Below is the difference given between Standalone, Pseudo Distributed and Fully Distributed Mode:
Standalone mode – Default mode of Hadoop which utilize local file system for input and output operations instead of HDFS and is mainly used for debugging purpose
Pseudo Distributed mode (Single node cluster) – Hadoop cluster will be set up on a single server running all Hadoop daemons on one node and is mainly used for real code to test in HDFS.
Fully Distributed mode (Multi node cluster) – Setting up Hadoop cluster on more than one server enabling a distributed environment for storage and processing which is mainly used for production phase