#video-understanding Repositories
Discover 51 active open source GitHub repositories tagged with #video-understanding, accumulating 114,089 total stars (averaging 2,237 stars per project). Leading this topic is QwenLM/Qwen3-VL with 20,240 stars.
51 repositories total
Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.
PySlowFast: video understanding codebase from FAIR for reproducing state-of-the-art video models.
One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks
VILA is a family of state-of-the-art vision language models (VLMs) for diverse multimodal AI tasks across the edge, data center, and cloud.
[CVPR2024 Highlight][VideoChatGPT] ChatGPT with video understanding! And many more supported LMs such as miniGPT4, StableLM, and MOSS.
๐ฅ๐ฅ๐ฅ [IEEE TCSVT] Latest Papers, Codes and Datasets on Vid-LLMs.
A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.
State-of-the-art Image & Video CLIP, Multimodal Large Language Models, and More!
GLM-4.6V/4.5V/4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning
[ICCV 2019] TSM: Temporal Shift Module for Efficient Video Understanding
[EMNLP2026] "VideoAgent: All-in-One Agentic Framework for Video Understanding and Editing, and Remaking"
NVIDIA AI Blueprint for video search and summarization (VSS) is a GPU-accelerated reference architecture for building video analytics agents with real-time verified alerts, visual Q&A, and automated reporting. The VSS Blueprint uses vision language models (VLMs) such as NVIDIA Cosmos, LLMs such as NVIDIA Nemotron, RAG, and NVIDIA NIMs.
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.
A 3B-active-parameter native unified multimodal model for image and video understanding, generation, and editing.
awesome grounding: A curated list of research papers in visual grounding
:fire: :fire: :fire: A paper list of some recent Computer Vision(CV) works
[CVPR 2024 Highlight๐ฅ] Chat-UniVi: Unified Visual Representation Empowers Large Language Models with Image and Video Understanding
Open multimodal decision models, 0.8B to 27B, for text, images and video. Decide with probabilities over your own candidates and locate with bounding boxes, in place of generated text. Free API, ONNX, GGUF, vLLM and SGLang; fine-tuning code included.
Video-R1: Reinforcing Video Reasoning in MLLMs [๐ฅthe first paper to explore R1 for video]
Recent LLM-based CV and related works. Welcome to comment/contribute!
LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models (ECCV 2024)
Official implementation of UnifiedReward & [NeurIPS 2025] UnifiedReward-Think & UnifiedReward-Flex
[CVPR 2023] VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking
VideoLLM-online: Online Video Large Language Model for Streaming Video (CVPR 2024)
The official repo for "Vidi: Large Multimodal Models for Video Understanding and Editing"
Official code for Goldfish model for long video understanding and MiniGPT4-video for short video understanding
Give AI agents eyes, ears, and verifiable results. Watch Skill turns video, audio, and screen activity into searchable, timestamped evidence, and proves work with deterministic contractsโnot model opinions. DeepWatch is the agent workspace built on DeepSeek Harness. Python + npm, MCP, CLI, REST, Web.
A collection of multimodal reasoning papers, codes, datasets, benchmarks and resources.
Temporal Action Detection & Weakly Supervised Temporal Action Detection & Temporal Action Proposal Generation
ComfyUI nodes for vision-language models: Qwen3-VL, Moondream 3, Florence-2, SmolVLM2, InternVL, Gemma 3, MiniCPM-V. Plus open-vocabulary detection, SAM2/SAM3 segmentation, video temporal reasoning, GGUF via llama.cpp, and hosted LLM/VLM APIs.
Scene-aware storyboard for DeepSeek Harness: 9โ16 numbered cells with timestamps + transcript + an audio envelope, packed into ONE image (384 tokens). Two tools: video_storyboard and video_zoom.
Official Pytorch Implementation for โDINO-Tracker: Taming DINO for Self-Supervised Point Tracking in a Single Videoโ (ECCV 2024)
Frequently Asked Questions
What is the most popular #video-understanding repository on GitHub?
The top ranked open source #video-understanding project on GitHub is QwenLM/Qwen3-VL with 20,240 stars. Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.
Which programming languages are most used in #video-understanding repositories?
Among the top ranked #video-understanding projects on GitHub, the most common programming languages are Python, Jupyter Notebook, C++, JavaScript.
How is the #video-understanding repository ranking calculated?
We track 51 active open source repositories tagged with #video-understanding (500+ stars), totaling 114,089 stars with an average of 2,237 stars per repository. Rankings update daily from the GitHub API.