[GitHub] [incubator-mxnet] ptrendx commented on a change in pull request #18490: MXNet-TRT: Add PrePartition param caching - move init_tensorrt_params logic

[GitBox <gi...@apache.org>]

ptrendx commented on a change in pull request #18490:
URL: https://github.com/apache/incubator-mxnet/pull/18490#discussion_r465250126



##########
File path: src/operator/subgraph/tensorrt/tensorrt-inl.h
##########
@@ -267,6 +267,23 @@ class TensorrtProperty : public SubgraphProperty {
     return std::make_shared<TensorrtProperty>();
   }
 
+  void PrePartition(const nnvm::Graph& g,
+    const std::vector<std::pair<std::string, std::string>>& options_map) override {
+    auto& in_arg_names = g.GetAttr<std::vector<std::string>>("in_arg_names");
+    auto& in_aux_names = g.GetAttr<std::vector<std::string>>("in_aux_names");
+    NDArray **in_args_ptr = g.GetAttr<NDArray**>("in_args");
+    NDArray **in_aux_ptr = g.GetAttr<NDArray**>("in_aux");
+    in_args_dict.clear();
+    in_aux_dict.clear();
+    // we trust the Python API, len(in_arg_names) == len(in_args_ptr)

Review comment:
       Why not `CHECK_EQ` them?




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org