masahi commented on a change in pull request #9108:
URL: https://github.com/apache/tvm/pull/9108#discussion_r727603521
##########
File path: src/runtime/pipeline/pipeline_executor.h
##########
@@ -36,25 +43,114 @@ namespace runtime {
*
* This executor can be accessed by various language via TVM runtime
PackedFunc API.
*/
-class TVM_DLL PipelineRuntime : public ModuleNode {
+class TVM_DLL PipelineExecutor : public ModuleNode {
public:
/*!
* \Return the type key of the executor.
*/
- const char* type_key() const final { return "PipelineRuntime"; }
+ const char* type_key() const final { return "PipelineExecutor"; }
/*!
- * \brief Initialize the pipeline executor with module array and json text.
+ * \brief Initialize the pipeline executor with module array and JSON text.
* \param modules The module list used for building pipeline.
* \param pipeline_json The configuration of modules dependencies.
*/
- void Init(const Array<tvm::runtime::Module>& modules, const std::string&
pipeline_json);
+ void Init(const std::vector<Module>& modules, const std::string&
pipeline_json);
+ /*!
+ * \brief Use the information of mod_config to create a graph executor list.
Review comment:
a list of graph executor
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
