webbrain
Datasets
All datasets matching “webbrain”food-dataset
Food Dataset
An image classification dataset of food photos organized into 201 categories (folders), with 35,046 images total (~924 MB).
Each top-level folder is a category (e.g. adana kebab, sushi, waffles, tiramisu, ...) containing JPEG images of that food/dish. This follows the standard Hugging Face imagefolder layout, so it loads directly with:
from datasets import load_dataset
ds = load_dataset("webbrain-one/food-dataset")
Structure
<category… See the full description on the dataset page: https://huggingface.co/datasets/webbrain-one/food-dataset.webbrain-vl-2-450M-dataset
webbrain-vl-2-450M-dataset
Browser viewport screenshots paired with the six-section observation format used
by WebBrain's vision subsystem. Labels were generated by a teacher VLM with
DOM/accessibility evidence, then filtered for structure and grounded exact text.
This dataset trains
webbrain-one/webbrain-vl-2-450M.
The browser-ready export is
webbrain-one/webbrain-vl-2-450M-onnx.
Schema
image: browser viewport screenshot
system_prompt: WebBrain production vision… See the full description on the dataset page: https://huggingface.co/datasets/webbrain-one/webbrain-vl-2-450M-dataset.gym-exercises
Gym Exercises
A video library of gym/fitness exercise demonstrations, originally packaged as HTML5-compatible assets (exercises_html5_videos.tar) for a workout-tracking web app. Each exercise is numbered by an ID and provided in up to three encodings for browser compatibility: .mp4, .webm, and .ogv.
1,042 video files, ~2.3 GB, split by performer gender and video resolution/size tier:
male/original/<id>.<ext> # 298 files — full-size/original quality
male/small/<id>.<ext>… See the full description on the dataset page: https://huggingface.co/datasets/webbrain-one/gym-exercises.alpaca-turkish