PLUMED-NEST

Project ID: plumID:19.060
Source: 3_ala4/simulation/plumed.dat
Originally used with PLUMED version: 2.5
Stable: zipped raw stdout - zipped raw stderr - stderr
Master: zipped raw stdout - zipped raw stderr - stderr

Click on the labels of the actions for more information on what each action computes

# vim:ft=plumed
Enables syntax highlighting for PLUMED files in vim. See here for more details. 

LOAD
Loads a library, possibly defining new actions. More details
 
FILE
file to be loaded
=../../0_code/NeuralNetworkVes.cpp

phi1: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=5,7,9,15
The TORSION action with label phi1 calculates a scalar quantityphi2: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=15,17,19,25
The TORSION action with label phi2 calculates a scalar quantityphi3: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=25,27,29,35
The TORSION action with label phi3 calculates a scalar quantitypsi1: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=7,9,15,17
The TORSION action with label psi1 calculates a scalar quantitypsi2: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=17,19,25,27
The TORSION action with label psi2 calculates a scalar quantitypsi3: 
TORSION
Calculate a torsional angle. More details
 
ATOMS
the four atoms involved in the torsional angle
=27,29,35,37

# the following is a fake example using the grid-based version
# see below ENDPLUMED for the MC-based version used in the paper

The TORSION action with label psi3 calculates a scalar quantity
NN_VES
This action is not part of PLUMED and was included by using a LOAD command More details
 ...
  TEMP=300.0
  LABEL=bias
  ARG=phi1,phi2,phi3,psi1,psi2,psi3
  NODES=96,48,24
  OPTIM=ADAM
  ACTIVATION=RELU
  GRID_MIN=-pi,-pi,-pi,-pi,-pi,-pi
  GRID_MAX=pi,pi,pi,pi,pi,pi
  GRID_BIN=10,10,10,10,10,10 # set to 10 to avoid std_bad_alloc
  AVE_STRIDE=500
  PRINT_STRIDE=100
  TARGET_STRIDE=1
  LRATE=0.001
  GAMMA=10
  TAU_KL=100000
  DECAY=2000
  ADAPTIVE_DECAY=20
... NN_VES


PRINT
Print quantities to a file. More details
 
ARG
the input for this action is the scalar output from one or more other actions
 
STRIDE
 the frequency with which the quantities of interest should be output
=100 
FILE
the name of the file on which to output these quantities
=COLVAR

ENDPLUMED
Terminate plumed input. More details


# The following input is related to the MC-based version used in the paper. 
This is not compatible with newer plumed versions, if you want to try it you should use the code NeuralNetworkVes_ORIGINAL.cpp. 

NN_VES ...
  TEMP=300.0
  LABEL=bias
  ARG=phi1,phi2,phi3,psi1,psi2,psi3
  NODES=96,48,24
  OPTIM=ADAM
  ACTIVATION=RELU
  GRID_MIN=-pi,-pi,-pi,-pi,-pi,-pi
  GRID_MAX=pi,pi,pi,pi,pi,pi
  GRID_BIN=300,300,300,300,300,300
  AVE_STRIDE=500
  PRINT_STRIDE=100
  TARGET_STRIDE=1
  LRATE=0.001
  GAMMA=10
  TAU_KL=100000
  MODE_DECAY=4
  DECAY_STEP_START=10000
  DECAY=2000
  ADAPTIVE_DECAY=20
  MC_POINTS=20000
  MC_SIGMA=0.5,0.5,0.5,0.5,0.5,0.5
  CALC_MC_ERR=0
  MC_CONVERGENCE_TEST=0
  L2_weight=0.2
  SERIAL
  CALC_RCT
... NN_VES