from peslearn.ml import NeuralNetwork
from peslearn import InputProcessor
import torch
import numpy as np
from itertools import combinations
nn = NeuralNetwork('PES.dat', InputProcessor(''), molecule_type='A2B')
params = {'layers': (16,), 'morse_transform': {'morse': False}, 'pip': {'degree_reduction': True, 'pip': True}, 'scale_X': {'activation': 'tanh', 'scale_X': 'mm11'}, 'scale_y': 'mm11', 'lr': 0.5}
X, y, Xscaler, yscaler = nn.preprocess(params, nn.raw_X, nn.raw_y)
model = torch.load('model.pt')
# How to use 'compute_energy()' function
# --------------------------------------
# E = compute_energy(geom_vectors, cartesian=bool)
# 'geom_vectors' is either:
# 1. A list or tuple of coordinates for a single geometry.
# 2. A column vector of one or more sets of 1d coordinate vectors as a list of lists or 2D NumPy array:
# [[ coord1, coord2, ..., coordn],
# [ coord1, coord2, ..., coordn],
# : : : ],
# [ coord1, coord2, ..., coordn]]
# In all cases, coordinates should be supplied in the exact same format and exact same order the model was trained on.
# If the coordinates format used to train the model was interatomic distances, each set of coordinates should be a 1d array of either interatom distances or cartesian coordinates.
# If cartesian coordinates are supplied, cartesian=True should be passed and it will convert them to interatomic distances.
# The order of coordinates matters. If PES-Learn datasets were used they should be in standard order;
# i.e. cartesians should be supplied in the order x,y,z of most common atoms first, with alphabetical tiebreaker.
# e.g., C2H3O2 --> H1x H1y H1z H2x H2y H2z H3x H3y H3z C1x C1y C1z C2x C2y C2z O1x O1y O1z O2x O2y O2z
# and interatom distances should be the row-wise order of the lower triangle of the interatom distance matrix, with standard order atom axes:
# H H H C C O O
# H
from peslearn.ml import GaussianProcess
from peslearn import InputProcessor
from GPy.core.model import Model
import numpy as np
import json
from itertools import combinations
gp = GaussianProcess('PES.dat', InputProcessor(''), molecule_type='A2B')
params = {
'morse_transform': {
'morse': True,
'morse_alpha': 1.6
},
'pip': {
'degree_reduction': False,
'pip': True
},
'scale_X': 'mm01',
'scale_y': 'std'
}
X, y, Xscaler, yscaler = gp.preprocess(params, gp.raw_X, gp.raw_y)
model = Model('mymodel')
with open('model.json', 'r') as f:
model_dict = json.load(f)
final = model.from_dict(model_dict)
# How to use 'compute_energy()' function
# --------------------------------------
# E = compute_energy(geom_vectors, cartesian=bool)
# 'geom_vectors' is either:
from compute_energy import pes
from peslearn.ml import NeuralNetwork
from peslearn import InputProcessor
import torch
import numpy as np
from itertools import combinations
from communicate.get_energy import read_eg
nn = NeuralNetwork('/home/luoshu/PycharmProjects/gaus_ipi/PES.dat',
InputProcessor(''),
molecule_type='A2B')
params = {
'layers': (128, 128, 128),
'morse_transform': {
'morse': True,
'morse_alpha': 1.3
},
'pip': {
'degree_reduction': False,
'pip': True
},
'scale_X': {
'activation': 'tanh',
'scale_X': 'mm11'
},
'scale_y': 'std',
'lr': 0.5
}
X, y, Xscaler, yscaler = nn.preprocess(params, nn.raw_X, nn.raw_y)
model = torch.load('/home/luoshu/PycharmProjects/gaus_ipi/model.pt')