, a temporal mixed-precision framework for diffusion models that assigns different numeric precision to different denoising timesteps.