Your objective is to construct action recognition models on a subset of [the UCF101 dataset](https://www.crcv.ucf.edu/data/UCF101.php). The UCF101 dataset consists of a total of 13,320 videos spanning diverse human action categories. However, for the final project, our focus narrows down to a subset of the dataset, comprising 2,355 training videos and 969 test videos distributed among 30 action categories. You will have to build one binary classifier and one multi-class classifier.

In this competition, your task is to build a classifier for the class *JugglingBalls*. Your goal is to train a model to classify JugglingBalls videos as *positive* and the remaining videos as *negative*. You are free to choose any machine learning models from the class. 

For more details regarding the tasks, including the dataset download, please refer to the information provided in [this Google Colab link](https://colab.research.google.com/drive/19IOk0hmokk726V2WRwNsX9TQDlcRECuU?usp=sharing).

Please note that **the deadline for project submission is Tuesday, March 5, 11:59 pm**.