def test_train_small_model_presets(subset_moddata, tf_session):
"""Tests the `fit_preset()` method."""
from copy import deepcopy
from modnet.model_presets import MODNET_PRESETS
from modnet.models import MODNetModel
modified_presets = deepcopy(MODNET_PRESETS)
for ind, preset in enumerate(modified_presets):
modified_presets[ind]["epochs"] = 5
data = subset_moddata
# set 'optimal' features manually
data.optimal_features = [
col for col in data.df_featurized.columns if col.startswith("ElementProperty")
]
model = MODNetModel(
[[["eform", "egap"]]],
weights={"eform": 1, "egap": 1},
num_neurons=([16], [8], [8], [4]),
n_feat=10,
)
model.fit_preset(data, presets=modified_presets, val_fraction=0.2)
def test_train_small_model_single_target_classif(subset_moddata, tf_session):
"""Tests the single target training."""
from modnet.models import MODNetModel
data = subset_moddata
# set 'optimal' features manually
data.optimal_features = [
col for col in data.df_featurized.columns
if col.startswith("ElementProperty")
]
def is_metal(egap):
if egap == 0:
return 1
else:
return 0
data.df_targets["is_metal"] = data.df_targets["egap"].apply(is_metal)
model = MODNetModel(
[[["is_metal"]]],
weights={"is_metal": 1},
num_neurons=[[16], [8], [8], [4]],
num_classes={"is_metal": 2},
n_feat=10,
)
model.fit(data, epochs=5)
def test_model_integration(subset_moddata, tf_session):
"""Tests training, saving, loading and predictions."""
from modnet.models import MODNetModel
data = subset_moddata
# set 'optimal' features manually
data.optimal_features = [
col for col in data.df_featurized.columns
if col.startswith("ElementProperty")
][:10]
model = MODNetModel(
[[["eform", "egap"]]],
weights={
"eform": 1,
"egap": 1
},
num_neurons=[[16], [8], [8], [4]],
n_feat=10,
)
model.fit(data, epochs=5)
model.save("test")
loaded_model = MODNetModel.load("test")
assert model.predict(data) == loaded_model.predict(data)
def test_train_small_model_presets(subset_moddata, tf_session):
"""Tests the `fit_preset()` method."""
from modnet.model_presets import gen_presets
from modnet.models import MODNetModel
modified_presets = gen_presets(100, 100)[:2]
for ind, preset in enumerate(modified_presets):
modified_presets[ind]["epochs"] = 5
data = subset_moddata
# set 'optimal' features manually
data.optimal_features = [
col for col in data.df_featurized.columns
if col.startswith("ElementProperty")
]
model = MODNetModel(
[[["eform", "egap"]]],
weights={
"eform": 1,
"egap": 1
},
num_neurons=[[16], [8], [8], [4]],
n_feat=10,
)
# nested=0/False -> no inner loop, so only 1 model
# nested=1/True -> inner loop, but default n_folds so 5
for num_nested, nested_option in zip([5, 1, 5], [5, 0, 1]):
results = model.fit_preset(
data,
presets=modified_presets,
nested=nested_option,
val_fraction=0.2,
n_jobs=1,
)
models = results[0]
assert len(models) == len(modified_presets)
assert len(models[0]) == num_nested
def test_train_small_model_multi_target(subset_moddata, tf_session):
"""Tests the multi-target training."""
from modnet.models import MODNetModel
data = subset_moddata
# set 'optimal' features manually
data.optimal_features = [
col for col in data.df_featurized.columns if col.startswith("ElementProperty")
]
model = MODNetModel(
[[["eform", "egap"]]],
weights={"eform": 1, "egap": 1},
num_neurons=([16], [8], [8], [4]),
n_feat=10,
)
model.fit(data, epochs=5)
model.predict(data)