Create the IntermediateOutputCapturer Class to Store the IntermediateOutput of the AOT Graph
#11202
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/11202
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit de271fe with merge base 6fafe7c ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D75492919 |
|
@pytorchbot label "release notes: none" |
pytorch-bot
bot
added
the
release notes: none
Juntian777
force-pushed
the
export-D75492919
branch
from
54b1e28 to
5a67d5b
Compare
Juntian777
added a commit
to Juntian777/executorch
that referenced
this pull request
May 29, 2025
…teOutput of the AOT Graph (pytorch#11202) Summary: This Diff introduces a new Python class, IntermediateOutputCapturer, which inherits from torch.fx.interpreter.Interpreter. The primary purpose of this class is to capture the output tensor(s) produced by each operator (node) for an EdgeProgramManager's GraphModule. We will use these stored outputs to compare with later runtime operator outputs to detect numerical discrepancies. The IntermediateOutputCapturer class overrides the run_node method to store the computed results in an instance dictionary. It checks for the presence of a debug_handle in the node's metadata and the type of the node and uses it as a key to store the result. Tensors are detached and cloned to prevent side effects, while non-tensor results are stored directly. A public method, run_and_capture, is implemented to call the base Interpreter's run method and return the dictionary containing the captured debug_handle -> output mappings. Additionally, an __init__ method is provided to accept an fx.GraphModule as input and a print_captured_outputs method is included for debugging purposes. Differential Revision: D75492919
|
This pull request was exported from Phabricator. Differential Revision: D75492919 |
Juntian777
added a commit
to Juntian777/executorch
that referenced
this pull request
May 29, 2025
…teOutput of the AOT Graph (pytorch#11202) Summary: Pull Request resolved: pytorch#11202 This Diff introduces a new Python class, IntermediateOutputCapturer, which inherits from torch.fx.interpreter.Interpreter. The primary purpose of this class is to capture the output tensor(s) produced by each operator (node) for an EdgeProgramManager's GraphModule. We will use these stored outputs to compare with later runtime operator outputs to detect numerical discrepancies. The IntermediateOutputCapturer class overrides the run_node method to store the computed results in an instance dictionary. It checks for the presence of a debug_handle in the node's metadata and the type of the node and uses it as a key to store the result. Tensors are detached and cloned to prevent side effects, while non-tensor results are stored directly. A public method, run_and_capture, is implemented to call the base Interpreter's run method and return the dictionary containing the captured debug_handle -> output mappings. Additionally, an __init__ method is provided to accept an fx.GraphModule as input and a print_captured_outputs method is included for debugging purposes. Differential Revision: D75492919
Juntian777
force-pushed
the
export-D75492919
branch
from
5a67d5b to
ef06235
Compare
Juntian777
added a commit
to Juntian777/executorch
that referenced
this pull request
May 29, 2025
…teOutput of the AOT Graph (pytorch#11202) Summary: This Diff introduces a new Python class, IntermediateOutputCapturer, which inherits from torch.fx.interpreter.Interpreter. The primary purpose of this class is to capture the output tensor(s) produced by each operator (node) for an EdgeProgramManager's GraphModule. We will use these stored outputs to compare with later runtime operator outputs to detect numerical discrepancies. The IntermediateOutputCapturer class overrides the run_node method to store the computed results in an instance dictionary. It checks for the presence of a debug_handle in the node's metadata and the type of the node and uses it as a key to store the result. Tensors are detached and cloned to prevent side effects, while non-tensor results are stored directly. A public method, run_and_capture, is implemented to call the base Interpreter's run method and return the dictionary containing the captured debug_handle -> output mappings. Additionally, an __init__ method is provided to accept an fx.GraphModule as input and a print_captured_outputs method is included for debugging purposes. Differential Revision: D75492919
Juntian777
force-pushed
the
export-D75492919
branch
from
ef06235 to
356d472
Compare
…teOutput of the AOT Graph (pytorch#11202) Summary: Pull Request resolved: pytorch#11202 This Diff introduces a new Python class, IntermediateOutputCapturer, which inherits from torch.fx.interpreter.Interpreter. The primary purpose of this class is to capture the output tensor(s) produced by each operator (node) for an EdgeProgramManager's GraphModule. We will use these stored outputs to compare with later runtime operator outputs to detect numerical discrepancies. The IntermediateOutputCapturer class overrides the run_node method to store the computed results in an instance dictionary. It checks for the presence of a debug_handle in the node's metadata and the type of the node and uses it as a key to store the result. Tensors are detached and cloned to prevent side effects, while non-tensor results are stored directly. A public method, run_and_capture, is implemented to call the base Interpreter's run method and return the dictionary containing the captured debug_handle -> output mappings. Additionally, an __init__ method is provided to accept an fx.GraphModule as input and a print_captured_outputs method is included for debugging purposes. Differential Revision: D75492919
|
This pull request was exported from Phabricator. Differential Revision: D75492919 |
Juntian777
force-pushed
the
export-D75492919
branch
from
356d472 to
de271fe
Compare
Gasoonjia
approved these changes
May 29, 2025
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 introduces a new Python class, IntermediateOutputCapturer, which inherits from torch.fx.interpreter.Interpreter. The primary purpose of this class is to capture the output tensor(s) produced by each operator (node) for an EdgeProgramManager's GraphModule. We will use these stored outputs to compare with later runtime operator outputs to detect numerical discrepancies.
The IntermediateOutputCapturer class overrides the run_node method to store the computed results in an instance dictionary. It checks for the presence of a debug_handle in the node's metadata and the type of the node and uses it as a key to store the result. Tensors are detached and cloned to prevent side effects, while non-tensor results are stored directly.
A public method, run_and_capture, is implemented to call the base Interpreter's run method and return the dictionary containing the captured debug_handle -> output mappings.
Additionally, an init method is provided to accept an fx.GraphModule as input and a print_captured_outputs method is included for debugging purposes.
Differential Revision: D75492919