datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
mini-VTAB-corruptions
mini-VTAB-C
A collection of VTAB (Visual Task Adaptation Benchmark) datasets. We sampled 1K training samples and 1K testing samples for each task. For each test set, we apply all 15 corruption types from ImageNet-C.
Tasks
datasets = [
"caltech101",
"cifar10",
"cifar100",
"dtd",
"flowers",
"pets",
"sun397",
"svhn",
"pcam",
"eurosat",
"resisc45",
"diabetic_retinopathy",
"clevr_count_all"… See the full description on the dataset page: https://huggingface.co/datasets/antofuller/mini-VTAB-corruptions.corruption-gaussian_noise
Corruption Dataset: Gaussian_Noise
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using gaussian_noise corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Gaussian_Noise
Adds… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-gaussian_noise.corruption-spatter
Corruption Dataset: Spatter
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using spatter corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Spatter
Adds spatter effects… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-spatter.corruption-snow
Corruption Dataset: Snow
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using snow corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Snow
Adds snow effects to images… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-snow.corruption-zoom_blur
Corruption Dataset: Zoom_Blur
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using zoom_blur corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Zoom_Blur
Applies zoom blur… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-zoom_blur.corruption-brightness
Corruption Dataset: Brightness
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using brightness corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Brightness
Adjusts image… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-brightness.corruption-contrast
Corruption Dataset: Contrast
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using contrast corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Contrast
Modifies image contrast… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-contrast.corruption-jpeg_compression
Corruption Dataset: Jpeg_Compression
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using jpeg_compression corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Jpeg_Compression… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-jpeg_compression.corruption-glass_blur
Corruption Dataset: Glass_Blur
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using glass_blur corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Glass_Blur
Applies glass blur… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-glass_blur.corruption-gaussian_blur
Corruption Dataset: Gaussian_Blur
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using gaussian_blur corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Gaussian_Blur
Applies… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-gaussian_blur.corruption-frost
Corruption Dataset: Frost
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using frost corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Frost
Adds frost effects, simulating… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-frost.bea-19-corruption
BEA 19 Shared task.
BEA 19 Shared task dataset already preprocessed to have the original and corrupted sentence.
I merged the def and train datasets into one and applied all the annotated edits.
Source: https://www.cl.cam.ac.uk/research/nl/bea2019st/
quest-corruption-393k
Quest-corruption 393K (CLM)
This is quest-corruption for training of CLM restorers. For RM training, use the data in quest-corruption-1m-rm.
corruption-impulse_noise
Corruption Dataset: Impulse_Noise
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using impulse_noise corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Impulse_Noise
Adds… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-impulse_noise.tabrepairbench-replayable-corruption-lineage
TabRepairBench: Replayable Corruption Lineage
This is a finite, wholly authored synthetic reference dataset for auditing
tabular corruption lineage. Public Lineage v1 contains 72 independent groups
and 6,912 clean/corrupt cell pairs across three authored structural generator
families, six corruption schedules, two severities, and two audit partitions.
It is intentionally not presented as real-world data. It makes no claim of
real-data representativeness, causal identification… See the full description on the dataset page: https://huggingface.co/datasets/haidang2405/tabrepairbench-replayable-corruption-lineage.ccnews-sentences-corruptioncorruption-shot_noise
Corruption Dataset: Shot_Noise
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using shot_noise corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Shot_Noise
Adds shot noise… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-shot_noise.ccnews-encoding-corruptionafrica-seychelles-seychelles-corruption-perceptions-index-702ea721
Seychelles - Corruption Perceptions Index | Africa (Seychelles official open data)
78 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Seychelles as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Seychelles - Corruption Perceptions… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-seychelles-seychelles-corruption-perceptions-index-702ea721.africa-sao-tome-and-principe-sao-tome-and-principe-corruption-perceptions-index-22ed4811
Sao Tome and Principe - Corruption Perceptions Index | Africa (Sao Tome and Principe official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Sao Tome and Principe as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Sao… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-sao-tome-and-principe-sao-tome-and-principe-corruption-perceptions-index-22ed4811.africa-drc-democratic-republic-of-the-congo-corruption-perceptions-in-9ed8b46b
Democratic Republic of the Congo - Corruption Perceptions Index | Africa (DRC official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from DRC as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Democratic Republic of the Congo… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-drc-democratic-republic-of-the-congo-corruption-perceptions-in-9ed8b46b.africa-equatorial-guinea-equatorial-guinea-corruption-perceptions-index-ab0a47ce
Equatorial Guinea - Corruption Perceptions Index | Africa (Equatorial Guinea official open data)
66 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Equatorial Guinea as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Equatorial Guinea -… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-equatorial-guinea-equatorial-guinea-corruption-perceptions-index-ab0a47ce.repro-finite-and-corruption-robust-regret-bounds-in-online-inverse-linear-optimization-traces
Agent traces
Agent sessions published from a Trackio Logbook.
africa-guinea-bissau-guinea-bissau-corruption-perceptions-index-a2b60290
Guinea-Bissau - Corruption Perceptions Index | Africa (Guinea-Bissau official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Guinea-Bissau as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Guinea-Bissau - Corruption… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-guinea-bissau-guinea-bissau-corruption-perceptions-index-a2b60290.africa-madagascar-madagascar-corruption-perceptions-index-4b1c5438
Madagascar - Corruption Perceptions Index | Africa (Madagascar official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Madagascar as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Madagascar - Corruption Perceptions… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-madagascar-madagascar-corruption-perceptions-index-4b1c5438.africa-eswatini-eswatini-corruption-perceptions-index-8ded8217
Eswatini - Corruption Perceptions Index | Africa (Eswatini official open data)
72 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Eswatini as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Eswatini - Corruption Perceptions Index… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-eswatini-eswatini-corruption-perceptions-index-8ded8217.corruption-speckle_noise
Corruption Dataset: Speckle_Noise
Dataset Description
This dataset contains corrupted versions of ImageNet-1K images using speckle_noise corruption. It is part of the ImageNet-C benchmark for evaluating model robustness to common image corruptions.
Dataset Structure
Train: 1,281,167 corrupted images
Validation: 50,000 corrupted images
Classes: 1000 ImageNet-1K classes
Format: Arrow (Hugging Face Datasets)
Corruption Type: Speckle_Noise
Adds… See the full description on the dataset page: https://huggingface.co/datasets/MarMaster/corruption-speckle_noise.africa-cabo-verde-cabo-verde-corruption-perceptions-index-2a312f9b
Cabo Verde - Corruption Perceptions Index | Africa (Cabo Verde official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Cabo Verde as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Cabo Verde - Corruption Perceptions… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-cabo-verde-cabo-verde-corruption-perceptions-index-2a312f9b.africa-lesotho-lesotho-corruption-perceptions-index-230f38cb
Lesotho - Corruption Perceptions Index | Africa (Lesotho official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Lesotho as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Lesotho - Corruption Perceptions Index
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-lesotho-lesotho-corruption-perceptions-index-230f38cb.africa-comoros-comoros-corruption-perceptions-index-2b183db3
Comoros - Corruption Perceptions Index | Africa (Comoros official open data)
84 rows - 1 Africa country - 2012-2025 - Repackaged by Electric Sheep Africa
TL;DR
This dataset packages one official CSV resource from Comoros as
ML-ready Parquet. The source file is the provenance boundary; all usable
indicators or tabular columns from the resource stay together in this repo.
About the source
Source: Comoros - Corruption Perceptions Index
Publisher:… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-comoros-comoros-corruption-perceptions-index-2b183db3.