Fix whitespace issues in documentation files

Co-authored-by: gramalingam <[email protected]>

Author: Copilot

docs/tutorial/rewriter/allow_other_inputs.md

@@ -2,7 +2,7 @@
 
 This section demonstrates the use of the `_allow_other_inputs` option in pattern-based rewriting.
 The `_allow_other_inputs` option allows the pattern to match nodes that have additional inputs
-beyond those specified in the pattern. If it is set to `False` (the default), then the node must 
+beyond those specified in the pattern. If it is set to `False` (the default), then the node must
 have exactly the specified inputs for a successful match. If set to `True`, the pattern will
 match nodes that have the specified inputs plus any number of additional inputs.
 
@@ -22,6 +22,6 @@ This is particularly useful when matching operations like `Conv` that can have o
 ```
 
 In this example, the pattern matches `Conv` operations with any number of inputs. A `Conv` operation
-might have 2 inputs (input and weight) or 3 inputs (input, weight, and bias). By setting 
+might have 2 inputs (input and weight) or 3 inputs (input, weight, and bias). By setting
 `_allow_other_inputs=True`, our pattern will match both cases even though we only specify 2 inputs
-in the pattern definition.
+in the pattern definition.

docs/tutorial/rewriter/domain_option.md

@@ -15,7 +15,7 @@ ONNX operators can belong to different domains:
 :pyobject: custom_relu_pattern
 ```
 
-In this pattern, `_domain="custom.domain"` ensures that only `Relu` operations from the 
+In this pattern, `_domain="custom.domain"` ensures that only `Relu` operations from the
 "custom.domain" domain will be matched, not standard ONNX `Relu` operations.
 
 ## Creating replacement operations in a specific domain
@@ -35,4 +35,4 @@ provide optimized implementations of standard operations.
 
 This example shows how domain-specific pattern matching can be used to migrate operations
 between different operator domains, such as replacing custom domain operations with
-standard ONNX operations or vice versa.
+standard ONNX operations or vice versa.

docs/tutorial/rewriter/examples/allow_other_inputs.py

@@ -68,4 +68,4 @@ def apply_rewrite(model):
 
 
 _model_with_rewrite = apply_rewrite(_model)
-onnx.checker.check_model(_model_with_rewrite)
+onnx.checker.check_model(_model_with_rewrite)

docs/tutorial/rewriter/examples/domain_option.py

@@ -83,4 +83,4 @@ def apply_rewrite(model):
 # The rewrite rule will now match the Relu operation in the custom domain
 # and replace it with a standard ONNX Relu operation
 _model_with_rewrite = apply_rewrite(_model)
-onnx.checker.check_model(_model_with_rewrite)
+onnx.checker.check_model(_model_with_rewrite)

docs/tutorial/rewriter/examples/outputs_option.py

@@ -73,4 +73,4 @@ def apply_rewrite(model):
 
 
 _model_with_rewrite = apply_rewrite(_model)
-onnx.checker.check_model(_model_with_rewrite)
+onnx.checker.check_model(_model_with_rewrite)

docs/tutorial/rewriter/outputs_option.md

@@ -26,7 +26,7 @@ specification ensures the pattern only matches operations with this specific out
 In the replacement, `_outputs=["first_half", "second_half"]` creates two outputs with
 descriptive names. This can make the replacement pattern more readable and maintainable.
 
-**Important**: The number of outputs in the replacement pattern must match the number of 
+**Important**: The number of outputs in the replacement pattern must match the number of
 outputs in the target pattern. Since the pattern specifies `_outputs=2`, the replacement
 must also produce exactly 2 outputs.
 
@@ -40,4 +40,4 @@ The `_outputs` option is particularly important when:
 - Working with operations that have variable numbers of outputs (like `Split`)
 - Creating custom operations that need specific output configurations
 - Ensuring pattern matching precision by specifying exact output counts
-- Improving code readability by naming outputs in replacement patterns
+- Improving code readability by naming outputs in replacement patterns