How to improve load time while loading from file through copy command?

Loading data with the advantage of parallelism will speed up the operation.
Distribute your files on several nodes to load in parallel, instead of loading on a single node. You can use a wildcard or glob (such as *.gz) to load multiple input files, combined with the ON ANY NODE clause, then COPY will distribute the list of files to all cluster nodes to spread the workload.

If you have only a single file, Vertica could divide it and load its portions (apportioned load) on several nodes in parallel, however, you cannot use apportioned load for a compressed file.