Load 9 inputs, push a button then run model, THEN allow user to filter model results without rerunning script

Using Streamlit
1

I have 9 inputs to a model that I allow my user to adjust, THEN they press a button to run the model (that way the model doesn’t try to run until every input has been updated). Once the model runs, I have a dataframe (model results) I display that I want to give the user the ability to filter with a multiselect, however, when I make a filter selection, it re-runs the script and resets back to before the button was pushed…

format of my code:

#function that runs my model
    @st.cache()
    def model_run()

#I have eight number inputs that take different values to enter into the model
    st.sidebar.number_input() X 8

#button to run the model
    if st.button('Run Model'):
        model_results = model_run()

#mutliselect to let user filter model run results 
        options = st.multiselect("select a filter")
        new_df = model_results_df[model_results_df['Entity'].isin(options)]
        st.dataframe(new_df)

However, when I run the code above, when my user makes a selection, it reruns the script back to before I clicked my button and my output is lost.