SpecKCast: Spectral Kalman Forecasting for Diffusion Acceleration
扩散模型推理加速算法，论文阶段未放权重
Diffusion Transformers (DiTs) achieve high-fidelity image and video generation, but they suffer from substantial inference costs due to repeated transformer evaluations across many denoising steps. Feature caching reduces this cost by reusing features across denoising steps. Rece
判断口径 agnes-3.0-flash · 入库 #56 · 回首页