There shouldn't be any need to shrink the log file while doing the inserts. This just adds extra overhead as the log has to auto-grow back out as records are inserted.

It probably would have been a lot easier if you just set the recovery model to SIMPLE initially and left it until after you are done inserting, that way after each batch of records is committed the log space used is cleared and can be re-used for the next insert.

After you are done, shrink the log down, change the recovery model to FULL, and do a full backup and log backup.