datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
severstal_steel_defects
Dataset Card for severstal_steel_defects
This is a FiftyOne dataset with 18074 samples.
Installation
If you haven't already, install FiftyOne:
pip install -U fiftyone
Usage
import fiftyone as fo
from fiftyone.utils.huggingface import load_from_hub
# Load the dataset
# Note: other available arguments include 'max_samples', etc
dataset = load_from_hub("Voxel51/severstal_steel_defects")
# Launch the App
session = fo.launch_app(dataset)… See the full description on the dataset page: https://huggingface.co/datasets/Voxel51/severstal_steel_defects.SteelDefectXsteel-ocr-dataset-960
Steel OCR Dataset (Resized 960px)
鉄骨の手書き製品コード認識のための学習データセット（リサイズ版）。
学習高速化のため、画像を長辺960pxにリサイズしたバージョン。
バウンディングボックス座標も同じスケールで変換済み。
オリジナル版との違い
項目
オリジナル
リサイズ版
総ピクセル数
5077.8 MP
215.1 MP
削減率
-
95.8%
用途
評価、高精度学習
高速学習
データセット構成
train / val はリーク無しで分割済み（画像レベルで互いに素、sha256 + dHash 監査で 0 leak）。
固定 14 枚 test セットとも画像・クロップが重複しない。
train_data/
├── det/ # 検出モデル用
│ ├── train.txt # 学習データラベル (222画像)
│ ├── val.txt… See the full description on the dataset page: https://huggingface.co/datasets/iput-tk230215/steel-ocr-dataset-960.SteelDefectXsteelball_detect
A yolo format dataset for steelball_detection nearly two thousand images
the steel balls is about one centimeter in three diferent environments.
notes: some labels are json format and becuase some images are taken by different dvices ,its resolution are different
steelball
钢球 / Steel Ball Dataset
A real-image dataset for instance segmentation and 2D apparent-silhouette center
estimation of polished steel balls.
The dataset targets the following tasks:
detecting and counting steel balls in an image;
separating touching or mutually occluding ball instances;
predicting the visible-region mask of each instance;
estimating the center of each ball's complete 2D apparent silhouette;
analysing performance across occlusion levels using visible_fraction.… See the full description on the dataset page: https://huggingface.co/datasets/xinyuanzaizhe/steelball.severstal-steel-detectionsteel-ocr-dataset
Steel OCR Dataset
鉄骨の手書き製品コード認識のための学習データセット。
データセット構成
train_data/
├── det/ # 検出モデル用
│ ├── train.txt # 学習データラベル
│ ├── val.txt # 検証データラベル
│ └── images/ # 画像ファイル (319枚)
└── rec/ # 認識モデル用
├── rec_gt_train.txt # 学習データラベル
├── rec_gt_val.txt # 検証データラベル
├── dict.txt # 文字辞書
└── crop_img/ # クロップ画像 (467枚)
test_data/
├── det/ # 検出モデル評価用
│… See the full description on the dataset page: https://huggingface.co/datasets/iput-tk230215/steel-ocr-dataset.SteelDefectXsteelman-sft-ada
Steelman SFT: Ada 2022 & SPARK Training and Evaluation Data
To our knowledge, the first publicly available instruction-tuning dataset for Ada 2022 and SPARK code generation. 6,110 compiler-verified instruction-output pairs across 9 task categories. Every example compiles cleanly with the GNAT Ada compiler under strict flags.
This dataset trained Steelman-14B-Ada v0.3, which scores 62.4% on a 754-prompt Ada eval -- outperforming Claude Opus 4.6 (12.7%), GPT-5.4 (12.9%), and every… See the full description on the dataset page: https://huggingface.co/datasets/the-clanker-lover/steelman-sft-ada.SteelDefectXdetails_Steelskull__Aethora-7b-v1
Dataset Card for Evaluation run of Steelskull/Aethora-7b-v1
Dataset automatically created during the evaluation run of model Steelskull/Aethora-7b-v1 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Aethora-7b-v1.details_Steelskull__Etheria-55b-v0.1
Dataset Card for Evaluation run of Steelskull/Etheria-55b-v0.1
Dataset automatically created during the evaluation run of model Steelskull/Etheria-55b-v0.1 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Etheria-55b-v0.1.anne-steele-steele-vol_1-3-1780
Document OCR using GLM-OCR
This dataset contains OCR results from images in willwhim/steele_vol_1-3 using GLM-OCR, a compact 0.9B OCR model achieving SOTA performance.
Processing Details
Source Dataset: willwhim/steele_vol_1-3 (DELETED)
Model: zai-org/GLM-OCR
Task: text recognition
Number of Samples: 820
Processing Time: 88.6 min
Processing Date: 2026-04-03 01:25 UTC
Results
See steele-poetry
Configuration
Image Column: image
Output… See the full description on the dataset page: https://huggingface.co/datasets/willwhim/anne-steele-steele-vol_1-3-1780.details_Steelskull__Aurora_base_test
Dataset Card for Evaluation run of Steelskull/Aurora_base_test
Dataset automatically created during the evaluation run of model Steelskull/Aurora_base_test on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Aurora_base_test.details_Steelskull__Umbra-MoE-4x10.7
Dataset Card for Evaluation run of Steelskull/Umbra-MoE-4x10.7
Dataset automatically created during the evaluation run of model Steelskull/Umbra-MoE-4x10.7 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Umbra-MoE-4x10.7.details_Steelskull__VerB-Etheria-55b
Dataset Card for Evaluation run of Steelskull/VerB-Etheria-55b
Dataset automatically created during the evaluation run of model Steelskull/VerB-Etheria-55b on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__VerB-Etheria-55b.steel-defect-leaderboard-resultsdetails_Steelskull__Lumosia-v2-MoE-4x10.7
Dataset Card for Evaluation run of Steelskull/Lumosia-v2-MoE-4x10.7
Dataset automatically created during the evaluation run of model Steelskull/Lumosia-v2-MoE-4x10.7 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Lumosia-v2-MoE-4x10.7.anne-steele-steele-sheppard-1868
Document OCR using GLM-OCR
This dataset contains OCR results from images in willwhim/steele-sheppard-1868-images using GLM-OCR, a compact 0.9B OCR model achieving SOTA performance.
Processing Details
Source Dataset: willwhim/steele-sheppard-1868-images (DELETED)
Model: zai-org/GLM-OCR
Task: text recognition
Number of Samples: 300
Processing Time: 17.7 min
Processing Date: 2026-04-02 20:14 UTC
Results
See steele-poetry
Configuration
Image… See the full description on the dataset page: https://huggingface.co/datasets/willwhim/anne-steele-steele-sheppard-1868.steel_plates
Landsat
The Steel Plates dataset from the UCI repository.
Configurations and tasks
Configuration
Task
Description
steel_plates
Multiclass classification.
steel_plates_0
Binary classification.
Is the input of class 0?
steel_plates_1
Binary classification.
Is the input of class 1?
steel_plates_2
Binary classification.
Is the input of class 2?
steel_plates_3
Binary classification.
Is the input of class 3?
steel_plates_4
Binary classification.
Is the… See the full description on the dataset page: https://huggingface.co/datasets/mstz/steel_plates.TacRich-Manip-LeRobot-umi-set-the-steel-plate
TacRich-Manip LeRobot v3 — set_the_steel_plate
Multimodal real-robot trajectories for gripper-based contact-rich manipulation,
published in the standard LeRobot v3 layout. This task repository contains
umi/set_the_steel_plate and is private during active collection.
Dataset summary
Property
Value
Repository
qingzhu-robotics/TacRich-Manip-LeRobot-umi-set-the-steel-plate
Collection method
umi
Robot
rm75b-pika-tachin
Episodes
98
Frames
86,392… See the full description on the dataset page: https://huggingface.co/datasets/qingzhu-robotics/TacRich-Manip-LeRobot-umi-set-the-steel-plate.addition_decimalSteelDefectXesab-filler-metal-recommendations-by-astm-steel-grade
ESAB recommended filler metals and suggested preheat group, by ASTM steel grade
Canonical, always-current version: https://referencesource.org/esab-filler-metal-recommendations-by-astm-steel-grade/
Machine-readable: https://referencesource.org/esab-filler-metal-recommendations-by-astm-steel-grade/data.json — this mirror is a point-in-time copy.
Last verified: 2026-08-19
Stale after: 2028-08-18 (past this date, prefer the canonical copy —
it re-verifies on a cadence this… See the full description on the dataset page: https://huggingface.co/datasets/referencesource/esab-filler-metal-recommendations-by-astm-steel-grade.sillytavern-scenarios-datasetSteelskull__L3.3-Nevoria-R1-70b-details
Dataset Card for Evaluation run of Steelskull/L3.3-Nevoria-R1-70b
Dataset automatically created during the evaluation run of model Steelskull/L3.3-Nevoria-R1-70b
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/Steelskull__L3.3-Nevoria-R1-70b-details.global_steel_hrc_price_daily
قیمت جهانی فولاد (ورق گرم) — روزانه
سریِ روزانهٔ قیمت جهانی فولاد (ورق گرم) از بازارهای جهانی.
پوشش: 1387-07-29 → 1405-07-01 (4,470 روز) · تناوب: روزانه · واحد: دلار بر تنِ کوتاه
هر ردیف شاملِ قیمتِ پایانی است؛ بیشترین و کمترینِ روز نیز ارائه میشود. فایلِ steel_hrc.parquet نمای کامل
(بازگشایی/بیشترین/کمترین/پایانی) را دارد.
منبع: بازارهای جهانی (Yahoo Finance)
دربارهٔ فرمانا
این مجموعهداده بخشی از بانک دادهٔ فرمانا است: گردآوری، پاکسازی و استانداردسازیِ… See the full description on the dataset page: https://huggingface.co/datasets/Farmaanaa/global_steel_hrc_price_daily.steel_strength_v1-1
Mechanical properties of some steels
Dataset containing compositions and mechanical properties (yield strength, tensile strength, elongation) of 312 steel alloys
Dataset Information
Source: Foundry-ML
DOI: 10.18126/524z-vd6m
Year: 2022
Authors: Conduit, Gareth
Data Type: tabular
Fields
Field
Role
Description
Units
formula
input
Steel composition
c
input
C weight percent
wt %
mn
input
Mn weight percent
wt %
si
input
Si weight percent
wt %… See the full description on the dataset page: https://huggingface.co/datasets/foundry-ml/steel_strength_v1-1.details_Steelskull__Lumosia-MoE-4x10.7
Dataset Card for Evaluation run of Steelskull/Lumosia-MoE-4x10.7
Dataset automatically created during the evaluation run of model Steelskull/Lumosia-MoE-4x10.7 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_Steelskull__Lumosia-MoE-4x10.7.