chore(settings): remove the bmm attention methods

Batch matrix-matrix and Dynamic Attention BMM applied a legacy Attention
processor to pipe.unet, which a diffusion transformer does not have, so
they served unet models alone and said nothing elsewhere. The choices, the
processor and its slice helper are removed, an unrecognized method now
warns rather than selecting nothing, and a stored value is rewritten to
Scaled-Dot-Product on load.

modules/shared_items.py

@@ -142,8 +142,6 @@ def list_crossattention():
         "Disabled",
         "Scaled-Dot-Product",
         "xFormers",
-        "Batch matrix-matrix",
-        "Dynamic Attention BMM"
     ]