Datasets › COIN

COIN

Introduced by Yansong Tang et al. in COIN: A Large-scale Dataset for Comprehensive Instructional Video Analysis archive 2025-07-28

The COIN dataset (a large-scale dataset for COmprehensive INstructional video analysis) consists of 11,827 videos related to 180 different tasks in 12 domains (e.g., vehicles, gadgets, etc.) related to our daily life. The videos are all collected from YouTube. The average length of a video is 2.36 minutes. Each video is labelled with 3.91 step segments, where each segment lasts 14.91 seconds on average. In total, the dataset contains videos of 476 hours, with 46,354 annotated segments.

Source: COIN: A Large-scale Dataset for Comprehensive Instructional Video Analysis

Homepage Benchmarks Papers JSON

Benchmarks archive 2025-07-28

All 2 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

		First row (archive order)	Paper	Code
Action Segmentation	COIN	UnLoc-L Frame accuracy 72.8	UnLoc: A Unified Framework for Video Localization Tasks	google-research/scenic	9	Compare
Video Classification	COIN	HERMES Accuracy (%) 93.5	HERMES: temporal-coHERent long-forM understanding with...	joslefaure/HERMES	7	Compare

Papers archive 2025-07-28

15 shown of 15 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 105. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

			Date	Samples run Syntology
HERMES: temporal-coHERent long-forM understanding with Episodes and Semantics	1	1	30 Aug 2024	ran 7 of 9 samples (2 unverified)
MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding	1	1	8 Apr 2024	ran 7 of 9 samples (2 unverified)
Multi-granularity Correspondence Learning from Long-term Noisy Videos	1	1	30 Jan 2024	ran 2 of 3 samples (1 unverified)
UnLoc: A Unified Framework for Video Localization Tasks	1	1	21 Aug 2023	not harvested
Selective Structured State-Spaces for Long-Form Video Understanding	0	1	25 Mar 2023	not harvested
Efficient Movie Scene Detection using State-Space Transformers	1	1	29 Dec 2022	ran 3 of 3 samples (0 unverified)
Long Movie Clip Classification with State-Space Video Models	1	1	4 Apr 2022	ran 11 of 18 samples (7 unverified; 2 pointer-only for licence)
Learning To Recognize Procedural Activities with Distant Supervision	1	1	26 Jan 2022	not harvested
VideoCLIP: Contrastive Pre-training for Zero-shot Video-Text Understanding	2	1	28 Sep 2021	not harvested
TACo: Token-aware Cascade Contrastive Learning for Video-Text Alignment	0	1	23 Aug 2021	not harvested
VLM: Task-agnostic Video-Language Model Pre-training for Video Understanding	1	1	20 May 2021	not harvested
ActBERT: Learning Global-Local Video-Text Representations	1	1	14 Nov 2020	not harvested
UniVL: A Unified Video and Language Pre-Training Model for Multimodal Understanding and Generation	2	1	15 Feb 2020	ran 1 of 4 samples (3 unverified; 1 pointer-only for licence)
End-to-End Learning of Visual Representations from Uncurated Instructional Videos	4	2	13 Dec 2019	ran 3 of 5 samples (2 unverified)
Temporal Segment Networks for Action Recognition in Videos	11	1	8 May 2017	not harvested

Dataset loaders archive 2025-07-28

No loader listed in the archive.

Tasks archive 2025-07-28

Action Recognition Action Localization Temporal Action Localization Action Segmentation Video Classification

License archive 2025-07-28

Custom

Modalities archive 2025-07-28

Videos

Languages archive 2025-07-28

No language tagged.

Variants archive 2025-07-28

COIN

1 variant name, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections