datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
reef-guidance-system
Dataset Card for Reef Guidance System
This dataset provides imagery used for training and evaluation of models in the Reef Guidance System. All imagery was collected by the Australian Institute of Marine Science using the ReefScan™ Transom Marine Monitoring System.
If you use this dataset in your work, please cite the associated paper: AI-driven dispensing of coral reseeding devices for broad-scale restoration of the Great Barrier Reef (citations provided at bottom of this… See the full description on the dataset page: https://huggingface.co/datasets/QCR-Underwater-Perception/reef-guidance-system.marrs-global-coral-reef-soundscapesThis version of the dataset has been parsed into Parquet format for easy access and experimentation.
Raw: 1 min Audio clips collected from coral reef restoration sites around the world.
Detections: 5 second clips of detected sonotypes extracted via human-in-the-loop Agile modelling as described in: Williams et al. (2025) Evidence of ecosystem process recovery across a large-scale coral reef restoration programme using AI accelerated soundscape analysis… See the full description on the dataset page: https://huggingface.co/datasets/bens-bots/marrs-global-coral-reef-soundscapes.reefnet_species_images
ReefNet Species Images
Full-resolution coral reef images with species-level annotations from the
ReefNet dataset.
Dataset Description
This dataset contains 51,080 images from 54 CoralNet sources
with 365,030 annotation patches across 245 species.
Each annotation in metadata.parquet specifies a patch center (Row, Column)
within a full-resolution image. Crop a 224×224 patch centered at these coordinates
for classification tasks.
Splits
The curated… See the full description on the dataset page: https://huggingface.co/datasets/TsinghuaCorals/reefnet_species_images.Reefknot
Reefknot
Comprehensive benchmark for evaluating relation hallucination in MLLMs. 21,560 samples across 3 evaluation formats, 11,084 unique Visual Genome images.
Fields
Field
Description
image
Visual Genome image
image_id
VG image identifier
query_prompt
Question about object relations
label
Ground truth answer
type
Question type (Yes/No, Multichoice, VQA)
relation_type
Relation category (perception/cognitive)
Configs
Config… See the full description on the dataset page: https://huggingface.co/datasets/MM-Hallu/Reefknot.ReefNet-1.0
ReefNet Dataset
The ReefNet dataset is a comprehensive, curated collection of annotated coral reef images sourced from the CoralNet platform. It is designed for training and evaluating machine learning models for coral classification on a global scale. ReefNet's consolidated label set and rich metadata facilitate the straightforward extraction of custom datasets to support coral classification tasks across different geographic regions.
Dataset Structure
The dataset is… See the full description on the dataset page: https://huggingface.co/datasets/ReefNet/ReefNet-1.0.reefReefSet-CReefSet-C (RS-C) serves as a benchmark for DHAuDS in the context of test-time adaptation for audio classification.
RS-C, a subset of ReefSet, utilizes only test-set samples to address domain shift. Each sample in RS-C has a duration
of 1.88 seconds, a sample rate of 16 kHz, and belongs to one of 37 classes. The dataset comprises two sets: adaptation and evaluation sets. Each set comprises
seven corruption categories and two levels. Seven corruption categories are defined: WHN, ENQ, END1… See the full description on the dataset page: https://huggingface.co/datasets/AndyShao90/ReefSet-C.ai-tfts-aims-reefscapes-marine-featuresReef-Sensor-Staging
Coastal sensor release manifest
Dataset: Reef-Sensor-Staging
Register: Sensor release register
Eligibility: approved + embargoed only; approved records >= 10
Order: released desc, embargoed asc, station asc
Representative: latest approved sample, smallest records, table order
Manifest: Wave Shelf [RS-508] => 16 released, 6 embargoed || Lagoon Rim [RS-109] => 14 released, 3 embargoed || Manta Pass [RS-201] => 13 released, 3 embargoed || Coral Gate [RS-011] => 10 released, 8… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/Reef-Sensor-Staging.ts-aims-reefscapes-satellite-segmentation
Dataset Card for "ts-aims-reefscapes-satellite-segmentation"
More Information needed
ReefEcho-Event-Captions
ReefEcho Event Captions
This dataset contains generated captions describing acoustic events in reef recordings. The captions are intended for evaluation of multimodal retrieval systems.
Clearance matrix
The matrix below is the authoritative permissions record for this release.
Material role
Decision
conditions score
decision date
Permission label
Entry
captioning model
cleared
5
2025-03-11
Apache-2.0
EC-207
captioning model
cleared
1
2025-05-09… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/ReefEcho-Event-Captions.reefscape-dataAurora-Reef-Imagery
Aurora Reef Imagery
Annotated underwater imagery captured during the Aurora reef survey.
Distribution license: CC0 1.0
A legal note records a second permitted release path:
Distribution license: MIT
Distribution license: CC0 1.0
Distribution license: Apache-2.0
so100_testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100",
"total_episodes": 2,
"total_frames": 495,
"total_tasks": 1,
"total_videos": 4,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:2"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Reef123/so100_test.coral-reef-bioacoustics
Coral Reef Bioacoustics Dataset
Underwater audio recordings from two contrasting coral reef sites in Moreton Bay, Queensland, Australia, labeled for binary reef health classification.
Dataset Description
This dataset contains 19 WAV recordings captured using a Cetacean Research Technology CRT-40P hydrophone (48 kHz / 24-bit) deployed 0.5-1 m above the seafloor in November 2025. The recordings capture the natural soundscape of sub-tropical turbid reefs, including… See the full description on the dataset page: https://huggingface.co/datasets/mikulrai/coral-reef-bioacoustics.ReefEcho-Hydrophone-Segments
ReefEcho Hydrophone Segments
ReefEcho Hydrophone Segments is the curated underwater-audio collection used by the BlueCurrent habitat-monitoring release.
Rights ledger
Asset
Relationship
Canonical license
Reuse statement
Coastal observatory hydrophone archive
direct input
CC-BY-SA-4.0
commercial and noncommercial adaptations and redistribution are permitted with attribution and share-alike
Waveform denoiser
processing utility
AGPL-3.0-only
software… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/ReefEcho-Hydrophone-Segments.ReefSense-DatabaseReefEcho-Habitat-Notes
ReefEcho Habitat Notes
This dataset contains concise habitat observations synthesized from the ReefEcho hydrophone segments.
Source assessment
Component
Relationship
Canonical license
Reuse statement
ReefEcho Hydrophone Segments
direct input
CC-BY-SA-4.0
commercial and noncommercial adaptations and redistribution are permitted with attribution and share-alike
MarineScribe 2B
generation model
MIT
commercial and noncommercial modification and… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/ReefEcho-Habitat-Notes.EDA-US-Bankruptcy-Prediction
Assignment 1: EDA - US Company Bankruptcy Prediction
Student Name: Reef Zehavi
Date: November 10, 2025
📹 Project Presentation Video
[(https://www.loom.com/share/6920e493e8654ef3bb4f67a10eb9b03d)]
1. Overview and Project Goal
The goal of this project is to perform Exploratory Data Analysis (EDA) on a fundamental dataset of American companies. The analysis focuses on understanding the financial characteristics that differentiate between companies that survived… See the full description on the dataset page: https://huggingface.co/datasets/reefzehavi/EDA-US-Bankruptcy-Prediction.tabuhan_island_reef_images
Tabuhan Island Reef Dataset 🌊🐠
This dataset contains underwater reef imagery and 3D point cloud data collected at Tabuhan Island, Indonesia, using a GoPro Hero 13.It is intended for use in marine research, computer vision, and reef monitoring applications (e.g., photogrammetry, habitat mapping, AI-based coral detection).
📂 Dataset Structure
1. raw.zip
Size: ~6590 images
Description:The original raw images captured at the reef site.
Organization:
left/ →… See the full description on the dataset page: https://huggingface.co/datasets/Ramzzz13/tabuhan_island_reef_images.asia-owid-endemic-reef-forming-coral-species
Endemic Reef Forming Coral Species | Asia (Our World in Data)
🌏 49 observations · 49 Asia countries · 2025–2025 · Repackaged by Electric Sheep Asia
TL;DR
This dataset contains 49 observations of Endemic Reef Forming Coral Species data across 49 Asia countries, spanning 2025–2025.
About the source
Source: Our World in Data
Publisher: Our World in Data
License: cc-by-4.0
Topic: Endemic Reef Forming Coral Species
Geographic coverage… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepasia/asia-owid-endemic-reef-forming-coral-species.maritime-reefer-temperature-power-event-coherence-risk-v0.1What this repo is for
Detect spoilage risk before the claim arrives.
You use it to flag
power instability driving excursions
slow alarm response creating compounding risk
temperature drift without clear power events
dwell exposure amplifying small failures
Why it matters
Reefer claims are expensive.
They also damage relationships fast.
imnet1k_coral_reefelsa-frozen-expandedelsa-frozen-testafrica-owid-endemic-reef-forming-coral-species
Endemic Reef Forming Coral Species | Africa (Our World in Data) | Africa (Electric Sheep Africa metadata inventory)
Size category: n<1K - Formats: parquet - Sector: other_unclassified - Engineered by Electric Sheep Africa
TL;DR
This dataset is part of the Electric Sheep Africa catalog on Hugging Face. It is indexed for African data discovery with standardized metadata, loading guidance, provenance notes, and analyst-oriented context.
What This… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-owid-endemic-reef-forming-coral-species.ts-aims-reefscapes-satellite-segmentationeurope-owid-endemic-reef-forming-coral-species
Endemic Reef Forming Coral Species | Europe (Our World in Data)
🇪🇺 44 observations · 44 Europe countries · 2025–2025 · Repackaged by Electric Sheep Europe
TL;DR
This dataset contains 44 observations of Endemic Reef Forming Coral Species data across 44 Europe countries, spanning 2025–2025.
About the source
Source: Our World in Data
Publisher: Our World in Data
License: cc-by-4.0
Topic: Endemic Reef Forming Coral Species
Geographic… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepeurope/europe-owid-endemic-reef-forming-coral-species.Tideglass-Reef-Detections
Tideglass Reef Detections
Tideglass-Reef-Detections provides reviewed detection labels derived from the Tideglass sonar observation workflow. Labels identify candidate reef structures without publishing precise sensitive-site coordinates.
Label fields
Published rows contain a sweep identifier, interval bounds, detection class, reviewer agreement, and confidence band.
Reuse permission docket
Permit reference
Review status
Territory
Asset… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/Tideglass-Reef-Detections.