---
title: Humaneness-Voice-Large-1.2-Alpha-Demo-Audio
canonical_url: "https://www.modelscope.cn/datasets/laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio"
md_url: "https://www.modelscope.cn/datasets/laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio.md"
repository: laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio
last_updated: 2026-10-22
license: cc-by-4.0
storage_size: "116 MB"
downloads: 177
stars: 0
---

# Humaneness-Voice-Large-1.2-Alpha-Demo-Audio

> Humaneness-Voice-Large-1.2-Alpha-Demo-Audio - laion 在 ModelScope 开源的数据集。Humaneness Voice Large 1.2 Alpha: listening audio

laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio 是 ModelScope 魔搭社区上的数据集，存储大小 116 MB，采用 cc-by-4.0 许可。

- **Repository**: laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio
- **License**: cc-by-4.0
- **Storage size**: 116 MB
- **Downloads**: 177
- **Stars**: 0
- **Last updated**: 2026-10-22

Source: https://www.modelscope.cn/datasets/laion/Humaneness-Voice-Large-1.2-Alpha-Demo-Audio

---

# Humaneness Voice Large 1.2 Alpha: listening audio

Seven newly written fictional acting scenes in five languages, with two
conditions per case: prompt only, or prompt plus a distinct speaker reference.
There are 70 selected predictions and five input reference clips. Players:
https://laion-humaneness-voice-large-1-2-alpha-listening.static.hf.space/

Model and full training/inference documentation:
https://huggingface.co/laion/Humaneness-Voice-Large-1.2-Alpha

The target is 34.8 seconds per scene. Actual durations, seeds, full prompts and
selection diagnostics are in scenes.json/predictions.json. If the first take
was outside 30–40 seconds, up to two additional deterministic seeds were tried;
the first duration-valid take is shown, otherwise the closest-duration take.
This is duration-only curation, not MOS/WER/reward selection. All attempts,
including unselected takes, are retained in all_attempts.tar. No clip is
trimmed or time-stretched to disguise a duration miss.

audio/ contains individual 128 kbps mono MP3s for browser streaming.
selected_audio.tar contains the same selected audio plus metadata and
12-codebook MOSS-v2 arrays. all_attempts.tar includes every candidate with
its codec array and JSON. Twelve indices per 80 ms frame = 150 indices/s.
The five references are codec-reconstructed crops from the public synthetic
Gemini corpus laion/german-english-voice-acting-challenges, CC BY 4.0.
references.json retains source UID, provider speaker, duration and attribution.
References are inputs, not model predictions or original lossless recordings.
The other scripts and predictions are newly authored release artifacts.

CC BY 4.0: attribute LAION contributors, Humaneness Voice Large 1.2 Alpha, 2026,
and the linked source corpus for references. Respect upstream terms and
personality rights. These examples are an experimental listening demo, not
a statistically representative quality benchmark or an impersonation tool.