Abstract
Classifier-free guidance (CFG) has become the standard method for enhancingthe quality of conditional diffusion models. However, employing CFG requireseither training an unconditional model alongside the main diffusion model ormodifying the training procedure by periodically inserting a null condition.There is also no clear extension of CFG to unconditional models. In this paper,we revisit the core principles of CFG and introduce a new method, independentcondition guidance (ICG), which provides the benefits of CFG without the needfor any special training procedures. Our approach streamlines the trainingprocess of conditional diffusion models and can also be applied duringinference on any pre-trained conditional model. Additionally, by leveraging thetime-step information encoded in all diffusion networks, we propose anextension of CFG, called time-step guidance (TSG), which can be applied to anydiffusion model, including unconditional ones. Our guidance techniques are easyto implement and have the same sampling cost as CFG. Through extensiveexperiments, we demonstrate that ICG matches the performance of standard CFGacross various conditional diffusion models. Moreover, we show that TSGimproves generation quality in a manner similar to CFG, without relying on anyconditional information.