Datasets:
Tasks:
Image Classification
Formats:
parquet
Sub-tasks:
multi-label-image-classification
Size:
10K - 100K
License:
Hi everyone! This dataset is still in development and collecting data
#1
by
stochastic
- opened
If you see a problem feel free to let us know! There are so many improvements that can be done....
- improving the combination of individual images to me a more beautiful panoramic
- adding more labels or object detection
- improving the coverage of countries
- adding splits (once the number of images and limitations are finalized)
- dropping bad data (some images are scraped and are completely blacked out)
Hello! Thanks for uploading this! Kinda of a dumb question, but do you know how to get the filename? I couldn't find the name from the parquet, and when I read the image from the parquet, the image.filename was an empty string.