add compute_intermediate_quantities to TracInCP
#1068
Closed
Conversation
This file contains 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: D40688327 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 16, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 7f8c95ccc761ba5e1a59d78d0eccea999f6e0edb
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
force-pushed
the
export-D40688327
branch
from
322c28a to
07bcbdc
Compare
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 18, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 505eddc34da93391975ba2579abd8dcc9c7560c5
99warriors
force-pushed
the
export-D40688327
branch
from
07bcbdc to
befd927
Compare
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 19, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: cfac6edbd03e62ffbf68ee075661480e0be3df98
99warriors
force-pushed
the
export-D40688327
branch
from
befd927 to
e6035b9
Compare
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 21, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 7fb76a6a8de16628993fb621accbb4337e02c987
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
force-pushed
the
export-D40688327
branch
from
e6035b9 to
639e751
Compare
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Nov 21, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: 0f57a7a73d91a49f615a117a5ac3c2fdb2fb1bf8
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Nov 21, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 3ce1d66f416b5c7e85dcb1392ac4412217cea811
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
force-pushed
the
export-D40688327
branch
from
639e751 to
62a28af
Compare
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Dec 3, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 675f9609e8a5f3612d6f7c9d046dc90d5e02ea72
99warriors
force-pushed
the
export-D40688327
branch
from
62a28af to
0df5253
Compare
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 3, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: d7a328a6227e5fae9e95b06188835524c2a5c86b
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Dec 8, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 7d67eb9e22f0d01763838223e86097ccb97cc66a
99warriors
force-pushed
the
export-D40688327
branch
from
0df5253 to
18dc5bf
Compare
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 8, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: 58e95b6da70406abb1536e83c3e63c07635deb08
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 8, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: cb8bc42468e3bd6799546522eb3e07ca52008c5b
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 8, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: 7cea23b0c3b55ca65a06b6d97e6145660b75c7b0
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 9, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: 06f1d22e57ce4001d4d4e8e29620a97dc734c8fc
99warriors
pushed a commit
to 99warriors/captum
that referenced
this pull request
Dec 9, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: 07ee5bb2ff7ac581d442a5621efb87cef6c7934d
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
99warriors
force-pushed
the
export-D40688327
branch
from
18dc5bf to
ecb9212
Compare
99warriors
added a commit
to 99warriors/captum
that referenced
this pull request
Dec 9, 2022
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Differential Revision: https://internalfb.com/D40688327 fbshipit-source-id: 02d1ecbe66e8092f1be07ca831040e72d1b7ae37
Summary: Pull Request resolved: pytorch#1073 - For all `TracInCPBase` implementations, this adds an additional `test_loss_fn` initialization argument, which is the loss function to apply to test examples when computing the influence of a training example on a test example. With this change,the influence score is a sum over terms for each checkpoint, where each term is the gradient of `loss_fn` for a given training example, multiplied with the gradient of `test_loss_fn` for a given test example. Before, `test_loss_fn` was assumed to be the same as `loss_fn`. - checks regarding the reduction type of both `loss_fn` and `test_loss_fn` are now handled by helper functions `_check_tracincp_loss_fn` and `_check_tracincp_fast_loss_fn`. - documentation is updated. one detail: for `TracInCP`, we assume that `sample_wise_grads_per_batch` is applied to both `loss_fn` and `test_loss_fn` (if provided), and this is mentioned in the documentation. - `test_tracin_regression.test_tracin_regression` is slightly modified - `DataInfluenceConstructor` now can explicitly pass in the same loss function for both `loss_fn` and `test_loss_fn` (done when `duplicate_loss_fn=True`). Doing so would have the same effect as not passing in `test_loss_fn`, so the original tests are also applied to the case when `duplicate_loss_fn=True`, as the expected behavior should be the same as before. - a new test, `test_tracin_regression.test_tracin_constant_test_loss_fn` is added. For all implementations of `TracInCPBase`, it checks that if `test_loss_fn` is a constant loss function, the influence scores are all 0's. This should be the case, because if `test_loss_fn` is constant, its gradients would all be 0's, so that training examples have 0 influence on test examples. Differential Revision: https://internalfb.com/D41202866 fbshipit-source-id: e9eaf630be6b54f7d0c86c681dcf0c9e6dddb6f1
Summary: Pull Request resolved: pytorch#1068 This diff adds the `compute_intermediate_quantities` method to `TracInCP`, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case of `TracInCP`, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints. There is also an `aggregate` option that if True, returns not the influence embeddings of each example in the given dataset, but instead their *sum*. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the *sum* of the influence embeddings over the validation dataset (i.e. with `aggregate=True`) For tests, the tests currently used for `TracInCPFastRandProj.compute_intermediate_quantities` (`test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api`, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent`) are applied to `TracInCP.compute_intermediate_quantities`. In addition, `test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregate` is added to test the `aggregate=True` option, checking that with `aggregate=True`, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset. Reviewed By: cyrjano Differential Revision: D40688327 fbshipit-source-id: b673984927bb0a17cc21fb627dd76c8f765eb43a
99warriors
force-pushed
the
export-D40688327
branch
from
ecb9212 to
c0db00f
Compare
|
This pull request was exported from Phabricator. Differential Revision: D40688327 |
|
This pull request has been merged in e7b58af. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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:
This diff adds the
compute_intermediate_quantitiesmethod toTracInCP, which returns influence embeddings such that the influence of one example on another is the dot-product of their respective influence embeddings. In the case ofTracInCP, its influence embeddings are simply the parameter-gradients for an example, concatenated over different checkpoints.There is also an
aggregateoption that if True, returns not the influence embeddings of each example in the given dataset, but instead their sum. This is useful for the validation diff workflow (which is the next diff in the stack), where we want to calculate the influence of a given training example on an entire validation dataset. This can be accomplished by taking the dot-product of the training example's influence embedding with the sum of the influence embeddings over the validation dataset (i.e. withaggregate=True)For tests, the tests currently used for
TracInCPFastRandProj.compute_intermediate_quantities(test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_api,test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_consistent) are applied toTracInCP.compute_intermediate_quantities. In addition,test_tracin_intermediate_quantities.test_tracin_intermediate_quantities_aggregateis added to test theaggregate=Trueoption, checking that withaggregate=True, the returned influence embedding is indeed the sum of the influence embeddings for the given dataset.Reviewed By: cyrjano
Differential Revision: D40688327