fixed qlinearconv conversion for freeze_params=True

Author: masahi

python/tvm/relay/frontend/onnx.py

@@ -3720,12 +3720,14 @@ def _impl_v10(cls, inputs, attr, params):
             if attr["auto_pad"] in ("SAME_UPPER", "SAME_LOWER"):
                 # Warning: Convolution does not yet support dynamic shapes,
                 # one will need to run dynamic_to_static on this model after import
+                zp = fold_constant(x_zero_point), "Zero point expected to be a constant"
+                assert isinstance(zp, relay.Constant)
                 data = autopad(
                     data,
                     attr.get("strides", [1] * (ndim - 2)),
                     attr["kernel_shape"],
                     attr.get("dilations", [1] * (ndim - 2)),
-                    pad_value=x_zero_point.data,
+                    pad_value=zp.data,
                     mode=attr["auto_pad"],
                 )
             elif attr["auto_pad"] == "VALID":