datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Workspace-Bench-Workspaces
Workspace-Bench-Workspaces
Workspace-Bench-Workspaces contains the large workspace filesystem archives used by Workspace-Bench and Workspace-Bench-Lite. These files provide the realistic directory trees that agents must explore when solving workspace tasks.
The task metadata is hosted separately:
Full task metadata: Workspace-Bench
Lite task metadata: Workspace-Bench-Lite
Contents
This repository contains two workspace archives:
filesys_en.zip: English… See the full description on the dataset page: https://huggingface.co/datasets/Workspace-Bench/Workspace-Bench-Workspaces.Workspace-Bench
Workspace-Bench
Workspace-Bench is a benchmark for evaluating AI agents on realistic workspace tasks with large-scale file dependencies. It is designed to measure whether an agent can discover, interpret, and use the right files inside a noisy professional workspace, rather than solving tasks from isolated inputs or fully pre-packaged evidence.
The benchmark is introduced in the paper "Workspace-Bench 1.0: Benchmarking AI Agents on Workspace Tasks with Large-Scale File… See the full description on the dataset page: https://huggingface.co/datasets/Workspace-Bench/Workspace-Bench.Workspace-Bench-Lite
Workspace-Bench-Lite
A Lightweight Subset of Workspace-Bench for Fast and Cost-Efficient Evaluation
Overview •
LeaderBoard •
Distribution •
Quick Start •
Changelog •
Citation
Overview
Workspace-Bench-Lite is the Lite split of Workspace-Bench 1.0, designed for fast iteration and lower-cost benchmarking while preserving the core evaluation setting of the full benchmark.
It contains 100 tasks selected from the full Workspace-Bench and is intended to… See the full description on the dataset page: https://huggingface.co/datasets/Workspace-Bench/Workspace-Bench-Lite.kitchen-workspace-understanding-safe-manipulation
Kitchen Workspace Understanding & Safe Manipulation
Generated by datapack-import.ts
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL to the primary render image uploaded under /data; image_path stores the relative repository path and data_commit_sha pins the Hugging Face dataset commit used by those URLs. Files are uploaded as downloaded unless optional PNG recompression is enabled by… See the full description on the dataset page: https://huggingface.co/datasets/physicl/kitchen-workspace-understanding-safe-manipulation.local-workspace
local-workspace — consolidated oracle-lens artifacts
Everything the oracle-lens (OLA) program published, in one repo. Consolidated 2026-08-06 from
three repos (oracle-lens-data, oracle-lens-ar-checkpoints, oracle-lens-ao-checkpoints)
by scripts/ola/migrate_hf_workspace.py in the global-workspace repo, which also holds the
exact old->new path map and a --verify audit mode.
Layout
data/ # training/eval data (was oracle-lens-data)… See the full description on the dataset page: https://huggingface.co/datasets/agu18dec/local-workspace.vsi-workspace
VSI-Bench spatial-code pipeline
Video → 3D spatial reasoning: Depth Anything 3 (metric depth + camera poses) and SAM 3
(text-prompted instance segmentation) are fused into a per-scene spatial code — a JSON of
object positions, sizes, distances, room geometry, appearance order, and camera trajectory. A
unified vision-language model (Qwen3.5-2B / 4B) then answers VSI-Bench questions from one of
three inputs: the spatial code alone (code), 32 uniform video frames alone (frames), or… See the full description on the dataset page: https://huggingface.co/datasets/AntonioJun/vsi-workspace.inertialtransformer-workspace-datasynthetic-workspace-gymtinygsm_fobinary_workspace_depth1to9_traindepth5argo_workspacesindia-e1-workspace-mirroromen-agent-workspacearena-download-workspaceworkspace
Spatial Code VSI-Bench Workspace
This workspace evaluates VSI-Bench question answering with several input regimes:
raw video frames, perceived spatial codes from SAM3 + Depth Anything 3 caches,
ground-truth spatial codes from dataset annotations, and a deterministic symbolic
solver. The code is organized so important outputs are reproducible from fixed
inputs, fixed packages, fixed model checkpoints, and fixed SAM3/DA3 caches.
The repository intentionally separates three… See the full description on the dataset page: https://huggingface.co/datasets/AntonioJun/workspace.nested-tkf-repro-workspace
Code for "Nested birth-death processes are competitive with neural networks as protein evolution models", ICML 2026.
This repo contains code to preprocess Pfam v36.0 into pairwise alignments, train and evaluate pairHMM and neural models of protein sequence evolution, and reproduce all reported log-likelihoods.
We also include details about the three train/dev/test partitions, as well as log-likelihood metrics for all models trained on the three training replicates.
The bioRxiv… See the full description on the dataset page: https://huggingface.co/datasets/latticetower/nested-tkf-repro-workspace.dedup-workspacetinygsm_fopython_workspace_depth1to9_traindepth5RESEARCH_DASHBOARD
Research Dashboard Data
Internal data store for the RACA experiments dashboard. Not meant for direct browsing.
Use the dashboard at your local URL or HF Space to view experiments.
Files: experiments.json, runs.json, sub_experiments.json, experiment_notes.json, activity_logs.json, artifacts.json, summary_findings.json
curiosity-workspace-large-files-20260708
Curiosity reproduction assets
This private dataset mirrors the non-Git inputs and selected final endpoints
needed to reproduce the yanhr21/Curiosity workspace as curated on
2026-08-12. Paths intentionally match the repository root, so the bundle can
be restored in place.
cd /path/to/Curiosity
hf download Railgun526/curiosity-workspace-large-files-20260708 \
--repo-type dataset --local-dir .
Active Plan 14 inputs
The current no-training Newton/IsaacLab universal… See the full description on the dataset page: https://huggingface.co/datasets/Railgun526/curiosity-workspace-large-files-20260708.NVIDIA_WORKSPACE_3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 41,
"total_frames": 30151,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:41"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/tremmelnicholas/NVIDIA_WORKSPACE_3.runpod-workspace-archive-2026-09-07
Qwen2.5 sparsification experiment checkpoints
Research artifacts from experiments that prune Qwen2.5 weights while constraining the change in the model's output distribution. The collection preserves intermediate pruning stages, recovery states, teacher probability tensors, and source-model cache objects for reproducibility and analysis. Checkpoints are experimental intermediate results, not validated production models.
The completed archive contains 123 objects totaling 277,823… See the full description on the dataset page: https://huggingface.co/datasets/Ionel2023/runpod-workspace-archive-2026-09-07.tinygsm_fopython_workspace_depth1to9_traindepth5_tokenizedworkspace-data-archive
workspace-data-archive
Data-artifact archive of projects under /mnt/nvme0/zhujiayi/workspace/ on the
Rosetta host, packaged 2026-09-08 before the machine was wiped. Code lives on GitHub;
this repo preserves the non-reproducible data products (task sets, agent
trajectories, annotations, checkpoints, evaluation results). Related:
foreverCuSO4/transfer-bench-Rosetta-runs-archive (self-play run archive),
foreverCuSO4/transfer-bench-intents-v2-packed (3619-task dataset).
Per-archive… See the full description on the dataset page: https://huggingface.co/datasets/foreverCuSO4/workspace-data-archive.kaggle-workspace
BDMCPF
BD motorcycle ride POV footage, collected to experiment with road condition analysis from vehicle POV footage and road condition mapping.
Dataset Description
This dataset contains first-person (POV) video recordings of motorcycle rides in Bangladesh. The footage is intended for research on analyzing road conditions directly from vehicle POV video.
Use Cases
Road condition classification and analysis
Road surface quality mapping… See the full description on the dataset page: https://huggingface.co/datasets/arissassina/kaggle-workspace.NVIDIA_WORKSPACEThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 27,
"total_frames": 19913,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:27"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/tremmelnicholas/NVIDIA_WORKSPACE.tinygsm_fobinary_workspace_depth1to9_traindepth5_tokenizedritual-mirror-workspace
Ritual Mirror Workspace
Dataset workspace for Ritual Mirror agent state.
workspace-bench
workspace-bench — frozen eval item banks
Family renames (2026-08-25): safety_cases→agentic_misalignment,
superposed→multi_concept_directed_modulation, ordered_association→role_bound_association,
compositional_association→conjunctive_association, relational→relational_multihop,
order_ops→arithmetic_intermediates, ethical_consequences→moral_rationale,
readout_coherence→readout_quality (sub-splits ethical-*→moral-*). Bank contents are
unchanged — only names/paths moved. Bundles… See the full description on the dataset page: https://huggingface.co/datasets/camilablank/workspace-bench.xp-ft-workspace-fr-016c
ml-models/ — Modeles ML en production
Derniere mise a jour : 28 mars 2026
Monte dans le container ml-light-spark sous /models/.
Modeles en production
Modele
Taille
Usage
Charge par
gliner-notarial-v4
1.7G
NER (entites nommees)
GLINER_MODEL env var
gliclass-notarial-v2
1.5G
Classification documents (45 classes)
GLICLASS_MODEL env var
bge-m3-notarial-v3
2.2G
Embeddings (dense + sparse + ColBERT)
BGE_M3_MODEL env var
gliner-relex-notarial-v1
3.6G… See the full description on the dataset page: https://huggingface.co/datasets/baconnier/xp-ft-workspace-fr-016c.workspace-backup-20260911
Research workspace snapshot
This is the user's public snapshot of /workspace after research was stopped on
2026-09-11. It contains the retained research datasets, code, model artifacts,
metadata, installed workspace tooling, and research records selected by the
privacy-filtered inventory. Existing datasets and their media/location metadata
were intentionally retained at the user's instruction.
Authentication tokens, credentials, account state, private model/browser sessions,
and… See the full description on the dataset page: https://huggingface.co/datasets/EISLab-Hwlee/workspace-backup-20260911.