# Spatiotemporal Feature Extraction

A video understanding model fine-tuned on a subset of UCF101 based on the VideoMAE base model, achieving 95.71% accuracy

A video understanding model based on the VideoMAE architecture, fine-tuned on the RWF-2000 dataset for violence detection tasks

VideoMAE is a self-supervised video pre-training model based on masked autoencoder, pre-trained for 2400 epochs on the Something-Something-v2 dataset.

VideoMAE is a self-supervised video pre-training model based on masked autoencoder, fine-tuned on the Kinetics-400 dataset for video classification tasks.