datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
reaction_data
🧠 Overview
Reaction Data v1.0 is a curated multimodal dataset built from YouTube human reaction videos, designed for studying singing voice synthesis (SVS) evaluation, audio–text alignment, and natural-language reaction modeling.Each included creator (e.g., Alex Hefner, Beth Roars, The Charismatic Voice) provides a distinct commentary style and persona, offering diverse reactions across genres and vocal styles.
Each reviewer’s folder contains processed data, transcripts, and raw… See the full description on the dataset page: https://huggingface.co/datasets/KimZoey/reaction_data.filtered_reaction_bilibiliexpress4d-reaction-nofilter-v1-partial-3632seq-2026-04-15
Express4D Reaction No-Filter Partial Snapshot
This is a partial snapshot exported from a reaction-video pipeline into an Express4D-like format.
Current snapshot status:
sequence csv count: 3632
text count: 3628
payload count: 3628
anchor image count: 7228
relative to Express4D sequence count (1205): about 3.01x
status: partial snapshot
Files:
Express4D_reaction_nofilter_v1_partial_3632seq_2026-04-15.tar.gz
SNAPSHOT_STATUS.json
This repo was uploaded automatically from a local… See the full description on the dataset page: https://huggingface.co/datasets/JintaoDong256/express4d-reaction-nofilter-v1-partial-3632seq-2026-04-15.