Commit 1b97292e authored by Bharath Ramsundar's avatar Bharath Ramsundar
Browse files

Low data updates

parent 4d579034

examples/low_data/muv_attn_one_fold.py

+21 −20
Original line number Diff line number Diff line
@@ -5,10 +5,11 @@ from __future__ import print_function 
from __future__ import division

from __future__ import unicode_literals



import tempfile

import numpy as np

import deepchem as dc

np.random.seed(123)

import tensorflow as tf

tf.set_random_seed(123)

import deepchem as dc

from datasets import load_muv_convmol



# Number of folds for split
@@ -16,8 +17,8 @@ K = 4 
# Depth of attention module

max_depth = 3

# number positive/negative ligands

n_pos = 10

n_neg = 10

n_pos = 1

n_neg = 1

# Set batch sizes for network

test_batch_size = 128

support_batch_size = n_pos + n_neg
@@ -32,8 +33,7 @@ n_feat = 75 
muv_tasks, dataset, transformers = load_muv_convmol()



# Define metric

metric = dc.metrics.Metric(

    dc.metrics.roc_auc_score, mode="classification")

metric = dc.metrics.Metric(dc.metrics.roc_auc_score, mode="classification")



task_splitter = dc.splits.TaskSplitter()

fold_datasets = task_splitter.k_fold_split(dataset, K)
@@ -46,25 +46,26 @@ test_dataset = fold_datasets[-1] 
support_model = dc.nn.SequentialSupportGraph(n_feat)



# Add layers

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, n_feat, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(128, activation='relu'))

support_model.add(dc.nn.GraphConv(128, 64, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, 128, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.Dense(128, activation='tanh'))

support_model.add(dc.nn.Dense(128, 64, activation='tanh'))



support_model.add_test(dc.nn.GraphGather(test_batch_size, activation='tanh'))

support_model.add_support(dc.nn.GraphGather(support_batch_size, activation='tanh'))



# Apply an attention lstm layer

support_model.join(dc.nn.AttnLSTMEmbedding(

    test_batch_size, support_batch_size, max_depth))

    test_batch_size, support_batch_size, 128, max_depth))



with tf.Session() as sess:

model = dc.models.SupportGraphClassifier(

    sess, support_model, test_batch_size=test_batch_size,

    support_batch_size=support_batch_size, learning_rate=learning_rate)

  support_model,

  test_batch_size=test_batch_size,

  support_batch_size=support_batch_size,

  learning_rate=learning_rate)



model.fit(train_dataset, nb_epochs=nb_epochs, 

          n_episodes_per_epoch=n_train_trials,

examples/low_data/muv_graph_conv_one_fold.py

+32 −33
Original line number Diff line number Diff line
@@ -40,32 +40,31 @@ task_scores = {task: [] for task in range(len(test_dataset.get_task_names()))} 


for trial_num, (task, support) in enumerate(support_generator):

  print("Starting trial %d" % trial_num)

  # Train model on support

  g = tf.Graph()

  sess = tf.Session(graph=g)



  # Do setup required for tf/keras models

  # Number of features on conv-mols

  n_feat = 75

  # Batch size of models

  batch_size = 50



  with g.as_default():

  graph_model = dc.nn.SequentialGraph(n_feat)

    graph_model.add(dc.nn.GraphConv(64, activation='relu'))

  graph_model.add(dc.nn.GraphConv(64, n_feat, activation='relu'))

  graph_model.add(dc.nn.GraphPool())

    graph_model.add(dc.nn.GraphConv(128, activation='relu'))

  graph_model.add(dc.nn.GraphConv(128, 64, activation='relu'))

  graph_model.add(dc.nn.GraphPool())

    graph_model.add(dc.nn.GraphConv(64, activation='relu'))

  graph_model.add(dc.nn.GraphConv(64, 128, activation='relu'))

  graph_model.add(dc.nn.GraphPool())

    graph_model.add(dc.nn.Dense(128, activation='tanh'))

  graph_model.add(dc.nn.Dense(128, 64, activation='tanh'))

  graph_model.add(dc.nn.GraphGather(batch_size, activation="tanh"))



    with tf.Session() as sess:

  model = dc.models.MultitaskGraphClassifier(

        sess, graph_model, 1, batch_size=batch_size,

        learning_rate=1e-3, learning_rate_decay_time=1000,

        optimizer_type="adam", beta1=.9, beta2=.999)

    graph_model,

    1,

    n_feat,

    batch_size=batch_size,

    learning_rate=1e-3,

    learning_rate_decay_time=1000,

    optimizer_type="adam",

    beta1=.9,

    beta2=.999)



  # Fit trained model

  model.fit(support, nb_epoch=10)

examples/low_data/muv_res_one_fold.py

+20 −18
Original line number Diff line number Diff line
@@ -5,10 +5,11 @@ from __future__ import print_function 
from __future__ import division

from __future__ import unicode_literals



import tempfile

import numpy as np

import deepchem as dc

np.random.seed(123)

import tensorflow as tf

tf.set_random_seed(123)

import deepchem as dc

from datasets import load_muv_convmol



# Number of folds for split
@@ -16,8 +17,8 @@ K = 4 
# Depth of attention module

max_depth = 3

# num positive/negative ligands

n_pos = 10

n_neg = 10

n_pos = 1

n_neg = 5

# Set batch sizes for network

test_batch_size = 128

support_batch_size = n_pos + n_neg
@@ -45,25 +46,26 @@ test_dataset = fold_datasets[-1] 
support_model = dc.nn.SequentialSupportGraph(n_feat)



# Add layers

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, n_feat, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(128, activation='relu'))

support_model.add(dc.nn.GraphConv(128, 64, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, 128, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.Dense(128, activation='tanh'))

support_model.add(dc.nn.Dense(128, 64, activation='tanh'))



support_model.add_test(dc.nn.GraphGather(test_batch_size, activation='tanh'))

support_model.add_support(dc.nn.GraphGather(support_batch_size, activation='tanh'))



# Apply a residual lstm layer

support_model.join(dc.nn.ResiLSTMEmbedding(

    test_batch_size, support_batch_size, max_depth))

    test_batch_size, support_batch_size, 128, max_depth))



with tf.Session() as sess:

model = dc.models.SupportGraphClassifier(

    sess, support_model, test_batch_size=test_batch_size,

    support_batch_size=support_batch_size, learning_rate=learning_rate)

  support_model,

  test_batch_size=test_batch_size,

  support_batch_size=support_batch_size,

  learning_rate=learning_rate)



model.fit(train_dataset, nb_epochs=nb_epochs,

          n_episodes_per_epoch=n_train_trials,

examples/low_data/muv_siamese_one_fold.py

+14 −13
Original line number Diff line number Diff line
@@ -43,21 +43,22 @@ test_dataset = fold_datasets[-1] 
support_model = dc.nn.SequentialSupportGraph(n_feat)



# Add layers

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, n_feat, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(128, activation='relu'))

support_model.add(dc.nn.GraphConv(128, 64, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, 128, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.Dense(128, activation='tanh'))

support_model.add(dc.nn.Dense(128, 64, activation='tanh'))



support_model.add_test(dc.nn.GraphGather(test_batch_size, activation='tanh'))

support_model.add_support(dc.nn.GraphGather(support_batch_size, activation='tanh'))



with tf.Session() as sess:

model = dc.models.SupportGraphClassifier(

    sess, support_model, test_batch_size=test_batch_size,

    support_batch_size=support_batch_size, learning_rate=learning_rate)

  support_model,

  test_batch_size=test_batch_size,

  support_batch_size=support_batch_size,

  learning_rate=learning_rate)



model.fit(train_dataset, nb_epochs=nb_epochs,

          n_episodes_per_epoch=n_train_trials,

examples/low_data/sider_attn_one_fold.py

+20 −18
Original line number Diff line number Diff line
@@ -5,10 +5,11 @@ from __future__ import print_function 
from __future__ import division

from __future__ import unicode_literals



import tempfile

import numpy as np

import deepchem as dc

np.random.seed(123)

import tensorflow as tf

tf.set_random_seed(123)

import deepchem as dc

from datasets import load_sider_convmol



# Number of folds for split
@@ -16,8 +17,8 @@ K = 4 
# Depth of attention module

max_depth = 3

# number positive/negative ligands

n_pos = 10

n_neg = 10

n_pos = 1

n_neg = 1

# Set batch sizes for network

test_batch_size = 128

support_batch_size = n_pos + n_neg
@@ -45,25 +46,26 @@ test_dataset = fold_datasets[-1] 
support_model = dc.nn.SequentialSupportGraph(n_feat)



# Add layers

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, n_feat, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(128, activation='relu'))

support_model.add(dc.nn.GraphConv(128, 64, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.GraphConv(64, activation='relu'))

support_model.add(dc.nn.GraphConv(64, 128, activation='relu'))

support_model.add(dc.nn.GraphPool())

support_model.add(dc.nn.Dense(128, activation='tanh'))

support_model.add(dc.nn.Dense(128, 64, activation='tanh'))



support_model.add_test(dc.nn.GraphGather(test_batch_size, activation='tanh'))

support_model.add_support(dc.nn.GraphGather(support_batch_size, activation='tanh'))



# Apply an attention lstm layer

support_model.join(dc.nn.AttnLSTMEmbedding(

    test_batch_size, support_batch_size, max_depth))

    test_batch_size, support_batch_size, 128, max_depth))



with tf.Session() as sess:

model = dc.models.SupportGraphClassifier(

    sess, support_model, test_batch_size=test_batch_size,

    support_batch_size=support_batch_size, learning_rate=learning_rate)

  support_model,

  test_batch_size=test_batch_size,

  support_batch_size=support_batch_size,

  learning_rate=learning_rate)



model.fit(train_dataset, nb_epochs=nb_epochs, 

          n_episodes_per_epoch=n_train_trials,

CRA Git | Maintained and supported by SUSTech CRA and CCSE