Conformational sampling in excited states

[rgora]

Describe the issue
Crest is capable of finding MECP's - is it possible to perform conformational sampling in excited states?
In the latter case, the excited determinants are used according to [[calculation.level]] config keyword (which seems to be undocumented btw). Would the following input perform sampling for S1 state?

To Reproduce
#This is a CREST input file
input = 'start.xyz'
runtype='imtd-gc'
threads = 8

[[calculation.level]]
method = "gfn0*"
uhf = 0
chrg = 0
config = [1,1]

[pprcht]

In principle, yes. The GFN-xTB methods can't really distinguish between S₁ and T₁ however.

Maybe remove the uhf line for the gfn0* case. For gfn1 and gfn2, which do not have the config option, you could use uhf=2 to achieve something similar. The uhf parameter specifies the difference amount of spin α-β electrons, so setting it to two would be a T₁