datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
aidovecl-vehicle-detection-classification-localization
AIDOVECL: AI-generated Dataset of Outpainted Vehicles for Eye-level Classification and Localization
We introduce an annotated AI-generated dataset of eye-level vehicle images using outpainting, offering versatile generation of diverse vehicle classes in varied contexts with pretrained models.
Citation Notice
Please ensure that all publications and presentations using this data reference the following paper:
Kazemi, A., Fatima, Q. ul A., Kindratenko, V., & Tessum, C. W.… See the full description on the dataset page: https://huggingface.co/datasets/amir-kazemi/aidovecl-vehicle-detection-classification-localization.lca-bug-localization
🏟️ Long Code Arena (Bug localization)
This is the benchmark for the Bug localization task as part of the
🏟️ Long Code Arena benchmark.
The bug localization problem can be formulated as follows: given an issue with a bug description and a repository snapshot in a state where the bug is reproducible, identify the files within the repository that need to be modified to address the reported bug.
The dataset provides all the required components for evaluation of bug localization… See the full description on the dataset page: https://huggingface.co/datasets/JetBrains-Research/lca-bug-localization.bug-localization
Bug Localization
This is the data for Bug Localization benchmark.
How-to
Since the dataset is private, if you haven't used HF Hub before, add your token via huggingface-cli first:
huggingface-cli login
List all the available configs via datasets.get_dataset_config_names and choose an appropriate one
Load the data via load_dataset:
from datasets import load_dataset
# Select a configuration from ["py", "java", "kt", "mixed"]
configuration = "py"
# Select a split from… See the full description on the dataset page: https://huggingface.co/datasets/tiginamaria/bug-localization.ramanv-image-real-localizationbenchname-bug-localization
🥷 BenchName (Bug localization)
This is the benchmark for the Bug localization task as part of the
🥷 BenchName benchmark.
The bug localization problem can be formulated as follows: given an issue with a bug description and a repository snapshot in a state where the bug is reproducible, identify the files within the repository that need to be modified to address the reported bug.
The dataset provides all the required components for evaluation of bug localization approaches in… See the full description on the dataset page: https://huggingface.co/datasets/anon-iclr-submission/benchname-bug-localization.biomap-research-localization_prediction
localization_prediction
Sourced from biomap-research/localization_prediction and prepared for Hugging Face datasets usage.
Data files
Parquet files are stored under data/ using Hugging Face split naming conventions
(train-*, validation-*, test-*).
Preparation
Preprocess mode: minimal.
Seed: 1957723.
No max sequence length filter was applied.
Renamed source columns: label -> targets, seq -> sequence.
Columns: id, sequence, targets, split.
Validation… See the full description on the dataset page: https://huggingface.co/datasets/swhitfield/biomap-research-localization_prediction.illusionVQA-Soft-Localization
IllusionVQA: Optical Illusion Dataset
Project Page |
Paper |
Github
TL;DR
IllusionVQA is a dataset of optical illusions and hard-to-interpret scenes designed to test the capability of Vision Language Models in comprehension and soft localization tasks. GPT4V achieved 62.99% accuracy on comprehension and 49.7% on localization, while humans achieved 91.03% and 100% respectively.
Usage
from datasets import load_dataset
import base64
from openai import OpenAI… See the full description on the dataset page: https://huggingface.co/datasets/csebuetnlp/illusionVQA-Soft-Localization.protein-localization
Overview
Protein localization is an annotation of the subcellular compartments that proteins are found in. This task is a multilabel classification -- proteins can be found in more than one compartment.
This dataset is redistributed as part of mRNABench: https://github.com/morrislab/mRNABench
Data Format
Description of data columns:
target: Multihot labelling of cellular components that a protein appear in.The label positions correspond to: [Cytosol, Nucleoli… See the full description on the dataset page: https://huggingface.co/datasets/morrislab/protein-localization.gasvid_plume_source_localization_yolo_datasetlocalization_prediction
Dataset Card for Localization Prediction Dataset
Dataset Summary
The task of Protein Subcellular Localization Prediction bears substantial relevance in bioinformatics, owing to its contributions to proteomics research and its potential to augment our comprehension of protein function and disease mechanisms. In this task, the input to the model is an amino acid sequence of a protein, which is transformed into an output comprising a probability distribution over 10 unique… See the full description on the dataset page: https://huggingface.co/datasets/biomap-research/localization_prediction.BeetleBox
Dataset Card for Dataset Name
BeetleBox
Dataset Details
The BeetleBox dataset is a comprehensive multi-language, multi-project dataset designed for bug localization research. It includes 26,321 bugs from 29 projects, covering five major programming languages: Java, Python, C++, JavaScript, and Go. The dataset was meticulously curated to ensure accuracy, with a manual analysis revealing an incorrect ground truth rate of only 0.06%.
Dataset Description
The… See the full description on the dataset page: https://huggingface.co/datasets/bug-localization/BeetleBox.wgo-bench-localization
WGO-Bench — Localization Given Labels
Self-contained eval for localization given labels: the model is given the gold event labels (shuffled, with multiplicity) and must return one time interval per occurrence. Videos and gold intervals are embedded in each row.
Derived from Macrodata Labs' WGO-Bench (blog). License: CC-BY-NC-SA-4.0. Keep downstream use consistent with Macrodata's attribution and non-commercial / share-alike terms.
The job
You get a robot video and… See the full description on the dataset page: https://huggingface.co/datasets/Nano1337/wgo-bench-localization.watermark_localizationThermal-Heatmap-Source-Localization
ThermBench 🔥 — Thermal Heatmap Source Localization Benchmark
📝 Summary
ThermBench is a physics-inspired synthetic dataset designed to evaluate algorithms that infer hidden thermal sources from an observed heat diffusion map.
Each data sample contains:
an observed heatmap (matrix of values),
and the ground-truth sources: (row, col, intensity).
Diffusion follows inverse Manhattan distance:
[
H(i,j) ;=; \sum_{s=1}^{K} \frac{I_s}{d(i,j,s)+1}
]
where (d) is the Manhattan… See the full description on the dataset page: https://huggingface.co/datasets/ZoneTwelve/Thermal-Heatmap-Source-Localization.Dataset-Subcellular_Localization-DeepLoc
Description
Subcellular Localization prediction is a 10-class classification task to predict where a protein locates in the cell, where each input protein x is mapped to a label y ∈ {0, 1, ..., 9}.
Splits
Protein Format: SA sequence (AF2)
The dataset is from DeepLoc: prediction of protein subcellular localization using deep learning. We employ all proteins (proteins that lack AF2 structures are removed), and split them based on 70% structure similarity (see ProteinShake)… See the full description on the dataset page: https://huggingface.co/datasets/SaProtHub/Dataset-Subcellular_Localization-DeepLoc.localization_multiRNAIf you use this dataset, please cite the paper below:
Citation:
Wang, Jun, Marc Horlacher, Lixin Cheng, and Ole Winther. ‘DeepLocRNA: An Interpretable Deep Learning Model for Predicting RNA Subcellular Localization with Domain-Specific Transfer-Learning’. Edited by Pier Luigi Martelli. Bioinformatics 40, no. 2 (1 February 2024): btae065. https://doi.org/10.1093/bioinformatics/btae065.
ver_localization_unbalanced_01full_pose_semantic_localization_dataset_gazebolocalizationlocalization_prediction
Dataset Card for Localization Prediction Dataset
Dataset Summary
The task of Protein Subcellular Localization Prediction bears substantial relevance in bioinformatics, owing to its contributions to proteomics research and its potential to augment our comprehension of protein function and disease mechanisms. In this task, the input to the model is an amino acid sequence of a protein, which is transformed into an output comprising a probability distribution over 10 unique… See the full description on the dataset page: https://huggingface.co/datasets/proteinglm/localization_prediction.Dataset-Binary_Localization-DeepLoc
Description
Binary Localization prediction is a binary classification task where each input protein x is mapped to a label y ∈ {0, 1}, corresponding to either "membrane-bound" or "soluble" .
Protein Format: SA sequence (AF2)
Splits
The dataset is from DeepLoc: prediction of protein subcellular localization using deep learning. We employ all proteins (proteins that lack AF2 structures are removed), and split them based on 70% structure similarity (see ProteinShake), with… See the full description on the dataset page: https://huggingface.co/datasets/SaProtHub/Dataset-Binary_Localization-DeepLoc.SWESwiss-SFT-Localization-5K
Overview
SFT dataset for training SWE-Swiss models on the localization task. Prompts are constructed from a subset of issues in SWE-Gym-Raw and the SWE-bench training set. To prevent data leakage, we've filtered out any repositories that also appear in the SWE-bench test set. The responses are generated by DeepSeek-R1-0528. An instance is included in the final dataset only if the model's prediction meets two conditions: the number of predicted files is five or fewer, and the recall… See the full description on the dataset page: https://huggingface.co/datasets/SWE-Swiss/SWESwiss-SFT-Localization-5K.uniprot_subcellular_localization
UniProt Subcellular Localization (Vertebrates) — ProVADA
A curated collection of vertebrate UniProt/Swiss‑Prot protein domains labeled for cytosolic and extracellular localization. We remove signal peptides, restrict domain lengths, and provide both the full set and a 30% identity‑clustered representative set with train/test/validation splits (70/20/10). This dataset underpins the subcellular localization oracle in ProVADA (preprint).
See the Files section for exact filenames and… See the full description on the dataset page: https://huggingface.co/datasets/Xiaowei0402/uniprot_subcellular_localization.mozgach_localizations
Mozgach Localizations Dataset
Dataset Description
This dataset contains localization strings for the Mozgach application, providing translations from Russian to multiple languages including Chinese, Arabic, and others. The dataset is formatted for instruction-following language models and translation tasks.
Languages
Source Language: Russian (ru)
Target Languages: Chinese (zh), Arabic (ar), and others
Dataset Structure
Each entry in the dataset… See the full description on the dataset page: https://huggingface.co/datasets/nativemind/mozgach_localizations.zhcn-zhtw-localization-pairs
zh-CN → zh-TW Localization Pairs
簡體中文 → 台灣繁體中文在地化平行語料:49,857 組句對,每組包含真實簡體中文
段落與對應的台灣繁中改寫(字形轉換 + 中國大陸用語 → 台灣慣用詞)。
輸出由內部的大型在地化模型生成(synthetic),非人工翻譯
每列附七道自動品質檢查的判定(c_pass),整體通過率 87.8%
用途:訓練/蒸餾輕量的簡轉繁在地化模型、簡繁轉換評測
欄位
欄位
型別
說明
id
str
sha1(src) 前 16 碼,全集唯一
src
str
簡體中文原文(真實網路文本,非合成簡中)
out
str
台灣繁中在地化輸出
simp_ratio
float
原文的簡體專屬字率(經驗字頻法判定)
c_pass
bool
七道檢查全數通過
c_reasons
str
未通過時的原因(分號分隔)
來源與生成
輸入:取自公開網路語料的真實簡體中文段落,以兩個判別器篩選:
簡體專屬字率… See the full description on the dataset page: https://huggingface.co/datasets/lianghsun/zhcn-zhtw-localization-pairs.m4-failure-topology-localization-v0.1What this dataset tests
Whether failures form clusters across models.
It labels failure topology:
clustered
isolated
diffuse
none
Why this exists
If failures clusterthey are predictable by neighborhood.
That beats per-model guessing.
Data format
Each row contains:
a failure type
model responses
per-model failure flags
Labels
clustered-failure
isolated-failure
diffuse-failure
no-failure
Typical clustered failures
correlation → causation
null → “ineffective”… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/m4-failure-topology-localization-v0.1.humanoid-spatial-localization-dataset
Humanoid Spatial Localization Dataset
Dataset for tracking and estimating humanoid position
within structured indoor environments.
Description
Contains position coordinates, map references,
and localization confidence scores.
File
spatial_localization_dataset.json
License
MIT
LoRNA_localizationbug-localization
BenchName (Bug localization)
This is the benchmark for the Bug localization task as part of the
BenchName benchmark.
The bug localization problem can be formulated as follows: given an issue with a bug description and a repository snapshot in a state where the bug is reproducible, identify the files within the repository that need to be modified to address the reported bug.
The dataset provides all the required components for evaluation of bug localization approaches in real… See the full description on the dataset page: https://huggingface.co/datasets/icmlbenchname/bug-localization.SURFI-p1_claude_localization_1500_enriched_reasoning_filtered