You have probably collected several full handshakes for the same network (all data for this set of handshakes is present in the hccapx file) and hashcat loads all of them.
This shouldn't make much speed difference and there might be the tiny possibility that some of the handshakes are invalid (e.g. the user entered the wrong password e.g. if not all messages M1-M2-M3-M4 were captured).
If you really want to you could also clean the hccapx file and just use 1 handshake (the file format is explained here: https://hashcat.net/wiki/doku.php?id=hccapx, you could just split the file into several smaller hccapx files, or just use the first 393 bytes), but it's probably not worth it.
This shouldn't make much speed difference and there might be the tiny possibility that some of the handshakes are invalid (e.g. the user entered the wrong password e.g. if not all messages M1-M2-M3-M4 were captured).
If you really want to you could also clean the hccapx file and just use 1 handshake (the file format is explained here: https://hashcat.net/wiki/doku.php?id=hccapx, you could just split the file into several smaller hccapx files, or just use the first 393 bytes), but it's probably not worth it.