From 8345b9db48133dc0a2782ccbf8ee038f8012fdf6 Mon Sep 17 00:00:00 2001
From: saienduri <77521230+saienduri@users.noreply.github.com>
Date: Thu, 23 May 2024 14:48:23 -0700
Subject: [PATCH] Update IREE onnx import to be in sync with Torch-MLIR
 (#17476)

This commit updates `iree-import-onnx` so that it behaves the same as
torch-mlir's version
(https://github.com/llvm/torch-mlir/blob/main/python/torch_mlir/tools/import_onnx/__main__.py).
Specifically, enabling the data propagation improves shape inference and
is leading to more models passing.

Related to #17021

---------

Signed-off-by: saienduri <saimanas.enduri@amd.com>
---
 .../compiler/tools/import_onnx/__main__.py    | 73 +++++++++++++++++--
 1 file changed, 65 insertions(+), 8 deletions(-)

diff --git a/compiler/bindings/python/iree/compiler/tools/import_onnx/__main__.py b/compiler/bindings/python/iree/compiler/tools/import_onnx/__main__.py
index 9e2fe1397073..e5aff51f990c 100644
--- a/compiler/bindings/python/iree/compiler/tools/import_onnx/__main__.py
+++ b/compiler/bindings/python/iree/compiler/tools/import_onnx/__main__.py
@@ -15,8 +15,10 @@
   python -m iree.compiler.tools.import_onnx ...
 """
 import argparse
+import os
 from pathlib import Path
 import sys
+import tempfile
 
 try:
     import onnx
@@ -38,8 +40,8 @@
 )
 
 
-def main(args):
-    model_proto = load_onnx_model(args.input_file)
+def main(args: argparse.Namespace):
+    model_proto = load_onnx_model(args)
     context = Context()
     model_info = onnx_importer.ModelInfo(model_proto)
     m = model_info.create_module(context=context).operation
@@ -58,13 +60,56 @@ def main(args):
         print(m.get_asm(assume_verified=not args.no_verify))
 
 
-def load_onnx_model(file_path: Path) -> onnx.ModelProto:
-    raw_model = onnx.load(file_path)
-    inferred_model = onnx.shape_inference.infer_shapes(raw_model)
-    return inferred_model
+def load_onnx_model(args: argparse.Namespace) -> onnx.ModelProto:
+    input_dir = os.path.dirname(os.path.abspath(args.input_file))
 
-
-def parse_arguments(argv=None):
+    # Load the model, with possible external data coming from the default
+    # location, or the location specified on the command line.
+    if args.data_dir is None:
+        raw_model = onnx.load(args.input_file)
+    else:
+        raw_model = onnx.load(args.input_file, load_external_data=False)
+        onnx.load_external_data_for_model(raw_model, args.data_dir)
+
+    # Do shape inference two ways.  First, attempt in-memory to avoid redundant
+    # loading and the need for writing a temporary file somewhere.  If that
+    # fails, typically because of the 2 GB protobuf size limit, try again via
+    # files.  See
+    # https://onnx.ai/onnx/repo-docs/PythonAPIOverview.html#shape-inference-a-large-onnx-model-2gb
+    # for details about the file-based technique.
+
+    # Run the checker to test whether the file is above the threshold for
+    # in-memory shape inference.  If not, go ahead and do the shape inference.
+    try:
+        onnx.checker.check_model(raw_model)
+        inferred_model = onnx.shape_inference.infer_shapes(
+            raw_model, data_prop=args.data_prop
+        )
+        return inferred_model
+    except ValueError:
+        pass
+
+    # Model is too big for in-memory inference: do file-based shape inference
+    # to a temp file.
+    # Make a temp dir for all the temp files we'll be generating as a side
+    # effect of infering shapes. For now, the only file is a new .onnx holding
+    # the revised model with shapes.
+    with tempfile.TemporaryDirectory(dir=input_dir) as temp_dir_name:
+        temp_dir_path = Path(temp_dir_name)
+        temp_inferred_file = temp_dir_path / "temp-inferred.onnx"
+        onnx.shape_inference.infer_shapes_path(
+            args.input_file, temp_inferred_file, data_prop=args.data_prop
+        )
+
+        # Load the temp file and the external data.
+        inferred_model = onnx.load(temp_inferred_file, load_external_data=False)
+        data_dir = Path(input_dir if args.data_dir is None else args.data_dir)
+        onnx.load_external_data_for_model(inferred_model, data_dir)
+
+        return inferred_model
+
+
+def parse_arguments(argv=None) -> argparse.Namespace:
     parser = argparse.ArgumentParser(description="IREE ONNX import tool")
     parser.add_argument("input_file", help="ONNX protobuf input", type=Path)
     parser.add_argument(
@@ -75,6 +120,18 @@ def parse_arguments(argv=None):
         action="store_true",
         help="Disable verification prior to printing",
     )
+    parser.add_argument(
+        "--data-prop",
+        default=True,
+        action=argparse.BooleanOptionalAction,
+        help="Toggle data propogation for onnx shape inference",
+    )
+    parser.add_argument(
+        "--data-dir",
+        help="Path to the base directory of the data."
+        " Defaults to the directory of the input file.",
+        type=Path,
+    )
     args = parser.parse_args(argv)
     return args