Add dataset card for TryAny-Bench
#1
by nielsr HF Staff - opened
README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
task_categories:
|
| 3 |
+
- image-to-video
|
| 4 |
+
tags:
|
| 5 |
+
- virtual-try-on
|
| 6 |
+
- video-try-on
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# TryAny-Bench
|
| 10 |
+
|
| 11 |
+
This repository contains **TryAny-Bench**, the benchmark dataset introduced in the paper [OmniTryOn: Video Try-On Anything at Once!](https://huggingface.co/papers/2606.08514).
|
| 12 |
+
|
| 13 |
+
[Project Page](https://xcltql666.github.io/OmniTryOnProj/) | [Paper](https://huggingface.co/papers/2606.08514) | [GitHub](https://github.com/xcltql666/OminTryOn)
|
| 14 |
+
|
| 15 |
+
## Overview
|
| 16 |
+
|
| 17 |
+
TryAny-Bench is a benchmark dataset designed for the **Try-On Anything** task, which aims to simultaneously transfer diverse wearable objects onto a person in a video in a single inference pass. It encompasses a paired video dataset alongside a tailored evaluation protocol.
|