Datasaur
Search
⌃K

Data Samples

As we mentioned, Datasaur has three primary task types: token-based, row-based, and document-based. Below are a series of data samples for all the tasks you can create on the Datasaur platform.

Token-based

The following zip files include sample label sets.
ner-samples (1).zip
5KB
Binary
NER samples
pos-samples.zip
4KB
Binary
POS samples
ocr-samples (1).zip
973KB
Binary
OCR samples

Token-based with arrows

The following zip files include sample label sets.
dependency-sample.zip
194KB
Binary
Dependency samples
coreference-sample (1).zip
832B
Binary
Coreference samples
relation-sample.zip
940B
Binary
Relation samples

Row-based

bookreview2020 (1) (1) (1) (1) (3).xlsx
5KB
Binary
Book Review
bookreview2020-labelset (1).csv
61B
Binary
Book Review question set
bookcover-multiplefiles.zip
1KB
Binary
Book Cover multiple files

Document-based

The following zip files including both question sets and answer sets.
imagesamplefiles (1).zip
1MB
Binary
Image sample files
pdfsamplefiles.zip
3MB
Binary
PDF samples
imagesamplefiles-hierarchical.zip
1MB
Binary
Hierarchical samples

Bounding box-based

sampleimages.zip
2MB
Binary
samplepdf.zip
3MB
Binary