Principal-timestep Restricted Init via Sparse Matrix-decomposition in Flow-matching
Flow-matching diffusion models have recently emerged as a strong paradigm for high-fidelity visual generation. However, their prohibitively high fine-tuning cost limits scalability to downstream tasks. While Low-Rank Adaptation (LoRA) combined with spectral initialization has demonstrated accelerated convergence and im...
Jian-Yang Gu, Zheng Fang, Li-Chuan Xiang et al.
· 0 citations