Arm backend: Add VGF grid_sampler integration test

Add a VGF export/integration test for grid_sampler using the
existing VgfPipeline.

The test is gated on model-converter availability and runs with
quantization disabled and Vulkan runtime execution disabled, so
it validates the converter/export handoff without depending on
runtime shader dispatch.

Also register grid_sampler as a recognized custom edge op for the
Arm test-name validator.

Signed-off-by: Baris Demir <baris.demir@arm.com>
Change-Id: I495d4283b09de7dedebce5d2f24b240e79d37dfc

Author: Baris Demir

backends/arm/test/ops/test_grid_sampler.py

@@ -0,0 +1,62 @@
+# Copyright 2026 Arm Limited and/or its affiliates.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+from typing import Tuple
+
+import torch
+import torch.nn.functional as F
+from executorch.backends.arm.test import common
+from executorch.backends.arm.test.tester.test_pipeline import VgfPipeline
+
+input_t = Tuple[torch.Tensor, torch.Tensor]
+aten_op = "torch.ops.aten.grid_sampler.default"
+exir_op = "executorch_exir_dialects_edge__ops_aten_grid_sampler_2d_default"
+
+test_data_suite = {
+    "2d_bilinear_zeros": lambda: (
+        torch.randn(1, 3, 8, 8),
+        torch.randn(1, 4, 4, 2),
+    ),
+}
+
+xfails = {
+    "2d_bilinear_zeros": (
+        "CI model_converter does not yet include Vulkan custom-shader "
+        "tosa.custom legalization",
+        RuntimeError,
+    ),
+}
+
+
+class GridSampler2d(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.interpolation_mode_ = 0
+        self.padding_mode_ = 0
+        self.align_corners_ = False
+
+    def forward(self, x, grid):
+        return F.grid_sample(
+            x,
+            grid,
+            mode="bilinear" if self.interpolation_mode_ == 0 else "nearest",
+            padding_mode="zeros" if self.padding_mode_ == 0 else "border",
+            align_corners=self.align_corners_,
+        )
+
+
+@common.parametrize("test_data", test_data_suite, xfails=xfails, strict=False)
+@common.SkipIfNoModelConverter
+def test_grid_sampler_vgf_no_quant(test_data):
+    test_data = test_data()
+    pipeline = VgfPipeline[input_t](
+        GridSampler2d(),
+        test_data,
+        aten_op,
+        exir_op,
+        quantize=False,
+        run_on_vulkan_runtime=False,
+    )
+    pipeline.run()