Refactor ONNX Runtime patching into capability modules

.pre-commit-config.yaml

@@ -105,7 +105,10 @@ repos:
               modelopt/torch/quantization/utils/calib_utils.py|
               modelopt/torch/quantization/ggml/codebooks.py|
               modelopt/onnx/quantization/operators.py|
-              modelopt/onnx/quantization/ort_patching.py|
+              modelopt/onnx/quantization/ort_calibration.py|
+              modelopt/onnx/quantization/ort_calibration_per_node.py|
+              modelopt/onnx/quantization/ort_quantization.py|
+              modelopt/onnx/quantization/ort_session.py|
               modelopt/torch/_deploy/utils/onnx_utils.py|
               modelopt/torch/export/transformer_engine.py|
               modelopt/torch/puzzletron/anymodel/models/gpt_oss/gpt_oss_pruned_to_mxfp4.py|

CHANGELOG.rst

@@ -23,6 +23,10 @@ Changelog
 
 **Backward Breaking Changes**
 
+- The ``modelopt.onnx.quantization.ort_patching`` module has been removed with no
+  compatibility shim; update direct imports to use ``ort_session`` for model loading
+  and session setup, ``ort_calibration`` or ``ort_calibration_per_node`` for calibration,
+  ``ort_quantization`` for static Q/DQ quantization, and ``ort_patches`` for patch composition.
 - The ``modelopt.onnx.quantization.graph_utils`` module has been removed with no
   compatibility shim; update direct imports using this migration map:
 

CONTRIBUTING.md

@@ -145,6 +145,8 @@ See [`modelopt/torch/quantization/utils/calib_utils.py`](./modelopt/torch/quanti
 
 We use [pytest](https://docs.pytest.org/) for all tests. For any new features / examples, make sure to add tests and that the coverage check in your PR passes. The tests are organized into the following directories:
 
+Test filenames must mirror the production module they primarily cover: tests for `foo.py` belong in `test_foo.py`. When a production module is renamed, split, or removed, rename or split its tests in the same change so test filenames do not preserve obsolete module names. A cross-module integration test may use a feature-oriented name only when no single production module owns the behavior.
+
 - `tests/unit`: Fast cpu-based unit tests for the core ModelOpt library. They should not take more than a few seconds to run.
 - `tests/gpu`: Fast GPU-based unit tests for the core ModelOpt library. In most cases, they should not take more than a few seconds to run.
 - `tests/gpu_megatron`: Fast GPU-based unit tests for the core ModelOpt library for Megatron-Core features. In most cases, they should not take more than a few seconds to run.

modelopt/onnx/quantization/fp8.py

@@ -39,7 +39,7 @@
     find_nodes_to_exclude,
 )
 from modelopt.onnx.quantization.int8 import _find_nodes_to_quantize
-from modelopt.onnx.quantization.ort_patching import _quantize_static as quantize_static
+from modelopt.onnx.quantization.ort_quantization import _quantize_static as quantize_static
 from modelopt.onnx.quantization.ort_utils import configure_ort
 from modelopt.onnx.quantization.precision_utils import (
     _convert_to_runtime_precision,

modelopt/onnx/quantization/int8.py

@@ -35,7 +35,7 @@
     find_nodes_from_matmul_to_exclude,
     find_nodes_to_exclude,
 )
-from modelopt.onnx.quantization.ort_patching import _quantize_static as quantize_static
+from modelopt.onnx.quantization.ort_quantization import _quantize_static as quantize_static
 from modelopt.onnx.quantization.ort_utils import configure_ort
 from modelopt.onnx.quantization.partitioning import (
     find_fusible_partitions,