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.

Sign up for GitHub

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

Jump to bottom

TrainingSample: refactor and encapsulate image handling, improving performance and reliability #374

Merged
merged 14 commits into from
May 2, 2024
Merged

TrainingSample: refactor and encapsulate image handling, improving performance and reliability #374

merged 14 commits into from
May 2, 2024

Commits on May 2, 2024

  1. wip: training sample handler

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    a4a780e View commit details
    Browse the repository at this point in the history

  2. fix typo in pixel alignment warning

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    ad441e9 View commit details
    Browse the repository at this point in the history

  3. training sample (wip) should set up resolution_type before the target… 

    … size is calculated. we will keep target sizes up to date on the object.
    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    d2f768d View commit details
    Browse the repository at this point in the history

  4. use the training sample wrapper for image preparation during metadata… 

    …/aspect bucket scan
    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    7044f86 View commit details
    Browse the repository at this point in the history

  5. base metadata backend should no longer call function to convert to str

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    f5a31c8 View commit details
    Browse the repository at this point in the history

  6. update debug logging for training loop to have better phrasing, more … 

    …information for DF training issues
    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    32632a9 View commit details
    Browse the repository at this point in the history

  7. update PIL type hint

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    2035fa4 View commit details
    Browse the repository at this point in the history

  8. disable torch anomaly detector

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    554c6d0 View commit details
    Browse the repository at this point in the history

  9. nuke prepare_image from existence

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    63dc4a1 View commit details
    Browse the repository at this point in the history

  10. mps: increase allowed batch size to 16 since anomaly detector eats a … 

    …lot of compute, and is now disabled.
    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    de23593 View commit details
    Browse the repository at this point in the history

  11. training sample should use a structured return for prepared samples

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    e48fe87 View commit details
    Browse the repository at this point in the history

  12. switch to using PreparedSample

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    51bb2a7 View commit details
    Browse the repository at this point in the history

  13. trainingsample: fix square and aspect-preserving crops.

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    2202ab0 View commit details
    Browse the repository at this point in the history

  14. remove prints

    bghira committed May 2, 2024
    Configuration menu
    Copy the full SHA
    cbf0f07 View commit details
    Browse the repository at this point in the history