def __init__(self, total_images, nsets, job):
self.total_images = total_images
self.nsets = nsets
if self.nsets == 1:
self.train_all = self.total_images[:]
else:
self.parts = divide_int(len(self.total_images), self.nsets)
def __init__(self, total_images, nsets, job):
self.total_images = total_images
if isinstance(nsets, int):
self.nsets=nsets
else if isinstance(nsets, list):
if
if self.nsets == 1:
self.train_all=self.total_images[:]
else:
self.parts = divide_int(len(self.total_images), self.nsets)
def __init__(self, total_images, nsets=None):
self.total_images = total_images
self.nsets = nsets
### in case nsets is defined
if isinstance(nsets, int):
if self.nsets == 1:
self.train_all = self.total_images[:]
else:
### defin self.parts
self.parts = divide_int(len(self.total_images), self.nsets)
def __init__(self, total_images, data_type, data_mining):
self.total_images = total_images
self.data_type = data_type
self.data_mine = data_mining
self.nsets = None
self.parts = None
if self.data_type == 'set':
self.nsets = data_mining
if self.nsets == 1:
self.train_all = self.total_images[:]
else:
### defin self.parts
self.parts = divide_int(len(self.total_images), self.nsets)
def get_val_train_images(self, n):
parts = divide_int(len(self.total_images), self.nsets)
print(parts, len(self.total_images))
if n == 0:
v_start = 0 # v for valence set area
v_stop = parts[0]
else:
v_start = parts[n-1]
v_stop = parts[n]
val_images = self.total_images[v_start:v_stop]
if n == 0:
tra_images = self.total_images[v_stop:parts[-2]]
else:
tra_images = self.total_images[:v_start]
if n != self.nsets-2:
tra_images.extend(self.total_images[v_stop:parts[-2]])
return tra_images, val_images
def get_val_train_images(self, ndiv, n):
#self.ndiv = ndiv
parts = divide_int(len(self.total_images), ndiv)
print(parts, len(self.total_images))
if n == 0:
v_start = 0
v_stop = parts[0]
else:
v_start = parts[n-1]
v_stop = parts[n]
val_images = self.total_images[v_start:v_stop]
if n == 0:
tra_images = self.total_images[v_stop:parts[-2]]
else:
tra_images = self.total_images[:v_start]
if n != ndiv-2:
tra_images.extend(self.total_images[v_stop:parts[-2]])
return tra_images, val_images
def get_test_images(self, ndiv):
self.ndiv = ndiv
self.parts = divide_int(len(self.total_images), self.ndiv)
print(self.parts, len(self.total_images))
return self.total_images[self.parts[-2]:]
def amp_jobs(fdata, job, ndata, HL, E_conv):
total_images = ase.io.read(fdata, index=':')
if job == "train":
if not ndata:
#images = ase.io.read(fdata, index=':')
images = total_images
print("Start training using all the data %d" % len(total_images))
else:
#images = ase.io.read(fdata, index=':ndata')
images = total_images[:ndata]
print("number of traing data %d/%d" %
(len(images), len(total_images)))
exe_train_images(images, HL, E_conv)
elif job == "test":
amp_pes = "amp.amp"
### get model and make title
title = fdata.split(".")[0] + "\n"
hl = '$\\times$'.join(str(x) for x in HL)
suptitle = "\n\nAMP Model(HL={}".format(hl) + ", "
suptitle += "E_rms={}):".format(E_conv) + " "
suptitle += "test/train={}/{}".format(len(total_images) - ndata, ndata)
exe_test_images(total_images[ndata:], amp_pes, title, suptitle)
elif job == "valid-test":
print("validation test")
if not ndata:
ndata = 5
parts = divide_int(len(total_images), ndata)
print(parts, "in total {}".format(len(total_images)))
for i in range(ndata - 1):
img_train, img_valid = divide_train_valid_images(
total_images[0:parts[-1]], parts, i)
exe_train_images(img_train, HL, E_conv)
amp_pes = "amp.amp"
### get model and make title
title = fdata.split(".")[0] + "\n"
hl = '$\\times$'.join(str(x) for x in HL)
suptitle = "AMP Model(HL={}".format(hl) + ", "
suptitle += "E_rms={}):".format(E_conv) + " "
suptitle += "train:validation:test=1:1:1\n"
exe_test_images(total_images[ndata:], amp_pes, title, suptitle)
# check divided image sets: plot 2d here
if False:
x_draw = []
y_draw = []
for n, atoms in enumerate(img_train):
pot = atoms.get_potential_energy()
x_draw.append(n)
y_draw.append(pot)
mplot_vector_two(x_draw,
y_draw,
Title="Extracted Training Set %d" % i,
Xtitle="serial number",
Ytitle="Epot")
elif job == 'md':
if not ndata:
atoms = ase.io.read(fdata, index='0')
else:
atoms = ase.io.read(fdata, index=ndata)
run_md(atoms)
return
