datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Multitask-National-Speech-Corpus-v1Multitask-National-Speech-Corpus (MNSC v1) is derived from IMDA's NSC Corpus.
MNSC is a multitask speech understanding dataset derived and further annotated from IMDA NSC Corpus. It focuses on the knowledge of Singapore's local accent, localised terms, and code-switching.
ASR: Automatic Speech Recognition
SQA: Speech Question Answering
SDS: Spoken Dialogue Summarization
PQA: Paralinguistic Question Answering
from datasets import load_dataset
data =… See the full description on the dataset page: https://huggingface.co/datasets/MERaLiON/Multitask-National-Speech-Corpus-v1.multi_task_multi_modal_knowledge_retrieval_benchmark_M2KR
PreFLMR M2KR Dataset Card
Dataset details
Dataset type:
M2KR is a benchmark dataset for multimodal knowledge retrieval. It contains a collection of tasks and datasets for training and evaluating multimodal knowledge retrieval models.
We pre-process the datasets into a uniform format and write several task-specific prompting instructions for each dataset. The details of the instruction can be found in the paper. The M2KR benchmark contains three types of tasks:… See the full description on the dataset page: https://huggingface.co/datasets/BByrneLab/multi_task_multi_modal_knowledge_retrieval_benchmark_M2KR.Multitask-National-Speech-Corpus-v1-extendmulti_task_multi_modal_knowledge_retrieval_benchmark_M2KR_CN
PreFLMR M2KR Dataset Card
Dataset details
Dataset type:
M2KR is a benchmark dataset for multimodal knowledge retrieval. It contains a collection of tasks and datasets for training and evaluating multimodal knowledge retrieval models.
We pre-process the datasets into a uniform format and write several task-specific prompting instructions for each dataset. The details of the instruction can be found in the paper. The M2KR benchmark contains three types of tasks:… See the full description on the dataset page: https://huggingface.co/datasets/BByrneLab/multi_task_multi_modal_knowledge_retrieval_benchmark_M2KR_CN.multitask_german_examples_32kmulti_task_multi_modal_knowledge_retrieval_benchmark_M2KRmultitask_vqa_benchmarkThis dataset is a part of .
🍈 MMT-47: Multimodal Multi-Task Benchmark
47 Tasks · 7 Categories · 3 Modalities (Image, Video, Text)
Cite our ICML-2026 paper for this dataset
@article{kowsher2026lime,
title={LiME: Lightweight Mixture of Experts for Efficient Multimodal Multi-task Learning},
author={Kowsher, Md and Mansoor, Haris and Prottasha, Nusrat Jahan and Garibay, Ozlem and Zhu, Victor and Ji, Zhengping and Chen, Chen},
journal={arXiv preprint… See the full description on the dataset page: https://huggingface.co/datasets/Kowsher/multitask_vqa_benchmark.VQA-MultiTaskSampling-Multitask-National-Speech-Corpus-v1
Sampling Multitask-National-Speech-Corpus-v1
Original dataset from https://huggingface.co/datasets/MERaLiON/Multitask-National-Speech-Corpus-v1, we only take Part 3 and do sampling.
how to prepare the dataset
huggingface-cli download \
mesolitica/Sampling-Multitask-National-Speech-Corpus-v1 \
--include "*.zip" \
--repo-type "dataset" \
--local-dir './'
wget… See the full description on the dataset page: https://huggingface.co/datasets/mesolitica/Sampling-Multitask-National-Speech-Corpus-v1.dynasample_multitasks_cleanconfigurable-system-prompt-multitask
Configurable System Prompt Multi-task Dataset 🛞
We release the synthetic dataset for the multi-task experiments from the paper "Configurable Safety Tuning of Language Models with Synthetic Preference Data", https://huggingface.co/papers/2404.00495. This dataset has two sources for the examples:
Self-critique on a safety task from Harmful Behaviours, using the SOLAR-Instruct model. It employs two system prompts to learn the different behaviors:
You are a helpful yet harmless… See the full description on the dataset page: https://huggingface.co/datasets/vicgalle/configurable-system-prompt-multitask.multi-task-dataset
Multi-Task Dataset
Description
A large-scale multi-task dataset designed for training and evaluating AI models across reasoning, mathematics, code, research, verification, data analysis, and general problem solving.
Content
100,000,001 examples
20+ task families
English + French
Train / Validation / Test splits
Structured reasoning and verification signals
Multiple difficulty levels
OOD and generalization-oriented examples
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/Lelonthecodeur/multi-task-dataset.sanskrit-multitask-devanagarimultilingual-multitask-refusal
Multilingual Multitask Refusal
A multilingual, multitask prompt dataset for analysing refusal behaviour across language, task wrapper, and harmful / harmless labels.
English seed spans and labels come from the previous Multitask Multilingual Refusal dataset. Non-English content is produced with Google Sheets GOOGLETRANSLATE. Task instructions are language-localized via templates_localized.json.
Rows
211,320
English seeds
1,761
Languages
15
Tasks
8
Product
1… See the full description on the dataset page: https://huggingface.co/datasets/kaustubhg73/multilingual-multitask-refusal.sanskrit-multitaskMultitask_Preplay_JaxMaze_models
Multitask Preplay — jaxmaze model data
Data for the paper "Multitask Preplay" (PNAS).
Analysis code: https://github.com/wcarvalho/multitask_preplay (branch pnas).
Splits: qlearning, usfa, dyna, preplay, her, bfs, dfs, greedy_euclidean, memory_based_euclidean.
Each split is also available as a top-level parquet file.
MultitaskPreplay_jaxmaze_human_dftmp-multitask-en-clinical
Dataset Card for "tmp-multitask-en-clinical"
More Information needed
MultiTasks-v2
MultiTasks-v2
MultiTasks-v2 is a collection of twelve multimodal benchmark subsets normalized
into a shared image-question-answer format. Each subset provides a train split
and a test split.
Dataset Structure
Each example contains:
id: a unique sample identifier in the form {Dataset}_{split}_{index}.
images: a list containing one image. Images are stored as JPEG bytes.
problem: the prompt shown to the model.
answer: the target answer.
RefAdv is the only subset… See the full description on the dataset page: https://huggingface.co/datasets/LoserLi/MultiTasks-v2.Multitask_Preplay_Craftax_models
Multitask Preplay — craftax model data
Data for the paper "Multitask Preplay" (PNAS).
Analysis code: https://github.com/wcarvalho/multitask_preplay (branch pnas).
Splits: qlearning, usfa, dyna, preplay, her, greedy_euclidean, memory_based_euclidean.
Each split is also available as a top-level parquet file.
UR7e_CaP_MultiTask_300epi_10fps
UR7e CaP MultiTask 300epi 10fps
This is a public LeRobot v3.0 multi-task dataset for UR7e Code-as-Policies manipulation. It merges three 100-episode 10fps datasets into one 300-episode training corpus while preserving the original numeric observations, actions, and task labels. The final uploaded copy stores all videos as 10fps H.264 MP4 files.
Dataset Summary
Robot: ur7e
Format: LeRobot v3.0
FPS: 10
Episodes: 300
Frames: 152,926
Cameras:… See the full description on the dataset page: https://huggingface.co/datasets/Cache-SCA/UR7e_CaP_MultiTask_300epi_10fps.multitask_v1_toySR-MultiTaskmultitask_v4_codingmultitask_textqa_benchmarkThis dataset is a part of .
🍈 MMT-47: Multimodal Multi-Task Benchmark
47 Tasks · 7 Categories · 3 Modalities (Image, Video, Text)
Cite our ICML-2026 paper for this dataset
@article{kowsher2026lime,
title={LiME: Lightweight Mixture of Experts for Efficient Multimodal Multi-task Learning},
author={Kowsher, Md and Mansoor, Haris and Prottasha, Nusrat Jahan and Garibay, Ozlem and Zhu, Victor and Ji, Zhengping and Chen, Chen},
journal={arXiv preprint… See the full description on the dataset page: https://huggingface.co/datasets/Kowsher/multitask_textqa_benchmark.UR7e_CaP_MultiTask_300epi_10fps_state_tplus1_action
UR7e CaP MultiTask 300epi 10fps
This is a public LeRobot v3.0 multi-task dataset for UR7e Code-as-Policies manipulation. It merges three 100-episode 10fps datasets into one 300-episode training corpus while preserving the original numeric observations, actions, and task labels. The final uploaded copy stores all videos as 10fps H.264 MP4 files.
Dataset Summary
Robot: ur7e
Format: LeRobot v3.0
FPS: 10
Episodes: 300
Frames: 152,926
Cameras:… See the full description on the dataset page: https://huggingface.co/datasets/Cache-SCA/UR7e_CaP_MultiTask_300epi_10fps_state_tplus1_action.multitask_v3_coding_11ksatquery-multitask-datasetru-multitask-toxicity
RU Multi-Task Toxicity Dataset
Описание
Датасет для обучения multi-task классификатора токсичности русскоязычных
текстов по трём независимым бинарным категориям:
profanity — ненормативная лексика (мат)
threat — угрозы в адрес пользователя/третьих лиц
illegal — запросы, связанные с нарушением закона
(например, "как сделать бомбу", "где купить наркотики")
Каждый текст может относиться сразу к нескольким категориям одновременно
(multi-label задача), либо ни к одной… See the full description on the dataset page: https://huggingface.co/datasets/AtesiT/ru-multitask-toxicity.tmp-multitask-fr-clinical
Dataset Card for "tmp-multitask-fr-clinical"
More Information needed