datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MLLM-Generated-Image-Detection-Dataset
MLLM-Generated Image Dataset
This dataset contains real and AI-generated image samples organized for binary MLLM-generated image detection.
Paper | Code
Dataset Summary
We construct an MLLM-generated image detection benchmark from GPT Image2 and Nano Banana2. This benchmark covers texture-dominated, structure-dominated, and hybrid-dominated. It is designed to evaluate detector performance under the new challenges introduced by large-scale image generation models.… See the full description on the dataset page: https://huggingface.co/datasets/zr-zhang/MLLM-Generated-Image-Detection-Dataset.newimagenet-samples
newimagenet-samples
Rare WordNet vocabulary, with images and the full hypernym chain from each synset up to
entity.n.01.
Where ImageNet covers 1000 common classes, these are the words nobody searches for:
sphacelotheca (a genus of smut fungus), anisogamete, calymmatobacterium,
griseofulvin, merostomata, supinator, wincey.
120 classes · 2,393 images · hierarchy depth 4-16
Sample classes
Every example below shows its complete field set — nothing truncated. All 120… See the full description on the dataset page: https://huggingface.co/datasets/elliot-mllm/newimagenet-samples.MLLMVisualEmotion
MLLMVisualEmotion
Reading Human Emotion Through Machine Eyes
This repository accompanies the study "Reading Human Emotion Through Machine Eyes" on visual emotion recognition by multimodal large language models (MLLMs). It contains the newly constructed FaceEmotion dataset, the preprocessed Emotion6 dataset, outcome tables, source data underlying all figures, and the complete model outputs for all tasks across all scenarios.
Repository Structure… See the full description on the dataset page: https://huggingface.co/datasets/YushuoSu/MLLMVisualEmotion.FLARE-MLLM-2D
FLARE25 Medical Multimodal Dataset
This repository contains a multimodal medical imaging dataset for FLARE 2025 with question-answer pairs across various medical imaging modalities.
Dataset Structure
The dataset is organized into the following main directories:
training/: Training data
validation-public/: Public validation data
validation-hidden/: Hidden validation data (answer not released)
testing/: Hidden testing data (not released)
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/FLARE-MedFM/FLARE-MLLM-2D.FLARE26-MLLM-3D
FLARE 2026: Multimodal Model for 3D Medical Image Parsing
The task is to train one multimodal model for report generation and vision QA.
Data Description
The dataset contains two subsets for abdomen and lung CT report generation and VQA.
FLARE-Task5-MLLM-3D/
├── README.md
├── train # training set
│ ├── CT-AMOS-1290 # source: https://era-ai-biomed.github.io/amos/
│ ├── CT-AMOS-Tr.json
│ ├── CT-RATE-2000 # source:… See the full description on the dataset page: https://huggingface.co/datasets/FLARE-MedFM/FLARE26-MLLM-3D.