def grada(direcs, v='valang_O', dft='siesta', batch=100):
''' variables like: bo1_C-H, boc5_C rosi_C-H boc1
'''
print('- grading ... ... ')
for m in direcs:
mol = m
rn = ReaxFF(libfile='ffield',
direcs=direcs,
dft=dft,
batch_size=batch,
pkl=True)
rn.initialize()
rn.session(learning_rate=1.0e-4, method='AdamOptimizer')
fg = open('gradient.txt', 'w')
l = 'eang'
grad = rn.get_gradient(rn.__dict__[l][mol], rn.p[v])
print('- the gradient of %s_%s/%s is: ' % (l, mol, v), grad, file=fg)
fg.close()
# anglit = ['D_ang','thet','theta0','expang','f_7','f_8','EANG','EPEN','ETC']
anglit = ['EANG']
angs = rn.angs
# angs = ['C-C-C']
for ang in angs:
if rn.nang[ang] > 0:
for l in anglit:
grad = rn.get_gradient(rn.__dict__[l][ang], rn.p[v])
print('- the gradient of %s/%s is: ' % (l + '_' + ang, v),
grad)
rn.sess.close()
def gradbm(direcs, v='bo5', bd='H-H', dft='siesta', batch=100):
''' variables like: bo1_C-H, boc5_C rosi_C-H boc1
'''
v_ = v
v = v + '_' + bd
print('- grading ... ...')
rn = ReaxFF(libfile='ffield',
direcs=direcs,
dft=dft,
batch_size=batch,
pkl=True)
rn.initialize()
rn.session(learning_rate=3.0e-4, method='AdamOptimizer')
bdlit = [
'bop', 'bop_si', 'bop_pi', 'bop_pp', 'f_1', 'f_2', 'f_3', 'f_4', 'f_5',
'bosi', 'bopi', 'bopp', 'powb', 'expb', 'sieng', 'EBD'
]
gl = rn.get_gradient(rn.Loss, rn.p[v])
logger.info('- the gradient of Loss/%s is %f.' % (v, gl))
bonds = rn.bonds
bonds = [bd]
for bd in bonds:
v = v_ + '_' + bd
if rn.nbd[bd] > 0:
grad = rn.get_gradient(rn.__dict__['EBD'][bd], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
('EBD' + '_' + bd, v, str(grad)))
if not grad is None:
if np.isnan(grad):
for l in bdlit:
grad = rn.get_gradient(rn.__dict__[l][bd], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(l + '_' + bd, v, str(grad)))
ml = [
'ebond', 'elone', 'eover', 'eunder', 'eang', 'epen', 'tconj', 'etor',
'efcon', 'evdw', 'ehb'
]
for bd in bonds:
v = v_ + '_' + bd
if rn.nbd[bd] > 0:
for m in direcs:
mol = m
grad = rn.get_gradient(rn.__dict__['loss'][mol], rn.p[v])
logger.info('- the gradient of %s_%s/%s is: %s' %
('loss', mol, v, str(grad)))
if not grad is None:
if np.isnan(grad):
for l in ml:
grad = rn.get_gradient(rn.__dict__[l][mol],
rn.p[v])
logger.info('- the gradient of %s_%s/%s is: %s' %
(l, mol, v, str(grad)))
rn.sess.close()
def gradt(v='bo2_C-N',
direcs={},
torlit=['f_10', 'f_11', 'expv2', 'ETOR', 'Efcon'],
batch=100):
''' variables like: bo1_C-H, boc5_C rosi_C-H boc1
torlit = ['f_10','f_11','expv2','ETOR','Efcon']
'''
print('- grading ... ... ')
rn = ReaxFF(libfile='ffield',
direcs=direcs,
dft='siesta',
batch_size=batch,
pkl=True)
rn.initialize()
rn.session(learning_rate=1.0e-4, method='AdamOptimizer')
fg = open('gradient.txt', 'w')
grad = rn.get_gradient(rn.Loss, rn.p[v])
print('- the gradient of Loss/%s is: ' % v, grad, file=fg)
molit = ['etor', 'efcon']
for mol in rn.direcs:
for l in molit:
grad = rn.get_gradient(rn.__dict__[l][mol], rn.p[v])
print('- the gradient of %s/%s is: ' % (l + '_' + mol, v),
grad,
file=fg)
fg.close()
i = 0
for tor in rn.tors:
if rn.ntor[tor] > 0:
# if tor=='C-H-H-C':
for l in torlit:
fg = open('gradient.txt', 'a')
grad = rn.get_gradient(rn.__dict__[l][tor], rn.p[v])
print('- the gradient of %s/%s is: ' % (l + '_' + tor, v),
grad,
file=fg)
fg.close()
i += 1
rn.sess.close()
def gradb(direcs,
v='bo5',
bd='H-H',
nn=False,
bo_layer=[9, 2],
debd=True,
deba=True,
deang=True,
dft='siesta',
batch=100):
''' variables like: bo1_C-H, boc5_C rosi_C-H boc1
'''
v_ = v
v = v + '_' + bd
print('- grading ... ...')
ffield = 'ffield.json' if nn else 'ffield'
rn = ReaxFF(libfile=ffield,
direcs=direcs,
dft=dft,
nn=nn,
bo_layer=bo_layer,
batch_size=batch,
pkl=True)
rn.initialize()
rn.session(learning_rate=3.0e-4, method='AdamOptimizer')
if nn:
bdlit = [
'bop', 'bop_si', 'bop_pi', 'bop_pp', 'F', 'bosi', 'bopi', 'bopp',
'powb', 'expb', 'sieng', 'EBD'
]
else:
bdlit = [
'bop', 'bop_si', 'bop_pi', 'bop_pp', 'f_1', 'f_2', 'f_3', 'f_4',
'f_5', 'bosi', 'bopi', 'bopp', 'powb', 'expb', 'sieng', 'EBD'
]
if debd:
bonds = rn.bonds
for b in bonds:
v = v_ + '_' + b
grad = rn.get_gradient(rn.Loss, rn.p[v])
text_ = '- the gradient of Loss/%s is ' % v
logger.info(text_ + str(grad))
if grad is None:
continue
if not np.isnan(grad):
continue
if rn.nbd[b] > 0:
grad = rn.get_gradient(rn.__dict__['EBD'][b], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
('EBD' + '_' + b, v, str(grad)))
if not grad is None:
if np.isnan(grad):
for l in bdlit:
grad = rn.get_gradient(rn.__dict__[l][bd], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(l + '_' + b, v, str(grad)))
v = v_ + '_' + bd
if deba:
sl = ['EL', 'EOV', 'EUN']
alist = {
'EL': ['Delta_lp', 'Delta_e', 'D', 'explp'],
'EOV': ['Delta_lpcorr', 'so', 'otrm1', 'otrm2'],
'EUN': []
}
for sp in rn.spec:
for l in sl:
if sp in rn.__dict__[l]:
grad = rn.get_gradient(rn.__dict__[l][sp], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(l + '_' + sp, v, str(grad)))
if not grad is None:
if np.isnan(grad):
for al in alist[l]:
grad = rn.get_gradient(rn.__dict__[al][sp],
rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(al + '_' + sp, v, str(grad)))
if deang:
al = ['EANG', 'EPEN', 'ETC']
for ang in rn.angs:
# v = 'val1'+'_'+ang
if rn.nang[ang] > 0:
for l in al:
grad = rn.get_gradient(rn.__dict__[l][ang], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(l + '_' + ang, v, str(grad)))
tl = ['ETOR', 'Efcon']
for tor in rn.tors:
# v = 'tor2' # +'_'+tor
if rn.ntor[tor] > 0:
for l in tl:
grad = rn.get_gradient(rn.__dict__[l][tor], rn.p[v])
logger.info('- the gradient of %s/%s is: %s' %
(l + '_' + tor, v, str(grad)))
rn.sess.close()