Please mention Yield for apoc.create.relationship(f,m.relship,{weight:10}, t)
i.e.,
CALL apoc.load.xls('file:///filename.xlsx','Nodes') YIELD map as m CALL
CALL apoc.create.node([m.fromNodeLabel], {name:m.fromNodeName,value:m.fromValue,instanceID:m.InstanceID}) YIELD node as f
CALL apoc.create.node([m.toNodeLabel], {name:m.toNodeName,instanceID:m.InstanceID}) YIELD node as t
CALL apoc.create.relationship(f,m.relship,{weight:10}, t) yield rel as rel return rel,f,t
If you are running a procedure all by itself then you don't need to do yield. If you are using somewhere in the middle of a cypher query you need to do yield.
If you need to use the output of a procedure use Yield. This use could be in Return statement or in other query etc.
Procedure apoc.create.relationship can just give you output node or relationship and that is what I did. In the neo4j documentation details of Yield is only given for generic procedures like CALL apoc.create.uuids(count) YIELD uuid, row.