I am looking at using xgboost’s predict_proba to output the probabilities of a binary classification problem.
My local machine has no issues doing this with Dask - it outputs two probabilities 0/1.
However putting this into production my model only outputs a 1-D array with probabilities (unsure if it is 0 or 1) and was expecting to see two probabilities 0/1.
Very eager for any suggestions or why this may be. Thank you!
predictions = model.predict_proba(X) #create 0/1 probabilities for use
predictions_df = predictions.to_dask_dataframe() #convert dask array to dask dataframe
xgboost version: xgboost==1.3.3
dask version: dask-bigquery==2021.10.1