Score Mismatching for Generative Modeling
This addresses efficiency issues in generative modeling for researchers and practitioners, though it appears incremental as it builds on existing score-based frameworks.
The paper tackles the computational burden of iterative sampling in score-based generative models by proposing a one-step sampling method, achieving faster training with only 10 diffusion steps and outperforming Consistency Model and Denoising Score Matching on CIFAR-10.
We propose a new score-based model with one-step sampling. Previously, score-based models were burdened with heavy computations due to iterative sampling. For substituting the iterative process, we train a standalone generator to compress all the time steps with the gradient backpropagated from the score network. In order to produce meaningful gradients for the generator, the score network is trained to simultaneously match the real data distribution and mismatch the fake data distribution. This model has the following advantages: 1) For sampling, it generates a fake image with only one step forward. 2) For training, it only needs 10 diffusion steps.3) Compared with consistency model, it is free of the ill-posed problem caused by consistency loss. On the popular CIFAR-10 dataset, our model outperforms Consistency Model and Denoising Score Matching, which demonstrates the potential of the framework. We further provide more examples on the MINIST and LSUN datasets. The code is available on GitHub.