fansunqi/VideoTool
Official Repository for NeurIPS'25 Paper "Tool-Augmented Spatiotemporal Reasoning for Streamlining Video Question Answering Task"
This project helps researchers and developers working with AI models to better understand and answer questions about video content. It takes a video and a question as input, intelligently analyzes the video using a suite of specialized tools, and outputs a more accurate answer to the question. It's designed for AI practitioners focused on improving video comprehension and reasoning in large multimodal models.
Use this if you are developing or researching advanced AI systems that need to perform complex spatiotemporal reasoning and answer detailed questions based on video content.
Not ideal if you are looking for a simple, off-the-shelf video analysis tool for general business or creative tasks.
Stars
17
Forks
—
Language
Python
License
MIT
Category
Last pushed
Mar 08, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/fansunqi/VideoTool"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
open-mmlab/mmpretrain
OpenMMLab Pre-training Toolbox and Benchmark
facebookresearch/mmf
A modular framework for vision & language multimodal research from Facebook AI Research (FAIR)
adambielski/siamese-triplet
Siamese and triplet networks with online pair/triplet mining in PyTorch
HuaizhengZhang/Awsome-Deep-Learning-for-Video-Analysis
Papers, code and datasets about deep learning and multi-modal learning for video analysis
KaiyangZhou/pytorch-vsumm-reinforce
Unsupervised video summarization with deep reinforcement learning (AAAI'18)