This project presents an advanced action recognitioner and Video Classifier program that leverages a parameterized SlowFast neural network to address the challenges of classifying actions in both static and dynamic datasets. By allowing users to fine-tune the model according to the dataset's characteristics, the program exhibits a high degree of adaptability and performance. Future enhancements could focus on automating parameter selection through meta-learning approaches or expanding the program's capability to real-time action recognition in mixed data environments