Conditional Generative Data-free Knowledge Distillation

Understand

Knowledge distillation has made remarkable achievements in model compression.

• However, most existing methods require the original training data, which is usually unavailable due to privacy and security issues.
• In this paper, we propose a conditional generative data-free knowledge distillation (CGDD) framework for training lightweight networks without any training data.
• This method realizes efficient knowledge distillation based on conditional image generation.