TorchServe

TorchServe is a flexible and easy to use tool for serving PyTorch models.

For full documentation, see Model Server for PyTorch Documentation.

TorchServe Architecture

Architecture Diagram

Terminology:

Contents of this Document

Install TorchServe

Conda instructions are provided in more detail, but you may also use pip and virtualenv if that is your preference. Note: Java 11 is required. Instructions for installing Java 11 for Ubuntu or macOS are provided in the Install with Conda section.

Install with pip

  1. Install Java 11

    sudo apt-get install openjdk-11-jdk
  2. Use pip to install TorchServe and the model archiver:

    pip install torch torchtext torchvision sentencepiece psutil future
    pip install torchserve torch-model-archiver

Install with Conda

Note: For Conda, Python 3.8 is required to run Torchserve

Ubuntu

  1. Install Java 11

    sudo apt-get install openjdk-11-jdk
  2. Install Conda (https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html)

  3. Create an environment and install torchserve and torch-model-archiver For CPU

    conda create --name torchserve torchserve torch-model-archiver psutil future pytorch torchtext torchvision -c pytorch -c powerai

    For GPU

    conda create --name torchserve torchserve torch-model-archiver psutil future pytorch torchtext torchvision cudatoolkit=10.1 -c pytorch -c powerai
  4. Activate the environment

    source activate torchserve
  5. Optional if using torchtext models

    pip install sentencepiece

macOS

  1. Install Java 11

    brew tap AdoptOpenJDK/openjdk
    brew cask install adoptopenjdk11
  2. Install Conda (https://docs.conda.io/projects/conda/en/latest/user-guide/install/macos.html)

  3. Create an environment and install torchserve and torch-model-archiver

    conda create --name torchserve torchserve torch-model-archiver psutil future pytorch torchtext torchvision -c pytorch -c powerai
  4. Activate the environment

    source activate torchserve
  5. Optional if using torchtext models

    pip install sentencepiece

Now you are ready to package and serve models with TorchServe.

Install TorchServe for development

If you plan to develop with TorchServe and change some of the source code, you must install it from source code.

Please deactivate any conda env that you might be within. Run the following script from the top of the source directory.

NOTE: This script uninstalls existing torchserve and torch-model-archiver installations

For Debian Based Systems

Verified on EC2 instances running Ubuntu DL AMI 28.x

./scripts/install_from_src_ubuntu

For macOS

./scripts/install_from_src_macos

For information about the model archiver, see detailed documentation.

Serve a model

This section shows a simple example of serving a model with TorchServe. To complete this example, you must have already installed TorchServe and the model archiver.

To run this example, clone the TorchServe repository:

git clone https://github.com/pytorch/serve.git

Then run the following steps from the parent directory of the root of the repository. For example, if you cloned the repository into /home/my_path/serve, run the steps from /home/my_path.

Store a Model

To serve a model with TorchServe, first archive the model as a MAR file. You can use the model archiver to package a model. You can also create model stores to store your archived models.

  1. Create a directory to store your models.

    mkdir model_store
  2. Download a trained model.

    wget https://download.pytorch.org/models/densenet161-8d451a50.pth
  3. Archive the model by using the model archiver. The extra-files param uses fa file from the TorchServe repo, so update the path if necessary.

    torch-model-archiver --model-name densenet161 --version 1.0 --model-file ./serve/examples/image_classifier/densenet_161/model.py --serialized-file densenet161-8d451a50.pth --export-path model_store --extra-files ./serve/examples/image_classifier/index_to_name.json --handler image_classifier

For more information about the model archiver, see Torch Model archiver for TorchServe

Start TorchServe to serve the model

After you archive and store the model, use the torchserve command to serve the model.

torchserve --start --ncs --model-store model_store --models densenet161.mar

After you execute the torchserve command above, TorchServe runs on your host, listening for inference requests.

Note: If you specify model(s) when you run TorchServe, it automatically scales backend workers to the number equal to available vCPUs (if you run on a CPU instance) or to the number of available GPUs (if you run on a GPU instance). In case of powerful hosts with a lot of compute resoures (vCPUs or GPUs), this start up and autoscaling process might take considerable time. If you want to minimize TorchServe start up time you should avoid registering and scaling the model during start up time and move that to a later point by using corresponding Management API, which allows finer grain control of the resources that are allocated for any particular model).

Get predictions from a model

To test the model server, send a request to the server's predictions API.

Complete the following steps:

kitten

The following code completes all three steps:

curl -O https://s3.amazonaws.com/model-server/inputs/kitten.jpg
curl http://127.0.0.1:8080/predictions/densenet161 -T kitten.jpg

The predict endpoint returns a prediction response in JSON. It will look something like the following result:

[
  {
    "tiger_cat": 0.46933549642562866
  },
  {
    "tabby": 0.4633878469467163
  },
  {
    "Egyptian_cat": 0.06456148624420166
  },
  {
    "lynx": 0.0012828214094042778
  },
  {
    "plastic_bag": 0.00023323034110944718
  }
]

You will see this result in the response to your curl call to the predict endpoint, and in the server logs in the terminal window running TorchServe. It's also being logged locally with metrics.

Now you've seen how easy it can be to serve a deep learning model with TorchServe! Would you like to know more?

Stop the running TorchServe

To stop the currently running TorchServe instance, run the following command:

torchserve --stop

You see output specifying that TorchServe has stopped.

Concurrency And Number of Workers

TorchServe exposes configurations that allow the user to configure the number of worker threads on CPU and GPUs. There is an important config property that can speed up the server depending on the workload. Note: the following property has bigger impact under heavy workloads. If TorchServe is hosted on a machine with GPUs, there is a config property called number_of_gpu that tells the server to use a specific number of GPUs per model. In cases where we register multiple models with the server, this will apply to all the models registered. If this is set to a low value (ex: 0 or 1), it will result in under-utilization of GPUs. On the contrary, setting to a high value (>= max GPUs available on the system) results in as many workers getting spawned per model. Clearly, this will result in unnecessary contention for GPUs and can result in sub-optimal scheduling of threads to GPU.

ValueToSet = (Number of Hardware GPUs) / (Number of Unique Models)

Quick Start with Docker

Refer torchserve docker for details.

Learn More

Contributing

We welcome all contributions!

To learn more about how to contribute, see the contributor guide here.

To file a bug or request a feature, please file a GitHub issue. For filing pull requests, please use the template here. Cheers!

TorchServe acknowledges the Multi Model Server (MMS) project from which it was derived