datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
windows_osworld_file_cache
OSWorld File Cache
This repository serves as a file cache for the OSWorld project, providing reliable and fast access to evaluation files that were previously hosted on Google Drive.
Overview
OSWorld is a scalable, real computer environment for multimodal agents, supporting task setup, execution-based evaluation, and interactive learning across various operating systems and applications. This cache repository ensures that all evaluation files are consistently accessible… See the full description on the dataset page: https://huggingface.co/datasets/xlangai/windows_osworld_file_cache.fractal_rawWindowsOfficial repository link: microsoft/SWE-bench-Live
Paper link: https://arxiv.org/abs/2603.05026
Developing projects compatible on Windows platform is important to expand the user market.
There are some bugs that would only occur on Windows. To migrate projects to Windows some codes need to be rewritten to have branching points across different os or use cross-platform compatible libraries...
To test LLM's knowledge of Windows-specific SWE knowledge AND powershell terminal operation capability… See the full description on the dataset page: https://huggingface.co/datasets/SWE-bench-Live/Windows.windows_osworld
Dataset Card for Dataset Name
This repository contains the task examples, retrieval documents (in the archive evaluation_examples.zip), and virtual machine snapshots for benchmark OSWorld (loaded by VMware/VirtualBox depending on the machine architecture x86 or arm64).
You can find more information from our paper OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
paper Arxiv link: https://arxiv.org/abs/2404.07972
project website:… See the full description on the dataset page: https://huggingface.co/datasets/xlangai/windows_osworld.shamela-bangla-windowsosatlas-windows-imagesWindows-powershellWindowsServer_21H2old-windows-48-5
European Architectural Details Dataset — 48 Authentic Heritage Shots
📋 Description
Curated collection of 48 high-resolution photographs documenting authentic European architectural details — rustic windows with wooden shutters, ornate stone facades, wrought-iron grilles, weathered doors, and charming building exteriors shot across Italy and Europe.
Each image includes comprehensive CSV metadata with 13 classification fields optimized for machine learning… See the full description on the dataset page: https://huggingface.co/datasets/Kos1976/old-windows-48-5.Windows_wheelschapman-ecg-windows
Chapman-Shaoxing ECG Windows
Preprocessed from the Chapman-Shaoxing 12-lead ECG Database.
45,150 patients, 180,600 windows. Lead II, 128Hz, 4s windows (512 samples), 50% overlap.
Split
Windows
Patients
Train
126,420
31,605
Val
27,088
6,772
Test
27,092
6,773
Schema
Column
Type
Description
patient_id
string
Record ID (e.g. JS00001)
record_id
string
Record + window index
window
list[float32]
512-sample ECG (4s@128Hz, Lead II)… See the full description on the dataset page: https://huggingface.co/datasets/dheraingoud/chapman-ecg-windows.9-facades-doors-windows-50-commercial
Architectural Facades and Details — 50 Authentic European Photographs
📋 Dataset Summary
Curated collection of 50 high-resolution photographs documenting authentic European architectural details — facades, doors, windows, walls, arches, gates, balconies, and columns. From rustic Italian farmhouses and Tuscan alleyways to ornate stone archways, Gothic cathedral entrances, medieval iron gates, historic French academy portals, and charming countryside barn doors.… See the full description on the dataset page: https://huggingface.co/datasets/Kos1976/9-facades-doors-windows-50-commercial.bnci-windows
EEG Dataset
This dataset was created using braindecode, a library for deep learning with EEG/MEG/ECoG signals.
Dataset Information
Number of recordings: 1
Number of channels: 26
Sampling frequency: 250.0 Hz
Data type: Windowed (from Epochs object)
Number of windows: 48
Total size: 0.04 MB
Storage format: zarr
Usage
To load this dataset:
from braindecode.datasets import BaseConcatDataset
# Load dataset from Hugging Face Hub
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/Kkuntal990/bnci-windows.windowsmisc-windows-latest-3.12old-windows-doors-48-6
European Architectural Details Dataset — 48 Authentic Heritage Shots
📋 Description
Curated collection of 48 high-resolution photographs documenting authentic European architectural details — rustic windows with wooden shutters, ornate stone facades, wrought-iron grilles, weathered doors, historic archways, and charming building exteriors shot across Italy and Europe.
Each image includes comprehensive CSV metadata with 13 classification fields optimized for… See the full description on the dataset page: https://huggingface.co/datasets/Kos1976/old-windows-doors-48-6.fractal_lerobotneurofeed-eeg-windows
Neurofeed EEG Windows
Ready-to-train 2-second, 256 Hz EEG windows derived from open EEG/PSG datasets and remapped onto consumer-headset
layouts: Muse (AF7, AF8, TP9, TP10) and Neurosity Crown (C3/C4-centred). Every config ships fixed
subject-level splits, per-recording manifests with per-source license and upstream URL, frozen-encoder
baselines, and honest negative results.
About 1.56 M windows in 8 configs, all from open sources (CC0-1.0, ODC-By-1.0, CC-BY-4.0).
Derived data… See the full description on the dataset page: https://huggingface.co/datasets/windwerfer/neurofeed-eeg-windows.cad-1000-hours
CAD-1K Open v2 - 1,018.1229 Hours
509 end-to-end, single-display Windows CAD task recordings across seven CAD software families.
Each task contains:
task_desc.json - task prompt, application, reference-input paths, and expected deliverables
input_files/ - reference inputs named input.ext or input_N.ext
output_files/ - submitted CAD deliverables and supplemental outputs named output.ext or output_N.ext
rubrics.json - task-specific evaluation criteria
task_overview.pdf - review… See the full description on the dataset page: https://huggingface.co/datasets/windows2t2/cad-1000-hours.10-facades-doors-windows-50-commercial
European Architectural Details — 50 Authentic Photographs
📋 Dataset Summary
Curated collection of 50 high-resolution photographs documenting authentic architectural details — doors, windows, structures, facades, arches, balconies, and gates. From rustic Italian farmhouses and Tuscan alleyways to ornate stone archways, historic portals, medieval iron gates, and charming countryside barn doors.
Each image includes comprehensive CSV metadata with 14 classification… See the full description on the dataset page: https://huggingface.co/datasets/Kos1976/10-facades-doors-windows-50-commercial.Windows11_26H1windows-rtx-4060ti-8gb-moe-offload-bench-2026-05
RTX 4060 Ti 8GB — Multi-Model Benchmark (2026-05)
practitioner benchmarks on consumer hardware (8GB VRAM, 32GB RAM). 10 models tested, covering MoE expert offload, hybrid SSM architectures, dense models, MLA, dense partial GPU offload, and the 1B speed ceiling. all runs on the same physical rig, same methodology.
current leaderboard (decode tok/s at sweet spot)
model
active params
GGUF size
sweet spot tok/s
quality (6 tests)
architecture
Llama 3.2 1B
1.24B
771… See the full description on the dataset page: https://huggingface.co/datasets/witcheer/windows-rtx-4060ti-8gb-moe-offload-bench-2026-05.chb-mit-seizure-windows-2to1
CHB-MIT Seizure Detection - Balanced Binary Classification Dataset
Dataset Description
Preprocessed, balanced version of the CHB-MIT Scalp EEG Database
for binary seizure classification tasks.
Key Features
Window Size: 8 seconds
Sample Rate: 200 Hz
Channels: 18 (standardized from original 23)
Samples per Window: 1600
Non-seizure ratio: 2:1 (non-seizure:seizure)
Balance Strategy
Only full seizure windows where the entire 8s is within a seizure
50%… See the full description on the dataset page: https://huggingface.co/datasets/smam/chb-mit-seizure-windows-2to1.Windows11_26H2bnci-windows-test
EEG Dataset
This dataset was created using braindecode, a library for deep learning with EEG/MEG/ECoG signals.
Dataset Information
Number of recordings: 1
Number of channels: 26
Sampling frequency: 250.0 Hz
Data type: Windowed (from Epochs object)
Number of windows: 48
Total size: 0.04 MB
Storage format: zarr
Usage
To load this dataset:
from braindecode.datasets import BaseConcatDataset
# Load dataset from Hugging Face Hub
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/Kkuntal990/bnci-windows-test.qwen3.8-conv1d-sigscalesync
The conv1d sigma outliers in Qwen3.8-27B are real. The quantization story attached to them is not — and the rescale variant I tested cost perplexity while fixing nothing measurable.
Scope revision (2026-08-24). An earlier version of this README said the rescale was applied
"exactly as published" and that the mechanism "cannot work". Both were too strong. What I tested
is a 7-tensor median-normalization variant (σ > 1.5× median → α = median/σ); the intervention
actually… See the full description on the dataset page: https://huggingface.co/datasets/windowsxp811203/qwen3.8-conv1d-sigscalesync.uav-v2v-windows93
UAV V2V windows93
12,369 paired 93-frame aerial FPV windows (640x640, 10 fps, 9.3 s) cut from
TartanAir-V2
Data_easy (561 trajectories, 25% overlap, stride 70), built for
Cosmos-Transfer2.5 control-branch post-training.
Contents
path
what
rgb_shards/rgb-NNN.tar (40 shards, 79 GB)
RGB windows, H.264
first_frame_shards/first_frame-NNN.tar (4 shards, 7.6 GB)
frame 0 of each window (image conditioning)
poses/<env>/<traj>.txt (561)
full-trajectory camera… See the full description on the dataset page: https://huggingface.co/datasets/greenery8848/uav-v2v-windows93.dataclaw-windows
Coding Agent Conversation Logs
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop anyone else from doing the same with their data - pulling up the ladder behind them. DataClaw lets you throw the ladder back down. The dataset it produces is yours to share.
Exported with DataClaw.
Tag: dataclaw - Browse all DataClaw datasets
Stats
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/woctordho/dataclaw-windows.osatlas-windows-combineddecodingtrust-windows-qcow2