Next, I will analyze mmaction2 (SlowFast-action recognition) for everyone without dead ends. The previous articles are as follows (the following are all the projects I work on, and I have done a 100% detailed interpretation of each project. As the number of projects increases, for convenience and not to be bloated, I give the following link) \color{blue}{attached at the end of the article }The official account is attachedat the end of the article − \color{blue}{official account-}Officialaccount - massive resources. \color{blue}{massive resource}.Massiveresources .
Highly recommended commercial grade project: \color{red}{Highly recommended commercial grade project:}Highlyrecommendedcommercial-level project: This is my own behavior analysis project, mainly including (1. Pedestrian detection, 2. Pedestrian tracking, 3. Behavior recognition three modules): Behavior analysis (commercial level) 00-Catalogue-The latest explanation without dead ends in history
I believe that my series of blogs may not be the earliest in China about the explanation of mmaction2 (SlowFast), but it is definitely the most detailed. The paper corresponding to this network is: SlowFast: SlowFast Networks for Video Recognition If my code has been modified a lot, in the last chapter, I will publish my modified source code\color{red}{If my code has been modified a lot, in the last chapter, I will publish my modified source code }IfIhave modifieda lot ofcode, in thelastchapter, I will publishthe source codeaftermy modification . Let's not talk nonsense, let's start directly !