Skip to content

Commit

Permalink
Atomic Convs Cleanup
Browse files Browse the repository at this point in the history
  • Loading branch information
@lilleswing
lilleswing committed Apr 27, 2017
1 parent 6bbe6a4 commit 74e8329
Show file tree
Hide file tree
Showing 35 changed files with 1,470 additions and 249 deletions.
57 changes: 42 additions & 15 deletions contrib/atomicconv/acnn/core/opt_random.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,31 +26,52 @@
test_dir = os.path.join(data_dir, "random_test")
model_dir = os.path.join(base_dir, "random_model")

frag1_num_atoms = 140
frag2_num_atoms = 821
complex_num_atoms = 908
frag1_num_atoms = 70
frag2_num_atoms = 634
complex_num_atoms = 701
max_num_neighbors = 12
neighbor_cutoff = 12.0

train_dataset = dc.data.DiskDataset(train_dir)
test_dataset = dc.data.DiskDataset(test_dir)
pdbbind_tasks = ["-logKd/Ki"]
transformers = [
dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)
]
for transformer in transformers:
train_dataset = transformer.transform(train_dataset)
test_dataset = transformer.transform(test_dataset)
transformers = []
#transformers = [dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)]
#for transformer in transformers:
# train_dataset = transformer.transform(train_dataset)
# test_dataset = transformer.transform(test_dataset)

y_train = train_dataset.y
y_train *= -1 * 2.479 / 4.184
train_dataset = dc.data.DiskDataset.from_numpy(
train_dataset.X,
y_train,
train_dataset.w,
train_dataset.ids,
tasks=pdbbind_tasks)

y_test = test_dataset.y
y_test *= -1 * 2.479 / 4.184
test_dataset = dc.data.DiskDataset.from_numpy(
test_dataset.X,
y_test,
test_dataset.w,
test_dataset.ids,
tasks=pdbbind_tasks)

at = [1., 6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[12.0], [0.0, 4.0, 8.0], [4.0]]
at = [6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[
1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5,
9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0
], [0.0, 4.0, 8.0], [0.4]]
#radial = [[12.0], [0.0, 4.0, 8.0], [0.4]]
rp = create_symmetry_parameters(radial)
layer_sizes = [32, 32, 16]
weight_init_stddevs = [
1 / np.sqrt(layer_sizes[0]), 1 / np.sqrt(layer_sizes[1]),
1 / np.sqrt(layer_sizes[2])
]
dropouts = [0., 0., 0.]
dropouts = [0.3, 0.3, 0.05]
penalty_type = "l2"
penalty = 0.
model = TensorflowFragmentRegressor(
Expand All @@ -76,17 +97,23 @@
boxsize=None,
verbose=True,
seed=seed)
model.fit(train_dataset, nb_epoch=100)
model.fit(train_dataset, nb_epoch=10)
metric = [
dc.metrics.Metric(dc.metrics.mean_absolute_error, mode="regression"),
dc.metrics.Metric(dc.metrics.pearson_r2_score, mode="regression")
]
train_evaluator = dc.utils.evaluate.Evaluator(model, train_dataset,
transformers)
train_scores = train_evaluator.compute_model_performance(metric)
train_scores = train_evaluator.compute_model_performance(
metric,
csv_out="train_predict_ac_random.csv",
stats_out="train_stats_ac_random.csv")
print("Train scores")
print(train_scores)
test_evaluator = dc.utils.evaluate.Evaluator(model, test_dataset, transformers)
test_scores = test_evaluator.compute_model_performance(metric)
test_scores = test_evaluator.compute_model_performance(
metric,
csv_out="test_predict_ac_random.csv",
stats_out="test_stats_ac_random.csv")
print("Test scores")
print(test_scores)
73 changes: 73 additions & 0 deletions contrib/atomicconv/acnn/core/opt_random_tensorgraph.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,73 @@
from __future__ import print_function
from __future__ import division
from __future__ import unicode_literals

import os
import sys
import deepchem as dc
import numpy as np
import tensorflow as tf
from deepchem.models.tensorgraph.models.atomic_conv import atomic_conv_model

sys.path.append("../../models")
from deepchem.models.tensorgraph.layers import Layer, Feature, Label, L2LossLayer, AtomicConvolution, Transpose, Dense
from deepchem.models import TensorGraph

import numpy as np
import tensorflow as tf
import itertools
import time

seed = 123
np.random.seed(seed)
tf.set_random_seed(seed)

base_dir = os.getcwd()
data_dir = os.path.join(base_dir, "datasets")
train_dir = os.path.join(data_dir, "random_train")
test_dir = os.path.join(data_dir, "random_test")
model_dir = os.path.join(base_dir, "random_model")

train_dataset = dc.data.DiskDataset(train_dir)
test_dataset = dc.data.DiskDataset(test_dir)
pdbbind_tasks = ["-logKd/Ki"]
transformers = []

y_train = train_dataset.y
y_train *= -1 * 2.479 / 4.184
train_dataset = dc.data.DiskDataset.from_numpy(
train_dataset.X,
y_train,
train_dataset.w,
train_dataset.ids,
tasks=pdbbind_tasks)

y_test = test_dataset.y
y_test *= -1 * 2.479 / 4.184
test_dataset = dc.data.DiskDataset.from_numpy(
test_dataset.X,
y_test,
test_dataset.w,
test_dataset.ids,
tasks=pdbbind_tasks)

batch_size = 24
tg, feed_dict_generator, label = atomic_conv_model()

print("Fitting")
metric = [
dc.metrics.Metric(dc.metrics.mean_absolute_error, mode="regression"),
dc.metrics.Metric(dc.metrics.pearson_r2_score, mode="regression")
]
tg.fit_generator(feed_dict_generator(train_dataset, batch_size, epochs=10))

train_evaluator = dc.utils.evaluate.GeneratorEvaluator(
tg, feed_dict_generator(train_dataset, batch_size), transformers, [label])
train_scores = train_evaluator.compute_model_performance(metric)
print("Train scores")
print(train_scores)
test_evaluator = dc.utils.evaluate.GeneratorEvaluator(
tg, feed_dict_generator(test_dataset, batch_size), transformers, [label])
test_scores = test_evaluator.compute_model_performance(metric)
print("Test scores")
print(test_scores)
57 changes: 42 additions & 15 deletions contrib/atomicconv/acnn/core/opt_scaffold.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,31 +26,52 @@
test_dir = os.path.join(data_dir, "scaffold_test")
model_dir = os.path.join(base_dir, "scaffold_model")

frag1_num_atoms = 140
frag2_num_atoms = 821
complex_num_atoms = 908
frag1_num_atoms = 70
frag2_num_atoms = 634
complex_num_atoms = 701
max_num_neighbors = 12
neighbor_cutoff = 12.0

train_dataset = dc.data.DiskDataset(train_dir)
test_dataset = dc.data.DiskDataset(test_dir)
pdbbind_tasks = ["-logKd/Ki"]
transformers = [
dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)
]
for transformer in transformers:
train_dataset = transformer.transform(train_dataset)
test_dataset = transformer.transform(test_dataset)
transformers = []
#transformers = [dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)]
#for transformer in transformers:
# train_dataset = transformer.transform(train_dataset)
# test_dataset = transformer.transform(test_dataset)

y_train = train_dataset.y
y_train *= -1 * 2.479 / 4.184
train_dataset = dc.data.DiskDataset.from_numpy(
train_dataset.X,
y_train,
train_dataset.w,
train_dataset.ids,
tasks=pdbbind_tasks)

y_test = test_dataset.y
y_test *= -1 * 2.479 / 4.184
test_dataset = dc.data.DiskDataset.from_numpy(
test_dataset.X,
y_test,
test_dataset.w,
test_dataset.ids,
tasks=pdbbind_tasks)

at = [1., 6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[12.0], [0.0, 4.0, 8.0], [4.0]]
at = [6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[
1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5,
9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0
], [0.0, 4.0, 8.0], [0.4]]
#radial = [[12.0], [0.0, 4.0, 8.0], [0.4]]
rp = create_symmetry_parameters(radial)
layer_sizes = [32, 32, 16]
weight_init_stddevs = [
1 / np.sqrt(layer_sizes[0]), 1 / np.sqrt(layer_sizes[1]),
1 / np.sqrt(layer_sizes[2])
]
dropouts = [0., 0., 0.]
dropouts = [0.3, 0.3, 0.05]
penalty_type = "l2"
penalty = 0.
model = TensorflowFragmentRegressor(
Expand All @@ -71,7 +92,7 @@
learning_rate=0.002,
momentum=0.8,
optimizer="adam",
batch_size=24,
batch_size=64,
conv_layers=1,
boxsize=None,
verbose=True,
Expand All @@ -83,10 +104,16 @@
]
train_evaluator = dc.utils.evaluate.Evaluator(model, train_dataset,
transformers)
train_scores = train_evaluator.compute_model_performance(metric)
train_scores = train_evaluator.compute_model_performance(
metric,
csv_out="train_predict_ac_scaffold.csv",
stats_out="train_stats_ac_scaffold.csv")
print("Train scores")
print(train_scores)
test_evaluator = dc.utils.evaluate.Evaluator(model, test_dataset, transformers)
test_scores = test_evaluator.compute_model_performance(metric)
test_scores = test_evaluator.compute_model_performance(
metric,
csv_out="test_predict_ac_scaffold.csv",
stats_out="test_stats_ac_scaffold.csv")
print("Test scores")
print(test_scores)
57 changes: 42 additions & 15 deletions contrib/atomicconv/acnn/core/opt_stratified.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,31 +26,52 @@
test_dir = os.path.join(data_dir, "stratified_test")
model_dir = os.path.join(base_dir, "stratified_model")

frag1_num_atoms = 140
frag2_num_atoms = 821
complex_num_atoms = 908
frag1_num_atoms = 70
frag2_num_atoms = 634
complex_num_atoms = 701
max_num_neighbors = 12
neighbor_cutoff = 12.0

train_dataset = dc.data.DiskDataset(train_dir)
test_dataset = dc.data.DiskDataset(test_dir)
pdbbind_tasks = ["-logKd/Ki"]
transformers = [
dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)
]
for transformer in transformers:
train_dataset = transformer.transform(train_dataset)
test_dataset = transformer.transform(test_dataset)
transformers = []
#transformers = [dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)]
#for transformer in transformers:
# train_dataset = transformer.transform(train_dataset)
# test_dataset = transformer.transform(test_dataset)

y_train = train_dataset.y
y_train *= -1 * 2.479 / 4.184
train_dataset = dc.data.DiskDataset.from_numpy(
train_dataset.X,
y_train,
train_dataset.w,
train_dataset.ids,
tasks=pdbbind_tasks)

y_test = test_dataset.y
y_test *= -1 * 2.479 / 4.184
test_dataset = dc.data.DiskDataset.from_numpy(
test_dataset.X,
y_test,
test_dataset.w,
test_dataset.ids,
tasks=pdbbind_tasks)

at = [1., 6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[12.0], [0.0, 4.0, 8.0], [4.0]]
at = [6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53.]
radial = [[
1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5,
9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0
], [0.0, 4.0, 8.0], [0.4]]
#radial = [[12.0], [0.0, 4.0, 8.0], [0.4]]
rp = create_symmetry_parameters(radial)
layer_sizes = [32, 32, 16]
weight_init_stddevs = [
1 / np.sqrt(layer_sizes[0]), 1 / np.sqrt(layer_sizes[1]),
1 / np.sqrt(layer_sizes[2])
]
dropouts = [0., 0., 0.]
dropouts = [0.3, 0.3, 0.05]
penalty_type = "l2"
penalty = 0.
model = TensorflowFragmentRegressor(
Expand All @@ -71,7 +92,7 @@
learning_rate=0.002,
momentum=0.8,
optimizer="adam",
batch_size=24,
batch_size=64,
conv_layers=1,
boxsize=None,
verbose=True,
Expand All @@ -83,10 +104,16 @@
]
train_evaluator = dc.utils.evaluate.Evaluator(model, train_dataset,
transformers)
train_scores = train_evaluator.compute_model_performance(metric)
train_scores = train_evaluator.compute_model_performance(
metric,
csv_out="train_predict_ac_stratified.csv",
stats_out="train_stats_ac_stratified.csv")
print("Train scores")
print(train_scores)
test_evaluator = dc.utils.evaluate.Evaluator(model, test_dataset, transformers)
test_scores = test_evaluator.compute_model_performance(metric)
test_scores = test_evaluator.compute_model_performance(
metric,
csv_out="test_predict_ac_stratified.csv",
stats_out="test_stats_ac_stratified.csv")
print("Test scores")
print(test_scores)
Loading

0 comments on commit 74e8329

Please sign in to comment.