Train a Lennard-Jones potential

In this tutorial, we train a Lennard-Jones potential that is build in KLIFF (i.e. not models archived on OpenKIM_). From a user’s perspective, a KLIFF built-in model is not different from a KIM model.

Compare this with tut_kim_sw.

from kliff.calculators import Calculator
from kliff.dataset import Dataset
from kliff.loss import Loss
from kliff.models import LennardJones
from kliff.utils import download_dataset

# training set
dataset_path = download_dataset(dataset_name="Si_training_set_4_configs")
tset = Dataset(dataset_path)
configs = tset.get_configs()

# calculator
model = LennardJones()
model.echo_model_params()

# fitting parameters
model.set_opt_params(sigma=[["default"]], epsilon=[["default"]])
model.echo_opt_params()

calc = Calculator(model)
calc.create(configs)

# loss
loss = Loss(calc, nprocs=1)
result = loss.minimize(method="L-BFGS-B", options={"disp": True, "maxiter": 10})


# print optimized parameters
model.echo_opt_params()
model.save("kliff_model.yaml")

2023-08-01 21:59:15.496 | INFO     | kliff.dataset.dataset:_read:398 - 4 configurations read from /Users/mjwen.admin/Packages/kliff/docs/source/tutorials/Si_training_set_4_configs
2023-08-01 21:59:15.499 | INFO     | kliff.calculators.calculator:create:107 - Create calculator for 4 configurations.
2023-08-01 21:59:15.499 | INFO     | kliff.loss:minimize:310 - Start minimization using method: L-BFGS-B.
2023-08-01 21:59:15.500 | INFO     | kliff.loss:_scipy_optimize:427 - Running in serial mode.
 This problem is unconstrained.

#================================================================================
# Available parameters to optimize.
# Parameters in `original` space.
# Model: LJ6-12
#================================================================================

name: epsilon
value: [1.]
size: 1

name: sigma
value: [2.]
size: 1

name: cutoff
value: [5.]
size: 1


#================================================================================
# Model parameters that are optimized.
# Note that the parameters are in the transformed space if 
# `params_transform` is provided when instantiating the model.
#================================================================================

sigma 1
  2.0000000000000000e+00 

epsilon 1
  1.0000000000000000e+00 


RUNNING THE L-BFGS-B CODE

           * * *

Machine precision = 2.220D-16
 N =            2     M =           10

At X0         0 variables are exactly at the bounds

At iterate    0    f=  6.40974D+00    |proj g|=  2.92791D+01

At iterate    1    f=  2.98676D+00    |proj g|=  3.18782D+01

At iterate    2    f=  1.56102D+00    |proj g|=  1.02614D+01

At iterate    3    f=  9.61567D-01    |proj g|=  8.00167D+00

At iterate    4    f=  3.20489D-02    |proj g|=  7.63379D-01

At iterate    5    f=  2.42400D-02    |proj g|=  5.96998D-01

At iterate    6    f=  1.49914D-02    |proj g|=  6.87782D-01

At iterate    7    f=  9.48615D-03    |proj g|=  1.59376D-01

At iterate    8    f=  6.69609D-03    |proj g|=  1.14378D-01

2023-08-01 21:59:16.968 | INFO     | kliff.loss:minimize:312 - Finish minimization using method: L-BFGS-B.

At iterate    9    f=  4.11024D-03    |proj g|=  3.20712D-01

At iterate   10    f=  2.97209D-03    |proj g|=  7.03411D-02

           * * *

Tit   = total number of iterations
Tnf   = total number of function evaluations
Tnint = total number of segments explored during Cauchy searches
Skip  = number of BFGS updates skipped
Nact  = number of active bounds at final generalized Cauchy point
Projg = norm of the final projected gradient
F     = final function value

           * * *

   N    Tit     Tnf  Tnint  Skip  Nact     Projg        F
    2     10     13      1     0     0   7.034D-02   2.972D-03
  F =   2.9720927488600178E-003

STOP: TOTAL NO. of ITERATIONS REACHED LIMIT                 
#================================================================================
# Model parameters that are optimized.
# Note that the parameters are in the transformed space if 
# `params_transform` is provided when instantiating the model.
#================================================================================

sigma 1
  2.0629054951532582e+00 

epsilon 1
  1.5614850326987884e+00