[ot][spam][crazy][data] transformer model 'attention' improvement

[Undiscussed Horrific Abuse, One Victim & Survivor of Many]

arright! this is hard!

maybe we can expand some test tensor to be really big and see if
memory allocation changes