"unorder_set" is really much faster, and I've never used "set<vector>" before, thanks for your advice, it's all new to me!

• Question D made me very sad. In the process of the competition, in order to avoid the single hash being hacked, I used double hash, which passed normally during the competition and took 1.6s. However, the retest after the competition was timed out due to the new data, but I passed all the samples after changing to single hash.
• (In the single hash case, obviously, it's easy to create a special sample hack when you know the hash value.)
• TLE-CODE:129383238
• AC-CODE:129474392