OpenS2V-Nexus: A Benchmark and Million-Scale Dataset for Subject-to-Video Generation

OpenS2V-Nexus is a detailed benchmark and dataset for subject-to-video generation. The released OpenS2V-5M subset contains approximately 5 million samples after filtering. It was created by author peterlrm and last updated on May 28, 2026.

Use Cases

• Benchmarking subject-to-video generation models based on the provided detailed benchmark.
• Training conditional video generation models using the million-scale dataset.
• Evaluating the quality of generated videos against a high-quality, filtered dataset.
• Researching filtering techniques for large-scale video datasets based on the described methodology.

Strengths

• Dataset contains approximately 5 million samples.
• High-quality subset was created by applying filtering across the entire dataset.
• Includes a detailed benchmark for evaluation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the specific released subset is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

huggingface

Freshness

Last updated 2026-05-28 09:02:26; freshness should be verified.