I usually prefer to get as much learn data captured before, performing a transfer. It's already part of the gross fuel table. I'll run one or two iterations and manually clean up the fuel map (peaks/valleys). As indicated above, your cam may or may not like low speed learn or closed loop for that matter, you'll need to verify if the data is valid. There are some cool features of the datalogger using the scatter plot that show how well areas of your run are mapped.