Edit on GitHub

serve

Create an API from model methods using a server implementation.

Synopsis

Description

An MLEM Model can be served via a server implementation (e.g. fastapi) and its methods exposed as API endpoints. This allows us to easily make requests (inference and others) against the served model.

For the common fastapi server implementation, the OpenAPI spec is available on the /docs endpoint.

HTTP Requests to the model-server can be made either with the corresponding built-in client, or common HTTP clients, such as curl and httpie CLIs, or the requests Python library.

Options

  • --request_serializer <str> - Serializer to use for all requests. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'file', 'lightgbm', 'ndarray', 'number', 'pil_numpy', 'primitive', 'series', 'tf_tensor', 'torch', 'torch_image', 'xgboost_dmatrix']. Run 'mlem types serializer ' for list of nested fields for each subtype
  • --response_serializer <str> - Serializer to use for all responses. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'file', 'lightgbm', 'ndarray', 'number', 'pil_numpy', 'primitive', 'series', 'tf_tensor', 'torch', 'torch_image', 'xgboost_dmatrix']. Run 'mlem types serializer ' for list of nested fields for each subtype
  • --standardize <boolean> - Use standard model interface [default: True]
  • --methods <dict> - Optional augmented interface [default: NOT_SET]
  • --methods.key.name <str> - If set, match only method with this name [default: NOT_SET]
  • --methods.key.returns.data_type <str> - Change data type. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'lightgbm', 'list', 'ndarray', 'number', 'primitive', 'series', 'tf_tensor', 'torch', 'tuple', 'unspecified', 'xgboost_dmatrix']. Run 'mlem types data_type ' for list of nested fields for each subtype [default: NOT_SET]
  • --methods.key.returns.ser <str> - Change serializer. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'file', 'lightgbm', 'ndarray', 'number', 'pil_numpy', 'primitive', 'series', 'tf_tensor', 'torch', 'torch_image', 'xgboost_dmatrix']. Run 'mlem types serializer ' for list of nested fields for each subtype [default: NOT_SET]
  • --methods.key.args <dict> - Change arguments options [default: NOT_SET]
  • --methods.key.args.key.data_type <str> - Change data type. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'lightgbm', 'list', 'ndarray', 'number', 'primitive', 'series', 'tf_tensor', 'torch', 'tuple', 'unspecified', 'xgboost_dmatrix']. Run 'mlem types data_type ' for list of nested fields for each subtype [default: NOT_SET]
  • --methods.key.args.key.ser <str> - Change serializer. One of ['array', 'binary', 'csr_matrix', 'd_dict', 'dataframe', 'dict', 'file', 'lightgbm', 'ndarray', 'number', 'pil_numpy', 'primitive', 'series', 'tf_tensor', 'torch', 'torch_image', 'xgboost_dmatrix']. Run 'mlem types serializer ' for list of nested fields for each subtype [default: NOT_SET]
  • --methods.key.args.key.name <str> - If set, match only argument with this name [default: NOT_SET]
  • -m <path>, --model <path> - Path to MLEM model [required]
  • -p <path>, --project <path> - Path to MLEM project [default: (none)]
  • --rev <commitish> - Repo revision to use [default: (none)]
  • -f <text>, --file_conf <text> - File with options for server in format field.name=path_to_config
  • -h, --help - Show this message and exit.

Examples

For examples, please refer to the Get Started or the Serving guide.

Content

­čÉŤ Found an issue? Let us know! Or fix it:

Edit on GitHub

ÔŁô Have a question? Join our chat, we will help you:

Discord Chat