Image dataset
10 datasets showcase how to configure and load image datasets
Viewer • Updated • 3Note If your dataset only consists of one column with images, you can use a simple setup: store your image files at the root.
datasets-examples/doc-image-2
Viewer • UpdatedNote or in a subdirectory
datasets-examples/doc-image-3
Viewer • UpdatedNote Multiple formats are supported at the same time, including PNG, JPEG, TIFF and WebP.
datasets-examples/doc-image-4
Viewer • Updated • 3Note If there is additional information you'd like to include about your dataset, like text captions or bounding boxes, add it as a `metadata.csv` file in your repository.
datasets-examples/doc-image-5
Viewer • Updated • 8Note You can also use a JSONL file `metadata.jsonl`.
datasets-examples/doc-image-6
Viewer • UpdatedNote You can also locate the images in a different subdirectory of the split
datasets-examples/doc-image-7
Viewer • UpdatedNote For image classification datasets, you can also use a simple setup: use directories to name the image classes.
datasets-examples/doc-image-8
Viewer • UpdatedNote You can also use provide multiple splits.
datasets-examples/doc-image-9
Viewer • UpdatedNote You can disable this automatic behavior in the YAML configuration. If your directory names have no special meaning, set `drop_labels: true` in the README header
datasets-examples/doc-image-10
Viewer • UpdatedNote Instead of uploading the images and metadata as individual files, you can embed everything inside a Parquet file.