aimagelab/VHS
[CVPR2026 Findings] VHS: Verifier on Hidden States, an efficient inference-time scaling verification framework for DiT-based image generation.
VHS helps creative professionals, marketers, or researchers quickly find the best image from multiple AI-generated options for a given text prompt. It takes a text description and generates several candidate images, then automatically selects the highest-quality one, saving time and computational resources compared to manually reviewing or fully evaluating every image. This is ideal for anyone generating many images from text who needs a fast, automated way to pick the best result.
Use this if you are generating multiple images from text descriptions and need an efficient, automated way to select the best one without extensive manual review or full evaluation of every candidate.
Not ideal if you need fine-grained control over image generation beyond selecting the best from a batch, or if your primary goal is generating unique images rather than optimizing quality from existing candidates.
Stars
11
Forks
—
Language
Python
License
—
Category
Last pushed
Mar 25, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/aimagelab/VHS"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
hao-ai-lab/FastVideo
A unified inference and post-training framework for accelerated video generation.
ModelTC/LightX2V
Light Image Video Generation Inference Framework
thu-ml/TurboDiffusion
TurboDiffusion: 100–200× Acceleration for Video Diffusion Models
PKU-YuanGroup/Helios
Helios: Real Real-Time Long Video Generation Model
PKU-YuanGroup/MagicTime
[TPAMI 2025🔥] MagicTime: Time-lapse Video Generation Models as Metamorphic Simulators