|
MLCompilerBridge
Tools for streamlining communication with ML models for compiler optimizations.
|
| ▼ CompilerInterface | |
| ▼ compilerinterface | |
| __init__.py | |
| BaseCompilerInterface.py | This file contains the abstract class for compiler interface |
| GrpcCompilerInterface.py | Compiler interface for gRPC |
| log_reader.py | Reader for training log |
| PipeCompilerInterface.py | Compiler interface for pipes |
| SerDes.py | SerDes for JSON and bitstream data |
| fetch_version.py | |
| ▼ include | |
| ▼ MLModelRunner | |
| ▼ C | |
| ONNXModelRunner.h | This file defines the C APIs for ONNXModelRunner |
| PipeModelRunner.h | This file defines the C API for PipeModelRunner |
| ▼ ONNXModelRunner | |
| agent.h | Agent class to support ML/RL model inference via ONNX |
| environment.h | Base Environment class to support ONNX based inference of RL models |
| onnx.h | This file defines the ONNXModel class, which is a wrapper around the ONNX C++ interface |
| ONNXModelRunner.h | ONNXModelRunner class supporting communication via ONNX C++ Runtime |
| utils.h | |
| ▼ Utils | |
| DataTypes.h | This file defines the bit widths of integral and floating point types supported by the MLCompilerBridge |
| Debug.h | This file defines the debug macros for the MLCompilerBridge |
| MLConfig.h | This file defines the MLConfig class, which is a wrapper around the MLConfig command line option for passing information like path of the models and other configuration to the compiler passes |
| gRPCModelRunner.h | GRPCModelRunner class supporting communication via gRPC |
| MLModelRunner.h | The MLModelRunner class is the main interface for interacting with the ML models |
| PipeModelRunner.h | PipeModelRunner class supporting communication via OS pipes between the compiler and an external ML agent |
| TFModelRunner.h | This file implements a model runner wrapping an AOT compiled ML model |
| ▼ SerDes | |
| baseSerDes.h | Supporting new SerDes: |
| bitstreamSerDes.h | Bitstream Serialization/Deserialization which sends header information followed by the raw data |
| jsonSerDes.h | Json Serialization/Deserialization using LLVM's json library |
| protobufSerDes.h | Protobuf Serialization/Deserialization to support gRPC communication |
| tensorflowSerDes.h | Serialization/Deserialization to support TF AOT models |
| TensorSpec.h | |
| ▼ MLModelRunner | |
| ▼ C | |
| ONNXModelRunnerCWrapper.cpp | This file defines the C APIs for ONNXModelRunner |
| PipeModelRunnerCWrapper.cpp | This file defines the C APIs for PipeModelRunner |
| gRPCModelRunner | |
| ▼ ONNXModelRunner | |
| agent.cpp | This file defines the Agent class, which is a wrapper around the ONNXModel class |
| onnx.cpp | This file defines the ONNXModel class, which is a wrapper around the ONNX C++ interface |
| ONNXModelRunner.cpp | This file defines the ONNXModelRunner class to support ML model inference via ONNX |
| ▼ Utils | |
| MLConfig.cpp | |
| PipeModelRunner.cpp | This file defines the PipeModelRunner class to interface with an external ML model during training and inference |
| ▼ SerDes | |
| bitstreamSerDes.cpp | This file defines the BitstreamSerDes class, which is a serializer for Bitstream format |
| jsonSerDes.cpp | This file defines the JsonSerDes class, which is a wrapper around the JSON C++ interface to serialize and deserialize data to and from JSON |
| protobufSerDes.cpp | This file defines the ProtobufSerDes class, which is a wrapper around the protobuf C++ interface to support gRPC communication between the client and server |
| tensorflowSerDes.cpp | This file defines the TensorflowSerDes class, to support interfacing with Tensorflow AOT models via TFModelRunner |
| TensorSpec.cpp | Implementation file for the abstraction of a tensor type, and JSON loading utils |
| tools.cpp |
1.10.0