pro2nit/STREAM
official implementation of 'STREAM : Spatio-TempoRal Evaluation and Analysis Metric for Video Generative Models'
This tool helps researchers and engineers who are developing video generative AI models to accurately assess how realistic and diverse their generated videos are. By inputting sets of real and AI-generated video clips, it produces distinct scores for the spatial quality (how good individual frames look) and the temporal quality (how smooth and natural the movement is). This helps you understand specific areas for improving your video generation AI.
Use this if you are developing or evaluating video generative AI models and need a robust, comprehensive metric to assess both spatial and temporal video quality independently.
Not ideal if you are evaluating image generative models or if you only need a simple, single-score evaluation that doesn't distinguish between spatial and temporal aspects.
Stars
28
Forks
2
Language
Python
License
MIT
Category
Last pushed
Dec 24, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/pro2nit/STREAM"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
roboflow/inference
Turn any computer or edge device into a command center for your computer vision projects.
roboflow/roboflow-python
The official Roboflow Python package. Manage your datasets, models, and deployments. Roboflow...
dusty-nv/jetson-inference
Hello AI World guide to deploying deep-learning inference networks and deep vision primitives...
hailo-ai/tappas
High-performance, optimized pre-trained template AI application pipelines for systems using Hailo devices
Apra-Labs/ApraPipes
A pipeline framework for developing video and image processing application. Supports multiple...