Note a cascaded diffusion model for 4D novel view synthesis conditioned on one or more images of a general scene, and a set of camera poses and timestamps joint training on 3D (with camera pose), 4D (pose+time) and video (time but no pose) data

Note encoding a complex, potentially multimodal data distribution into a single continuous Gaussian distribution arguably represents an unnecessarily challenging learning problem. 【问题都没看懂】

Note [2R] This paper presents Diffusion Forcing, a new training paradigm where a diffusion model is trained to denoise a set of tokens with independent per-token noise levels.

Note 【2R】 Classifier-free guidance (CFG) has become the standard method for enhancing the quality of conditional diffusion models. However, employing CFG requires either training an unconditional model alongside the main diffusion model or modifying the training procedure by periodically inserting a null condition. A new method, independent condition guidance (ICG), which provides the benefits of CFG without the need for any special training procedures.

Note 【2R】 utilizing the CLIP image embedding space for more visually-oriented tasks through methods such as IP-Adapter.