Datasets:
PR to add segmentation train/test/val splits and correct identification CSVs
@ksmehrab , I added the README that goes along with these CSV updates. Note that no subsets will actually show the images as it is now.
I'm adding the large metadata.csv file generated with the notebook in this PR on the GH repo to #5, since that should match the status of the repo used to generate it.
The CSVs from my update used those from #7 (a9464827ff5c608f4079d24d2c24b939b036aa4e), so they don't incorporate changes from #11 (c9d096b1081c7e20adb80b1ca4bff9394ea039cd); however, the CSV combination process can easily be re-run for those -- assuming they're the correct CSVs to use at this point. @ksmehrab , could you clarify the status, since I know there were updates at some point? Is it just the README that needs updating to reflect the proper stats?
- README.md
- classification_test.csv
- classification_train.csv
- classification_val.csv
- identification_test_insp.csv
- identification_test_lvsp.csv
- identification_train.csv
- identification_val.csv
- segmentation_test.csv
- segmentation_train.csv
- segmentation_val.csv