class-conditional generation
A technique in generative modeling where the generation process is conditioned on specific class labels, allowing for the production of samples that conform to desired categories, enhancing control over generated outputs.
- Cross-fluctuation phase transitions reveal sampling dynamics in diffusion models
- DiCo: Revitalizing ConvNets for Scalable and Efficient Diffusion Modeling
- Diffusion Tree Sampling: Scalable inference‑time alignment of diffusion models
- Entropy Rectifying Guidance for Diffusion and Flow Models
- Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials
- Plug-and-Play Context Feature Reuse for Efficient Masked Generation
- STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis
- Sparse Image Synthesis via Joint Latent and RoI Flow