Commit History

  1. chore: clean model comment wording⁠

    Harsh Yadav
    f19b67a
  2. feat: update image analysis prompts and training assets⁠

    Harsh Yadav
    22fdceb
  3. chore: mock ML endpoints with Gemini

    Harsh Yadav
    da71fe4
  4. fix: improve image analysis reliability and error handling

    Harsh Yadav
    c6daa99

  1. feat: add digital ink detection to ELA heuristic

    Harsh Yadav
    67e7ef3
  2. fix: remove CNN from build (OOM), inline tabular training, ELA heuristic fallback for image

    Harsh Yadav
    165fd8b

  1. fix: remove XGBoost use_label_encoder (removed in v2+), reduce CNN to 3k imgs/5 epochs for HF timeout, pre-download ResNet18

    Harsh Yadav
    ba6d7cd
  2. fix: load all HF dataset splits (validation+test), fix joblib CPU cores in container

    Harsh Yadav
    2a34453
  3. fix: pin numpy<2.0 and transformers==4.40.2 for torch 2.2.2 compatibility

    Harsh Yadav
    d156d1d
  4. fix: remove saved_models from git (trained at Docker build time)

    Harsh Yadav
    15e8801
  5. feat: hybrid CNN training with real HF cert images + fix Dockerfile offline vars bug

    Harsh Yadav
    1a7bfd2

  1. feat: scale training data to 25k rows + 5k CNN images

    Harsh Yadav
    844042d
  2. feat: migrate to HF Spaces — ResNet-18 CNN + BERT + DistilBERT + tabular ensemble

    Harsh Yadav
    8e68c3a

  1. feat: optimize for 512MB memory (Render free tier)

    Harsh Yadav
    19934f1
  2. FIX: optimize training models

    Harsh Yadav
    e3c3682
  3. UPADTE : created a endpoint to prevent sleep

    Harsh Yadav
    95a1208
  4. CREATE : Trained the model

    Harsh Yadav
    6de2f28