Next, I will analyze MUNIT (Multimodal Unsupervised) for everyone without dead ends, which belongs to the style transfer series. The previous articles are as follows (the following are all the projects I work on, each of which is, I have done a 100% detailed interpretation. As the number of projects increases, for convenience and not bloated, so give the following link) \color{blue}{attached at the end of the article }The official account is attached at the end of the article − \color{blue}{official account-}Official account - massive resources. \color{blue}{massive resource}.Massive resources .
Visual Work Project - Provide a helping hand for you in the future!
I believe that my series of blogs may not be the earliest in China about the explanation of MUNIT proposed by NVIDIA in 2018, but it must be the most detailed. The title of the paper corresponding to this network is: MUNIT: Multimodal UNsupervised Image-to-image
Translation
The paper was accepted by CVPR 2018, so let's not talk nonsense, let's start directly!