← 输入框|github.com|刷新
Skip to content

Data not available in benchmark-tutorial.ipynb #210

Open

Description

Working through competitions/cloud-cover/benchmark-tutorial.ipynb on a hub instance, the notebook states the data should be available in a volume but this is not the case:

---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
Input In [4], in <cell line: 5>()
      2 TRAIN_FEATURES = DATA_DIR / "train_features"
      3 TRAIN_LABELS = DATA_DIR / "train_labels"
----> 5 assert TRAIN_FEATURES.exists()

AssertionError: 

Activity

  1. TomAugspurger commented on Aug 15, 2022

    @TomAugspurger

    Thanks for the report. I was cleaning some things up in preparation for a Hub migration and completely forgot that this notebook existed :/

    I'll need to think a bit about how to adjust for that. The notebook will fail to run in the meantime, unless you signed up for the competition: https://www.drivendata.org/competitions/83/cloud-cover/data/

  2. robmarkcole commented on Aug 15, 2022

    @robmarkcole
    Author

    No problem! I did not sign up for the competition but appears as it has finished I cannot.
    I tried downloading but as this is 53GB I guess I wont have space. I only want to train on the NIR band, is there a way to request just that band?

  3. TomAugspurger commented on Aug 15, 2022

    @TomAugspurger

    I tried downloading but as this is 53GB I guess I wont have space.

    You might have space outside of your home directory (e.g. /tmp) but that's reset each time the notebook server restarts (https://planetarycomputer.microsoft.com/docs/overview/environment/#understanding-the-file-system).

    is there a way to request just that band?

    I'm not sure. With the assets hosted by the Planetary Computer you can access single bands. I don't recall how these assets were distributed, but it might have been a large gz or ZIP file.

  4. robmarkcole commented on Aug 15, 2022

    @robmarkcole
    Author

    yes they are large gz. NP I will download elsewhere and create a NIR version for use here, thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions