datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Cambrian-10M
Cambrian-10M Dataset
Please see paper & website for more information:
https://cambrian-mllm.github.io/
https://arxiv.org/abs/2406.16860
Overview
Cambrian-10M is a comprehensive dataset designed for instruction tuning, particularly in multimodal settings involving visual interaction data. The dataset is crafted to address the scarcity of high-quality multimodal instruction-tuning data and to maintain the language abilities of multimodal large language models (LLMs).… See the full description on the dataset page: https://huggingface.co/datasets/nyu-visionx/Cambrian-10M.xcopa
Dataset Card for "xcopa"
Dataset Summary
XCOPA: A Multilingual Dataset for Causal Commonsense Reasoning
The Cross-lingual Choice of Plausible Alternatives dataset is a benchmark to evaluate the ability of machine learning models to transfer commonsense reasoning across
languages. The dataset is the translation and reannotation of the English COPA (Roemmele et al. 2011) and covers 11 languages from 11 families and several areas around
the globe. The dataset is… See the full description on the dataset page: https://huggingface.co/datasets/cambridgeltl/xcopa.Cambrian-Alignment
Cambrian-Alignment Dataset
Please see paper & website for more information:
https://cambrian-mllm.github.io/
https://arxiv.org/abs/2406.16860
Overview
Cambrian-Alignment is an question-answering alignment dataset comprised of alignment data from LLaVA, Mini-Gemini, Allava, and ShareGPT4V.
Getting Started with Cambrian Alignment Data
Before you start, ensure you have sufficient storage space to download and process the data.
Download the Data Repository… See the full description on the dataset page: https://huggingface.co/datasets/nyu-visionx/Cambrian-Alignment.cbt
Dataset Card for CBT
Dataset Summary
The Children’s Book Test (CBT) is designed to measure directly how well language models can exploit wider linguistic context. The CBT is built from books that are freely available.
This dataset contains four different configurations:
V: where the answers to the questions are verbs.
P: where the answers to the questions are pronouns.
NE: where the answers to the questions are named entities.
CN: where the answers to the questions are… See the full description on the dataset page: https://huggingface.co/datasets/cam-cst/cbt.vsr_random
VSR: Visual Spatial Reasoning
This is the random set of VSR: Visual Spatial Reasoning (TACL 2023) [paper].
Usage
from datasets import load_dataset
data_files = {"train": "train.jsonl", "dev": "dev.jsonl", "test": "test.jsonl"}
dataset = load_dataset("cambridgeltl/vsr_random", data_files=data_files)
Note that the image files still need to be downloaded separately. See data/ for details.
Go to our github repo for more introductions.
Citation
If you find VSR… See the full description on the dataset page: https://huggingface.co/datasets/cambridgeltl/vsr_random.vsr_zeroshot
VSR: Visual Spatial Reasoning
This is the zero-shot set of VSR: Visual Spatial Reasoning (TACL 2023) [paper].
Usage
from datasets import load_dataset
data_files = {"train": "train.jsonl", "dev": "dev.jsonl", "test": "test.jsonl"}
dataset = load_dataset("cambridgeltl/vsr_zeroshot", data_files=data_files)
Note that the image files still need to be downloaded separately. See data/ for details.
Go to our github repo for more introductions.
Citation
If you find… See the full description on the dataset page: https://huggingface.co/datasets/cambridgeltl/vsr_zeroshot.loong
Additional Information
Project Loong Dataset
This dataset is part of Project Loong, a collaborative effort to explore whether reasoning-capable models can bootstrap themselves from small, high-quality seed datasets.
Dataset Description
This comprehensive collection contains problems across multiple domains, each split is determined by the domain.
Available Domains:
Advanced Math
Advanced mathematics problems including calculus, algebra… See the full description on the dataset page: https://huggingface.co/datasets/camel-ai/loong.parliamentrag-camera-leg19
ParliamentRAG — Italian Chamber of Deputies, 19th legislature
Full proceedings of the Italian Chamber of Deputies (Camera dei deputati) for the 19th legislature, from the first sitting on 13 October 2022 through 6 August 2026: verbatim speech transcripts, roll-call votes with every individual ballot, parliamentary acts with EuroVoc subjects, and the deputies' group, committee and government memberships over time. The dataset is refreshed as new sittings are ingested.
The tables… See the full description on the dataset page: https://huggingface.co/datasets/emeierkeio/parliamentrag-camera-leg19.fieldopsbench
FieldOpsBench
Multimodal benchmark for AI systems acting in real-world
field-operations contexts across sixteen trades.
194 active public cases · 16 trades · 27 code bodies · 7 categories ·
8 scoring dimensions · 5-layer contamination defense.
This is the HuggingFace dataset mirror of FieldOpsBench. The full
harness, scorers, runners, and CI live at
github.com/camerasearch/fieldopsbench.
Quick start
from datasets import load_dataset
# Load the public dev split (one… See the full description on the dataset page: https://huggingface.co/datasets/CameraSearch/fieldopsbench.OWL-SFT
OWL SFT (Planner) Dataset
Dataset Summary
OWL SFT is a supervised fine‑tuning dataset designed for training the planner agent in the Optimized Workforce Learning (OWL) framework – a system for multi‑agent assistance in real‑world task automation.
The dataset contains 1,564 multi‑turn conversations, focusing on task decomposition, sequencing, and coordination skills that are crucial for high‑level planning.
Languages
All conversation turns are written in… See the full description on the dataset page: https://huggingface.co/datasets/camel-ai/OWL-SFT.Verified-Camel
This is the Official Verified Camel dataset. Just over 100 verified examples, and many more coming soon!
Comprised of over 100 highly filtered and curated examples from specific portions of CamelAI stem datasets.
These examples are verified to be true by experts in the specific related field, with atleast a bachelors degree in the subject.
Roughly 30-40% of the originally curated data from CamelAI was found to have atleast minor errors and/or incoherent questions(as determined… See the full description on the dataset page: https://huggingface.co/datasets/LDJnr/Verified-Camel.Verified-Camel-KO
Verified-Camel-KO
이 데이터셋은 https://huggingface.co/datasets/LDJnr/Verified-Camel 의 한국어 번역입니다.
GPT4 Turbo로 번역한 뒤, 약간의 수정을 거쳤습니다.
이 데이터에 대한 방침은 전부 원 저자의 방침을 따릅니다.
This is the Official Verified Camel dataset. Just over 100 verified examples, and many more coming soon!
Comprised of over 100 highly filtered and curated examples from specific portions of CamelAI stem datasets.
These examples are verified to be true by experts in the specific related field, with atleast a… See the full description on the dataset page: https://huggingface.co/datasets/kuotient/Verified-Camel-KO.Verified-Camel-zhThis is a direct Chinese translation using GPT4 of the Verified-Camel dataset. I hope you find it useful.
https://huggingface.co/datasets/LDJnr/Verified-Camel
Citation:
@article{daniele2023amplify-instruct,
title={Amplify-Instruct: Synthetically Generated Diverse Multi-turn Conversations for Effecient LLM Training.},
author={Daniele, Luigi and Suphavadeeprasit},
journal={arXiv preprint arXiv:(comming soon)},
year={2023}
}
Camel-Milk-in-Gastrointestinal-Therapy
Dataset Card: Qualitative Data Extraction Matrix - Camel Milk in Gastrointestinal Pathology
Dataset Description
This dataset provides a Qualitative Data Extraction Matrix derived from the comprehensive 2026 clinical research report: "Advanced Therapeutic Applications of Camel Milk in Gastrointestinal Pathology: Microbiome Modulation, Mucosal Regeneration, and the Critical Role of Processing Technologies".
The tabular data reflects categorized molecular outcomes… See the full description on the dataset page: https://huggingface.co/datasets/camelway/Camel-Milk-in-Gastrointestinal-Therapy.SCQA_FirstTurnadaption-multichannel-campaign-optimizer-dataset
Multichannel Campaign Optimizer Dataset — Adaption AutoScientist Challenge
Powered by Adaptive Data — Adaption Labs
What This Dataset Teaches
This dataset trains a model to make data-grounded marketing optimization decisions — not just look up single metrics, but perform compound reasoning:
Task
Example
Budget Reallocation
"Given 4 campaigns, rank by ROAS, decide which to cut and which to scale"
A/B Test Significance
"Given test vs. control metrics… See the full description on the dataset page: https://huggingface.co/datasets/narendarcodes/adaption-multichannel-campaign-optimizer-dataset.posqaThis dataset is based on our publication POSQA: Probe the World Models of LLMs with Size Comparisons (PDF).
More details: GitHub Repo
Camildae
merge of some datasets from Alpaca Cot
cameldataCAMP
CAMP: Contextualized Arithmetic with Minimal Pairs
CAMP is a benchmark for studying how natural language framing affects LLM
accuracy on unit conversion tasks. Each item is a minimal pair: the same
underlying computation presented in multiple surface forms, allowing direct
comparison of model behavior across prompt conditions while holding the
math constant.
Key Finding
LLMs answer unit conversion questions significantly less accurately when the
problem is framed in… See the full description on the dataset page: https://huggingface.co/datasets/janewarrenn/CAMP.camel_qa_dataseta-env-seed2synth-synth
SETA Env Seed-to-Synth Synthetic Data
Synthetically generated terminal agent tasks derived from the SETA seed dataset. Each task contains a Docker-based environment, an instruction, a reference solution, and automated tests.
Dataset Structure
{source}/
├── summary.csv # task index with status, verdict, and timing info
└── {task_id}/
├── task.toml # task metadata (id, source, category, title)
├── instruction.md # natural language task… See the full description on the dataset page: https://huggingface.co/datasets/camel-ai/seta-env-seed2synth-synth.ShareGPT4Video
ShareGPT4Video 4.8M Dataset Card
Dataset details
Dataset type:
ShareGPT4Video Captions 4.8M is a set of GPT4-Vision-powered multi-modal captions data of videos.
It is constructed to enhance modality alignment and fine-grained visual concept perception in Large Video-Language Models (LVLMs) and Text-to-Video Models (T2VMs). This advancement aims to bring LVLMs and T2VMs towards the capabilities of GPT4V and Sora.
sharegpt4video_40k.jsonl is generated by GPT4-Vision… See the full description on the dataset page: https://huggingface.co/datasets/Camellia054/ShareGPT4Video.cameo_newsDataset used in my thesis (https://github.com/valentinwerner1/Thesis_RelationExtraction_PoliticsNews)
Reformatted for training with LLMs, experimenting whether these can improve performance
ShareGPT4V
News
[2024/5/8] We released ShareGPT4Video, a large-scale video-caption dataset, with 40K captions annotated by GPT4V and 4.8M captions annotated by our ShareCaptioner-Video. The total videos last with 300 hours and 3000 hours separately!
ShareGPT4V 1.2M Dataset Card
Dataset details
Dataset type:
ShareGPT4V Captions 1.2M is a set of GPT4-Vision-powered multi-modal captions data.
It is constructed to enhance modality alignment and fine-grained visual concept… See the full description on the dataset page: https://huggingface.co/datasets/Camellia054/ShareGPT4V.seta-env-seed2synth-seed
SETA Env Seed-to-Synth Seed Data
Seed data collected from technical Q&A platforms and programming communities, used as source material for the SETA (Synthetic Environment Terminal Agent) data synthesis pipeline.
Dataset Structure
The dataset is organised by source, then by seed ID:
{source}/
└── {seed_id}/
├── main.json # primary Q&A pair with metadata
├── related_1.json # related question/post #1
├── related_2.json # related question/post… See the full description on the dataset page: https://huggingface.co/datasets/camel-ai/seta-env-seed2synth-seed.