add conv validator

Author: zewenli98

py/torch_tensorrt/dynamo/conversion/aten_ops_converters.py

@@ -814,21 +814,20 @@ def aten_ops_isinf(
     )
 
 
-@dynamo_tensorrt_converter(torch.ops.aten.convolution.default)
+def conv_param_validator(conv_node: Node) -> bool:
+    return (not conv_node.args[6]) and (conv_node.args[7] in ([0], [0, 0], [0, 0, 0]))
+
+
+@dynamo_tensorrt_converter(
+    torch.ops.aten.convolution.default, capability_validator=conv_param_validator
+)
 def aten_ops_convolution(
     network: TRTNetwork,
     target: Target,
     args: Tuple[Argument, ...],
     kwargs: Dict[str, Argument],
     name: str,
 ) -> Union[TRTTensor, Sequence[TRTTensor]]:
-    # we do not handle transposed.
-    if args[6] is True:
-        raise RuntimeError(f"Target {target} does not support `transposed=True` ")
-    # we do not handle output_padding.
-    if args[7] not in ([0], [0, 0], [0, 0, 0]):
-        raise RuntimeError(f"Target {target} has non-0 output_padding")
-
     return impl.conv.convNd(
         network,
         target,