datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
safedocs-1M-muse-spark-1.3-judged
SafeDocs: Muse Spark 1.3 judge annotations
Incrementally published, one complete shard per commit. All original source columns,
images, complete Paddle JSON, rows and row order are preserved. No language or quality
filtering. New columns: judge_verdict (PERFECT/ERROR), judge_reason, judge_status,
and judge_error. Operational failures retain the original page with a null verdict
and reason, status failed, and a diagnostic in judge_error; they are not OCR ERRORs.
Direct Meta API… See the full description on the dataset page: https://huggingface.co/datasets/albertklorer/safedocs-1M-muse-spark-1.3-judged.MUSE-News
MUSE-News
MUSE is a comprehensive machine unlearning evaluation benchmark that assesses six key properties for unlearned models: (1) no verbatim memorization, (2) no knowledge memorization, (3) no privacy leakage, (4) utility preservation on data not intended for removal, (5) scalability with respect to the size of removal requests, and (6) sustainability over sequential unlearning requests. MUSE focuses on two types of textual data that commonly require unlearning: news articles… See the full description on the dataset page: https://huggingface.co/datasets/muse-bench/MUSE-News.MUSE-Books
MUSE-Books
MUSE is a comprehensive machine unlearning evaluation benchmark that assesses six key properties for unlearned models: (1) no verbatim memorization, (2) no knowledge memorization, (3) no privacy leakage, (4) utility preservation on data not intended for removal, (5) scalability with respect to the size of removal requests, and (6) sustainability over sequential unlearning requests. MUSE focuses on two types of textual data that commonly require unlearning: news articles… See the full description on the dataset page: https://huggingface.co/datasets/muse-bench/MUSE-Books.MPIE-Bench
MPIE-Bench
Official 2,500-sample test set for multi-person interaction-aware image editing evaluation.
GitHub (code + protocol): https://github.com/AnnLin0628/mpie-bench
Org: muset-ai
Dataset Viewer
The default config (default / test) is one row per evaluation sample:
Column
Meaning
cat
Interaction category (filter / group by this)
gt
Held-out ground-truth image
prompt
Edit instruction
ref_paths
Reference image paths under images/
sample_id… See the full description on the dataset page: https://huggingface.co/datasets/muset-ai/MPIE-Bench.met_museumMUSE-VA
MUSE-VA Dataset
English | 中文
MUSE-VA (Multimodal MUSic Emotion Dataset with Balanced VA) is a large-scale multimodal music emotion dataset designed for music emotion understanding, emotion-controllable music generation, and cross-modal affective modeling. The dataset starts from target coordinates sampled in the continuous Valence-Arousal (VA) space and uses a five-stage LLM agent pipeline with affective and musical knowledge injection to construct music, text, images, and… See the full description on the dataset page: https://huggingface.co/datasets/jiahaomei/MUSE-VA.MUSE-VA-A2IMUSE-VA-I2AMuse-Glimmer-SWE-Gym-2k
Muse-Glimmer-SWE-Gym-2k
Agentic coding traces from meta-models/Muse-Glimmer-30B, recorded for training a
speculative-decoding drafter. 1,981 mini-swe-agent trajectories over SWE-Gym and
SWE-bench-extra instances, and the 159,999 individual chat-completion calls behind them.
Configs
Config
Rows
Size
What it is
train
1,981
57 MB
One row per trajectory: the full conversation as messages.
raw
159,999
2.7 GB
One row per recorded API call: request and… See the full description on the dataset page: https://huggingface.co/datasets/Satgoy152/Muse-Glimmer-SWE-Gym-2k.MUSES
MUSES: a benchmark for Marked Unevenly Spaced Event Sequences
MUSES, a benchmark for Marked Unevenly Spaced Event Sequences, is a collection of unevenly spaced time series datasets from various domains, containing marked events for training and evaluating prediction approaches.
Languages
All the columns and classes (when textual) in MUSES are in English (BCP-47 en)
Dataset Structure
Data Instances
All datasets formatting follows… See the full description on the dataset page: https://huggingface.co/datasets/ddrg/MUSES.parallel-sentences-muse
Dataset Card for Parallel Sentences - MUSE
This dataset contains parallel sentences (i.e. English sentence + the same sentences in another language) for numerous other languages. Most of the sentences originate from the OPUS website.
In particular, this dataset contains the MUSE dataset.
Related Datasets
The following datasets are also a part of the Parallel Sentences collection:
parallel-sentences-europarl
parallel-sentences-global-voices
parallel-sentences-muse… See the full description on the dataset page: https://huggingface.co/datasets/sentence-transformers/parallel-sentences-muse.UncertaintyGym
UncertaintyGym
A Standardized Benchmark for LLM Epistemic Calibration & Uncertainty Expression
Abstract
UncertaintyGym evaluates whether language models recognize the boundaries of their knowledge. Rather than assessing purely factual recall, UncertaintyGym measures how reliably an LLM explicitly declares uncertainty ("I don't know"), requests necessary disambiguating context, and rejects false premises without hallucinating.
Benchmark Taxonomy… See the full description on the dataset page: https://huggingface.co/datasets/Muse-Ltd/UncertaintyGym.safedocs-1M-muse-spark-1.3-first3
SafeDocs first three shards: Muse Spark 1.3
Source: albertklorer/safedocs-1M, revision 87faff9053aa50c745f1359bef3592219ccb8c8b.
PaddleOCR-VL 1.6 teacher labels are compared to original pages using the existing side-by-side renderer and binary Muse Spark 1.3 contributor judge. Quality verdicts are only PERFECT or ERROR, with no quality reason. Operational failures have no verdict. These are model labels, not human ground truth. Native Paddle block list order is preserved. A… See the full description on the dataset page: https://huggingface.co/datasets/albertklorer/safedocs-1M-muse-spark-1.3-first3.MUSE-VA-A2IMUSE-News-Train
MUSE-News-Train
This dataset is a simple merger of the pretraining data from the original MUSE-News dataset.
Dataset Details
Dataset Sources [optional]
Repository: https://huggingface.co/datasets/muse-bench/MUSE-News
Paper: https://arxiv.org/pdf/2407.06460
Dataset Creation
To create this dataset, we simply started from the muse-bench dataset and selected the train subset.
Then, by merging the retain1 and retain2 splits we get the actual retain… See the full description on the dataset page: https://huggingface.co/datasets/tamarsonha/MUSE-News-Train.MUSE-Books-Train
MUSE-Books-Train
This dataset is a simple merger of the pretraining data from the original MUSE-Books dataset.
Dataset Details
Dataset Sources [optional]
Repository: https://huggingface.co/datasets/muse-bench/MUSE-Books
Paper: https://arxiv.org/pdf/2407.06460
Dataset Creation
To create this dataset, we simply started from the muse-bench dataset and selected the train subset.
Then, by merging the retain1 and retain2 splits we get the actual… See the full description on the dataset page: https://huggingface.co/datasets/tamarsonha/MUSE-Books-Train.Muse-Glimmer-Terminal-Bench-Eval
Muse Glimmer — Terminal-Bench eval traces
This is an EVALUATION set. Do not train on it.
These traces measure baseline speculative-decoding behaviour (acceptance length,
draft acceptance rate, throughput) for the Muse Glimmer speculator project on
agentic coding work, so that a fine-tuned speculator can be compared against
them later. Training data for that project is SWE-Gym and is deliberately
repo-disjoint from Terminal-Bench: it excludes every repository referenced by
any of… See the full description on the dataset page: https://huggingface.co/datasets/Satgoy152/Muse-Glimmer-Terminal-Bench-Eval.MUSE-Bench
MUSE-Bench: Memory Utilization Evaluation Benchmark
Official dataset for the paper "Beyond Memorization: Benchmarking Memory
Utilization in Conversational LLM Agents."
Anonymous release. This repository is an anonymized copy provided for
double-blind peer review. Author and affiliation information is withheld
until the review process concludes.
Motivation
LLM agents increasingly rely on persistent cross-session memory to support
long-horizon and personalized… See the full description on the dataset page: https://huggingface.co/datasets/anonymous111111111/MUSE-Bench.muse_textbooks_debate_onlymet-museum-no-imagesLiterally @miccull's dataset minus the images
Original source is the Met museum's open dataset of public domain works in their collection.
https://console.cloud.google.com/marketplace/product/the-metropolitan-museum-of-art/the-met-public-domain-art-works?project=smartmaps-423802
aave_transactions_blockchain_copychart-museum-samplesMUSE-News-2parallel-muse-deduplicatedmuse_mucodec_chordmuse-sarcasm-explanation
MuSe: Multimodal Sarcasm Explanation (Reformatted)
This repository provides a Hugging Face-compatible version of the MuSe (MORE) dataset.
Modifications in this version
To make the dataset easier to use with the datasets library, the following changes were made:
Unified Schema: Merged separate OCR and Non-OCR files into a single test split.
Metadata Flags: Added an is_ocr (boolean) column to distinguish between image types.
Image Integration: Converted image paths into a… See the full description on the dataset page: https://huggingface.co/datasets/alita9/muse-sarcasm-explanation.6971f242
Dataset Card for "6971f242"
More Information needed
mnist_muse2
Dataset Card for "mnist_muse2"
More Information needed
muse512
Dataset Card for "muse_512"
```py
from PIL import Image
import torch
from muse import PipelineMuse, MaskGiTUViT
from datasets import Dataset, Features
from datasets import Image as ImageFeature
from datasets import Value, load_dataset
device = "cuda" if torch.cuda.is_available() else "cpu"
pipe = PipelineMuse.from_pretrained(
transformer_path="valhalla/research-run",
text_encoder_path="openMUSE/clip-vit-large-patch14-text-enc"… See the full description on the dataset page: https://huggingface.co/datasets/diffusers-parti-prompts/muse512.africa-cote-d-ivoire-personnel-travaillant-dans-les-musees-et-institutions-assi-cb8f3e48
Personnel Travaillant Dans Les Musees Et Institutions Assi | Africa (Cote d'Ivoire DataFair)
125 rows - 1 Africa country/area - 2003-2006 - 1 indicator - Engineered by Electric Sheep Africa
TL;DR
This dataset contains 125 rows from Cote d'Ivoire DataFair, covering Personnel Travaillant Dans Les Musees Et Institutions Assi. It is published as ML-ready Parquet with consistent Hugging Face metadata, source provenance, and analysis-friendly loading examples.… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-cote-d-ivoire-personnel-travaillant-dans-les-musees-et-institutions-assi-cb8f3e48.