OpenMMLab's Next Generation Video Understanding Toolbox and Benchmark

Awesome video understanding toolkits based on PaddlePaddle. It supports video data annotation tools, lightweight RGB and skeleton based action recognition model, practical applications for video tagging and sport action detection.

Yolov5+SlowFast: Realtime Action Detection Based on PytorchVideo

Video Feature Extraction Code for EMNLP 2020 paper "HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training"

A PyTorch implementation of "SlowFast Networks for Video Recognition"

Video classification exercise using UCF101 data for training an early-fusion and SlowFast architecture model, both using the PyTorch Lightning framework.

A PyTorch implementation of SlowFast based on ICCV 2019 paper "SlowFast Networks for Video Recognition"

Yolov11 for object detection and Deepsort_oh for object tracking together with official SlowFast for action detection

Code for discovering slow timescale dynamics using the SINDy method

PyTorch implementation of SlowFast network.

DIVA-VQA: Detecting Inter-Frame Variations in UGC Video Quality

[CVPR 2019] SlowFast Networks for Video Recognition

Human Action Recognition

Make AVADataset custom dataset.

An advanced action recognition and Video Classifier

SignifAI is an engineering-driven project that translates French Sign Language (LSF) into French text and speech. It combines gesture recognition and language modeling, reflecting both linguistic and technical innovation.

Implementation of FICC 2024 paper "Benchmarking Jetson Edge Devices with an End-to-end Video-based Anomaly Detection System"

PySlowFast, the official video understanding framework from Facebook AI Research (FAIR), to train, evaluate, and reproduce state-of-the-art video models on the UCF24 action detection dataset. It supports customizable training pipelines, model fine-tuning, and evaluation for video-based action recognition and spatio-temporal localization tasks.

Multimodal version of SlowFast (Vision + Audio inputs)

Load a SlowFast model, load the weights and modify the head to be used in another project