Merge branch 'main' into justinchu/ser-initializer

Author: justinchuby

.github/workflows/lint.yaml

@@ -6,7 +6,7 @@ on:
       - main
       - 'gh/**/base' # ghstack base branches
   pull_request:
-    types: [opened, synchronize, reopened, ready_for_review]
+  merge_group:
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }}

.github/workflows/main.yaml

@@ -13,6 +13,7 @@ on:
 
   # Allows you to run this workflow manually from the Actions tab
   workflow_dispatch:
+  merge_group:
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }}

docs/tutorial/rewriter/examples/erfgelu.py

@@ -87,7 +87,7 @@ def erf_gelu_pattern_2(op, x):
 
 
 def gelu(op, x: ir.Value):
-    return op.Gelu(x, domain="com.microsoft")
+    return op.Gelu(x, _domain="com.microsoft")
 
 
 ####################################

onnxscript/backend/onnx_export_test.py

@@ -100,6 +100,10 @@ def skip(pattern: str | Pattern, reason: str, *, condition: bool = True):
             "cannot import module, import_module does not work",
         ),
         skip("^test_bitwise_not_3d", "cannot import module, import_module does not work"),
+        skip(
+            "^test_resize_upsample_scales_linear_half_pixel_symmetric",
+            "cannot import module, import_module does not work",
+        ),
     )
 
 

onnxscript/function_libs/torch_lib/graph_building/_graph_building_torch.py

@@ -459,8 +459,18 @@ def _add_attribute_to_torchscript_node(
             return node.fs_(key, list(value))  # type: ignore[arg-type]
         if isinstance(value[0], int):
             return node.is_(key, list(value))  # type: ignore[attr-defined]
-        raise TypeError(f"Unsupported sequence type '{type(value)}' for attribute '{key}'")
-    raise TypeError(f"Unsupported attribute type '{type(value)}' for attribute '{key}'")
+        raise TypeError(
+            f"Unsupported sequence type '{type(value)}' for attribute '{key}' in "
+            f"node={node!r}, value is {value!r}"
+        )
+    if "TensorProtoDataType" in str(type(value)):
+        # torch._C._onnx.TensorProtoDataType
+        return node.i_(key, int(value))
+
+    raise TypeError(
+        f"Unsupported attribute type '{type(value)}' for attribute '{key}' "
+        f"in node={node!r}, value is {value!r}"
+    )
 
 
 @runtime_typing.checked