Ahren W. Jasper
Ahren Jasper [[email protected]]
./comp.x in autofit/src
For pyfit wrapper input generator:
Mako Templates module for Python is required
see [https://www.makotemplates.org/]
The Fortran input file contains a series of records, each a single line of input parameters (see example fit.in files) Alternatively, input can be formatted in more user-friendly file and converted with accompanying Python wrapper (located in /pyfit) Below are listed each record and associated parameters, with data type in parantheses
Record 1: EnergyRanges
cut0 cut1 cut2 cut3 (dp)
-Values of energy ranges in kcal/mol with cut0 > cut1 > cut2 > cut3
Record 2: Epsilon
epsilon (dp)
-Range parameter in kcal/mol for the weight function (see )
Record 3: NumBatches
batches (int)
-Number of user defined batches of training set data in ai.all
Record 4: BatchZeroes
z1 z2... (dp)
-Sets the "zeroes" for the weight function for each batch
Record 5: BatchWeights
w1 w2... (dp)
-Sets the relative weights for each batch -- leave as 1 1 1 for default weighting
Record 6: NumAtoms
natom (int)
-Number of atoms in system
Record 7: AtomGroups
iagroup(natom) (int)
-Atom group of each atom
Record 8: Symbols
symb(natom) (char * 2)
-Atom symbol for each atom
Record 9: FactorOrder, TotalOrder
ipow ipowt (int)
-Maximum allowed order for each factor and total order of the term
Record 10: ReadBasis
lreadbasis (logical)
= TRUE, read basis from basis.dat
= FALSE, compute basis and write to basis.dat
Record 11: FindDisconnected
lreaddisc (logical)
= TRUE, determine disconnected atom groups to remove from basis
= FALSE, do not consider disconnected groups
if lreaddisc=TRUE:
Record 12: DisconnectedGroups
dgroup(natom) (int)
-Disconnected molecular groups in the product system. Currently only supports a single product channel.