A key feature of MMS is the ability to package all model artifacts into a single model archive file. It is a separate command line interface (CLI), model-archiver
, that can take model checkpoints and package them into a .mar
file. This file can then be redistributed and served by anyone using MMS. It takes in the following model artifacts: a model composed of one or more files, the description of the model's inputs in the form of a signature file, a service file describing how to handle inputs and outputs, and other optional assets that may be required to serve the model. The CLI creates a .mar
file that MMS's server CLI uses to serve the models.
Important: Make sure you try the Quick Start: Creating a Model Archive tutorial for a short example of using model-archiver
.
MMS can support any arbitrary model file. It is the custom service code's responsibility to locate and load the model files. The following information is required to create a standalone model archive:
Now let's cover the details on using the CLI tool: model-archiver
.
Here is an example usage with the squeezenet_v1.1 model archive which you can download or create by following the example in the main README:
model-archiver --model-name squeezenet_v1.1 --model-path squeezenet --handler mxnet_vision_service:handle
$ model-archiver -h
usage: model-archiver [-h] --model-name MODEL_NAME --model-path MODEL_PATH
--handler HANDLER [--runtime {python,python2,python3}]
[--export-path EXPORT_PATH] [-f]
Model Archiver Tool
optional arguments:
-h, --help show this help message and exit
--model-name MODEL_NAME
Exported model name. Exported file will be named as
model-name.mar and saved in current working directory
if no --export-path is specified, else it will be
saved under the export path
--model-path MODEL_PATH
Path to the folder containing model related files.
--handler HANDLER Handler path to handle custom MMS inference logic.
--runtime {python,python2,python3}
The runtime specifies which language to run your
inference code on. The default runtime is
RuntimeType.PYTHON. At the present moment we support
the following runtimes python, python2, python3
--export-path EXPORT_PATH
Path where the exported .mar file will be saved. This
is an optional parameter. If --export-path is not
specified, the file will be saved in the current
working directory.
--archive-format {tgz,default}
The format in which the model artifacts are archived.
"tgz": This creates the model-archive in <model-name>.tar.gz format.
If platform hosting MMS requires model-artifacts to be in ".tar.gz"
use this option.
"no-archive": This option creates an non-archived version of model artifacts
at "export-path/{model-name}" location. As a result of this choice,
MANIFEST file will be created at "export-path/{model-name}" location
without archiving these model files
"default": This creates the model-archive in <model-name>.mar format.
This is the default archiving format. Models archived in this format
will be readily hostable on native MMS.
-f, --force When the -f or --force flag is specified, an existing
.mar file with same name as that provided in --model-
name in the path specified by --export-path will
overwritten
MAR-INF is a reserved folder name that will be used inside .mar
file. This folder contains the model archive metadata files. Users should avoid using MAR-INF in their model path.
A valid model name must begin with a letter of the alphabet and can only contains letters, digits, underscores (_), dashes (-) and periods (.).
Note: The model name can be overridden when you register the model with Register Model API.
A folder that contains all necessary files needed to run inference code for the model. All the files and sub-folders (except excluded files) will be packaged into the .mar
file.
The following types of file will be excluded during model archive packaging:
- hidden files
- Mac system files: __MACOSX and .DS_Store
- MANIFEST.json
- python compiled byte code (.pyc) files and cache folder pycache
A handler is a python entry point that MMS can invoke to execute inference code. The format of a Python handler is:
- python_module_name[:function_name] (for example: lstm-service:handle).
The function name is optional if the provided python module follows one of predefined conventions:
- There is a
handle()
function available in the module - The module contains only one Class and that class contains a
handle()
function.
Further details and specifications are found on the custom service page.
1. Download these sample SqueezeNet model artifacts (if you don't have them handy)
mkdir squeezenet
curl -o squeezenet https://s3.amazonaws.com/model-server/model_archive_1.0/examples/squeezenet_v1.1/squeezenet_v1.1-symbol.json
curl -o squeezenet https://s3.amazonaws.com/model-server/model_archive_1.0/examples/squeezenet_v1.1/squeezenet_v1.1-0000.params
curl -o squeezenet https://s3.amazonaws.com/model-server/model_archive_1.0/examples/squeezenet_v1.1/signature.json
curl -o squeezenet https://s3.amazonaws.com/model-server/model_archive_1.0/examples/squeezenet_v1.1/synset.txt
The downloaded model artifact files are:
- Model Definition (json file) - contains the layers and overall structure of the neural network.
- Model Params and Weights (params file) - contains the parameters and the weights.
- Model Signature (json file) - defines the inputs and outputs that MMS is expecting to hand-off to the API.
- assets (text files) - auxiliary files that support model inference such as vocabularies, labels, etc. These vary depending on the model.
2. Download the model archiver source
git clone https://github.com/awslabs/mxnet-model-server.git
3. Prepare your model custom service code
You can implement your own model customer service code with a model archive entry point.
Here we are going to use the MXNet vision service model_service_template
.
This template is one of several provided with MMS.
Download the template and place it in your squeezenet
folder.
cp -r mxnet-model-server/examples/model_service_template/* squeezenet/
4. Package your model
With the model artifacts available locally, you can use the model-archiver
CLI to generate a .mar
file that can be used to serve an inference API with MMS.
In this next step we'll run model-archiver
and tell it our model's prefix is squeezenet_v1.1
with the model-name
argument. Then we're giving it the model-path
to the model's assets.
Note: For mxnet models, model-name
must match prefix of the symbol and param file name.
model-archiver --model-name squeezenet_v1.1 --model-path squeezenet --handler mxnet_vision_service:handle
This will package all the model artifacts files located in the squeezenet
directory and output squeezenet_v1.1.mar
in the current working directory. This .mar
file is all you need to run MMS, serving inference requests for a simple image recognition API. Go back to the Serve a Model tutorial and try to run this model archive that you just created!