Constarium
← Search

Data · dataset · 2019

YouTube-8M Dataset

Listed in BERD@NFDI Data Portal

YouTube-8M is a large-scale labeled video dataset that consists of millions of YouTube video IDs and with high-quality machine-generated & partially human-verified annotations from a diverse vocabulary of 3,800+ visual entities.

Description

It comprises two subsets: 8M Segments Dataset: 230K human-verified segment labels, 1000 classes, 5 segments/video 8M Dataset: May 2018 version (current): 6.1M videos, 3862 classes, 3.0 labels/video, 2.6B audio-visual features Thus, it comes with precomputed audio-visual features from billions of frames and audio segments, designed to fit on a single hard disk.

This makes it possible to train a strong baseline model on this dataset in less than a day on a single GPU! At the same time, the dataset's scale and diversity can enable deep exploration of complex audio-visual models that can take weeks to train even in a distributed fashion. YouTube offers the YouTube8M dataset for download as TensorFlow Record files on their website.

Read the rest (1 more)

Starter code for the dataset can be found on their GitHub page.

Links

Where it is published

Catalogue records · 1

Topics

Stated by source
Audio-visual · Labels · Large-scale · Segments · Video · YouTube
From keywords
Economics & Finance · Video
Inferred from text
Audio 75%
Provenance · 1 source records, 15 field assertions
SourceKeyLast seenRaw
BERD@NFDI Data Portal3m5k1-zhn699 d agoJSON v1
FieldAssertionExtractorEvidence
access_levelsource · berd platform deconnector:berd_platform_de@1.0.0
concepts[field].local:field:economics-financemapping · berd platform deconnector:berd_platform_de@1.0.0
concepts[modality].local:modality:audioenrichment · berd platform dekeyword-concept-rules@1.0.0title+description (75%)
concepts[modality].local:modality:videomapping · berd platform devocabulary-mapper@1.0.0keywords['Video']
concepts[topic].invenio_subject:berd_platform_de:audio-visualsource · berd platform deconnector:berd_platform_de@1.0.0
concepts[topic].invenio_subject:berd_platform_de:labelssource · berd platform deconnector:berd_platform_de@1.0.0
concepts[topic].invenio_subject:berd_platform_de:large-scalesource · berd platform deconnector:berd_platform_de@1.0.0
concepts[topic].invenio_subject:berd_platform_de:segmentssource · berd platform deconnector:berd_platform_de@1.0.0
concepts[topic].invenio_subject:berd_platform_de:videosource · berd platform deconnector:berd_platform_de@1.0.0
concepts[topic].invenio_subject:berd_platform_de:youtubesource · berd platform deconnector:berd_platform_de@1.0.0
descriptionsource · berd platform deconnector:berd_platform_de@1.0.0/metadata/description
licensesource · berd platform deconnector:berd_platform_de@1.0.0
publication_datesource · berd platform deconnector:berd_platform_de@1.0.0
titlesource · berd platform deconnector:berd_platform_de@1.0.0/metadata/title
updated_datesource · berd platform deconnector:berd_platform_de@1.0.0