Deploy Models and Integrate TVM

This page contains guidelines on how to deploy TVM to various platforms as well as how to integrate it with your project.

Build the TVM runtime library

Unlike traditional deep learning frameworks. TVM stack is divided into two major components:

• TVM compiler, which does all the compilation and optimizations of the model

• TVM runtime, which runs on the target devices.

In order to integrate the compiled module, we do not need to build entire TVM on the target device. You only need to build the TVM compiler stack on your desktop and use that to cross-compile modules that are deployed on the target device.

We only need to use a light-weight runtime API that can be integrated into various platforms.

For example, you can run the following commands to build the runtime API on a Linux based embedded system such as Raspberry Pi:

git clone --recursive https://github.com/apache/tvm tvm
cd tvm
mkdir build
cp cmake/config.cmake build
cd build
cmake ..
make runtime

Note that we type make runtime to only build the runtime library.

It is also possible to cross compile the runtime. Cross compiling the runtime library should not be confused with cross compiling models for embedded devices.

If you want to include additional runtime such as OpenCL, you can modify config.cmake to enable these options. After you get the TVM runtime library, you can link the compiled library

A model (optimized or not by TVM) can be cross compiled by TVM for different architectures such as aarch64 on a x64_64 host. Once the model is cross compiled it is necessary to have a runtime compatible with the target architecture to be able to run the cross compiled model.

Cross compile the TVM runtime for other architectures

In the example above the runtime library was compiled on a Raspberry Pi. Producing the runtime library can be done much faster on hosts that have high performace processors with ample resources (such as laptops, workstation) compared to a target devices such as a Raspberry Pi. In-order to cross compile the runtime the toolchain for the target device must be installed. After installing the correct toolchain, the main difference compared to compiling natively is to pass some additional command line argument to cmake that specify a toolchain to be used. For reference building the TVM runtime library on a modern laptop (using 8 threads) for aarch64 takes around 20 seconds vs ~10 min to build the runtime on a Raspberry Pi 4.

cross-compile for aarch64

sudo apt-get update
sudo apt-get install gcc-aarch64-linux-gnu g++-aarch64-linux-gnu

cmake .. \
    -DCMAKE_SYSTEM_NAME=Linux \
    -DCMAKE_SYSTEM_VERSION=1 \
    -DCMAKE_C_COMPILER=/usr/bin/aarch64-linux-gnu-gcc \
    -DCMAKE_CXX_COMPILER=/usr/bin/aarch64-linux-gnu-g++ \
    -DCMAKE_FIND_ROOT_PATH=/usr/aarch64-linux-gnu \
    -DCMAKE_FIND_ROOT_PATH_MODE_PROGRAM=NEVER \
    -DCMAKE_FIND_ROOT_PATH_MODE_LIBRARY=ONLY \
    -DMACHINE_NAME=aarch64-linux-gnu

make -j$(nproc) runtime

For bare metal ARM devices the following toolchain is quite handy to install instead of gcc-aarch64-linux-*

sudo apt-get install gcc-multilib-arm-linux-gnueabihf g++-multilib-arm-linux-gnueabihf

cross-compile for RISC-V

sudo apt-get update
sudo apt-get install gcc-riscv64-linux-gnu g++-riscv64-linux-gnu

cmake .. \
    -DCMAKE_SYSTEM_NAME=Linux \
    -DCMAKE_SYSTEM_VERSION=1 \
    -DCMAKE_C_COMPILER=/usr/bin/riscv64-linux-gnu-gcc \
    -DCMAKE_CXX_COMPILER=/usr/bin/riscv64-linux-gnu-g++ \
    -DCMAKE_FIND_ROOT_PATH=/usr/riscv64-linux-gnu \
    -DCMAKE_FIND_ROOT_PATH_MODE_PROGRAM=NEVER \
    -DCMAKE_FIND_ROOT_PATH_MODE_LIBRARY=ONLY \
    -DMACHINE_NAME=riscv64-linux-gnu

make -j$(nproc) runtime

The file command can be used to query the architecture of the produced runtime.

file libtvm_runtime.so
libtvm_runtime.so: ELF 64-bit LSB shared object, UCB RISC-V, version 1 (GNU/Linux), dynamically linked, BuildID[sha1]=e9ak845b3d7f2c126dab53632aea8e012d89477e, not stripped

Optimize and tune models for target devices

The easiest and recommended way to test, tune and benchmark TVM kernels on embedded devices is through TVM’s RPC API. Here are the links to the related tutorials.

• Cross Compilation and RPC

• Deploy the Pretrained Model on Raspberry Pi

Deploy optimized model on target devices

After you finished tuning and benchmarking, you might need to deploy the model on the target device without relying on RPC. See the following resources on how to do so.

• Deploy TVM Module using C++ API
  • Get TVM Runtime Library
  • Dynamic Library vs. System Module
• Deploy to Android
  • Build model for Android Target
  • TVM Runtime for Android Target
• Deploy to Adreno GPU
  • Introduction
  • Advantages of the Textures
  • Building TVM for Adreno
  • Build and deploy model for Adreno
• Integrate TVM into Your Project
  • DLPack Support
  • Integrate User Defined C++ Array
  • Integrate User Defined Python Array
• HLS Backend Example
  • Setup
  • Emulation
  • Synthesis
  • Run
• Relay Arm^® Compute Library Integration
  • Introduction
  • Installing Arm Compute Library
  • Building with ACL support
  • Usage
  • More examples
  • Operator support
  • Adding a new operator
• Relay TensorRT Integration
  • Introduction
  • Installing TensorRT
  • Building TVM with TensorRT support
  • Build and Deploy ResNet-18 with TensorRT
  • Partitioning and Compilation Settings
  • Runtime Settings
  • Operator support
  • Adding a new operator
• Vitis AI Integration
  • System Requirements
  • Setup instructions
  • Compiling a Model
  • Inference
• Relay BNNS Integration
  • Introduction
  • Building TVM with BNNS support
  • BNNS partitioning of Relay graph
  • Input data layout for operations to be offloaded to BNNS execution
  • Example: Build and Deploy Mobilenet v2 1.0 with BNNS
  • Operator support

Additional Deployment How-Tos

We have also developed a number of how-tos targeting specific devices, with working Python code that can be viewed in a Jupyter notebook. These how-tos describe how to prepare and deploy models to many of the supported backends.

• Deploy Deep Learning Models