Question about training and inference models

[hosytuyen]

Hi

Thank you for sharing your amazing work. I have a few questions and concerns:

1. Is it true that the accuracy of the training model and the re-parameterized model (inference model) is the same? As the output is similar?

2. Is it possible to fine-tune a VGG-like model on a different dataset using a pre-trained inference model, such as a very deep network like RepVGG-D2se? Is there a risk of vanishing gradients?

Thank you very much.