datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MarineEVT
MarineEVT Dataset
MarineEVT: Advancing Event-Centric Marine Video Understanding via Visual Tool Reasoning
📖 Description
MarineEVT is a comprehensive event-centric dataset and benchmark for marine video understanding. It comprises 20,000 richly annotated underwater video question-answer pairs spanning 20 fine-grained dimensions, designed to support semantic, contextualized, spatial-temporal, and causal reasoning in marine environments.
The dataset addresses the… See the full description on the dataset page: https://huggingface.co/datasets/AnTo2209/MarineEVT.lattice-marines-wins
Lattice Marines Eternal Ledger
Public vs-AI win book for Lattice Marines.
Hall: https://chatagent.ca/games/lattice-marines/ledger.html
Writer Space: https://huggingface.co/spaces/DeepSeekOracle/lattice-marines-ledger
File: ledger.json
Only adaptive-AI victories are stored (commander name, score, difficulty, map size, seed, turns, AI profile, date). Hot-seat is excluded.
marine-animals-multimodal-dataset
Marine Animals Multimodal Dataset 🐋
A comprehensive multimodal dataset combining audio recordings and images of 32 marine species.
Dataset Summary
Total samples: 24,911
Species: 32
Audio files: 1,357 unique recordings
Images: 581 (309 matched + 272 from iNaturalist)
Features
species (string): Species name
label (int32): Numeric label (0–31)
audio (Audio): Audio recording of the species
image (Image): Species image
image_index (int32): Image number… See the full description on the dataset page: https://huggingface.co/datasets/Hariprasath5128/marine-animals-multimodal-dataset.Open-o3-Video
Open-o3 Video
TL; DR: Open-o3 Video integrates explicit spatio-temporal evidence into video reasoning through curated STGR datasets and a two-stage SFT–RL training strategy, achieving state-of-the-art results on V-STAR and delivering verifiable, reliable reasoning for video understanding.
Data
To provide unified spatio-temporal supervision for grounded video reasoning, we build two datasets: STGR-CoT-30k for supervised fine-tuning and STGR-RL-36k for reinforcement… See the full description on the dataset page: https://huggingface.co/datasets/marinero4972/Open-o3-Video.marine_mammal_labelsmarine-figuresMarineLife-16K
Dataset Card for MarineLife-16K
We introduce MarineLife-16K, a marine-domain video benchmark designed to evaluate the video understanding capabilities of Vision-Language Models (VLMs). MarineLife-16K contains 2,000 video-text pairs and 16,080 video-question-answer pairs across a collection of 2,000 marine videos, including 12,080 multiple-choice questions and 4,000 open-ended questions. The benchmark emphasizes specialized marine knowledge, visual reasoning, temporal… See the full description on the dataset page: https://huggingface.co/datasets/MarineLife-16K/MarineLife-16K.marine-ecological-feature-dataset
Marine Ecological Feature Dataset Assets
"
"This dataset repository contains currently available assets for the typical marine ecological environment "
"feature recognition project.
"
"Current uploaded assets are partial and include GF1/GF2 scene inputs available locally plus the GF6 full-scene "
"inference mask/preview generated by the current prototype model. The normalized manifest workflow is defined "
"in dataset_standard.md; a… See the full description on the dataset page: https://huggingface.co/datasets/cuibinge/marine-ecological-feature-dataset.VideoZeroBench
VideoZeroBench: Probing the Limits of Video MLLMs with Spatio-Temporal Evidence Verification
Project Page: https://marinero4972.github.io/projects/VideoZeroBench/
Dataset Summary
VideoZeroBench is a challenging long-video understanding benchmark designed to evaluate whether video multimodal large language models (Video MLLMs) can not only answer difficult questions, but also identify the precise temporal and spatial evidence that supports their answers.
Unlike standard… See the full description on the dataset page: https://huggingface.co/datasets/marinero4972/VideoZeroBench.watkins-marine-mammal-full-cuts
Watkins Marine Mammal Sound Database
Dataset Description
The Watkins Marine Mammal Sound Database (WMMSD) is one of the largest historical collections of marine mammal vocalizations. It contains 15,248 recordings spanning nearly seven decades from 54 marine mammal species, including whales, dolphins, porpoises, seals, sea lions, manatees, sea otters, and other marine mammals.
This repository provides the complete dataset in a format fully compatible with the… See the full description on the dataset page: https://huggingface.co/datasets/ivangtorre/watkins-marine-mammal-full-cuts.marin-eval-policy-2026-09-24
Marin eval-policy artifacts (2026-09-24)
This directory is the durable publication bundle for the September 2026 Marin
evaluation-policy campaign. It contains the exact launch and serving configs,
score provenance, analysis reports, release tables, figures, and supporting
trace evidence used for the published results.
Contents
ABLATION_CONTEXT_TRACKER.md: context-length ablation, including the
explicitly marked estimate for the final Nemotron Terminal-Bench 2… See the full description on the dataset page: https://huggingface.co/datasets/open-athena/marin-eval-policy-2026-09-24.MarineEval
Dataset Card for MarineEval
The work introduces MarineEval, the first large-scale benchmark specifically designed to evaluate the marine understanding capabilities of Vision-Language Models (VLMs). MarineEval contains 2,000 expert-verified image-based question–answer pairs across 7 task dimensions and 20 domain-specific capacity dimensions, emphasizing specialized marine knowledge, visual reasoning, and real-world complexity. Through comprehensive benchmarking of 17 state-of-the-art… See the full description on the dataset page: https://huggingface.co/datasets/WongYukKwan/MarineEval.CyberV_ASRMARINER
MARINER
A maritime ship object detection dataset with 63 fine-grained ship categories.
Dataset Structure
test/
├── test.json # Bounding box annotations
├── *.jpg # Images (1000 total)
Annotation Format
Each entry in test.json:
{
"image_name": "054A_109.jpg",
"grounding_info": {
"text_prompt": "ship .",
"objects": [
{
"label": "054A",
"score": 0.9291,
"box_norm": [x_min, y_min, x_max… See the full description on the dataset page: https://huggingface.co/datasets/moinn1/MARINER.marine_ocean_mammal_sound
Marine Ocean Mammal Sound Dataset
Sound files on this website are free to download for personal or academic (not commercial) use.
In this database version, the audio archive includes sounds of 32 species:
Atlantic_Spotted_Dolphin
Bearded_Seal
Beluga,_White_Whale
Bottlenose_Dolphin
Bowhead_Whale
Clymene_Dolphin
Common_Dolphin
False_Killer_Whale
Fin,_Finback_Whale
Frasers_Dolphin
Grampus,_Rissos_Dolphin
Harp_Seal
Humpback_Whale
Killer_Whale
Leopard_Seal
Long-Finned_Pilot_Whale… See the full description on the dataset page: https://huggingface.co/datasets/ardavey/marine_ocean_mammal_sound.LLM-Vision-Marine-Animals
Dataset Card for Benchmarking Large Language Models for Image Classification of Marine Mammals
As Artificial Intelligence (AI) has developed rapidly over the past few decades, the new generation of AI, Large Language Models (LLMs) trained on massive datasets, has achieved ground-breaking performance in many applications. Further progress has been made in multimodal LLMs, with many datasets created to evaluate LLMs with vision abilities. However, none of those datasets focuses… See the full description on the dataset page: https://huggingface.co/datasets/yeyimilk/LLM-Vision-Marine-Animals.marin_exp1729__angiosperm_16_genomes__tokenizedMARINER
MARINER
A maritime ship object detection dataset with 63 fine-grained ship categories.
Dataset Structure
test/
├── test.json # Bounding box annotations
├── *.jpg # Images (1000 total)
Annotation Format
Each entry in test.json:
{
"image_name": "054A_109.jpg",
"grounding_info": {
"text_prompt": "ship .",
"objects": [
{
"label": "054A",
"score": 0.9291,
"box_norm": [x_min, y_min, x_max, y_max]… See the full description on the dataset page: https://huggingface.co/datasets/viviwang/MARINER.marine-animals-multimodalvn-provinces-marine-fishing-vessels-ge90cv
Vietnam provinces marine fishing vessels >=90 CV
Number of marine fishing vessels with engine power of 90 CV or more. Coverage 2010-2024. Coastal provinces only. Year 2024 is preliminary. Geographic labels are English (UN/GSO style ASCII romanization). Tables cover provinces, regions and national total where present. Province names follow ar_core.vn_geo (historical 63-province system).
Figures
Hero
Comparison
Color key
Files
provinces (435 rows)… See the full description on the dataset page: https://huggingface.co/datasets/letrinhan/vn-provinces-marine-fishing-vessels-ge90cv.vn-provinces-marine-fishing-vessel-power-ge90cv
Vietnam provinces marine fishing vessel power >=90 CV
Total engine power of marine fishing vessels with 90 CV or more (thousand CV). Coverage 2010-2024. Coastal provinces only. Year 2024 is preliminary. Geographic labels are English (UN/GSO style ASCII romanization). Tables cover provinces, regions and national total where present. Province names follow ar_core.vn_geo (historical 63-province system).
Figures
Hero
Comparison
Color key
Files… See the full description on the dataset page: https://huggingface.co/datasets/letrinhan/vn-provinces-marine-fishing-vessel-power-ge90cv.vn-provinces-marine-fish-capture-production
Vietnam provinces marine fish capture production
Marine fish capture production (thousand tons). Coastal provinces. Coverage varies by year. Year 2024 is preliminary. Geographic labels are English (UN/GSO style ASCII romanization). Tables cover provinces, regions and national total where present. Province names follow ar_core.vn_geo (historical 63-province system).
Figures
Hero
Hero (continued)
Comparison
Color key
Files
provinces (870 rows)… See the full description on the dataset page: https://huggingface.co/datasets/letrinhan/vn-provinces-marine-fish-capture-production.MarineEval
Dataset Card for MarineEval
The work introduces MarineEval, the first large-scale benchmark specifically designed to evaluate the marine understanding capabilities of Vision-Language Models (VLMs). MarineEval contains 2,000 expert-verified image-based question–answer pairs across 7 task dimensions and 20 domain-specific capacity dimensions, emphasizing specialized marine knowledge, visual reasoning, and real-world complexity. Through comprehensive benchmarking of 17 state-of-the-art… See the full description on the dataset page: https://huggingface.co/datasets/Animas1024/MarineEval.RVC_datasetsmarine_vla_dataset
Marine VLA Dataset
Vision-Language-Action dataset for autonomous marine vessel navigation using SmolVLA.
Dataset Structure
LeRobot-style format with 37 episodes, 12175 frames:
data/
episode_000000/
episode_data.json # frame-by-frame labels + metadata
observation.images.camera_0/
000000.jpg # 640x480 RGB frames
000001.jpg
...
episode_000001/
...
dataset_info.json # schema, label names, stats… See the full description on the dataset page: https://huggingface.co/datasets/MSaalaamaa/marine_vla_dataset.TW_Marine_2cls_datasetts-aims-reefscapes-marine-featuresTunisian-Mediterranean-Coastal-Marine-Litter-Trash-Dataset-Drone-Based-Real-World-ImagesMariner-Photo-Register
Mariner Photo Register
A descriptive register for scanned photographs from the Mariner waterfront collections.
Mariner publication register
Accession
Images
Publish
Rights
Shelf
Collection
MPR/041
160
release
cleared
Gallery
Harbor Works
MPR/118
190
RELEASE
partner permission
reading room
Island Ferries
MPR/330
240
release
cleared
gallery
Harbor Works
MPR/207
155
release
cleared
Reading Room
Lighthouse Files
MPR/012
75
release
onsite
Vault… See the full description on the dataset page: https://huggingface.co/datasets/SOTAagi2030/Mariner-Photo-Register.TypeCare-Datasets