datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
tcg-mtg-cards
Dataset
This dataset contains high-quality images of Magic: The Gathering cards, processed and formatted for machine learning tasks. The images are standardized to 500x700 pixels in PNG format.
Each image is labeled with its unique Scryfall ID, which can be used to fetch additional card information from the Scryfall API.
Usage
You can load the dataset using the Hugging Face datasets library:
from datasets import load_dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/acidtib/tcg-mtg-cards.mtg-urna-benchmark
mtg-urna-benchmark
38,627 magic: the gathering cards, one per oracle id, the scan and the rules text of each, packed into single .urna files that answer text and image queries from memory-mapped bytes: no server, no python at read time. ten such files live here. they carry the same cards, the same text and the same content hash; what differs is how the 4 GB of jpeg was encoded inside, and which image models embedded it.
the file to start with is release/v0.3/stills-5models. it… See the full description on the dataset page: https://huggingface.co/datasets/brennercruvinel/mtg-urna-benchmark.mt_genevalThe MT-GenEval benchmark evaluates gender translation accuracy on English -> {Arabic, French, German, Hindi, Italian,
Portuguese, Russian, Spanish}. The dataset contains individual sentences with annotations on the gendered target words,
and contrastive original-invertend translations with additional preceding context.MARBLEGenreClassification_MTG-Genre-Fold1
Dataset Card for "MARBLEGenreClassification_MTG-Genre-Fold1"
More Information needed
MARBLEGenreClassification_MTG-Genre-Fold3
Dataset Card for "MARBLEGenreClassification_MTG-Genre-Fold3"
More Information needed
mtg-datasetThe data set for Mind-the-Glitch: Visual Correspondence for Detecting Inconsistencies in Subject-Driven Generation
mtg-scryfall-cropped-art-embeddings-siglip-so400m-patch14-384MTGeval_mtg_clirmatrix_beirmtgjsonv5-11-19-24-AllPrintingsParquetFilesmtg-scryfall-unique-artwork-20240809-with-card-art-descriptions-and-images-with-embeddingsmtg-scryfall-unique-artwork-20240809-combined-elements-embeddingsmtg_face_objects_classifcation
mtg face objects dataset
This dataset does NOT contian ALL the highres card scan images only about 11,700 in the form of a zip file, more than enough for object classifcation
This dataset contains high-resolution annotations for Magic: The Gathering card scans. The annotations were created using LabelImg and include raw XML labels in labellmg PascalVoc. The dataset is organized into training and validation portions, enabling custom object detection model training. Around 90 labeled… See the full description on the dataset page: https://huggingface.co/datasets/JakeTurner616/mtg_face_objects_classifcation.mtg_cards-2025-04-04MTG cards labeled with similarity score.
Can be used for sentence and/or image similarity tasks.
I hearby declare that I don't own any rights to the data, I only added annotations.
Annotation
Thebdata was automatically annotated, the code will be added here soon.
Interesting samples:
2df7b947-bdb2-4204-8eb0-92fe66411613_8c5bd289-6069-4d99-83cd-d1dda24bc224
eval_clir_mtgmtg-embeddings
Dataset Card for Dataset Name
Text embeddings of all Magic: The Gathering card until Aetherdrift (2024-02-14). The text embeddings are centered around card mechanics (i.e. no flavor text/card art embeddings) in order to identify similar cards mathematically.
This dataset also includes zero-mean-centered 2D UMAP coordinates for all the cards, in columns x_2d and y_2d.
Dataset Details
How The Embeddings Were Created
Using the data exports from MTGJSON, the data… See the full description on the dataset page: https://huggingface.co/datasets/minimaxir/mtg-embeddings.mtg-scryfall-cropped-artmtg-scryfall-cropped-art-with-descriptionsallmusiccaps
AllMusicCaps
Music caption dataset built from professional AllMusic album reviews, cross-referenced with Discogs
releases and YouTube tracks. Captions are generated in two complementary styles by a two-stage LLM
pipeline.
Released with the ISMIR 2026 paper AllMusicCaps: Album Reviews as Complementary Supervision for
Music CLAP. Code and models: github.com/MTG/allmusiccaps.
This dataset contains no audio, only identifiers and captions. Recover audio from the
youtube_id field, or… See the full description on the dataset page: https://huggingface.co/datasets/mtg-upf/allmusiccaps.mtg-print-distribution
MTG Print Distribution
Expected pull rates for every Magic: The Gathering booster product
(Play Booster, Collector Booster, Draft Booster, Set Booster, Theme
Booster, Jumpstart, ...) across 191 sets — for each product, the expected
percentage of a pack made up of every distinguishable card variant (rarity,
finish, frame effects, borderless, and promo-tag combination), plus how
many distinct cards are eligible for each one. Each product also carries
its real current market price.… See the full description on the dataset page: https://huggingface.co/datasets/CXu0630/mtg-print-distribution.cmd-mtg_jamendo-metadataMARBLEInstrumentsClassification_MTGInstrument-Fold5
Dataset Card for "MARBLEInstrumentClassification_MTGInstrument-Fold5"
More Information needed
MARBLEMusicTagging_MTGTop50-Fold3
Dataset Card for "MARBLEMusicTagging_MTGTop50-Fold3"
More Information needed
mtg-scryfall-unique-artwork-20240809-with-card-art-descriptions-and-imagesMARBLEMusicTagging_MTGTop50-Fold2
Dataset Card for "MARBLEMusicTagging_MTGTop50-Fold2"
More Information needed
MARBLEMusicTagging_MTGTop50-Fold5
Dataset Card for "MARBLEMusicTagging_MTGTop50-Fold5"
More Information needed
MTG-EDH-Combos-QAmtg-forge-simulations-commander
mtg-forge-simulations-commander
A collection of simulated games gennerated with a modified copy of Forge.
Matches were run using the built in simulations AI.
Modifications were made to the logging engine to output cards in the format <<CARD owner="OWNER"|card="CARD NAME (id)"|id=ID>>
The cards column parses out card references in the order in which they appear in the logs.
mtg-image-data
Dataset Card for "mtg-image-data"
More Information needed
mtg-data
Dataset Card for "mtg-data"
Dataset Summary
The "mtg-data" dataset is a collection of prompts and responses related to Magic: The Gathering (MTG), a popular collectible card game.
The dataset contains various types of question and answer pairs, including official rulings as responses and corresponding questions generated by GPT-3.5,
Q&A data scraped from the web, glossary terms alongside their descriptions, and official rules formatted into Q/A pairs.
This dataset is… See the full description on the dataset page: https://huggingface.co/datasets/nelsntk/mtg-data.