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

Build gradient graph starting at the loss alone #17240

Merged
merged 4 commits into from
Aug 24, 2023
Merged

Build gradient graph starting at the loss alone #17240

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
0b9d730
Build gradient graph starting at the loss alone
baijumeswani Aug 21, 2023
d0be3ab
Address pull request review comments
baijumeswani Aug 22, 2023
ca4bf69
Merge branch 'main' of https://github.com/microsoft/onnxruntime into …
baijumeswani Aug 22, 2023
8eac9f7
Avoid the new line
baijumeswani Aug 22, 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
5 changes: 2 additions & 3 deletions orttraining/orttraining/python/training/onnxblock/_training_graph_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,13 +70,12 @@ def _move_initializers_to_inputs(model: onnx.ModelProto, initializer_names: Opti
def _gradient_model_for(
model: onnx.ModelProto,
requires_grad: Set[str],
output_names: List[str],
loss_name: str,
options: Optional[SessionOptions] = None,
) -> onnx.ModelProto:
"""Builds the gradient graph on top of the given input forward only graph."""

builder = GradientGraphBuilder(model.SerializeToString(), set(output_names), requires_grad, loss_name, options)
builder = GradientGraphBuilder(model.SerializeToString(), {loss_name}, requires_grad, loss_name, options)
builder.build()
return onnx.load_from_string(builder.get_model())

Expand Down Expand Up @@ -123,7 +122,7 @@ def build_gradient_graph(
optimized_model = onnx.load_from_string(get_optimized_model(model.SerializeToString(), requires_grad, options))

# Assumption is that the first graph output is the loss output
gradient_model = _gradient_model_for(optimized_model, requires_grad, output_names, output_names[0], options)
gradient_model = _gradient_model_for(optimized_model, requires_grad, output_names[0], options)

_reorder_outputs(gradient_model, output_names, requires_grad)

Expand Down