Instructions to use fsabiu/keras-modelscan-torchmodulewrapper-coverage-gap with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Keras
How to use fsabiu/keras-modelscan-torchmodulewrapper-coverage-gap with Keras:
# Available backend options are: "jax", "torch", "tensorflow". import os os.environ["KERAS_BACKEND"] = "jax" import keras model = keras.saving.load_model("hf://fsabiu/keras-modelscan-torchmodulewrapper-coverage-gap") - Notebooks
- Google Colab
- Kaggle
| # Keras ModelScan Coverage Validation Artifacts | |
| This repository contains benign Keras V3 `.keras` files for validating scanner behavior. | |
| ## Files | |
| - `benign_torch_module.keras`: benign Keras model containing `TorchModuleWrapper`. | |
| - `benign_lambda.keras`: benign Keras model containing a `Lambda` layer as a positive scanner control. | |
| - `modelscan-benign_torch_module.json`: ModelScan 0.8.8 JSON output for `benign_torch_module.keras`. | |
| - `modelscan-benign_lambda.json`: ModelScan 0.8.8 JSON output for `benign_lambda.keras`. | |
| - `detector-summary.md`: local static detector comparison summary. | |
| ## Reproduce | |
| ```sh | |
| python3 -m venv .venv | |
| .venv/bin/python -m pip install 'keras==3.14.0' torch h5py | |
| .venv/bin/python -m pip install 'modelscan==0.8.8' 'modelscan[tensorflow]==0.8.8' | |
| .venv/bin/modelscan scan -p benign_lambda.keras -r json | |
| .venv/bin/modelscan scan -p benign_torch_module.keras -r json | |
| ``` | |
| Expected result: | |
| - `benign_lambda.keras`: ModelScan reports one Medium Lambda issue. | |
| - `benign_torch_module.keras`: ModelScan reports zero issues. | |
| The files are benign and are provided only for scanner validation. | |