add influence gpu tests not using DataParallel
#1185
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
e26b4df to
84ac67c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
84ac67c to
d97eab6
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
d97eab6 to
9d8a091
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 14, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 26, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 26, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: NarineK Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 26, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 26, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 27, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
9d8a091 to
53ea863
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 27, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
53ea863 to
9bac757
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
|
@NarineK has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
99warriors
force-pushed
the
export-D47190429
branch
from
9bac757 to
e6d5aa5
Compare
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 27, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429 Pulled By: NarineK
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
1 similar comment
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 27, 2023
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Differential Revision: D47190429 Pulled By: NarineK
99warriors
force-pushed
the
export-D47190429
branch
from
e6d5aa5 to
f5fff22
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
Summary: Currently, when testing implementations of `TracInCPBase`, if the model to be tested is on gpu, we always wrap it in `DataParallel`. However, it is also worth testing when the model is on gpu, but is *not* wrapped in `DataParallel`. Whether the model is on gpu is currently specified by a `use_gpu` flag, which is boolean. In this diff, we change `use_gpu` to have type `Union[bool, str]`, which allowable values of `False` (model on cpu), `'cuda'` (model on gpu, not using `DataParallel`, and `'cuda_data_parallel'` (model on gpu, using `DataParallel`). This has backwards compatibility with classes like `ExplicitDataset`, which moves data to gpu `if use_gpu`, as strings are interpreted as being true. In further detail, the changes are as follows: - for tests (`TestTracInSelfInfluence`, `TestTracInKMostInfluential`) where `use_gpu` was called with `True`, now call them with values of `'cuda'` and `'cuda_parallel'` (in addition to `False`) - in those tests, make the layer names have the 'module' prefix only when `use_gpu='cuda_data_parallel'` - change `get_random_model_and_data`, which is where the `use_gpu` flag is used to create model and data, to reflect the new logic Reviewed By: vivekmig Differential Revision: D47190429
99warriors
force-pushed
the
export-D47190429
branch
from
f5fff22 to
1464e5c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
99warriors
force-pushed
the
export-D47190429
branch
from
1464e5c to
437ae82
Compare
|
This pull request was exported from Phabricator. Differential Revision: D47190429 |
|
This pull request has been merged in 5398892. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Currently, when testing implementations of
TracInCPBase, if the model to be tested is on gpu, we always wrap it inDataParallel. However, it is also worth testing when the model is on gpu, but is not wrapped inDataParallel. Whether the model is on gpu is currently specified by ause_gpuflag, which is boolean. In this diff, we changeuse_gputo have typeUnion[bool, str], which allowable values ofFalse(model on cpu),'cuda'(model on gpu, not usingDataParallel, and'cuda_data_parallel'(model on gpu, usingDataParallel). This has backwards compatibility with classes likeExplicitDataset, which moves data to gpuif use_gpu, as strings are interpreted as being true. In further detail, the changes are as follows:TestTracInSelfInfluence,TestTracInKMostInfluential) whereuse_gpuwas called withTrue, now call them with values of'cuda'and'cuda_parallel'(in addition toFalse)use_gpu='cuda_data_parallel'get_random_model_and_data, which is where theuse_gpuflag is used to create model and data, to reflect the new logicReviewed By: NarineK
Differential Revision: D47190429