Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SDK: Improve the PyTorch adapter layer #5455

Merged
merged 2 commits into from
Dec 16, 2022

Commits on Dec 13, 2022

  1. SDK: change extractor transforms to return tensors instead of primiti…

    …ve types
    
    This is more consistent with torchvision models, and lets the
    `ExtractBoundingBoxes` outputs to be used directly with the `torchmetrics`
    package (or at least the `MeanAveragePrecision` class).
    SpecLad committed Dec 13, 2022
    Configuration menu
    Copy the full SHA
    2277dfd View commit details
    Browse the repository at this point in the history
  2. TaskVisionDataset: add a way to specify a custom label-name-to-index …

    …mapping
    
    The default numbering is generally unpredictable, so it's not especially
    useful outside of basic use cases. Given that there's no way to specify
    custom indexes for labels on the server side, make it possible on the client
    side.
    
    The input dictionary maps label _names_ to indexes, because names is what a
    CVAT user would be familiar with. The output dictionary in the `Target`
    class still maps label _IDs_, because that's what CVAT uses in the
    annotation data models.
    SpecLad committed Dec 13, 2022
    Configuration menu
    Copy the full SHA
    d06e041 View commit details
    Browse the repository at this point in the history