generative performance

3 papers

• Adaptive Discretization for Consistency Models
• DiCo: Revitalizing ConvNets for Scalable and Efficient Diffusion Modeling
• SparseDiT: Token Sparsification for Efficient Diffusion Transformer