Increasing the number of index partitions is the only way to increase the rate at with you can build an FTS index.
This allows for the index to be partitioned and each partition is built concurrently on a subset of the data coming from your couchbase bucket. Note that the resource consumption (CPU, RAM) by the index will increase proportionally during build.
Just to add to what is already pointed here by Abhinav,
Having a higher/sufficient FTS memory quota helps to speed up the indexing too.
( If there isn’t enough RAM memory quota set aside for FTS to operate, then the indexing would be slow
and that would be evident in the FTS logs with the log lines prefixed with
“[INFO] app_herder: indexing over indexQuota:”
So, if you are seeing a lot of similar log lines in the fts logs, then you know it is time to bump the RAM quota for FTS.