Learning Energy-Based Generative Models via Coarse-to-Fine Expanding and Sampling

By Yang Zhao, Jianwen Xie, Ping Li. 2021

52 citations recorded by the source corpus.

Citations in the corpus

• COLD Decoding: Energy-based Constrained Text Generation with Langevin Dynamics (2022, unknown): “Instead, Langevin dynamics is most widely used on EBMs of other modalities with continuous values, such as images (Song & Ermon, 2019; Du & Mordatch, 2019; Zhao et al., 2021), 3D shapes (Xie et al.”