Hugging Face's logo

Spaces:
deep-learning
/
image-forgery-detector-v2
Running

App Files Community
image-forgery-detector-v2

Commit History

  1. Update app.py from main Running

    usamaalam
    a3b3a42

  1. Add model comparison script⁠

    usamaalam +1: Claude Haiku 4.5
    37a3120

  1. Enlarge tab caption font size⁠

    usamaalam +1: Claude Opus 4.8
    e09dc9f

  2. Switch to complete dark professional theme⁠

    usamaalam +1: Claude Opus 4.8
    d7ae285

  3. Fix header/footer consistency — full-bleed banner, unified page background⁠

    usamaalam +1: Claude Sonnet 4.6
    68cf859

  4. Add Model & Training tab with professional header/footer⁠

    usamaalam +1: Claude Sonnet 4.6
    ae857da

  5. Add CASIA scope note to UI

    usamaalam
    62b2f30

  6. Add debug instrumentation (build version, input order, tensor stats, raw pred)

    usamaalam
    53d4919

  7. Fix ELA preprocessing to match training notebook (amplify + JPEG + tf.image decode)

    usamaalam
    e3ec511

  8. Fix ELA computation to match train.py (ImageChops.multiply, not Brightness)⁠

    usamaalam
    0ceb664

  9. Fix: update sdk_version to 1.40.2, remove streamlit from requirements

    usamaalam
    54ee03a

  10. Reference M3_best_v2.h5 (full model saved from TF 2.20)

    usamaalam
    2a2cecd

  11. Upgrade to Streamlit 1.40 + TF 2.20 (resolves protobuf conflict); fix use_container_width

    usamaalam
    93ab4cb

  12. Revert to TF 2.15 (protobuf-compatible with Streamlit); use cpu build

    usamaalam
    48027b6

  13. Load full H5 model via load_model; pin TF 2.20 to match training runtime

    usamaalam
    88283ca

  14. Fix build_model to exactly match train.py architecture and preprocessing

    usamaalam
    11cc0f4

  15. Fix missing layers import in build_model

    usamaalam
    84cddaa

  16. Clean deployment: model loaded from HF Hub at runtime⁠

    usamaalam +1: Claude Sonnet 4.6
    55ae7dd