datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
d-firegavin-dfire-raw
D-Fire (raw mirror) — GAVIN project
This is a raw, unmodified mirror of the third-party D-Fire dataset, republished here as part of the GAVIN (Garuda Agni Vahini) wildfire-detection research project. It is not GAVIN's own processed/derived data — see "License" below for what that distinction means.
What this is
D-Fire: an image dataset for fire and smoke detection. ~21,527 ground-level images with YOLO-format bounding box annotations for fire and smoke classes.… See the full description on the dataset page: https://huggingface.co/datasets/hanvithSai/gavin-dfire-raw.dfir-validation-data
DFIR Validation Data
Reference data used for Volatility 3 / DFIR tool validation (tool-testing corpus).
cridex.vmem
Description: Public memory-forensics training sample — a RAM image of a Windows XP SP3 x86 host infected with the Cridex banking trojan. Long-standing Volatility Foundation reference sample.
Acquired: 2012-07-22
Size: 536,870,912 bytes
SHA-256: 02a63be2fcf3a63446c3c8ca9151aff963f888204d141e46c6be60ddde7c3e8d
Original provenance: Volatility… See the full description on the dataset page: https://huggingface.co/datasets/jhenning/dfir-validation-data.article-dfir-augmente-ia-artefacts-windows
AI-Augmented DFIR: Windows Artifact Analysis
DFIR Augmente par IA : Analyse Artefacts Windows
This dataset contains a technical article available in both French and English.
Cet article technique est disponible en francais et en anglais.
Navigation
Version Francaise
English Version
title: "DFIR Augmente par IA : Analyse Automatisee des Artefacts Windows avec LLM"
author: "AYI-NEDJIMI Consultants"
date: "2026-02-21"
language: "fr"
tags:
-… See the full description on the dataset page: https://huggingface.co/datasets/AYI-NEDJIMI/article-dfir-augmente-ia-artefacts-windows.DFIR-MetricDFIR-Metric: A Benchmark Dataset for Evaluating Large Language Models in Digital Forensics and Incident Response
Description
DFIR-Metric is a comprehensive benchmark developed to assess the performance of Large Language Models (LLMs) in the field of Digital Forensics and Incident Response (DFIR), aiming to fill the gap in standardized evaluation methods. The benchmark comprises three key components: (a) MODULE I: expert-validated knowledge-based questions , (b) MODULE II: realistic forensic… See the full description on the dataset page: https://huggingface.co/datasets/Neo111x/DFIR-Metric.