I'm building my first interface with Gradio and I've encountered an issue. The data I used for training my XGB classifier contains NaN values, and I want to enable possibility of NaN input in Gradio interface. As for now I've tried to set default field value to NaN, but it's not working.
Interface code: Ferr = gr.Number(label="Ferrytyna [ug/l]", value=float("NaN"))
Error message:
ValueError: DataFrame.dtypes for data must be int, float, bool or category. When categorical type is supplied, The experimental DMatrix parameterenable_categorical must be set to True. Invalid columns:ferrytyna_krew_ug_l: object
Does anyone knows how to solve this? Thanks
