diff --git a/orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py b/orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py
index 79efcb5587..b81e0746a1 100644
--- a/orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py
+++ b/orttraining/orttraining/test/python/orttraining_test_ortmodule_api.py
@@ -510,8 +510,13 @@ def test_gradient_correctness():
         assert torch.allclose(ort_prediction, pt_prediction)
         _test_helpers.assert_gradients_match_and_reset_gradient(ort_model, pt_model)
 
-@pytest.mark.parametrize("use_fp16", [False, True])
-def test_gradient_correctness_conv1d(use_fp16):
+@pytest.mark.parametrize("use_fp16, input_requires_grad", [
+    (False, False),
+    (False, True),
+    (True, False),
+    (True, True),
+    ])
+def test_gradient_correctness_conv1d(use_fp16, input_requires_grad):
     class NeuralNetConv1D(torch.nn.Module):
         def __init__(self, in_channels, out_channels, kernel_size, padding=0, groups=1):
             super(NeuralNetConv1D, self).__init__()
@@ -536,7 +541,7 @@ def test_gradient_correctness_conv1d(use_fp16):
         return prediction
 
     for step in range(10):
-        x = torch.randn(N, seq_len, C_in, device=device, requires_grad=True)
+        x = torch.randn(N, seq_len, C_in, device=device, requires_grad=input_requires_grad)
         pt_prediction = run_step(pt_model, x)
         ort_prediction = run_step(ort_model, x)
         
diff --git a/orttraining/orttraining/training_ops/cuda/nn/conv_grad.cc b/orttraining/orttraining/training_ops/cuda/nn/conv_grad.cc
index b1e9f7a6df..ca742ad897 100644
--- a/orttraining/orttraining/training_ops/cuda/nn/conv_grad.cc
+++ b/orttraining/orttraining/training_ops/cuda/nn/conv_grad.cc
@@ -129,7 +129,7 @@ Status ConvGrad<T>::ComputeInternal(OpKernelContext* context) const {
   Tensor* dW = context->Output(1, W->Shape());
   Tensor* dB = context->Output(2, {M});
 
-  ORT_RETURN_IF_ERROR(PrepareArgs(*dX, *dY, *dW, dB));
+  ORT_RETURN_IF_ERROR(PrepareArgs(*X, *dY, *W, dB));
 
   ORT_RETURN_IF_ERROR(ComputeWeightGradient(dW, dY, X));
   ORT_RETURN_IF_ERROR(ComputeInputGradient(dX, dY, W));