Improve dataset card: Add task category, abstract, links, and sample usage

README.md

@@ -2,19 +2,79 @@
 language:
 - en
 papers:
-- title: "Stable Video Infinity: Infinite-Length Video Generation with Error Recycling"
+- title: 'Stable Video Infinity: Infinite-Length Video Generation with Error Recycling'
   authors:
-    - Wuyang Li
-    - Wentao Pan
-    - Po-Chien Luan
-    - Yang Gao
-    - Alexandre Alahi
-  url: "https://arxiv.org/abs/2510.09212"
-  conference: "arXiv preprint, 2025"
+  - Wuyang Li
+  - Wentao Pan
+  - Po-Chien Luan
+  - Yang Gao
+  - Alexandre Alahi
+  url: https://huggingface.co/papers/2510.09212
+  conference: arXiv preprint, 2025
 tags:
 - video generation
 - benchmark
+task_categories:
+- image-to-video
 ---
-This Benchmark is introduced in the paper  
-**[Stable Video Infinity: Infinite-Length Video Generation with Error Recycling](https://arxiv.org/abs/2510.09212)**  
+
+# Stable Video Infinity Benchmark Dataset
+
+This repository contains the benchmark datasets introduced in the paper:
+
+**[Stable Video Infinity: Infinite-Length Video Generation with Error Recycling](https://huggingface.co/papers/2510.09212)**
 by Wuyang Li, Wentao Pan, Po-Chien Luan, Yang Gao, Alexandre Alahi (2025).
+
+**Project Page:** [https://stable-video-infinity.github.io/homepage/](https://stable-video-infinity.github.io/homepage/)
+**Code:** [https://github.com/vita-epfl/Stable-Video-Infinity](https://github.com/vita-epfl/Stable-Video-Infinity)
+
+## Abstract
+We propose **Stable Video Infinity (SVI)** that is able to generate infinite-length videos with high temporal consistency, plausible scene transitions, and controllable streaming storylines. While existing long-video methods attempt to *mitigate accumulated errors* via handcrafted anti-drifting (e.g., modified noise scheduler, frame anchoring), they remain limited to single-prompt extrapolation, producing homogeneous scenes with repetitive motions. We identify that the fundamental challenge extends beyond error accumulation to a critical discrepancy between the training assumption (seeing clean data) and the test-time autoregressive reality (conditioning on self-generated, error-prone outputs). To bridge this hypothesis gap, SVI incorporates **Error-Recycling Fine-Tuning**, a new type of efficient training that recycles the Diffusion Transformer (DiT)'s self-generated errors into supervisory prompts, thereby encouraging DiT to *actively identify and correct its own errors*. This is achieved by injecting, collecting, and banking errors through closed-loop recycling, autoregressively learning from error-injected feedback. Specifically, we (i) inject historical errors made by DiT to intervene on clean inputs, simulating error-accumulated trajectories in flow matching; (ii) efficiently approximate predictions with one-step bidirectional integration and calculate errors with residuals; (iii) dynamically bank errors into replay memory across discretized timesteps, which are resampled for new input. SVI is able to scale videos from seconds to infinite durations with no additional inference cost, while remaining compatible with diverse conditions (e.g., audio, skeleton, and text streams). We evaluate SVI on three benchmarks, including consistent, creative, and conditional settings, thoroughly verifying its versatility and state-of-the-art role.
+
+## Description
+This benchmark dataset is designed to evaluate models for infinite-length video generation, featuring high temporal consistency, plausible scene transitions, and controllable storylines. The datasets within this repository include various test and training components used for the Stable Video Infinity (SVI) family of models.
+
+## Sample Usage
+
+You can download the SVI benchmark datasets locally using the Hugging Face CLI:
+
+```bash
+huggingface-cli download --repo-type dataset vita-video-gen/svi-benchmark --local-dir ./data/svi-benchmark
+```
+
+## Datasets
+
+This repository contains benchmark datasets created by our Automatic Prompt Stream Engine (see Appendix. A of the paper for more details). These datasets include images and associated prompt streams according to specific storylines, suitable for various evaluation settings.
+
+### Benchmark Datasets (Test)
+
+| Data                                               | Use  | HuggingFace Link                                                                                            | Comment                                                                                           |
+| :------------------------------------------------- | :--- | :---------------------------------------------------------------------------------------------------------- | :------------------------------------------------------------------------------------------------ |
+| **Consistent Video Generation**              | Test | [🤗 Dataset](https://huggingface.co/datasets/vita-video-gen/svi-benchmark/tree/main/consisent_video_gen)       | Generate 1 long video using 1 text prompt                                                         |
+| **Creative Video Generation**                | Test | [🤗 Dataset](https://huggingface.co/datasets/vita-video-gen/svi-benchmark/tree/main/creative_video_gen)        | Generate 1 long video using 1 text prompt stream according to storyline (1 prompt for 5 sec clip) |
+| **Creative Video Generation (More prompts)** | Test | [🤗 Dataset](https://huggingface.co/datasets/vita-video-gen/svi-benchmark/tree/main/creative_video_gen_longer) | Generate 1 long video using 1 text prompt stream according to storyline (1 prompt for 5 sec clip) |
+
+### Training Datasets
+
+The following training datasets were also used for the SVI family of models and are linked here for convenience:
+
+| Data                                           | Use   | HuggingFace Link                                                                                     | Comment                                                 |
+| :--------------------------------------------- | :---- | :--------------------------------------------------------------------------------------------------- | :------------------------------------------------------ |
+| **Customized Datasets**                  | Train | [🤗 Dataset](https://huggingface.co/datasets/vita-video-gen/svi-benchmark/tree/main/customized_dataset) | You can make your customized datasets using this format |
+| **Consistent/Creative Video Generation** | Train | [🤗 Dataset](https://huggingface.co/datasets/LanguageBind/Open-Sora-Plan-v1.1.0/tree/main/all_mixkit)   | MixKit Dataset                                          |
+| **Consistent/Creative Video Generation** | Train | [🤗 Dataset](https://huggingface.co/datasets/APRIL-AIGC/UltraVideo-Long)                                | UltraVideo Dataset                                      |
+| **Human Talking**                        | Train | [🤗 Dataset](https://huggingface.co/fudan-generative-ai/hallo3)                                         | 5k subset from Hallo v                                  |
+| **Human Dancing**                        | Train | [🤗 Dataset](https://www.kaggle.com/datasets/yasaminjafarian/tiktokdataset)                             | TikTok                                                  |
+
+## Citation
+
+If you find our work helpful for your research, please consider citing our paper:
+
+```bibtex
+@article{li2025stable,
+      title={Stable Video Infinity: Infinite-Length Video Generation with Error Recycling}, 
+      author={Wuyang Li and Wentao Pan and Po-Chien Luan and Yang Gao and Alexandre Alahi},
+      journal={arXiv preprint arXiv: arXiv:2510.09212},
+      year={2025},
+}
+```