datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
iclr-wm-backup-public
ICLR Watermark Benchmark — backup overflow (public part)
Companion to the private repo Aak975/iclr-wm-backup, which reached its
storage quota. Together the two repos form ONE backup — every file exists in
exactly one of them, with the same layout:
archives/<sub>/part-0000 ... part-NNNN, MANIFEST.json
restore one archive: cat part-* | zstd -d | tar -x
MANIFEST.json = {"parts": N, "sha256": <whole-stream>, "total_bytes": M}
This public part holds only shareable image data… See the full description on the dataset page: https://huggingface.co/datasets/Aak975/iclr-wm-backup-public.cybergymict_s2s_refactoredicrm-hitek-full-db-mixed
ICMR + HITEK Full DB (Mixed) — Prebuilt Indexes + One-Click Setup
Prebuilt sorted indexes for the Kzr0xx/Icmr-and-hitek dataset (2.5B rows, 11 columns, ~104 GB raw parquet).
Building these indexes took ~17 hours of compute. This repo saves you that work: download + run = API live in ~1-2 hours (download speed dependent).
Contents
The indexes are stored as sorted parts (each < 50 GB, split at row-group boundaries, order preserved) because HuggingFace's classic HTTP… See the full description on the dataset page: https://huggingface.co/datasets/MRSHREY197/icrm-hitek-full-db-mixed.ICMR-HITEK-FULL-MIXED-DBInstructS2S-200K
InstructS2S-200K
Dataset Description
InstructS2S-200K is a multi-turn speech-to-speech conversation dataset containing approximately 200,000 dialogues, developed for the LLaMA-Omni and LLaMA-Omni 2 research projects on real-time spoken chatbots.
Usage
The dataset is split into multiple parts and needs to be reconstructed:
# Combine the parts and extract
cat en_part_* > instructs2s_200k.tar.gz
tar -xzf instructs2s_200k.tar.gz
License
This dataset is… See the full description on the dataset page: https://huggingface.co/datasets/ICTNLP/InstructS2S-200K.IC-VCO-Dataset
IC-VCO-Dataset
This dataset package contains the two IC-VCO training subsets:
sft: supervised fine-tuning examples.
preference: visual contrastive preference examples.
The two subsets intentionally use different schemas, so they are represented as separate Hugging Face dataset configurations instead of separate splits under a single configuration. Each configuration has a train split.
Loading From Hugging Face… See the full description on the dataset page: https://huggingface.co/datasets/OPPOer/IC-VCO-Dataset.llava_v1_5_mix665k
LLaVA v1.5 Mix 665K Dataset
This dataset contains 665,298 multimodal instruction-following samples used for fine-tuning the LLaVA v1.5 model.
Dataset Structure
id: Unique identifier for the sample
model: Model name (if applicable)
conversations: JSON string containing conversation turns in original format
image: List of PIL Image objects (embedded in parquet)
image_path: List of strings containing original relative paths to images
Load the Dataset
from… See the full description on the dataset page: https://huggingface.co/datasets/Icey444/llava_v1_5_mix665k.ChatDoctor-iCliniq
Dataset Card for "ChatDoctor-iCliniq"
More Information needed
icrm-hitek-fulldbicrm-hitek-fulldbTreeUQ
TreeUQ — Geographically-Tiled Bavaria EO Benchmark
TreeUQ is a large-scale Earth observation benchmark for tree species mapping and tree structure estimation (height, count, density, variance) over Bavaria, Germany.
Each 128×128 pixel patch (10 m resolution, EPSG:25832) contains:
Sentinel-2 — 4 seasonal composites (spring/summer/autumn/winter 2025), 10 bands
Sentinel-1 GRD — 4 seasonal composites, VV + VH polarisation, linear gamma-0
Tree species raster — Bavarian species… See the full description on the dataset page: https://huggingface.co/datasets/iclr2027kiwi/TreeUQ.icmr-parquetIconStack-48M-Rendered-Trainicrm-hitek-full-db-mixed
ICMR + HITEK Full DB (Mixed) — Prebuilt Indexes + One-Click Setup
Prebuilt sorted indexes for the Kzr0xx/Icmr-and-hitek dataset (2.5B rows, 11 columns, ~104 GB raw parquet).
Building these indexes took ~17 hours of compute. This repo saves you that work: download + run = API live in ~1-2 hours (download speed dependent).
Contents
The indexes are stored as sorted parts (each < 50 GB, split at row-group boundaries, order preserved) because HuggingFace's classic HTTP… See the full description on the dataset page: https://huggingface.co/datasets/sauravsingh2111/icrm-hitek-full-db-mixed.icl-dataset-joint-space
icl-dataset-fixed-action
Derived from adityx23/icl-dataset
(lerobot v2.1 format). Every existing column, task, episode flag
(success/valid/keep), and episode_uid is carried through unchanged.
What's added
One new feature, action.q_target (float32, shape [14], names
lj0..lj6, rj0..rj6): the joint-space reconstruction of each frame's
action.left_ee / action.right_ee cartesian targets, via the mink-based
IK procedure documented in vr_teleop_ik.md (available in the… See the full description on the dataset page: https://huggingface.co/datasets/Hannibal52Barca/icl-dataset-joint-space.icl-dataset
ICL Dataset
Bimanual robot manipulation data collected by human teleoperation on the YOR robot,
in LeRobot v2.1 format.
3,149 episodes · 2,385,871 frames · 22.1 hours · 36 tasks · 3 camera views
Each episode is a single teleoperated attempt at a manipulation task, recorded at 30 fps
with synchronised proprioception, commanded actions and three camera streams.
⚠️ Read this before training on it
This dataset contains failures, trimmed successes and unusable… See the full description on the dataset page: https://huggingface.co/datasets/adityx23/icl-dataset.ts-icl-pretraining-corpus
TS-ICL Pretraining Corpus (community reconstruction)
A unified, cleaned reconstruction of the univariate pretraining corpus described in
Table 5 of TS-ICL: A Flexible Time-Indexed Foundation Model for Time Series via
In-Context Learning (Le Naour, Nabil & Petralia, EDF R&D; arXiv:2606.05878). The TS-ICL
authors did not release their pretraining data pipeline, so this corpus is rebuilt from the
named upstream sources (LOTSA, Chronos, and the TempoPFN synthetic generators) and… See the full description on the dataset page: https://huggingface.co/datasets/JuaAI/ts-icl-pretraining-corpus.ICDAR2019-SROIE
ICDAR2019's Scanned Receipts OCR and Information Extraction (SROIE)
The ICDAR2019 SROIE dataset was originally published by Huang et al. for the
15th International Conference on Document Analysis and Recognition (ICDAR2019)
Robust Reading Challenge on Scanned Receipts OCR and Information Extraction
(SROIE).
This work presents an extension of the original ICDAR2019 SROIE dataset, including 14
receipt annotations missing from the original Task 3 test dataset, in a format
integrated… See the full description on the dataset page: https://huggingface.co/datasets/jsdnrs/ICDAR2019-SROIE.icl-dataset-end-effector-space
icl-dataset-fixed-obs
Derived from adityx23/icl-dataset
(lerobot v2.1 format). Every existing column, task, episode flag
(success/valid/keep), and episode_uid is carried through unchanged.
What's added
Two new features, observation.left_ee / observation.right_ee (float32,
shape [7], names qw, qx, qy, qz, x, y, z): the Cartesian end-effector
pose of each arm, forward-kinematics'd from that frame's recorded
observation.state (real joint encoders) through the same… See the full description on the dataset page: https://huggingface.co/datasets/Hannibal52Barca/icl-dataset-end-effector-space.IconArt🖼️ The dataset IconArt dataset was introduced in the following paper : "Weakly Supervised Object Detection in Artworks" Gonthier et al. ECCV 2018 Workshop Computer Vision for Art Analysis - VISART 2018.
This datasest is designed to evaluate Weakly Supervised object detection methods in paintings.
You can also find project page for the paper here.
This dataset contains 5955 images (from WikiCommons) : a train set of 2978 images and a test set of 2977 images (for classification task). 1480 of… See the full description on the dataset page: https://huggingface.co/datasets/NGonthier/IconArt.icrm-hitek-full-db-mixed
ICMR + HITEK Full DB (Mixed) — Prebuilt Indexes + One-Click Setup
Prebuilt sorted indexes for the Kzr0xx/Icmr-and-hitek dataset (2.5B rows, 11 columns, ~104 GB raw parquet).
Building these indexes took ~17 hours of compute. This repo saves you that work: download + run = API live in ~1-2 hours (download speed dependent).
Contents
The indexes are stored as sorted parts (each < 50 GB, split at row-group boundaries, order preserved) because HuggingFace's classic HTTP… See the full description on the dataset page: https://huggingface.co/datasets/sckeptic/icrm-hitek-full-db-mixed.x_dataset_53985
Bittensor Subnet 13 X (Twitter) Dataset
Dataset Summary
This dataset is part of the Bittensor Subnet 13 decentralized network, containing preprocessed data from X (formerly Twitter). The data is continuously updated by network miners, providing a real-time stream of tweets for various analytical and machine learning tasks.
For more information about the dataset, please visit the official repository.
Supported Tasks
The versatility of this… See the full description on the dataset page: https://huggingface.co/datasets/icedwind/x_dataset_53985.icl-demo-dataset
ICL Demo Dataset
A small, recent collection of bimanual manipulation demonstrations on the YOR
robot, in LeRobot v2.1 format. Recorded 3 and 6 August 2026.
285 episodes · 254,171 frames · 2.35 hours · 27 tasks · 3 camera views
This is a companion to adityx23/icl-dataset
— same robot, same schema, same task vocabulary, but a separate and much smaller
collection. Roughly ten demonstrations per task, several of them tasks that do not
appear in the larger dataset at all.… See the full description on the dataset page: https://huggingface.co/datasets/adityx23/icl-demo-dataset.icrm-hitek-full-db-mixed
ICMR + HITEK Full DB (Mixed) — Prebuilt Indexes + One-Click Setup
Prebuilt sorted indexes for the Kzr0xx/Icmr-and-hitek dataset (2.5B rows, 11 columns, ~104 GB raw parquet).
Building these indexes took ~17 hours of compute. This repo saves you that work: download + run = API live in ~1-2 hours (download speed dependent).
Contents
The indexes are stored as sorted parts (each < 50 GB, split at row-group boundaries, order preserved) because HuggingFace's classic HTTP… See the full description on the dataset page: https://huggingface.co/datasets/Craige113/icrm-hitek-full-db-mixed.x_dataset_34576
Bittensor Subnet 13 X (Twitter) Dataset
Dataset Summary
This dataset is part of the Bittensor Subnet 13 decentralized network, containing preprocessed data from X (formerly Twitter). The data is continuously updated by network miners, providing a real-time stream of tweets for various analytical and machine learning tasks.
For more information about the dataset, please visit the official repository.
Supported Tasks
The versatility of this… See the full description on the dataset page: https://huggingface.co/datasets/icedwind/x_dataset_34576.PsychEval
PsychEval: A Multi-Session and Multi-Therapy Benchmark for High-Realism AI Psychological Counselor
PsychEval is a comprehensive benchmark designed to evaluate Large Language Models (LLMs) in the context of psychological counseling. Unlike existing benchmarks that focus on single-turn interactions or single-session assessments, PsychEval emphasizes longitudinal, multi-session counseling processes and multi-therapy capabilities.
🌟 Key Features
Multi-Session… See the full description on the dataset page: https://huggingface.co/datasets/ecnu-icalk/PsychEval.Icarus-dataset
Icarus
A unified multi-modal curriculum dataset for evolutionary neural architecture search. Every row is one self-contained Task = {meta, support, query}, where support and query are lists of (input_Field, output_Field) pairs. The inner loop trains on support; fitness is scored on query. Support is non-empty for every task. Encoders read the Field descriptor (axes, value_type, n_classes, value_range, mask); mask is True where a value is padding/ignored. meta.class_names, when… See the full description on the dataset page: https://huggingface.co/datasets/Ardea/Icarus-dataset.ICPC_Data
ICPC World Finals — a discriminative subset, with model traces
24 ICPC World Finals problems (2021–2025), together with the full transcripts of an
LLM attempting each of them three times under simulated contest rules.
Selection
The model
Every run in this dataset comes from:
nvidia/Nemotron-Cascade-2-30B-A3B
The partitions
Every one of the 53 problems was run 3 times (seeds 1, 2, 3). Each problem was then
placed by its pass rate and… See the full description on the dataset page: https://huggingface.co/datasets/xupy21/ICPC_Data.Sombench-Ice-Prospectivity-Regression
SomBench Benchmark: Polar Ice Prospectivity Regression
Science theme: Polar volatiles
Task: Regression
Dataset Summary
A polar, multi-layer benchmark for predicting near-surface water-ice
prospectivity within ~10° latitude of each pole at 240 m/pixel. Following
the ice-prospectivity workflow of Coyan et al. (2025), the dataset includes a
group of physically motivated evidential layers (thermophysical,
illumination, and terrain) alongside a continuous prospectivity… See the full description on the dataset page: https://huggingface.co/datasets/nasa-ibm-ai4science/Sombench-Ice-Prospectivity-Regression.