datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
8k-video-game-datasetDataset Name: Kim2091's 8k Dataset V3
License: CC BY-NC-SA 4.0
Purpose: Video Game Upscaling
Description: This is the third revision of my 8k dataset. These games were all run at absolute maximum settings on RTX 3090 or 4090 at native 8k, with TAA disabled when possible.
This dataset is mostly compressed with ZSTD. You will have to use a tool like NanaZip to extract it if your main extraction tool does not support it
I tried to reduce the amount of blank space in the dataset, but due to the… See the full description on the dataset page: https://huggingface.co/datasets/Kim2091/8k-video-game-dataset.Game-Oracle_DOTA2-Match-Prediction-Dataset
Game Oracle: DOTA 2 Match Prediction Dataset
Executive Summary
Game Oracle is a comprehensive data analysis project focused on DOTA 2 match prediction and professional gameplay patterns. The project addresses the growing demand for data-driven insights in esports, particularly in competitive gaming strategy and outcome prediction.
Motivation
The exponential growth of esports has created a need for sophisticated analytical tools to understand game dynamics and… See the full description on the dataset page: https://huggingface.co/datasets/howl-anderson/Game-Oracle_DOTA2-Match-Prediction-Dataset.game-world-model-dataset-visual-guide8k-video-game-datasetDataset Name: Kim2091's 8k Dataset V3
License: CC BY-NC-SA 4.0
Purpose: Video Game Upscaling
Description: This is the third revision of my 8k dataset. These games were all run at absolute maximum settings on RTX 3090 or 4090 at native 8k, with TAA disabled when possible.
This dataset is mostly compressed with ZSTD. You will have to use a tool like NanaZip to extract it if your main extraction tool does not support it
I tried to reduce the amount of blank space in the dataset, but due to the… See the full description on the dataset page: https://huggingface.co/datasets/Falln87/8k-video-game-dataset.gamed-game-classification-dataset
Gamed — Game Classification Dataset
Train and held-out eval frames for a 3-class game classifier
(cs2, dota2, valorant). Used to train and evaluate
ybashir/gamed-game-classification,
the §1.1 hot-path detector in coach-api.
The source .mp4 files are not included — see urls.txt for the
original YouTube links. All frames here are derived at 0.4 fps via
ffmpeg and resized to 224×224 (squashed 1:1).
Structure
train/
manifest.csv 6450 rows — path,class,video_id… See the full description on the dataset page: https://huggingface.co/datasets/ybashir/gamed-game-classification-dataset.gamedatasetarcade_game_datasetrpg_game_datasetgame_dataset_final