Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix missing callbacks #57

Closed
wants to merge 19 commits into from
Closed

Fix missing callbacks #57

Show file tree
Hide file tree
Changes from 9 commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
440e641
Adding the APIs
tatianacv Nov 29, 2023
a8e2ddc
Adding other test
tatianacv Nov 29, 2023
ebc3b34
Adding documentation
tatianacv Nov 29, 2023
ae7078c
Getting the callbacks in the CG
tatianacv Nov 30, 2023
db9f985
Adding spotless
tatianacv Nov 30, 2023
7b6717b
Adding experimental
tatianacv Nov 30, 2023
7131070
Addressing requested changes
tatianacv Nov 30, 2023
72610a6
Merge branch 'master' into 92-missing_callbacks
khatchad Nov 30, 2023
5e29516
Add correct params, but test will fail
tatianacv Nov 30, 2023
95df2eb
Let's focus on one test for now.
khatchad Nov 30, 2023
3ec179a
Remove comment.
khatchad Nov 30, 2023
e5d5a3a
Lose a URL.
khatchad Nov 30, 2023
6b0af76
Format.
khatchad Nov 30, 2023
6bcb8ae
Change tests.
khatchad Nov 30, 2023
761b231
Merge branch 'master' into 92-missing_callbacks
khatchad Dec 1, 2023
b0c13ec
The value number here is actually 2 not 1.
khatchad Dec 1, 2023
873c7f0
Also update the comment.
khatchad Dec 1, 2023
82969d6
Took a crack at sending the argument.
khatchad Dec 1, 2023
835548c
Yet another crack.
khatchad Dec 1, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions ...wala.cast.python.ml.test/source/com/ibm/wala/cast/python/ml/test/TestTensorflowModel.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -216,6 +216,10 @@ public void testTf2()
// once https://github.com/wala/ML/issues/106 is fixed.
testTf2("tf2_test_model_call3.py", "SequentialModel.call", 1, 4, 2);
testTf2("tf2_test_model_call4.py", "SequentialModel.__call__", 1, 4, 2);
testTf2("tf2_test_callbacks.py", "get_dataset", 0, 0);
khatchad marked this conversation as resolved.
Show resolved Hide resolved
testTf2("tf2_test_callbacks2.py", "replica_fn", 1, 2, 2);
testTf2("tf2_test_callbacks3.py", "get_dataset", 0, 0);
testTf2("tf2_test_callbacks4.py", "replica_fn", 1, 2, 2);
khatchad marked this conversation as resolved.
Show resolved Hide resolved
}

private void testTf2(
Expand Down
58 changes: 58 additions & 0 deletions com.ibm.wala.cast.python.ml/data/tensorflow.xml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,9 @@
<new def="estimator" class="Lobject" />
<putfield class="LRoot" field="estimator" fieldType="LRoot" ref="x" value="estimator" />

<new def="distribute" class="Lobject" />
<putfield class="LRoot" field="distribute" fieldType="LRoot" ref="x" value="distribute" />

<new def="nn" class="Lobject" />
<putfield class="LRoot" field="nn" fieldType="LRoot" ref="x" value="nn" />
<new def="random" class="Lobject" />
Expand All @@ -58,6 +61,9 @@
<new def="Estimator" class="Ltensorflow/estimator/Estimator" />
<putfield class="LRoot" field="Estimator" fieldType="LRoot" ref="estimator" value="Estimator" />

<new def="MirroredStrategy" class="Ltensorflow/distribute/MirroredStrategy" />
<putfield class="LRoot" field="MirroredStrategy" fieldType="LRoot" ref="distribute" value="MirroredStrategy" />

<new def="inputs" class="Lobject" />
<putfield class="LRoot" field="inputs" fieldType="LRoot" ref="estimator" value="inputs" />

Expand Down Expand Up @@ -786,6 +792,58 @@
</class>
</package>

<package name="tensorflow/distribute">
<class name="MirroredStrategy" allocatable="true">
khatchad marked this conversation as resolved.
Show resolved Hide resolved
<method name="do" descriptor="()LRoot;" numArgs="3" paramNames="self function">
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why are there three arguments but only two parameter names?

<new def="x" class="Ltensorflow/distribute/distribute_datasets_from_function/distribute_datasets_from_function" />
<putfield class="LRoot" field="distribute_datasets_from_function" fieldType="LRoot" ref="self" value="x" />
<return value="arg0" />
<new def="y" class="Ltensorflow/distribute/experimental_distribute_datasets_from_function/experimental_distribute_datasets_from_function" />
<putfield class="LRoot" field="experimental_distribute_datasets_from_function" fieldType="LRoot" ref="self" value="y" />
<return value="arg1" />
<new def="z" class="Ltensorflow/distribute/run/run" />
<putfield class="LRoot" field="run" fieldType="LRoot" ref="self" value="z" />
<return value="arg2" />
khatchad marked this conversation as resolved.
Show resolved Hide resolved
</method>
</class>
</package>

<package name="tensorflow/distribute/distribute_datasets_from_function">
<class name="distribute_datasets_from_function" allocatable="true">
<method name="do" descriptor="()LRoot;" numArgs="2" paramNames="self dataset_fn">
<new def="x" class="Lobject" />
<putfield class="LRoot" field="MirroredStrategy" fieldType="LRoot" ref="self" value="x" />
<putfield class="LRoot" field="$callback" fieldType="LRoot" ref="x" value="dataset_fn" />
<call class="LRoot" name="do" descriptor="()LRoot;" type="virtual" arg0="dataset_fn" arg1="1" numArgs="2" def="xx" />
<return value="xx" />
</method>
</class>
</package>

<package name="tensorflow/distribute/experimental_distribute_datasets_from_function">
<class name="experimental_distribute_datasets_from_function" allocatable="true">
<method name="do" descriptor="()LRoot;" numArgs="2" paramNames="self dataset_fn">
<new def="x" class="Lobject" />
<putfield class="LRoot" field="MirroredStrategy" fieldType="LRoot" ref="self" value="x" />
<putfield class="LRoot" field="$callback" fieldType="LRoot" ref="x" value="dataset_fn" />
<call class="LRoot" name="do" descriptor="()LRoot;" type="virtual" arg0="dataset_fn" arg1="1" numArgs="2" def="xx" />
<return value="xx" />
</method>
</class>
</package>

<package name="tensorflow/distribute/run">
<class name="run" allocatable="true">
<method name="do" descriptor="()LRoot;" numArgs="2" paramNames="self fn">
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking at https://www.tensorflow.org/api_docs/python/tf/distribute/Strategy#run, run() takes also args=(). Looks like that needs to be passed to the callback.

<new def="x" class="Lobject" />
<putfield class="LRoot" field="MirroredStrategy" fieldType="LRoot" ref="self" value="x" />
<putfield class="LRoot" field="$callback" fieldType="LRoot" ref="x" value="fn" />
khatchad marked this conversation as resolved.
Show resolved Hide resolved
<call class="LRoot" name="do" descriptor="()LRoot;" type="virtual" arg0="fn" arg1="1" numArgs="2" def="xx" />
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why does arg1 take 1 here?

<return value="xx" />
</method>
</class>
</package>

<package name="tensorflow/app">
<class name="run" allocatable="true">
<method name="do" descriptor="()LRoot;" numArgs="3" paramNames="self main argv">
Expand Down
15 changes: 15 additions & 0 deletions com.ibm.wala.cast.python.test/data/tf2_test_callbacks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
import tensorflow as tf
khatchad marked this conversation as resolved.
Show resolved Hide resolved

# Testing API https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/MirroredStrategy#distribute_datasets_from_function
# Making sure that function `get_dataset` is in the CG
khatchad marked this conversation as resolved.
Show resolved Hide resolved

def get_dataset(input_context):
khatchad marked this conversation as resolved.
Show resolved Hide resolved
batch_size = input_context.get_per_replica_batch_size(2)
return tf.data.Dataset.range(4).batch(batch_size)

global_batch_size = 2

strategy = tf.distribute.MirroredStrategy(devices=["GPU:0", "GPU:1"])

input_context = tf.distribute.InputContext()
dist_dataset = strategy.distribute_datasets_from_function(get_dataset)
14 changes: 14 additions & 0 deletions com.ibm.wala.cast.python.test/data/tf2_test_callbacks2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,14 @@
import tensorflow as tf

# Code from https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/Strategy#example_usage_2
# Testing API https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/MirroredStrategy#run
# Making sure that function `replica_fn` is in the CG

strategy = tf.distribute.MirroredStrategy(["GPU:0", "GPU:1"])
tensor_input = tf.constant(3.0)

@tf.function
def replica_fn(input):
khatchad marked this conversation as resolved.
Show resolved Hide resolved
return input*2.0

result = strategy.run(replica_fn, args=(tensor_input,))
khatchad marked this conversation as resolved.
Show resolved Hide resolved
15 changes: 15 additions & 0 deletions com.ibm.wala.cast.python.test/data/tf2_test_callbacks3.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
import tensorflow as tf

# Testing API https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/MirroredStrategy#distribute_datasets_from_function
# Making sure that function `get_dataset` is in the CG

def get_dataset(input_context):
batch_size = input_context.get_per_replica_batch_size(2)
return tf.data.Dataset.range(4).batch(batch_size)

global_batch_size = 2

strategy = tf.distribute.MirroredStrategy(devices=["GPU:0", "GPU:1"])

input_context = tf.distribute.InputContext()
dist_dataset = strategy.experimental_distribute_datasets_from_function(get_dataset)
14 changes: 14 additions & 0 deletions com.ibm.wala.cast.python.test/data/tf2_test_callbacks4.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,14 @@
import tensorflow as tf

# Code from https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/Strategy#example_usage_2
# Testing API https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/distribute/MirroredStrategy#run
# Making sure that function `replica_fn` is in the CG

strategy = tf.distribute.MirroredStrategy(["GPU:0", "GPU:1"])
tensor_input = tf.constant(3.0)

@tf.function
def replica_fn(input):
return input*2.0

result = strategy.run(replica_fn, (tensor_input,))
Loading