diff --git a/lib/Conversion/TorchToLinalg/DataMovement.cpp b/lib/Conversion/TorchToLinalg/DataMovement.cpp
index a18c0bae01fc..bd0215d302e7 100644
--- a/lib/Conversion/TorchToLinalg/DataMovement.cpp
+++ b/lib/Conversion/TorchToLinalg/DataMovement.cpp
@@ -403,17 +403,40 @@ class ConvertAtenReflectionPad2dOp
     int64_t vDim = numDims - 2;
     Value hDimSize = inputShape[hDim];
     Value vDimSize = inputShape[vDim];
+    Type indexType = rewriter.getIndexType();
 
-    assert(getHPadArgument(LEFT) < inputType.getShape()[hDim] &&
-           "Left padding too large");
-    assert(getHPadArgument(RIGHT) < inputType.getShape()[hDim] &&
-           "Right padding too large");
-    assert(getVPadArgument(TOP) < inputType.getShape()[vDim] &&
-           "Top padding too large");
-    assert(getVPadArgument(BOTTOM) < inputType.getShape()[vDim] &&
-           "Bottom padding too large");
+    auto leftPadAssertMsg = "Left padding too large";
+    auto rightPadAssertMsg = "Right padding too large";
+    auto topPadAssertMsg = "Top padding too large";
+    auto bottomPadAssertMsg = "Bottom padding too large";
+
+    auto addPadDynAssert = [&](int64_t pad, Value dimSize,
+                               const llvm::Twine &msg) {
+      Value padValue = getConstant(rewriter, loc, pad, indexType);
+      Value pred = rewriter.create<arith::CmpIOp>(
+          loc, arith::CmpIPredicate::slt, padValue, dimSize);
+      rewriter.create<cf::AssertOp>(loc, pred, rewriter.getStringAttr(msg));
+    };
+
+    if (inputType.getShape()[hDim] != ShapedType::kDynamic) {
+      assert(getHPadArgument(LEFT) < inputType.getShape()[hDim] &&
+             leftPadAssertMsg);
+      assert(getHPadArgument(RIGHT) < inputType.getShape()[hDim] &&
+             rightPadAssertMsg);
+    } else {
+      addPadDynAssert(getHPadArgument(LEFT), hDimSize, leftPadAssertMsg);
+      addPadDynAssert(getHPadArgument(RIGHT), hDimSize, rightPadAssertMsg);
+    }
+    if (inputType.getShape()[vDim] != ShapedType::kDynamic) {
+      assert(getVPadArgument(TOP) < inputType.getShape()[vDim] &&
+             topPadAssertMsg);
+      assert(getVPadArgument(BOTTOM) < inputType.getShape()[vDim] &&
+             bottomPadAssertMsg);
+    } else {
+      addPadDynAssert(getVPadArgument(TOP), vDimSize, topPadAssertMsg);
+      addPadDynAssert(getVPadArgument(BOTTOM), vDimSize, bottomPadAssertMsg);
+    }
 
-    Type indexType = rewriter.getIndexType();
     Value zero = getConstant(rewriter, loc, 0, indexType);
     Value one = getConstant(rewriter, loc, 1, indexType);
 
diff --git a/projects/pt1/python/torch_mlir_e2e_test/test_suite/padding.py b/projects/pt1/python/torch_mlir_e2e_test/test_suite/padding.py
index a97d7f09eda6..64d6e46a2a72 100644
--- a/projects/pt1/python/torch_mlir_e2e_test/test_suite/padding.py
+++ b/projects/pt1/python/torch_mlir_e2e_test/test_suite/padding.py
@@ -36,6 +36,29 @@ def ReflectionPad2dModule_basic(module, tu: TestUtils):
 # ==============================================================================
 
 
+class ReflectionPad2dDynamicSizesModule(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    @export
+    @annotate_args(
+        [
+            None,
+            ([1, -1, -1], torch.float32, True),
+        ]
+    )
+    def forward(self, x):
+        return torch.ops.aten.reflection_pad2d(x, (10, 10, 10, 10))
+
+
+@register_test_case(module_factory=lambda: ReflectionPad2dDynamicSizesModule())
+def ReflectionPad2dDynamicSizesModule_basic(module, tu: TestUtils):
+    module.forward(tu.rand(1, 20, 20, low=-1))
+
+
+# ==============================================================================
+
+
 class ReflectionPad2dModuleTop(torch.nn.Module):
     def __init__(self):
         super().__init__()