Ploba logo

Find the right AI tool for the job: apps, agents, APIs, MCP servers and datasets, explained plainly.

Get the AI Digest

Discover

  • Explore
  • Top Charts
  • Categories
  • Agents
  • MCP servers
  • Developer tools

Learn & community

  • Learn AI
  • AI Glossary
  • Blog
  • Hackathons
  • People
  • AI Digest

For makers

  • List your tool
  • Docs
  • Claim a listing
  • Advertise
  • Write for us

Company

  • About
  • Careers
  • Changelog
  • Trust
  • Security
  • Contact

© 2026 Ploba. All rights reserved.

  • Privacy
  • Terms
  • Cookies
  • Advertising
  • All policies
llms.txt
Ploba logo
DiscoverHackathonsPeopleLearnAI Digest
Sign inGet started
HomeTop ChartsExploreCategories
TypesWeb apps1,894Agents585Plugins416APIs376Mobile apps152Datasets85
AdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise hereAdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise here
  1. Explore
  2. /
  3. AI
  4. /
  5. Text-to-Speech Human Preferences 315K
AI
TH

Text-to-Speech Human Preferences 315K

by DatapointAI

A gated Datapoint AI dataset of 315,000 blind pairwise human votes comparing 15 text-to-speech models on 300 English voice-agent prompts, with all 4,500 audio samples included.

Dataset·Free

View on Hugging Face(opens in a new tab)

About Text-to-Speech Human Preferences 315K

Key features of Text-to-Speech Human Preferences 315K

  • 315,000 pairwise human votes
  • 15 text-to-speech models compared in a full round-robin
  • 300 English prompts across eight voice-agent categories
  • 4,500 audio renders stored as FLAC
  • Blind comparison with balanced candidate order
  • Trust-weighted vote totals for sensitivity analysis
  • Separate configs for audio, pairs, responses, prompts and models

Who Text-to-Speech Human Preferences 315K is for

  • Researchers training TTS reward models
  • Teams evaluating speech models for voice agents
  • Preference learning researchers

How to use Text-to-Speech Human Preferences 315K

  1. 1

    Step 1: Sign in to Hugging Face and accept the dataset's access conditions.

  2. 2

    Step 2: Read the dataset structure section to choose a config: audio, pairs, responses, prompts or models.

  3. 3

    Step 3: Load the config you need with the Hugging Face datasets library.

  4. 4

    Step 4: Join audio and comparison rows using the audio keys.

  5. 5

    Step 5: Use raw votes or the trust-weighted totals to train or evaluate a preference model.

Text-to-Speech Human Preferences 315K FAQ

Is the dataset open to everyone?

The repository is public, but it is gated. You must agree to share your Hugging Face information and accept the conditions to access the files.

What license applies?

CC-BY-4.0 with attribution to Datapoint AI. You also agree not to re-identify annotators, and audio usage rights follow each model provider's terms.

How many votes and models are included?

315,000 eligible votes comparing 15 text-to-speech models over 300 prompts.

What kinds of prompts are used?

Eight voice-agent categories, including empathy and de-escalation, greetings and brand delivery, and instructions and troubleshooting.

Details

Pricing
FreeCC-BY-4.0 with attribution to Datapoint AI. Access is gated: accept the conditions on Hugging Face and agree not to re-identify annotators. Usage rights for the audio follow each model provider's terms.
Type
Dataset
Categories
AIDataAudio
Added
Sep 17, 2026
DaPublisherDatapointAI

Built from public sources

We put this page together from Text-to-Speech Human Preferences 315K’s own website and docs so you can compare it with other tools. Ploba isn’t affiliated with Text-to-Speech Human Preferences 315K.

Work at Text-to-Speech Human Preferences 315K? Claim the listing to manage it yourself.

Something wrong here?

Related tools

See more
Hermes Agent Reasoning Traces iconHermes Agent Reasoning TracesLambda · FreeHermes Agent Reasoning Traces is Lambda's dataset of multi-turn tool-calling agent conversations with step-by-step reasoning and real tool execution results. It has two configs, one per source model.Developer Tools · FreeHLHumanity's Last ExamCAIS · FreeHumanity's Last Exam (HLE) is a benchmark of 2,500 expert-written questions across dozens of subjects, built to test AI models at the frontier of human knowledge. It is for researchers and model builders evaluating frontier AI.AI · FreeS2Spark 234KOpenDataArena · FreeSpark-234K is a dataset of 234K scientific reasoning question-answer pairs from OpenDataArena, built from frontier research papers. It is meant for training language models on hard, self-contained science problems.AI · Free