.. _example_gradcam_fnf:

Example: Doing Grad-CAM on the Food-Non-Food (FNF) model
========================================================

The code below is used to perform Grad-CAM on the FNF model.

You can refer to :ref:`tutorial_gradcam_fnf` for an in-depth explanation for intuition
behind the code.

.. code:: python3
    :number-lines:
    
    import XAI

    ds = XAI.datasets.DatasetFromCSV(path_to_csv="/path/to/desired/csv/data.csv")

    fnf_gc = XAI.FNFGradCam(
        model=model,
        target_layer_name="conv2d_7b",
    )

    fnf_results_df = fnf_gc.do_gradcam(
        image_dataset=ds,
        batch_size=2,
        relu_attributions=False,
        return_metadata=True,
        output_folder="/output_folder",
        plot_gradcam=True
    )

You can also execute the ``plot_gradcam()`` function independently from the
``do_gradcam()`` function by setting ``plot_gradcam=False`` and doing the
following:

.. code:: python3
    :number-lines:
    
    import XAI

    ds = XAI.datasets.DatasetFromCSV(path_to_csv="/path/to/desired/csv/data.csv")

    fnf_gc = XAI.FNFGradCam(
        model=model,
        target_layer_name="conv2d_7b",
    )

    fnf_results_df = fnf_gc.do_gradcam(
        image_dataset=ds,
        batch_size=2,
        relu_attributions=False,
        return_metadata=True,
        output_folder="/output_folder",
        plot_gradcam=False
    )

    fnf_results_df = fnf_gc.plot_gradcam(
        do_gradcam_df=fnf_results_df,
        savefig_dir="/output_folder",
        savefig=True,
        plt_show=True,
        save_csv=True
    )

After performing Grad-CAM on the FNF model as shown above, you can generate a
PowerPoint report of the Grad-CAM results using the
``XAI.report_generation.generate_pptx()`` function. You can refer to
:ref:`example_pptx_report` for the code.