diff options
author | Grant Watson <grant.watson@arm.com> | 2022-11-16 15:32:39 +0000 |
---|---|---|
committer | Eric Kunze <eric.kunze@arm.com> | 2022-12-15 16:41:27 +0000 |
commit | 64285a1f25e2c7b85ed1f00b7947403e92baea00 (patch) | |
tree | 6d29c54f6497741449339e808508c854ba6a2267 /reference_model/include/model_runner.h | |
parent | b45db9a696f5df7b233f374248f329c16ee7ae64 (diff) | |
download | reference_model-64285a1f25e2c7b85ed1f00b7947403e92baea00.tar.gz |
Extend reference model API with eager operator execution entrypoints
- Adds a script to generate operators.h and operators.cc
- Adds jinja2 templates for generating operators.h and operators.cc
- Adds unit tests for a subset of the operators generated
- Includes the TOSA specification as a submodule
- Adds supporting C++ and header files
Signed-off-by: Grant Watson <grant.watson@arm.com>
Change-Id: I5b60db4c56113110d8e75fe1152525d258233f9c
Diffstat (limited to 'reference_model/include/model_runner.h')
-rw-r--r-- | reference_model/include/model_runner.h | 14 |
1 files changed, 14 insertions, 0 deletions
diff --git a/reference_model/include/model_runner.h b/reference_model/include/model_runner.h index 4335794..86d0056 100644 --- a/reference_model/include/model_runner.h +++ b/reference_model/include/model_runner.h @@ -71,6 +71,13 @@ public: int setInput(std::string input_name, std::vector<T>& vals); /* + * Set the input tensors for the model through a raw byte buffer. + * The input_name much match the input tensor name in the model. + * NOTE: setInput() must be called for each input tensor before run() is called. + */ + int setInput(std::string input_name, uint8_t* raw_ptr, size_t size); + + /* * Retrieve the output tensors from the graph after running. * The output_name much match the output tensor name in the model. * NOTE: run() must be called before outputs are retrieved. @@ -78,6 +85,13 @@ public: template <typename T> std::vector<T> getOutput(std::string output_name); + /* + * Retrieve the output tensors from the graph after running in a raw byte buffer. + * The output_name much match the output tensor name in the model. + * NOTE: run() must be called before outputs are retrieved. + */ + int getOutput(std::string output_name, uint8_t* raw_ptr, size_t size); + private: std::unique_ptr<ModelRunnerImpl> model_runner_impl; }; |