docs: update generated docs (#1053)
Updates for both discovery docs and epydoc API Documentation
Fixes: #1049
diff --git a/docs/dyn/ml_v1.projects.models.versions.html b/docs/dyn/ml_v1.projects.models.versions.html
index 1971b65..a0ccb63 100644
--- a/docs/dyn/ml_v1.projects.models.versions.html
+++ b/docs/dyn/ml_v1.projects.models.versions.html
@@ -75,372 +75,115 @@
<h1><a href="ml_v1.html">AI Platform Training & Prediction API</a> . <a href="ml_v1.projects.html">projects</a> . <a href="ml_v1.projects.models.html">models</a> . <a href="ml_v1.projects.models.versions.html">versions</a></h1>
<h2>Instance Methods</h2>
<p class="toc_element">
+ <code><a href="#close">close()</a></code></p>
+<p class="firstline">Close httplib2 connections.</p>
+<p class="toc_element">
<code><a href="#create">create(parent, body=None, x__xgafv=None)</a></code></p>
-<p class="firstline">Creates a new version of a model from a trained TensorFlow model.</p>
+<p class="firstline">Creates a new version of a model from a trained TensorFlow model. If the version created in the cloud by this call is the first deployed version of the specified model, it will be made the default version of the model. When you add a version to a model that already has one or more versions, the default version does not automatically change. If you want a new version to be the default, you must call projects.models.versions.setDefault.</p>
<p class="toc_element">
<code><a href="#delete">delete(name, x__xgafv=None)</a></code></p>
-<p class="firstline">Deletes a model version.</p>
+<p class="firstline">Deletes a model version. Each model can have multiple versions deployed and in use at any given time. Use this method to remove a single version. Note: You cannot delete the version that is set as the default version of the model unless it is the only remaining version.</p>
<p class="toc_element">
<code><a href="#get">get(name, x__xgafv=None)</a></code></p>
-<p class="firstline">Gets information about a model version.</p>
+<p class="firstline">Gets information about a model version. Models can have multiple versions. You can call projects.models.versions.list to get the same information that this method returns for all of the versions of a model.</p>
<p class="toc_element">
- <code><a href="#list">list(parent, filter=None, pageToken=None, pageSize=None, x__xgafv=None)</a></code></p>
-<p class="firstline">Gets basic information about all the versions of a model.</p>
+ <code><a href="#list">list(parent, filter=None, pageSize=None, pageToken=None, x__xgafv=None)</a></code></p>
+<p class="firstline">Gets basic information about all the versions of a model. If you expect that a model has many versions, or if you need to handle only a limited number of results at a time, you can request that the list be retrieved in batches (called pages). If there are no versions that match the request parameters, the list request returns an empty response body: {}.</p>
<p class="toc_element">
<code><a href="#list_next">list_next(previous_request, previous_response)</a></code></p>
<p class="firstline">Retrieves the next page of results.</p>
<p class="toc_element">
<code><a href="#patch">patch(name, body=None, updateMask=None, x__xgafv=None)</a></code></p>
-<p class="firstline">Updates the specified Version resource.</p>
+<p class="firstline">Updates the specified Version resource. Currently the only update-able fields are `description`, `requestLoggingConfig`, `autoScaling.minNodes`, and `manualScaling.nodes`.</p>
<p class="toc_element">
<code><a href="#setDefault">setDefault(name, body=None, x__xgafv=None)</a></code></p>
-<p class="firstline">Designates a version to be the default for the model.</p>
+<p class="firstline">Designates a version to be the default for the model. The default version is used for prediction requests made against the model that don't specify a version. The first version to be created for a model is automatically set as the default. You must make any subsequent changes to the default version setting manually using this method.</p>
<h3>Method Details</h3>
<div class="method">
- <code class="details" id="create">create(parent, body=None, x__xgafv=None)</code>
- <pre>Creates a new version of a model from a trained TensorFlow model.
+ <code class="details" id="close">close()</code>
+ <pre>Close httplib2 connections.</pre>
+</div>
-If the version created in the cloud by this call is the first deployed
-version of the specified model, it will be made the default version of the
-model. When you add a version to a model that already has one or more
-versions, the default version does not automatically change. If you want a
-new version to be the default, you must call
-projects.models.versions.setDefault.
+<div class="method">
+ <code class="details" id="create">create(parent, body=None, x__xgafv=None)</code>
+ <pre>Creates a new version of a model from a trained TensorFlow model. If the version created in the cloud by this call is the first deployed version of the specified model, it will be made the default version of the model. When you add a version to a model that already has one or more versions, the default version does not automatically change. If you want a new version to be the default, you must call projects.models.versions.setDefault.
Args:
parent: string, Required. The name of the model. (required)
body: object, The request body.
The object takes the form of:
-{ # Represents a version of the model.
- #
- # Each version is a trained model deployed in the cloud, ready to handle
- # prediction requests. A model can have multiple versions. You can get
- # information about all of the versions of a given model by calling
- # projects.models.versions.list.
- "labels": { # Optional. One or more labels that you can add, to organize your model
- # versions. Each label is a key-value pair, where both the key and the value
- # are arbitrary strings that you supply.
- # For more information, see the documentation on
- # <a href="/ml-engine/docs/tensorflow/resource-labels">using labels</a>.
- "a_key": "A String",
+{ # Represents a version of the model. Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions. You can get information about all of the versions of a given model by calling projects.models.versions.list.
+ "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to create the version. See the [guide to model deployment](/ml-engine/docs/tensorflow/deploying-models) for more information. When passing Version to projects.models.versions.create the model service uses the specified location as the source of the model. Once deployed, the model version is hosted by the prediction service, so this location is useful only as a historical record. The total number of model files can't exceed 1000.
+ "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. Online prediction requests to a model version and the responses to these requests are converted to raw strings and saved to the specified BigQuery table. Logging is constrained by [BigQuery quotas and limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits, AI Platform Prediction does not log request-response pairs, but it continues to serve predictions. If you are using [continuous evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to specify this configuration manually. Setting up continuous evaluation automatically enables logging of request-response pairs. # Optional. *Only* specify this field in a projects.models.versions.patch request. Specifying it in a projects.models.versions.create request has no effect. Configures the request-response pair logging on predictions from this Version.
+ "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format: " project_id.dataset_name.table_name" The specified table must already exist, and the "Cloud ML Service Agent" for your project must have permission to write to it. The table must have the following [schema](/bigquery/docs/schemas): Field nameType Mode model STRING REQUIRED model_version STRING REQUIRED time TIMESTAMP REQUIRED raw_data STRING REQUIRED raw_prediction STRING NULLABLE groundtruth STRING NULLABLE
+ "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1. For example, if you want to log 10% of requests, enter `0.1`. The sampling window is the lifetime of the model version. Defaults to 0.
},
- "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only
- # applies to online prediction service. If this field is not specified, it
- # defaults to `mls1-c1-m2`.
- #
- # Online prediction supports the following machine types:
- #
- # * `mls1-c1-m2`
- # * `mls1-c4-m2`
- # * `n1-standard-2`
- # * `n1-standard-4`
- # * `n1-standard-8`
- # * `n1-standard-16`
- # * `n1-standard-32`
- # * `n1-highmem-2`
- # * `n1-highmem-4`
- # * `n1-highmem-8`
- # * `n1-highmem-16`
- # * `n1-highmem-32`
- # * `n1-highcpu-2`
- # * `n1-highcpu-4`
- # * `n1-highcpu-8`
- # * `n1-highcpu-16`
- # * `n1-highcpu-32`
- #
- # `mls1-c1-m2` is generally available. All other machine types are available
- # in beta. Learn more about the [differences between machine
- # types](/ml-engine/docs/machine-types-online-prediction).
- "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom
- # prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines)
- # or [scikit-learn pipelines with custom
- # code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code).
- #
- # For a custom prediction routine, one of these packages must contain your
- # Predictor class (see
- # [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally,
- # include any dependencies used by your Predictor or scikit-learn pipeline
- # uses that are not already included in your selected [runtime
- # version](/ml-engine/docs/tensorflow/runtime-version-list).
- #
- # If you specify this field, you must also set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
- "A String",
- ],
- "acceleratorConfig": { # Represents a hardware accelerator request config. # Optional. Accelerator config for using GPUs for online prediction (beta).
- # Only specify this field if you have specified a Compute Engine (N1) machine
- # type in the `machineType` field. Learn more about [using GPUs for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
- # Note that the AcceleratorConfig can be used in both Jobs and Versions.
- # Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and
- # [accelerators for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
+ "predictionClass": "A String", # Optional. The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the [`packageUris` field](#Version.FIELDS.package_uris). Specify this field if and only if you are deploying a [custom prediction routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify this field, you must set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and you must set `machineType` to a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction). The following code sample provides the Predictor interface: class Predictor(object): """Interface for constructing custom predictors.""" def predict(self, instances, **kwargs): """Performs custom prediction. Instances are the decoded values from the request. They have already been deserialized from JSON. Args: instances: A list of prediction input instances. **kwargs: A dictionary of keyword args provided as additional fields on the predict request body. Returns: A list of outputs containing the prediction results. This list must be JSON serializable. """ raise NotImplementedError() @classmethod def from_path(cls, model_dir): """Creates an instance of Predictor using the given path. Loading of the predictor should be done in this method. Args: model_dir: The local directory that contains the exported model file along with any additional files uploaded when creating the version resource. Returns: An instance implementing this Predictor class. """ raise NotImplementedError() Learn more about [the Predictor interface and custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
+ "framework": "A String", # Optional. The machine learning framework AI Platform uses to train this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`, `XGBOOST`. If you do not specify a framework, AI Platform will analyze files in the deployment_uri to determine a framework. If you choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version of the model to 1.4 or greater. Do **not** specify a framework if you're deploying a [custom prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction) in the `machineType` field, you must specify `TENSORFLOW` for the framework.
+ "description": "A String", # Optional. The description specified for the version when it was created.
+ "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify `manual_scaling`.
+ "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed. Therefore, the cost of operating this model will be at least `rate` * `min_nodes` * number of hours since last billing cycle, where `rate` is the cost per node-hour as documented in the [pricing guide](/ml-engine/docs/pricing), even if no predictions are performed. There is additional cost for each prediction performed. Unlike manual scaling, if the load gets too heavy for the nodes that are up, the service will automatically add nodes to handle the increased load as well as scale back as traffic drops, always maintaining at least `min_nodes`. You will be charged for the time in which additional nodes are used. If `min_nodes` is not specified and AutoScaling is used with a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 0, in which case, when traffic to a model stops (and after a cool-down period), nodes will be shut down and no charges will be incurred until traffic to the model resumes. If `min_nodes` is not specified and AutoScaling is used with a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a Compute Engine machine type. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use ManualScaling. You can set `min_nodes` when creating the model version, and you can also update `min_nodes` for an existing version: update_body.json: { 'autoScaling': { 'minNodes': 5 } } HTTP request: PATCH https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes -d @./update_body.json
+ },
+ "isDefault": True or False, # Output only. If true, this version will be used to handle prediction requests that do not specify a version. You can change the default version by calling projects.methods.versions.setDefault.
+ "createTime": "A String", # Output only. The time the version was created.
+ "routes": { # RouteMap is used to override HTTP paths sent to a Custom Container. If specified, the HTTP server implemented in the ContainerSpec must support the route. If unspecified, standard HTTP paths will be used.
+ "predict": "A String", # HTTP path to send prediction requests.
+ "health": "A String", # HTTP path to send health check requests.
+ },
+ "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
+ "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only applies to online prediction service. If this field is not specified, it defaults to `mls1-c1-m2`. Online prediction supports the following machine types: * `mls1-c1-m2` * `mls1-c4-m2` * `n1-standard-2` * `n1-standard-4` * `n1-standard-8` * `n1-standard-16` * `n1-standard-32` * `n1-highmem-2` * `n1-highmem-4` * `n1-highmem-8` * `n1-highmem-16` * `n1-highmem-32` * `n1-highcpu-2` * `n1-highcpu-4` * `n1-highcpu-8` * `n1-highcpu-16` * `n1-highcpu-32` `mls1-c1-m2` is generally available. All other machine types are available in beta. Learn more about the [differences between machine types](/ml-engine/docs/machine-types-online-prediction).
+ "acceleratorConfig": { # Represents a hardware accelerator request config. Note that the AcceleratorConfig can be used in both Jobs and Versions. Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and [accelerators for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus). # Optional. Accelerator config for using GPUs for online prediction (beta). Only specify this field if you have specified a Compute Engine (N1) machine type in the `machineType` field. Learn more about [using GPUs for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
"type": "A String", # The type of accelerator to use.
"count": "A String", # The number of accelerators to attach to each machine running the job.
},
- "state": "A String", # Output only. The state of a version.
- "name": "A String", # Required. The name specified for the version when it was created.
- #
- # The version name must be unique within the model it is created in.
- "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in
- # response to increases and decreases in traffic. Care should be
- # taken to ramp up traffic according to the model's ability to scale
- # or you will start seeing increases in latency and 429 response codes.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify
- # `manual_scaling`.
- "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These
- # nodes are always up, starting from the time the model is deployed.
- # Therefore, the cost of operating this model will be at least
- # `rate` * `min_nodes` * number of hours since last billing cycle,
- # where `rate` is the cost per node-hour as documented in the
- # [pricing guide](/ml-engine/docs/pricing),
- # even if no predictions are performed. There is additional cost for each
- # prediction performed.
- #
- # Unlike manual scaling, if the load gets too heavy for the nodes
- # that are up, the service will automatically add nodes to handle the
- # increased load as well as scale back as traffic drops, always maintaining
- # at least `min_nodes`. You will be charged for the time in which additional
- # nodes are used.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [legacy
- # (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 0, in which case, when traffic to a model stops
- # (and after a cool-down period), nodes will be shut down and no charges will
- # be incurred until traffic to the model resumes.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [Compute
- # Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a
- # Compute Engine machine type.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use
- # ManualScaling.
- #
- # You can set `min_nodes` when creating the model version, and you can also
- # update `min_nodes` for an existing version:
- # <pre>
- # update_body.json:
- # {
- # 'autoScaling': {
- # 'minNodes': 5
- # }
- # }
- # </pre>
- # HTTP request:
- # <pre style="max-width: 626px;">
- # PATCH
- # https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes
- # -d @./update_body.json
- # </pre>
- },
- "explanationConfig": { # Message holding configuration options for explaining model predictions. # Optional. Configures explainability features on the model's version.
- # Some explanation features require additional metadata to be loaded
- # as part of the model payload.
- # There are two feature attribution methods supported for TensorFlow models:
- # integrated gradients and sampled Shapley.
- # [Learn more about feature
- # attributions.](/ai-platform/prediction/docs/ai-explanations/overview)
- "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage # Attributes credit by computing the Aumann-Shapley value taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: http://proceedings.mlr.press/v70/sundararajan17a.html
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1703.01365
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help prevent simultaneous updates of a model from overwriting each other. It is strongly suggested that systems make use of the `etag` in the read-modify-write cycle to perform model updates in order to avoid race conditions: An `etag` is returned in the response to `GetVersion`, and systems are expected to put that etag in the request to `UpdateVersion` to ensure that their change will be applied to the model as intended.
+ "explanationConfig": { # Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. [Learn more about feature attributions.](/ai-platform/prediction/docs/ai-explanations/overview) # Optional. Configures explainability features on the model's version. Some explanation features require additional metadata to be loaded as part of the model payload.
+ "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs. # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage # Attributes credit by computing the XRAI taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365 # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that # An attribution method that approximates Shapley values for features that
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- "numPaths": 42, # The number of feature permutations to consider when approximating the
- # Shapley values.
+ "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features. # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
+ "numPaths": 42, # The number of feature permutations to consider when approximating the Shapley values.
},
},
- "pythonVersion": "A String", # Required. The version of Python used in prediction.
- #
- # The following Python versions are available:
- #
- # * Python '3.7' is available when `runtime_version` is set to '1.15' or
- # later.
- # * Python '3.5' is available when `runtime_version` is set to a version
- # from '1.4' to '1.14'.
- # * Python '2.7' is available when `runtime_version` is set to '1.15' or
- # earlier.
- #
- # Read more about the Python versions available for [each runtime
- # version](/ml-engine/docs/runtime-version-list).
- "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. # Optional. *Only* specify this field in a
- # projects.models.versions.patch
- # request. Specifying it in a
- # projects.models.versions.create
- # request has no effect.
- #
- # Configures the request-response pair logging on predictions from this
- # Version.
- # Online prediction requests to a model version and the responses to these
- # requests are converted to raw strings and saved to the specified BigQuery
- # table. Logging is constrained by [BigQuery quotas and
- # limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits,
- # AI Platform Prediction does not log request-response pairs, but it continues
- # to serve predictions.
- #
- # If you are using [continuous
- # evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to
- # specify this configuration manually. Setting up continuous evaluation
- # automatically enables logging of request-response pairs.
- "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1.
- # For example, if you want to log 10% of requests, enter `0.1`. The sampling
- # window is the lifetime of the model version. Defaults to 0.
- "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format:
- # "<var>project_id</var>.<var>dataset_name</var>.<var>table_name</var>"
- #
- # The specified table must already exist, and the "Cloud ML Service Agent"
- # for your project must have permission to write to it. The table must have
- # the following [schema](/bigquery/docs/schemas):
- #
- # <table>
- # <tr><th>Field name</th><th style="display: table-cell">Type</th>
- # <th style="display: table-cell">Mode</th></tr>
- # <tr><td>model</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>model_version</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>time</td><td>TIMESTAMP</td><td>REQUIRED</td></tr>
- # <tr><td>raw_data</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>raw_prediction</td><td>STRING</td><td>NULLABLE</td></tr>
- # <tr><td>groundtruth</td><td>STRING</td><td>NULLABLE</td></tr>
- # </table>
+ "labels": { # Optional. One or more labels that you can add, to organize your model versions. Each label is a key-value pair, where both the key and the value are arbitrary strings that you supply. For more information, see the documentation on using labels.
+ "a_key": "A String",
},
- "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the
- # model. You should generally use `auto_scaling` with an appropriate
- # `min_nodes` instead, but this option is available if you want more
- # predictable billing. Beware that latency and error rates will increase
- # if the traffic exceeds that capability of the system to serve it based
- # on the selected number of nodes.
- "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up,
- # starting from the time the model is deployed, so the cost of operating
- # this model will be proportional to `nodes` * number of hours since
- # last billing cycle plus the cost for each prediction performed.
+ "container": { # Specify a custom container to deploy. Our ContainerSpec is a subset of the Kubernetes Container specification. https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#container-v1-core
+ "ports": [ # Immutable. List of ports to expose from the container. Exposing a port here gives the system additional information about the network connections a container uses, but is primarily informational. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network.
+ { # ContainerPort represents a network port in a single container.
+ "containerPort": 42, # Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
+ },
+ ],
+ "env": [ # Immutable. List of environment variables to set in the container.
+ { # EnvVar represents an environment variable present in a Container.
+ "name": "A String", # Name of the environment variable. Must be a C_IDENTIFIER.
+ "value": "A String", # Variable references $(VAR_NAME) are expanded using the previous defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
+ },
+ ],
+ "command": [ # Immutable. Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
+ "image": "A String", # Docker image name. More info: https://kubernetes.io/docs/concepts/containers/images
+ "args": [ # Immutable. Arguments to the entrypoint. The docker image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
},
- "createTime": "A String", # Output only. The time the version was created.
"lastUseTime": "A String", # Output only. The time the version was last used for prediction.
- "framework": "A String", # Optional. The machine learning framework AI Platform uses to train
- # this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`,
- # `XGBOOST`. If you do not specify a framework, AI Platform
- # will analyze files in the deployment_uri to determine a framework. If you
- # choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version
- # of the model to 1.4 or greater.
- #
- # Do **not** specify a framework if you're deploying a [custom
- # prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines).
- #
- # If you specify a [Compute Engine (N1) machine
- # type](/ml-engine/docs/machine-types-online-prediction) in the
- # `machineType` field, you must specify `TENSORFLOW`
- # for the framework.
- "predictionClass": "A String", # Optional. The fully qualified name
- # (<var>module_name</var>.<var>class_name</var>) of a class that implements
- # the Predictor interface described in this reference field. The module
- # containing this class should be included in a package provided to the
- # [`packageUris` field](#Version.FIELDS.package_uris).
- #
- # Specify this field if and only if you are deploying a [custom prediction
- # routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines).
- # If you specify this field, you must set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and
- # you must set `machineType` to a [legacy (MLS1)
- # machine type](/ml-engine/docs/machine-types-online-prediction).
- #
- # The following code sample provides the Predictor interface:
- #
- # <pre style="max-width: 626px;">
- # class Predictor(object):
- # """Interface for constructing custom predictors."""
- #
- # def predict(self, instances, **kwargs):
- # """Performs custom prediction.
- #
- # Instances are the decoded values from the request. They have already
- # been deserialized from JSON.
- #
- # Args:
- # instances: A list of prediction input instances.
- # **kwargs: A dictionary of keyword args provided as additional
- # fields on the predict request body.
- #
- # Returns:
- # A list of outputs containing the prediction results. This list must
- # be JSON serializable.
- # """
- # raise NotImplementedError()
- #
- # @classmethod
- # def from_path(cls, model_dir):
- # """Creates an instance of Predictor using the given path.
- #
- # Loading of the predictor should be done in this method.
- #
- # Args:
- # model_dir: The local directory that contains the exported model
- # file along with any additional files uploaded when creating the
- # version resource.
- #
- # Returns:
- # An instance implementing this Predictor class.
- # """
- # raise NotImplementedError()
- # </pre>
- #
- # Learn more about [the Predictor interface and custom prediction
- # routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
- "isDefault": True or False, # Output only. If true, this version will be used to handle prediction
- # requests that do not specify a version.
- #
- # You can change the default version by calling
- # projects.methods.versions.setDefault.
- "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help
- # prevent simultaneous updates of a model from overwriting each other.
- # It is strongly suggested that systems make use of the `etag` in the
- # read-modify-write cycle to perform model updates in order to avoid race
- # conditions: An `etag` is returned in the response to `GetVersion`, and
- # systems are expected to put that etag in the request to `UpdateVersion` to
- # ensure that their change will be applied to the model as intended.
"serviceAccount": "A String", # Optional. Specifies the service account for resource access control.
- "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
- "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to
- # create the version. See the
- # [guide to model
- # deployment](/ml-engine/docs/tensorflow/deploying-models) for more
- # information.
- #
- # When passing Version to
- # projects.models.versions.create
- # the model service uses the specified location as the source of the model.
- # Once deployed, the model version is hosted by the prediction service, so
- # this location is useful only as a historical record.
- # The total number of model files can't exceed 1000.
- "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment.
- #
- # For more information, see the
- # [runtime version list](/ml-engine/docs/runtime-version-list) and
- # [how to manage runtime versions](/ml-engine/docs/versioning).
- "description": "A String", # Optional. The description specified for the version when it was created.
+ "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment. For more information, see the [runtime version list](/ml-engine/docs/runtime-version-list) and [how to manage runtime versions](/ml-engine/docs/versioning).
+ "name": "A String", # Required. The name specified for the version when it was created. The version name must be unique within the model it is created in.
+ "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the model. You should generally use `auto_scaling` with an appropriate `min_nodes` instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
+ "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed, so the cost of operating this model will be proportional to `nodes` * number of hours since last billing cycle plus the cost for each prediction performed.
+ },
+ "state": "A String", # Output only. The state of a version.
+ "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines) or [scikit-learn pipelines with custom code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code). For a custom prediction routine, one of these packages must contain your Predictor class (see [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally, include any dependencies used by your Predictor or scikit-learn pipeline uses that are not already included in your selected [runtime version](/ml-engine/docs/tensorflow/runtime-version-list). If you specify this field, you must also set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
+ "A String",
+ ],
+ "pythonVersion": "A String", # Required. The version of Python used in prediction. The following Python versions are available: * Python '3.7' is available when `runtime_version` is set to '1.15' or later. * Python '3.5' is available when `runtime_version` is set to a version from '1.4' to '1.14'. * Python '2.7' is available when `runtime_version` is set to '1.15' or earlier. Read more about the Python versions available for [each runtime version](/ml-engine/docs/runtime-version-list).
}
x__xgafv: string, V1 error format.
@@ -451,65 +194,33 @@
Returns:
An object of the form:
- { # This resource represents a long-running operation that is the result of a
- # network API call.
- "error": { # The `Status` type defines a logical error model that is suitable for # The error result of the operation in case of failure or cancellation.
- # different programming environments, including REST APIs and RPC APIs. It is
- # used by [gRPC](https://github.com/grpc). Each `Status` message contains
- # three pieces of data: error code, error message, and error details.
- #
- # You can find out more about this error model and how to work with it in the
- # [API Design Guide](https://cloud.google.com/apis/design/errors).
- "details": [ # A list of messages that carry the error details. There is a common set of
- # message types for APIs to use.
+ { # This resource represents a long-running operation that is the result of a network API call.
+ "name": "A String", # The server-assigned name, which is only unique within the same service that originally returns it. If you use the default HTTP mapping, the `name` should be a resource name ending with `operations/{unique_id}`.
+ "response": { # The normal response of the operation in case of success. If the original method returns no data on success, such as `Delete`, the response is `google.protobuf.Empty`. If the original method is standard `Get`/`Create`/`Update`, the response should be the resource. For other methods, the response should have the type `XxxResponse`, where `Xxx` is the original method name. For example, if the original method name is `TakeSnapshot()`, the inferred response type is `TakeSnapshotResponse`.
+ "a_key": "", # Properties of the object. Contains field @type with type URL.
+ },
+ "done": True or False, # If the value is `false`, it means the operation is still in progress. If `true`, the operation is completed, and either `error` or `response` is available.
+ "error": { # The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors). # The error result of the operation in case of failure or cancellation.
+ "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "details": [ # A list of messages that carry the error details. There is a common set of message types for APIs to use.
{
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
],
- "message": "A String", # A developer-facing error message, which should be in English. Any
- # user-facing error message should be localized and sent in the
- # google.rpc.Status.details field, or localized by the client.
- "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "message": "A String", # A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the google.rpc.Status.details field, or localized by the client.
},
- "done": True or False, # If the value is `false`, it means the operation is still in progress.
- # If `true`, the operation is completed, and either `error` or `response` is
- # available.
- "response": { # The normal response of the operation in case of success. If the original
- # method returns no data on success, such as `Delete`, the response is
- # `google.protobuf.Empty`. If the original method is standard
- # `Get`/`Create`/`Update`, the response should be the resource. For other
- # methods, the response should have the type `XxxResponse`, where `Xxx`
- # is the original method name. For example, if the original method name
- # is `TakeSnapshot()`, the inferred response type is
- # `TakeSnapshotResponse`.
+ "metadata": { # Service-specific metadata associated with the operation. It typically contains progress information and common metadata such as create time. Some services might not provide such metadata. Any method that returns a long-running operation should document the metadata type, if any.
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
- "metadata": { # Service-specific metadata associated with the operation. It typically
- # contains progress information and common metadata such as create time.
- # Some services might not provide such metadata. Any method that returns a
- # long-running operation should document the metadata type, if any.
- "a_key": "", # Properties of the object. Contains field @type with type URL.
- },
- "name": "A String", # The server-assigned name, which is only unique within the same service that
- # originally returns it. If you use the default HTTP mapping, the
- # `name` should be a resource name ending with `operations/{unique_id}`.
}</pre>
</div>
<div class="method">
<code class="details" id="delete">delete(name, x__xgafv=None)</code>
- <pre>Deletes a model version.
-
-Each model can have multiple versions deployed and in use at any given
-time. Use this method to remove a single version.
-
-Note: You cannot delete the version that is set as the default version
-of the model unless it is the only remaining version.
+ <pre>Deletes a model version. Each model can have multiple versions deployed and in use at any given time. Use this method to remove a single version. Note: You cannot delete the version that is set as the default version of the model unless it is the only remaining version.
Args:
- name: string, Required. The name of the version. You can get the names of all the
-versions of a model by calling
-projects.models.versions.list. (required)
+ name: string, Required. The name of the version. You can get the names of all the versions of a model by calling projects.models.versions.list. (required)
x__xgafv: string, V1 error format.
Allowed values
1 - v1 error format
@@ -518,59 +229,30 @@
Returns:
An object of the form:
- { # This resource represents a long-running operation that is the result of a
- # network API call.
- "error": { # The `Status` type defines a logical error model that is suitable for # The error result of the operation in case of failure or cancellation.
- # different programming environments, including REST APIs and RPC APIs. It is
- # used by [gRPC](https://github.com/grpc). Each `Status` message contains
- # three pieces of data: error code, error message, and error details.
- #
- # You can find out more about this error model and how to work with it in the
- # [API Design Guide](https://cloud.google.com/apis/design/errors).
- "details": [ # A list of messages that carry the error details. There is a common set of
- # message types for APIs to use.
+ { # This resource represents a long-running operation that is the result of a network API call.
+ "name": "A String", # The server-assigned name, which is only unique within the same service that originally returns it. If you use the default HTTP mapping, the `name` should be a resource name ending with `operations/{unique_id}`.
+ "response": { # The normal response of the operation in case of success. If the original method returns no data on success, such as `Delete`, the response is `google.protobuf.Empty`. If the original method is standard `Get`/`Create`/`Update`, the response should be the resource. For other methods, the response should have the type `XxxResponse`, where `Xxx` is the original method name. For example, if the original method name is `TakeSnapshot()`, the inferred response type is `TakeSnapshotResponse`.
+ "a_key": "", # Properties of the object. Contains field @type with type URL.
+ },
+ "done": True or False, # If the value is `false`, it means the operation is still in progress. If `true`, the operation is completed, and either `error` or `response` is available.
+ "error": { # The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors). # The error result of the operation in case of failure or cancellation.
+ "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "details": [ # A list of messages that carry the error details. There is a common set of message types for APIs to use.
{
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
],
- "message": "A String", # A developer-facing error message, which should be in English. Any
- # user-facing error message should be localized and sent in the
- # google.rpc.Status.details field, or localized by the client.
- "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "message": "A String", # A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the google.rpc.Status.details field, or localized by the client.
},
- "done": True or False, # If the value is `false`, it means the operation is still in progress.
- # If `true`, the operation is completed, and either `error` or `response` is
- # available.
- "response": { # The normal response of the operation in case of success. If the original
- # method returns no data on success, such as `Delete`, the response is
- # `google.protobuf.Empty`. If the original method is standard
- # `Get`/`Create`/`Update`, the response should be the resource. For other
- # methods, the response should have the type `XxxResponse`, where `Xxx`
- # is the original method name. For example, if the original method name
- # is `TakeSnapshot()`, the inferred response type is
- # `TakeSnapshotResponse`.
+ "metadata": { # Service-specific metadata associated with the operation. It typically contains progress information and common metadata such as create time. Some services might not provide such metadata. Any method that returns a long-running operation should document the metadata type, if any.
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
- "metadata": { # Service-specific metadata associated with the operation. It typically
- # contains progress information and common metadata such as create time.
- # Some services might not provide such metadata. Any method that returns a
- # long-running operation should document the metadata type, if any.
- "a_key": "", # Properties of the object. Contains field @type with type URL.
- },
- "name": "A String", # The server-assigned name, which is only unique within the same service that
- # originally returns it. If you use the default HTTP mapping, the
- # `name` should be a resource name ending with `operations/{unique_id}`.
}</pre>
</div>
<div class="method">
<code class="details" id="get">get(name, x__xgafv=None)</code>
- <pre>Gets information about a model version.
-
-Models can have multiple versions. You can call
-projects.models.versions.list
-to get the same information that this method returns for all of the
-versions of a model.
+ <pre>Gets information about a model version. Models can have multiple versions. You can call projects.models.versions.list to get the same information that this method returns for all of the versions of a model.
Args:
name: string, Required. The name of the version. (required)
@@ -582,361 +264,89 @@
Returns:
An object of the form:
- { # Represents a version of the model.
- #
- # Each version is a trained model deployed in the cloud, ready to handle
- # prediction requests. A model can have multiple versions. You can get
- # information about all of the versions of a given model by calling
- # projects.models.versions.list.
- "labels": { # Optional. One or more labels that you can add, to organize your model
- # versions. Each label is a key-value pair, where both the key and the value
- # are arbitrary strings that you supply.
- # For more information, see the documentation on
- # <a href="/ml-engine/docs/tensorflow/resource-labels">using labels</a>.
- "a_key": "A String",
+ { # Represents a version of the model. Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions. You can get information about all of the versions of a given model by calling projects.models.versions.list.
+ "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to create the version. See the [guide to model deployment](/ml-engine/docs/tensorflow/deploying-models) for more information. When passing Version to projects.models.versions.create the model service uses the specified location as the source of the model. Once deployed, the model version is hosted by the prediction service, so this location is useful only as a historical record. The total number of model files can't exceed 1000.
+ "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. Online prediction requests to a model version and the responses to these requests are converted to raw strings and saved to the specified BigQuery table. Logging is constrained by [BigQuery quotas and limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits, AI Platform Prediction does not log request-response pairs, but it continues to serve predictions. If you are using [continuous evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to specify this configuration manually. Setting up continuous evaluation automatically enables logging of request-response pairs. # Optional. *Only* specify this field in a projects.models.versions.patch request. Specifying it in a projects.models.versions.create request has no effect. Configures the request-response pair logging on predictions from this Version.
+ "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format: " project_id.dataset_name.table_name" The specified table must already exist, and the "Cloud ML Service Agent" for your project must have permission to write to it. The table must have the following [schema](/bigquery/docs/schemas): Field nameType Mode model STRING REQUIRED model_version STRING REQUIRED time TIMESTAMP REQUIRED raw_data STRING REQUIRED raw_prediction STRING NULLABLE groundtruth STRING NULLABLE
+ "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1. For example, if you want to log 10% of requests, enter `0.1`. The sampling window is the lifetime of the model version. Defaults to 0.
},
- "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only
- # applies to online prediction service. If this field is not specified, it
- # defaults to `mls1-c1-m2`.
- #
- # Online prediction supports the following machine types:
- #
- # * `mls1-c1-m2`
- # * `mls1-c4-m2`
- # * `n1-standard-2`
- # * `n1-standard-4`
- # * `n1-standard-8`
- # * `n1-standard-16`
- # * `n1-standard-32`
- # * `n1-highmem-2`
- # * `n1-highmem-4`
- # * `n1-highmem-8`
- # * `n1-highmem-16`
- # * `n1-highmem-32`
- # * `n1-highcpu-2`
- # * `n1-highcpu-4`
- # * `n1-highcpu-8`
- # * `n1-highcpu-16`
- # * `n1-highcpu-32`
- #
- # `mls1-c1-m2` is generally available. All other machine types are available
- # in beta. Learn more about the [differences between machine
- # types](/ml-engine/docs/machine-types-online-prediction).
- "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom
- # prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines)
- # or [scikit-learn pipelines with custom
- # code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code).
- #
- # For a custom prediction routine, one of these packages must contain your
- # Predictor class (see
- # [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally,
- # include any dependencies used by your Predictor or scikit-learn pipeline
- # uses that are not already included in your selected [runtime
- # version](/ml-engine/docs/tensorflow/runtime-version-list).
- #
- # If you specify this field, you must also set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
- "A String",
- ],
- "acceleratorConfig": { # Represents a hardware accelerator request config. # Optional. Accelerator config for using GPUs for online prediction (beta).
- # Only specify this field if you have specified a Compute Engine (N1) machine
- # type in the `machineType` field. Learn more about [using GPUs for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
- # Note that the AcceleratorConfig can be used in both Jobs and Versions.
- # Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and
- # [accelerators for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
+ "predictionClass": "A String", # Optional. The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the [`packageUris` field](#Version.FIELDS.package_uris). Specify this field if and only if you are deploying a [custom prediction routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify this field, you must set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and you must set `machineType` to a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction). The following code sample provides the Predictor interface: class Predictor(object): """Interface for constructing custom predictors.""" def predict(self, instances, **kwargs): """Performs custom prediction. Instances are the decoded values from the request. They have already been deserialized from JSON. Args: instances: A list of prediction input instances. **kwargs: A dictionary of keyword args provided as additional fields on the predict request body. Returns: A list of outputs containing the prediction results. This list must be JSON serializable. """ raise NotImplementedError() @classmethod def from_path(cls, model_dir): """Creates an instance of Predictor using the given path. Loading of the predictor should be done in this method. Args: model_dir: The local directory that contains the exported model file along with any additional files uploaded when creating the version resource. Returns: An instance implementing this Predictor class. """ raise NotImplementedError() Learn more about [the Predictor interface and custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
+ "framework": "A String", # Optional. The machine learning framework AI Platform uses to train this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`, `XGBOOST`. If you do not specify a framework, AI Platform will analyze files in the deployment_uri to determine a framework. If you choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version of the model to 1.4 or greater. Do **not** specify a framework if you're deploying a [custom prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction) in the `machineType` field, you must specify `TENSORFLOW` for the framework.
+ "description": "A String", # Optional. The description specified for the version when it was created.
+ "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify `manual_scaling`.
+ "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed. Therefore, the cost of operating this model will be at least `rate` * `min_nodes` * number of hours since last billing cycle, where `rate` is the cost per node-hour as documented in the [pricing guide](/ml-engine/docs/pricing), even if no predictions are performed. There is additional cost for each prediction performed. Unlike manual scaling, if the load gets too heavy for the nodes that are up, the service will automatically add nodes to handle the increased load as well as scale back as traffic drops, always maintaining at least `min_nodes`. You will be charged for the time in which additional nodes are used. If `min_nodes` is not specified and AutoScaling is used with a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 0, in which case, when traffic to a model stops (and after a cool-down period), nodes will be shut down and no charges will be incurred until traffic to the model resumes. If `min_nodes` is not specified and AutoScaling is used with a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a Compute Engine machine type. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use ManualScaling. You can set `min_nodes` when creating the model version, and you can also update `min_nodes` for an existing version: update_body.json: { 'autoScaling': { 'minNodes': 5 } } HTTP request: PATCH https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes -d @./update_body.json
+ },
+ "isDefault": True or False, # Output only. If true, this version will be used to handle prediction requests that do not specify a version. You can change the default version by calling projects.methods.versions.setDefault.
+ "createTime": "A String", # Output only. The time the version was created.
+ "routes": { # RouteMap is used to override HTTP paths sent to a Custom Container. If specified, the HTTP server implemented in the ContainerSpec must support the route. If unspecified, standard HTTP paths will be used.
+ "predict": "A String", # HTTP path to send prediction requests.
+ "health": "A String", # HTTP path to send health check requests.
+ },
+ "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
+ "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only applies to online prediction service. If this field is not specified, it defaults to `mls1-c1-m2`. Online prediction supports the following machine types: * `mls1-c1-m2` * `mls1-c4-m2` * `n1-standard-2` * `n1-standard-4` * `n1-standard-8` * `n1-standard-16` * `n1-standard-32` * `n1-highmem-2` * `n1-highmem-4` * `n1-highmem-8` * `n1-highmem-16` * `n1-highmem-32` * `n1-highcpu-2` * `n1-highcpu-4` * `n1-highcpu-8` * `n1-highcpu-16` * `n1-highcpu-32` `mls1-c1-m2` is generally available. All other machine types are available in beta. Learn more about the [differences between machine types](/ml-engine/docs/machine-types-online-prediction).
+ "acceleratorConfig": { # Represents a hardware accelerator request config. Note that the AcceleratorConfig can be used in both Jobs and Versions. Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and [accelerators for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus). # Optional. Accelerator config for using GPUs for online prediction (beta). Only specify this field if you have specified a Compute Engine (N1) machine type in the `machineType` field. Learn more about [using GPUs for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
"type": "A String", # The type of accelerator to use.
"count": "A String", # The number of accelerators to attach to each machine running the job.
},
- "state": "A String", # Output only. The state of a version.
- "name": "A String", # Required. The name specified for the version when it was created.
- #
- # The version name must be unique within the model it is created in.
- "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in
- # response to increases and decreases in traffic. Care should be
- # taken to ramp up traffic according to the model's ability to scale
- # or you will start seeing increases in latency and 429 response codes.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify
- # `manual_scaling`.
- "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These
- # nodes are always up, starting from the time the model is deployed.
- # Therefore, the cost of operating this model will be at least
- # `rate` * `min_nodes` * number of hours since last billing cycle,
- # where `rate` is the cost per node-hour as documented in the
- # [pricing guide](/ml-engine/docs/pricing),
- # even if no predictions are performed. There is additional cost for each
- # prediction performed.
- #
- # Unlike manual scaling, if the load gets too heavy for the nodes
- # that are up, the service will automatically add nodes to handle the
- # increased load as well as scale back as traffic drops, always maintaining
- # at least `min_nodes`. You will be charged for the time in which additional
- # nodes are used.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [legacy
- # (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 0, in which case, when traffic to a model stops
- # (and after a cool-down period), nodes will be shut down and no charges will
- # be incurred until traffic to the model resumes.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [Compute
- # Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a
- # Compute Engine machine type.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use
- # ManualScaling.
- #
- # You can set `min_nodes` when creating the model version, and you can also
- # update `min_nodes` for an existing version:
- # <pre>
- # update_body.json:
- # {
- # 'autoScaling': {
- # 'minNodes': 5
- # }
- # }
- # </pre>
- # HTTP request:
- # <pre style="max-width: 626px;">
- # PATCH
- # https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes
- # -d @./update_body.json
- # </pre>
- },
- "explanationConfig": { # Message holding configuration options for explaining model predictions. # Optional. Configures explainability features on the model's version.
- # Some explanation features require additional metadata to be loaded
- # as part of the model payload.
- # There are two feature attribution methods supported for TensorFlow models:
- # integrated gradients and sampled Shapley.
- # [Learn more about feature
- # attributions.](/ai-platform/prediction/docs/ai-explanations/overview)
- "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage # Attributes credit by computing the Aumann-Shapley value taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: http://proceedings.mlr.press/v70/sundararajan17a.html
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1703.01365
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help prevent simultaneous updates of a model from overwriting each other. It is strongly suggested that systems make use of the `etag` in the read-modify-write cycle to perform model updates in order to avoid race conditions: An `etag` is returned in the response to `GetVersion`, and systems are expected to put that etag in the request to `UpdateVersion` to ensure that their change will be applied to the model as intended.
+ "explanationConfig": { # Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. [Learn more about feature attributions.](/ai-platform/prediction/docs/ai-explanations/overview) # Optional. Configures explainability features on the model's version. Some explanation features require additional metadata to be loaded as part of the model payload.
+ "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs. # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage # Attributes credit by computing the XRAI taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365 # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that # An attribution method that approximates Shapley values for features that
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- "numPaths": 42, # The number of feature permutations to consider when approximating the
- # Shapley values.
+ "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features. # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
+ "numPaths": 42, # The number of feature permutations to consider when approximating the Shapley values.
},
},
- "pythonVersion": "A String", # Required. The version of Python used in prediction.
- #
- # The following Python versions are available:
- #
- # * Python '3.7' is available when `runtime_version` is set to '1.15' or
- # later.
- # * Python '3.5' is available when `runtime_version` is set to a version
- # from '1.4' to '1.14'.
- # * Python '2.7' is available when `runtime_version` is set to '1.15' or
- # earlier.
- #
- # Read more about the Python versions available for [each runtime
- # version](/ml-engine/docs/runtime-version-list).
- "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. # Optional. *Only* specify this field in a
- # projects.models.versions.patch
- # request. Specifying it in a
- # projects.models.versions.create
- # request has no effect.
- #
- # Configures the request-response pair logging on predictions from this
- # Version.
- # Online prediction requests to a model version and the responses to these
- # requests are converted to raw strings and saved to the specified BigQuery
- # table. Logging is constrained by [BigQuery quotas and
- # limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits,
- # AI Platform Prediction does not log request-response pairs, but it continues
- # to serve predictions.
- #
- # If you are using [continuous
- # evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to
- # specify this configuration manually. Setting up continuous evaluation
- # automatically enables logging of request-response pairs.
- "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1.
- # For example, if you want to log 10% of requests, enter `0.1`. The sampling
- # window is the lifetime of the model version. Defaults to 0.
- "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format:
- # "<var>project_id</var>.<var>dataset_name</var>.<var>table_name</var>"
- #
- # The specified table must already exist, and the "Cloud ML Service Agent"
- # for your project must have permission to write to it. The table must have
- # the following [schema](/bigquery/docs/schemas):
- #
- # <table>
- # <tr><th>Field name</th><th style="display: table-cell">Type</th>
- # <th style="display: table-cell">Mode</th></tr>
- # <tr><td>model</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>model_version</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>time</td><td>TIMESTAMP</td><td>REQUIRED</td></tr>
- # <tr><td>raw_data</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>raw_prediction</td><td>STRING</td><td>NULLABLE</td></tr>
- # <tr><td>groundtruth</td><td>STRING</td><td>NULLABLE</td></tr>
- # </table>
+ "labels": { # Optional. One or more labels that you can add, to organize your model versions. Each label is a key-value pair, where both the key and the value are arbitrary strings that you supply. For more information, see the documentation on using labels.
+ "a_key": "A String",
},
- "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the
- # model. You should generally use `auto_scaling` with an appropriate
- # `min_nodes` instead, but this option is available if you want more
- # predictable billing. Beware that latency and error rates will increase
- # if the traffic exceeds that capability of the system to serve it based
- # on the selected number of nodes.
- "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up,
- # starting from the time the model is deployed, so the cost of operating
- # this model will be proportional to `nodes` * number of hours since
- # last billing cycle plus the cost for each prediction performed.
+ "container": { # Specify a custom container to deploy. Our ContainerSpec is a subset of the Kubernetes Container specification. https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#container-v1-core
+ "ports": [ # Immutable. List of ports to expose from the container. Exposing a port here gives the system additional information about the network connections a container uses, but is primarily informational. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network.
+ { # ContainerPort represents a network port in a single container.
+ "containerPort": 42, # Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
+ },
+ ],
+ "env": [ # Immutable. List of environment variables to set in the container.
+ { # EnvVar represents an environment variable present in a Container.
+ "name": "A String", # Name of the environment variable. Must be a C_IDENTIFIER.
+ "value": "A String", # Variable references $(VAR_NAME) are expanded using the previous defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
+ },
+ ],
+ "command": [ # Immutable. Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
+ "image": "A String", # Docker image name. More info: https://kubernetes.io/docs/concepts/containers/images
+ "args": [ # Immutable. Arguments to the entrypoint. The docker image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
},
- "createTime": "A String", # Output only. The time the version was created.
"lastUseTime": "A String", # Output only. The time the version was last used for prediction.
- "framework": "A String", # Optional. The machine learning framework AI Platform uses to train
- # this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`,
- # `XGBOOST`. If you do not specify a framework, AI Platform
- # will analyze files in the deployment_uri to determine a framework. If you
- # choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version
- # of the model to 1.4 or greater.
- #
- # Do **not** specify a framework if you're deploying a [custom
- # prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines).
- #
- # If you specify a [Compute Engine (N1) machine
- # type](/ml-engine/docs/machine-types-online-prediction) in the
- # `machineType` field, you must specify `TENSORFLOW`
- # for the framework.
- "predictionClass": "A String", # Optional. The fully qualified name
- # (<var>module_name</var>.<var>class_name</var>) of a class that implements
- # the Predictor interface described in this reference field. The module
- # containing this class should be included in a package provided to the
- # [`packageUris` field](#Version.FIELDS.package_uris).
- #
- # Specify this field if and only if you are deploying a [custom prediction
- # routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines).
- # If you specify this field, you must set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and
- # you must set `machineType` to a [legacy (MLS1)
- # machine type](/ml-engine/docs/machine-types-online-prediction).
- #
- # The following code sample provides the Predictor interface:
- #
- # <pre style="max-width: 626px;">
- # class Predictor(object):
- # """Interface for constructing custom predictors."""
- #
- # def predict(self, instances, **kwargs):
- # """Performs custom prediction.
- #
- # Instances are the decoded values from the request. They have already
- # been deserialized from JSON.
- #
- # Args:
- # instances: A list of prediction input instances.
- # **kwargs: A dictionary of keyword args provided as additional
- # fields on the predict request body.
- #
- # Returns:
- # A list of outputs containing the prediction results. This list must
- # be JSON serializable.
- # """
- # raise NotImplementedError()
- #
- # @classmethod
- # def from_path(cls, model_dir):
- # """Creates an instance of Predictor using the given path.
- #
- # Loading of the predictor should be done in this method.
- #
- # Args:
- # model_dir: The local directory that contains the exported model
- # file along with any additional files uploaded when creating the
- # version resource.
- #
- # Returns:
- # An instance implementing this Predictor class.
- # """
- # raise NotImplementedError()
- # </pre>
- #
- # Learn more about [the Predictor interface and custom prediction
- # routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
- "isDefault": True or False, # Output only. If true, this version will be used to handle prediction
- # requests that do not specify a version.
- #
- # You can change the default version by calling
- # projects.methods.versions.setDefault.
- "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help
- # prevent simultaneous updates of a model from overwriting each other.
- # It is strongly suggested that systems make use of the `etag` in the
- # read-modify-write cycle to perform model updates in order to avoid race
- # conditions: An `etag` is returned in the response to `GetVersion`, and
- # systems are expected to put that etag in the request to `UpdateVersion` to
- # ensure that their change will be applied to the model as intended.
"serviceAccount": "A String", # Optional. Specifies the service account for resource access control.
- "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
- "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to
- # create the version. See the
- # [guide to model
- # deployment](/ml-engine/docs/tensorflow/deploying-models) for more
- # information.
- #
- # When passing Version to
- # projects.models.versions.create
- # the model service uses the specified location as the source of the model.
- # Once deployed, the model version is hosted by the prediction service, so
- # this location is useful only as a historical record.
- # The total number of model files can't exceed 1000.
- "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment.
- #
- # For more information, see the
- # [runtime version list](/ml-engine/docs/runtime-version-list) and
- # [how to manage runtime versions](/ml-engine/docs/versioning).
- "description": "A String", # Optional. The description specified for the version when it was created.
+ "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment. For more information, see the [runtime version list](/ml-engine/docs/runtime-version-list) and [how to manage runtime versions](/ml-engine/docs/versioning).
+ "name": "A String", # Required. The name specified for the version when it was created. The version name must be unique within the model it is created in.
+ "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the model. You should generally use `auto_scaling` with an appropriate `min_nodes` instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
+ "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed, so the cost of operating this model will be proportional to `nodes` * number of hours since last billing cycle plus the cost for each prediction performed.
+ },
+ "state": "A String", # Output only. The state of a version.
+ "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines) or [scikit-learn pipelines with custom code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code). For a custom prediction routine, one of these packages must contain your Predictor class (see [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally, include any dependencies used by your Predictor or scikit-learn pipeline uses that are not already included in your selected [runtime version](/ml-engine/docs/tensorflow/runtime-version-list). If you specify this field, you must also set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
+ "A String",
+ ],
+ "pythonVersion": "A String", # Required. The version of Python used in prediction. The following Python versions are available: * Python '3.7' is available when `runtime_version` is set to '1.15' or later. * Python '3.5' is available when `runtime_version` is set to a version from '1.4' to '1.14'. * Python '2.7' is available when `runtime_version` is set to '1.15' or earlier. Read more about the Python versions available for [each runtime version](/ml-engine/docs/runtime-version-list).
}</pre>
</div>
<div class="method">
- <code class="details" id="list">list(parent, filter=None, pageToken=None, pageSize=None, x__xgafv=None)</code>
- <pre>Gets basic information about all the versions of a model.
-
-If you expect that a model has many versions, or if you need to handle
-only a limited number of results at a time, you can request that the list
-be retrieved in batches (called pages).
-
-If there are no versions that match the request parameters, the list
-request returns an empty response body: {}.
+ <code class="details" id="list">list(parent, filter=None, pageSize=None, pageToken=None, x__xgafv=None)</code>
+ <pre>Gets basic information about all the versions of a model. If you expect that a model has many versions, or if you need to handle only a limited number of results at a time, you can request that the list be retrieved in batches (called pages). If there are no versions that match the request parameters, the list request returns an empty response body: {}.
Args:
parent: string, Required. The name of the model for which to list the version. (required)
filter: string, Optional. Specifies the subset of versions to retrieve.
- pageToken: string, Optional. A page token to request the next page of results.
-
-You get the token from the `next_page_token` field of the response from
-the previous call.
- pageSize: integer, Optional. The number of versions to retrieve per "page" of results. If
-there are more remaining results than this number, the response message
-will contain a valid value in the `next_page_token` field.
-
-The default value is 20, and the maximum page size is 100.
+ pageSize: integer, Optional. The number of versions to retrieve per "page" of results. If there are more remaining results than this number, the response message will contain a valid value in the `next_page_token` field. The default value is 20, and the maximum page size is 100.
+ pageToken: string, Optional. A page token to request the next page of results. You get the token from the `next_page_token` field of the response from the previous call.
x__xgafv: string, V1 error format.
Allowed values
1 - v1 error format
@@ -946,340 +356,81 @@
An object of the form:
{ # Response message for the ListVersions method.
- "nextPageToken": "A String", # Optional. Pass this token as the `page_token` field of the request for a
- # subsequent call.
"versions": [ # The list of versions.
- { # Represents a version of the model.
- #
- # Each version is a trained model deployed in the cloud, ready to handle
- # prediction requests. A model can have multiple versions. You can get
- # information about all of the versions of a given model by calling
- # projects.models.versions.list.
- "labels": { # Optional. One or more labels that you can add, to organize your model
- # versions. Each label is a key-value pair, where both the key and the value
- # are arbitrary strings that you supply.
- # For more information, see the documentation on
- # <a href="/ml-engine/docs/tensorflow/resource-labels">using labels</a>.
- "a_key": "A String",
+ { # Represents a version of the model. Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions. You can get information about all of the versions of a given model by calling projects.models.versions.list.
+ "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to create the version. See the [guide to model deployment](/ml-engine/docs/tensorflow/deploying-models) for more information. When passing Version to projects.models.versions.create the model service uses the specified location as the source of the model. Once deployed, the model version is hosted by the prediction service, so this location is useful only as a historical record. The total number of model files can't exceed 1000.
+ "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. Online prediction requests to a model version and the responses to these requests are converted to raw strings and saved to the specified BigQuery table. Logging is constrained by [BigQuery quotas and limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits, AI Platform Prediction does not log request-response pairs, but it continues to serve predictions. If you are using [continuous evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to specify this configuration manually. Setting up continuous evaluation automatically enables logging of request-response pairs. # Optional. *Only* specify this field in a projects.models.versions.patch request. Specifying it in a projects.models.versions.create request has no effect. Configures the request-response pair logging on predictions from this Version.
+ "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format: " project_id.dataset_name.table_name" The specified table must already exist, and the "Cloud ML Service Agent" for your project must have permission to write to it. The table must have the following [schema](/bigquery/docs/schemas): Field nameType Mode model STRING REQUIRED model_version STRING REQUIRED time TIMESTAMP REQUIRED raw_data STRING REQUIRED raw_prediction STRING NULLABLE groundtruth STRING NULLABLE
+ "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1. For example, if you want to log 10% of requests, enter `0.1`. The sampling window is the lifetime of the model version. Defaults to 0.
},
- "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only
- # applies to online prediction service. If this field is not specified, it
- # defaults to `mls1-c1-m2`.
- #
- # Online prediction supports the following machine types:
- #
- # * `mls1-c1-m2`
- # * `mls1-c4-m2`
- # * `n1-standard-2`
- # * `n1-standard-4`
- # * `n1-standard-8`
- # * `n1-standard-16`
- # * `n1-standard-32`
- # * `n1-highmem-2`
- # * `n1-highmem-4`
- # * `n1-highmem-8`
- # * `n1-highmem-16`
- # * `n1-highmem-32`
- # * `n1-highcpu-2`
- # * `n1-highcpu-4`
- # * `n1-highcpu-8`
- # * `n1-highcpu-16`
- # * `n1-highcpu-32`
- #
- # `mls1-c1-m2` is generally available. All other machine types are available
- # in beta. Learn more about the [differences between machine
- # types](/ml-engine/docs/machine-types-online-prediction).
- "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom
- # prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines)
- # or [scikit-learn pipelines with custom
- # code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code).
- #
- # For a custom prediction routine, one of these packages must contain your
- # Predictor class (see
- # [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally,
- # include any dependencies used by your Predictor or scikit-learn pipeline
- # uses that are not already included in your selected [runtime
- # version](/ml-engine/docs/tensorflow/runtime-version-list).
- #
- # If you specify this field, you must also set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
- "A String",
- ],
- "acceleratorConfig": { # Represents a hardware accelerator request config. # Optional. Accelerator config for using GPUs for online prediction (beta).
- # Only specify this field if you have specified a Compute Engine (N1) machine
- # type in the `machineType` field. Learn more about [using GPUs for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
- # Note that the AcceleratorConfig can be used in both Jobs and Versions.
- # Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and
- # [accelerators for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
+ "predictionClass": "A String", # Optional. The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the [`packageUris` field](#Version.FIELDS.package_uris). Specify this field if and only if you are deploying a [custom prediction routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify this field, you must set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and you must set `machineType` to a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction). The following code sample provides the Predictor interface: class Predictor(object): """Interface for constructing custom predictors.""" def predict(self, instances, **kwargs): """Performs custom prediction. Instances are the decoded values from the request. They have already been deserialized from JSON. Args: instances: A list of prediction input instances. **kwargs: A dictionary of keyword args provided as additional fields on the predict request body. Returns: A list of outputs containing the prediction results. This list must be JSON serializable. """ raise NotImplementedError() @classmethod def from_path(cls, model_dir): """Creates an instance of Predictor using the given path. Loading of the predictor should be done in this method. Args: model_dir: The local directory that contains the exported model file along with any additional files uploaded when creating the version resource. Returns: An instance implementing this Predictor class. """ raise NotImplementedError() Learn more about [the Predictor interface and custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
+ "framework": "A String", # Optional. The machine learning framework AI Platform uses to train this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`, `XGBOOST`. If you do not specify a framework, AI Platform will analyze files in the deployment_uri to determine a framework. If you choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version of the model to 1.4 or greater. Do **not** specify a framework if you're deploying a [custom prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction) in the `machineType` field, you must specify `TENSORFLOW` for the framework.
+ "description": "A String", # Optional. The description specified for the version when it was created.
+ "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify `manual_scaling`.
+ "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed. Therefore, the cost of operating this model will be at least `rate` * `min_nodes` * number of hours since last billing cycle, where `rate` is the cost per node-hour as documented in the [pricing guide](/ml-engine/docs/pricing), even if no predictions are performed. There is additional cost for each prediction performed. Unlike manual scaling, if the load gets too heavy for the nodes that are up, the service will automatically add nodes to handle the increased load as well as scale back as traffic drops, always maintaining at least `min_nodes`. You will be charged for the time in which additional nodes are used. If `min_nodes` is not specified and AutoScaling is used with a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 0, in which case, when traffic to a model stops (and after a cool-down period), nodes will be shut down and no charges will be incurred until traffic to the model resumes. If `min_nodes` is not specified and AutoScaling is used with a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a Compute Engine machine type. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use ManualScaling. You can set `min_nodes` when creating the model version, and you can also update `min_nodes` for an existing version: update_body.json: { 'autoScaling': { 'minNodes': 5 } } HTTP request: PATCH https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes -d @./update_body.json
+ },
+ "isDefault": True or False, # Output only. If true, this version will be used to handle prediction requests that do not specify a version. You can change the default version by calling projects.methods.versions.setDefault.
+ "createTime": "A String", # Output only. The time the version was created.
+ "routes": { # RouteMap is used to override HTTP paths sent to a Custom Container. If specified, the HTTP server implemented in the ContainerSpec must support the route. If unspecified, standard HTTP paths will be used.
+ "predict": "A String", # HTTP path to send prediction requests.
+ "health": "A String", # HTTP path to send health check requests.
+ },
+ "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
+ "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only applies to online prediction service. If this field is not specified, it defaults to `mls1-c1-m2`. Online prediction supports the following machine types: * `mls1-c1-m2` * `mls1-c4-m2` * `n1-standard-2` * `n1-standard-4` * `n1-standard-8` * `n1-standard-16` * `n1-standard-32` * `n1-highmem-2` * `n1-highmem-4` * `n1-highmem-8` * `n1-highmem-16` * `n1-highmem-32` * `n1-highcpu-2` * `n1-highcpu-4` * `n1-highcpu-8` * `n1-highcpu-16` * `n1-highcpu-32` `mls1-c1-m2` is generally available. All other machine types are available in beta. Learn more about the [differences between machine types](/ml-engine/docs/machine-types-online-prediction).
+ "acceleratorConfig": { # Represents a hardware accelerator request config. Note that the AcceleratorConfig can be used in both Jobs and Versions. Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and [accelerators for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus). # Optional. Accelerator config for using GPUs for online prediction (beta). Only specify this field if you have specified a Compute Engine (N1) machine type in the `machineType` field. Learn more about [using GPUs for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
"type": "A String", # The type of accelerator to use.
"count": "A String", # The number of accelerators to attach to each machine running the job.
},
- "state": "A String", # Output only. The state of a version.
- "name": "A String", # Required. The name specified for the version when it was created.
- #
- # The version name must be unique within the model it is created in.
- "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in
- # response to increases and decreases in traffic. Care should be
- # taken to ramp up traffic according to the model's ability to scale
- # or you will start seeing increases in latency and 429 response codes.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify
- # `manual_scaling`.
- "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These
- # nodes are always up, starting from the time the model is deployed.
- # Therefore, the cost of operating this model will be at least
- # `rate` * `min_nodes` * number of hours since last billing cycle,
- # where `rate` is the cost per node-hour as documented in the
- # [pricing guide](/ml-engine/docs/pricing),
- # even if no predictions are performed. There is additional cost for each
- # prediction performed.
- #
- # Unlike manual scaling, if the load gets too heavy for the nodes
- # that are up, the service will automatically add nodes to handle the
- # increased load as well as scale back as traffic drops, always maintaining
- # at least `min_nodes`. You will be charged for the time in which additional
- # nodes are used.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [legacy
- # (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 0, in which case, when traffic to a model stops
- # (and after a cool-down period), nodes will be shut down and no charges will
- # be incurred until traffic to the model resumes.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [Compute
- # Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a
- # Compute Engine machine type.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use
- # ManualScaling.
- #
- # You can set `min_nodes` when creating the model version, and you can also
- # update `min_nodes` for an existing version:
- # <pre>
- # update_body.json:
- # {
- # 'autoScaling': {
- # 'minNodes': 5
- # }
- # }
- # </pre>
- # HTTP request:
- # <pre style="max-width: 626px;">
- # PATCH
- # https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes
- # -d @./update_body.json
- # </pre>
- },
- "explanationConfig": { # Message holding configuration options for explaining model predictions. # Optional. Configures explainability features on the model's version.
- # Some explanation features require additional metadata to be loaded
- # as part of the model payload.
- # There are two feature attribution methods supported for TensorFlow models:
- # integrated gradients and sampled Shapley.
- # [Learn more about feature
- # attributions.](/ai-platform/prediction/docs/ai-explanations/overview)
- "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage # Attributes credit by computing the Aumann-Shapley value taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: http://proceedings.mlr.press/v70/sundararajan17a.html
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1703.01365
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help prevent simultaneous updates of a model from overwriting each other. It is strongly suggested that systems make use of the `etag` in the read-modify-write cycle to perform model updates in order to avoid race conditions: An `etag` is returned in the response to `GetVersion`, and systems are expected to put that etag in the request to `UpdateVersion` to ensure that their change will be applied to the model as intended.
+ "explanationConfig": { # Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. [Learn more about feature attributions.](/ai-platform/prediction/docs/ai-explanations/overview) # Optional. Configures explainability features on the model's version. Some explanation features require additional metadata to be loaded as part of the model payload.
+ "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs. # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage # Attributes credit by computing the XRAI taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365 # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that # An attribution method that approximates Shapley values for features that
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- "numPaths": 42, # The number of feature permutations to consider when approximating the
- # Shapley values.
+ "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features. # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
+ "numPaths": 42, # The number of feature permutations to consider when approximating the Shapley values.
},
},
- "pythonVersion": "A String", # Required. The version of Python used in prediction.
- #
- # The following Python versions are available:
- #
- # * Python '3.7' is available when `runtime_version` is set to '1.15' or
- # later.
- # * Python '3.5' is available when `runtime_version` is set to a version
- # from '1.4' to '1.14'.
- # * Python '2.7' is available when `runtime_version` is set to '1.15' or
- # earlier.
- #
- # Read more about the Python versions available for [each runtime
- # version](/ml-engine/docs/runtime-version-list).
- "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. # Optional. *Only* specify this field in a
- # projects.models.versions.patch
- # request. Specifying it in a
- # projects.models.versions.create
- # request has no effect.
- #
- # Configures the request-response pair logging on predictions from this
- # Version.
- # Online prediction requests to a model version and the responses to these
- # requests are converted to raw strings and saved to the specified BigQuery
- # table. Logging is constrained by [BigQuery quotas and
- # limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits,
- # AI Platform Prediction does not log request-response pairs, but it continues
- # to serve predictions.
- #
- # If you are using [continuous
- # evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to
- # specify this configuration manually. Setting up continuous evaluation
- # automatically enables logging of request-response pairs.
- "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1.
- # For example, if you want to log 10% of requests, enter `0.1`. The sampling
- # window is the lifetime of the model version. Defaults to 0.
- "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format:
- # "<var>project_id</var>.<var>dataset_name</var>.<var>table_name</var>"
- #
- # The specified table must already exist, and the "Cloud ML Service Agent"
- # for your project must have permission to write to it. The table must have
- # the following [schema](/bigquery/docs/schemas):
- #
- # <table>
- # <tr><th>Field name</th><th style="display: table-cell">Type</th>
- # <th style="display: table-cell">Mode</th></tr>
- # <tr><td>model</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>model_version</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>time</td><td>TIMESTAMP</td><td>REQUIRED</td></tr>
- # <tr><td>raw_data</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>raw_prediction</td><td>STRING</td><td>NULLABLE</td></tr>
- # <tr><td>groundtruth</td><td>STRING</td><td>NULLABLE</td></tr>
- # </table>
+ "labels": { # Optional. One or more labels that you can add, to organize your model versions. Each label is a key-value pair, where both the key and the value are arbitrary strings that you supply. For more information, see the documentation on using labels.
+ "a_key": "A String",
},
- "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the
- # model. You should generally use `auto_scaling` with an appropriate
- # `min_nodes` instead, but this option is available if you want more
- # predictable billing. Beware that latency and error rates will increase
- # if the traffic exceeds that capability of the system to serve it based
- # on the selected number of nodes.
- "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up,
- # starting from the time the model is deployed, so the cost of operating
- # this model will be proportional to `nodes` * number of hours since
- # last billing cycle plus the cost for each prediction performed.
+ "container": { # Specify a custom container to deploy. Our ContainerSpec is a subset of the Kubernetes Container specification. https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#container-v1-core
+ "ports": [ # Immutable. List of ports to expose from the container. Exposing a port here gives the system additional information about the network connections a container uses, but is primarily informational. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network.
+ { # ContainerPort represents a network port in a single container.
+ "containerPort": 42, # Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
+ },
+ ],
+ "env": [ # Immutable. List of environment variables to set in the container.
+ { # EnvVar represents an environment variable present in a Container.
+ "name": "A String", # Name of the environment variable. Must be a C_IDENTIFIER.
+ "value": "A String", # Variable references $(VAR_NAME) are expanded using the previous defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
+ },
+ ],
+ "command": [ # Immutable. Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
+ "image": "A String", # Docker image name. More info: https://kubernetes.io/docs/concepts/containers/images
+ "args": [ # Immutable. Arguments to the entrypoint. The docker image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
},
- "createTime": "A String", # Output only. The time the version was created.
"lastUseTime": "A String", # Output only. The time the version was last used for prediction.
- "framework": "A String", # Optional. The machine learning framework AI Platform uses to train
- # this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`,
- # `XGBOOST`. If you do not specify a framework, AI Platform
- # will analyze files in the deployment_uri to determine a framework. If you
- # choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version
- # of the model to 1.4 or greater.
- #
- # Do **not** specify a framework if you're deploying a [custom
- # prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines).
- #
- # If you specify a [Compute Engine (N1) machine
- # type](/ml-engine/docs/machine-types-online-prediction) in the
- # `machineType` field, you must specify `TENSORFLOW`
- # for the framework.
- "predictionClass": "A String", # Optional. The fully qualified name
- # (<var>module_name</var>.<var>class_name</var>) of a class that implements
- # the Predictor interface described in this reference field. The module
- # containing this class should be included in a package provided to the
- # [`packageUris` field](#Version.FIELDS.package_uris).
- #
- # Specify this field if and only if you are deploying a [custom prediction
- # routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines).
- # If you specify this field, you must set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and
- # you must set `machineType` to a [legacy (MLS1)
- # machine type](/ml-engine/docs/machine-types-online-prediction).
- #
- # The following code sample provides the Predictor interface:
- #
- # <pre style="max-width: 626px;">
- # class Predictor(object):
- # """Interface for constructing custom predictors."""
- #
- # def predict(self, instances, **kwargs):
- # """Performs custom prediction.
- #
- # Instances are the decoded values from the request. They have already
- # been deserialized from JSON.
- #
- # Args:
- # instances: A list of prediction input instances.
- # **kwargs: A dictionary of keyword args provided as additional
- # fields on the predict request body.
- #
- # Returns:
- # A list of outputs containing the prediction results. This list must
- # be JSON serializable.
- # """
- # raise NotImplementedError()
- #
- # @classmethod
- # def from_path(cls, model_dir):
- # """Creates an instance of Predictor using the given path.
- #
- # Loading of the predictor should be done in this method.
- #
- # Args:
- # model_dir: The local directory that contains the exported model
- # file along with any additional files uploaded when creating the
- # version resource.
- #
- # Returns:
- # An instance implementing this Predictor class.
- # """
- # raise NotImplementedError()
- # </pre>
- #
- # Learn more about [the Predictor interface and custom prediction
- # routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
- "isDefault": True or False, # Output only. If true, this version will be used to handle prediction
- # requests that do not specify a version.
- #
- # You can change the default version by calling
- # projects.methods.versions.setDefault.
- "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help
- # prevent simultaneous updates of a model from overwriting each other.
- # It is strongly suggested that systems make use of the `etag` in the
- # read-modify-write cycle to perform model updates in order to avoid race
- # conditions: An `etag` is returned in the response to `GetVersion`, and
- # systems are expected to put that etag in the request to `UpdateVersion` to
- # ensure that their change will be applied to the model as intended.
"serviceAccount": "A String", # Optional. Specifies the service account for resource access control.
- "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
- "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to
- # create the version. See the
- # [guide to model
- # deployment](/ml-engine/docs/tensorflow/deploying-models) for more
- # information.
- #
- # When passing Version to
- # projects.models.versions.create
- # the model service uses the specified location as the source of the model.
- # Once deployed, the model version is hosted by the prediction service, so
- # this location is useful only as a historical record.
- # The total number of model files can't exceed 1000.
- "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment.
- #
- # For more information, see the
- # [runtime version list](/ml-engine/docs/runtime-version-list) and
- # [how to manage runtime versions](/ml-engine/docs/versioning).
- "description": "A String", # Optional. The description specified for the version when it was created.
+ "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment. For more information, see the [runtime version list](/ml-engine/docs/runtime-version-list) and [how to manage runtime versions](/ml-engine/docs/versioning).
+ "name": "A String", # Required. The name specified for the version when it was created. The version name must be unique within the model it is created in.
+ "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the model. You should generally use `auto_scaling` with an appropriate `min_nodes` instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
+ "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed, so the cost of operating this model will be proportional to `nodes` * number of hours since last billing cycle plus the cost for each prediction performed.
+ },
+ "state": "A String", # Output only. The state of a version.
+ "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines) or [scikit-learn pipelines with custom code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code). For a custom prediction routine, one of these packages must contain your Predictor class (see [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally, include any dependencies used by your Predictor or scikit-learn pipeline uses that are not already included in your selected [runtime version](/ml-engine/docs/tensorflow/runtime-version-list). If you specify this field, you must also set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
+ "A String",
+ ],
+ "pythonVersion": "A String", # Required. The version of Python used in prediction. The following Python versions are available: * Python '3.7' is available when `runtime_version` is set to '1.15' or later. * Python '3.5' is available when `runtime_version` is set to a version from '1.4' to '1.14'. * Python '2.7' is available when `runtime_version` is set to '1.15' or earlier. Read more about the Python versions available for [each runtime version](/ml-engine/docs/runtime-version-list).
},
],
+ "nextPageToken": "A String", # Optional. Pass this token as the `page_token` field of the request for a subsequent call.
}</pre>
</div>
@@ -1299,365 +450,87 @@
<div class="method">
<code class="details" id="patch">patch(name, body=None, updateMask=None, x__xgafv=None)</code>
- <pre>Updates the specified Version resource.
-
-Currently the only update-able fields are `description`,
-`requestLoggingConfig`, `autoScaling.minNodes`, and `manualScaling.nodes`.
+ <pre>Updates the specified Version resource. Currently the only update-able fields are `description`, `requestLoggingConfig`, `autoScaling.minNodes`, and `manualScaling.nodes`.
Args:
name: string, Required. The name of the model. (required)
body: object, The request body.
The object takes the form of:
-{ # Represents a version of the model.
- #
- # Each version is a trained model deployed in the cloud, ready to handle
- # prediction requests. A model can have multiple versions. You can get
- # information about all of the versions of a given model by calling
- # projects.models.versions.list.
- "labels": { # Optional. One or more labels that you can add, to organize your model
- # versions. Each label is a key-value pair, where both the key and the value
- # are arbitrary strings that you supply.
- # For more information, see the documentation on
- # <a href="/ml-engine/docs/tensorflow/resource-labels">using labels</a>.
- "a_key": "A String",
+{ # Represents a version of the model. Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions. You can get information about all of the versions of a given model by calling projects.models.versions.list.
+ "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to create the version. See the [guide to model deployment](/ml-engine/docs/tensorflow/deploying-models) for more information. When passing Version to projects.models.versions.create the model service uses the specified location as the source of the model. Once deployed, the model version is hosted by the prediction service, so this location is useful only as a historical record. The total number of model files can't exceed 1000.
+ "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. Online prediction requests to a model version and the responses to these requests are converted to raw strings and saved to the specified BigQuery table. Logging is constrained by [BigQuery quotas and limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits, AI Platform Prediction does not log request-response pairs, but it continues to serve predictions. If you are using [continuous evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to specify this configuration manually. Setting up continuous evaluation automatically enables logging of request-response pairs. # Optional. *Only* specify this field in a projects.models.versions.patch request. Specifying it in a projects.models.versions.create request has no effect. Configures the request-response pair logging on predictions from this Version.
+ "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format: " project_id.dataset_name.table_name" The specified table must already exist, and the "Cloud ML Service Agent" for your project must have permission to write to it. The table must have the following [schema](/bigquery/docs/schemas): Field nameType Mode model STRING REQUIRED model_version STRING REQUIRED time TIMESTAMP REQUIRED raw_data STRING REQUIRED raw_prediction STRING NULLABLE groundtruth STRING NULLABLE
+ "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1. For example, if you want to log 10% of requests, enter `0.1`. The sampling window is the lifetime of the model version. Defaults to 0.
},
- "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only
- # applies to online prediction service. If this field is not specified, it
- # defaults to `mls1-c1-m2`.
- #
- # Online prediction supports the following machine types:
- #
- # * `mls1-c1-m2`
- # * `mls1-c4-m2`
- # * `n1-standard-2`
- # * `n1-standard-4`
- # * `n1-standard-8`
- # * `n1-standard-16`
- # * `n1-standard-32`
- # * `n1-highmem-2`
- # * `n1-highmem-4`
- # * `n1-highmem-8`
- # * `n1-highmem-16`
- # * `n1-highmem-32`
- # * `n1-highcpu-2`
- # * `n1-highcpu-4`
- # * `n1-highcpu-8`
- # * `n1-highcpu-16`
- # * `n1-highcpu-32`
- #
- # `mls1-c1-m2` is generally available. All other machine types are available
- # in beta. Learn more about the [differences between machine
- # types](/ml-engine/docs/machine-types-online-prediction).
- "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom
- # prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines)
- # or [scikit-learn pipelines with custom
- # code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code).
- #
- # For a custom prediction routine, one of these packages must contain your
- # Predictor class (see
- # [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally,
- # include any dependencies used by your Predictor or scikit-learn pipeline
- # uses that are not already included in your selected [runtime
- # version](/ml-engine/docs/tensorflow/runtime-version-list).
- #
- # If you specify this field, you must also set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
- "A String",
- ],
- "acceleratorConfig": { # Represents a hardware accelerator request config. # Optional. Accelerator config for using GPUs for online prediction (beta).
- # Only specify this field if you have specified a Compute Engine (N1) machine
- # type in the `machineType` field. Learn more about [using GPUs for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
- # Note that the AcceleratorConfig can be used in both Jobs and Versions.
- # Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and
- # [accelerators for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
+ "predictionClass": "A String", # Optional. The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the [`packageUris` field](#Version.FIELDS.package_uris). Specify this field if and only if you are deploying a [custom prediction routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify this field, you must set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and you must set `machineType` to a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction). The following code sample provides the Predictor interface: class Predictor(object): """Interface for constructing custom predictors.""" def predict(self, instances, **kwargs): """Performs custom prediction. Instances are the decoded values from the request. They have already been deserialized from JSON. Args: instances: A list of prediction input instances. **kwargs: A dictionary of keyword args provided as additional fields on the predict request body. Returns: A list of outputs containing the prediction results. This list must be JSON serializable. """ raise NotImplementedError() @classmethod def from_path(cls, model_dir): """Creates an instance of Predictor using the given path. Loading of the predictor should be done in this method. Args: model_dir: The local directory that contains the exported model file along with any additional files uploaded when creating the version resource. Returns: An instance implementing this Predictor class. """ raise NotImplementedError() Learn more about [the Predictor interface and custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
+ "framework": "A String", # Optional. The machine learning framework AI Platform uses to train this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`, `XGBOOST`. If you do not specify a framework, AI Platform will analyze files in the deployment_uri to determine a framework. If you choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version of the model to 1.4 or greater. Do **not** specify a framework if you're deploying a [custom prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction) in the `machineType` field, you must specify `TENSORFLOW` for the framework.
+ "description": "A String", # Optional. The description specified for the version when it was created.
+ "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify `manual_scaling`.
+ "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed. Therefore, the cost of operating this model will be at least `rate` * `min_nodes` * number of hours since last billing cycle, where `rate` is the cost per node-hour as documented in the [pricing guide](/ml-engine/docs/pricing), even if no predictions are performed. There is additional cost for each prediction performed. Unlike manual scaling, if the load gets too heavy for the nodes that are up, the service will automatically add nodes to handle the increased load as well as scale back as traffic drops, always maintaining at least `min_nodes`. You will be charged for the time in which additional nodes are used. If `min_nodes` is not specified and AutoScaling is used with a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 0, in which case, when traffic to a model stops (and after a cool-down period), nodes will be shut down and no charges will be incurred until traffic to the model resumes. If `min_nodes` is not specified and AutoScaling is used with a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a Compute Engine machine type. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use ManualScaling. You can set `min_nodes` when creating the model version, and you can also update `min_nodes` for an existing version: update_body.json: { 'autoScaling': { 'minNodes': 5 } } HTTP request: PATCH https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes -d @./update_body.json
+ },
+ "isDefault": True or False, # Output only. If true, this version will be used to handle prediction requests that do not specify a version. You can change the default version by calling projects.methods.versions.setDefault.
+ "createTime": "A String", # Output only. The time the version was created.
+ "routes": { # RouteMap is used to override HTTP paths sent to a Custom Container. If specified, the HTTP server implemented in the ContainerSpec must support the route. If unspecified, standard HTTP paths will be used.
+ "predict": "A String", # HTTP path to send prediction requests.
+ "health": "A String", # HTTP path to send health check requests.
+ },
+ "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
+ "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only applies to online prediction service. If this field is not specified, it defaults to `mls1-c1-m2`. Online prediction supports the following machine types: * `mls1-c1-m2` * `mls1-c4-m2` * `n1-standard-2` * `n1-standard-4` * `n1-standard-8` * `n1-standard-16` * `n1-standard-32` * `n1-highmem-2` * `n1-highmem-4` * `n1-highmem-8` * `n1-highmem-16` * `n1-highmem-32` * `n1-highcpu-2` * `n1-highcpu-4` * `n1-highcpu-8` * `n1-highcpu-16` * `n1-highcpu-32` `mls1-c1-m2` is generally available. All other machine types are available in beta. Learn more about the [differences between machine types](/ml-engine/docs/machine-types-online-prediction).
+ "acceleratorConfig": { # Represents a hardware accelerator request config. Note that the AcceleratorConfig can be used in both Jobs and Versions. Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and [accelerators for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus). # Optional. Accelerator config for using GPUs for online prediction (beta). Only specify this field if you have specified a Compute Engine (N1) machine type in the `machineType` field. Learn more about [using GPUs for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
"type": "A String", # The type of accelerator to use.
"count": "A String", # The number of accelerators to attach to each machine running the job.
},
- "state": "A String", # Output only. The state of a version.
- "name": "A String", # Required. The name specified for the version when it was created.
- #
- # The version name must be unique within the model it is created in.
- "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in
- # response to increases and decreases in traffic. Care should be
- # taken to ramp up traffic according to the model's ability to scale
- # or you will start seeing increases in latency and 429 response codes.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify
- # `manual_scaling`.
- "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These
- # nodes are always up, starting from the time the model is deployed.
- # Therefore, the cost of operating this model will be at least
- # `rate` * `min_nodes` * number of hours since last billing cycle,
- # where `rate` is the cost per node-hour as documented in the
- # [pricing guide](/ml-engine/docs/pricing),
- # even if no predictions are performed. There is additional cost for each
- # prediction performed.
- #
- # Unlike manual scaling, if the load gets too heavy for the nodes
- # that are up, the service will automatically add nodes to handle the
- # increased load as well as scale back as traffic drops, always maintaining
- # at least `min_nodes`. You will be charged for the time in which additional
- # nodes are used.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [legacy
- # (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 0, in which case, when traffic to a model stops
- # (and after a cool-down period), nodes will be shut down and no charges will
- # be incurred until traffic to the model resumes.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [Compute
- # Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a
- # Compute Engine machine type.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use
- # ManualScaling.
- #
- # You can set `min_nodes` when creating the model version, and you can also
- # update `min_nodes` for an existing version:
- # <pre>
- # update_body.json:
- # {
- # 'autoScaling': {
- # 'minNodes': 5
- # }
- # }
- # </pre>
- # HTTP request:
- # <pre style="max-width: 626px;">
- # PATCH
- # https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes
- # -d @./update_body.json
- # </pre>
- },
- "explanationConfig": { # Message holding configuration options for explaining model predictions. # Optional. Configures explainability features on the model's version.
- # Some explanation features require additional metadata to be loaded
- # as part of the model payload.
- # There are two feature attribution methods supported for TensorFlow models:
- # integrated gradients and sampled Shapley.
- # [Learn more about feature
- # attributions.](/ai-platform/prediction/docs/ai-explanations/overview)
- "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage # Attributes credit by computing the Aumann-Shapley value taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: http://proceedings.mlr.press/v70/sundararajan17a.html
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1703.01365
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help prevent simultaneous updates of a model from overwriting each other. It is strongly suggested that systems make use of the `etag` in the read-modify-write cycle to perform model updates in order to avoid race conditions: An `etag` is returned in the response to `GetVersion`, and systems are expected to put that etag in the request to `UpdateVersion` to ensure that their change will be applied to the model as intended.
+ "explanationConfig": { # Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. [Learn more about feature attributions.](/ai-platform/prediction/docs/ai-explanations/overview) # Optional. Configures explainability features on the model's version. Some explanation features require additional metadata to be loaded as part of the model payload.
+ "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs. # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage # Attributes credit by computing the XRAI taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365 # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that # An attribution method that approximates Shapley values for features that
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- "numPaths": 42, # The number of feature permutations to consider when approximating the
- # Shapley values.
+ "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features. # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
+ "numPaths": 42, # The number of feature permutations to consider when approximating the Shapley values.
},
},
- "pythonVersion": "A String", # Required. The version of Python used in prediction.
- #
- # The following Python versions are available:
- #
- # * Python '3.7' is available when `runtime_version` is set to '1.15' or
- # later.
- # * Python '3.5' is available when `runtime_version` is set to a version
- # from '1.4' to '1.14'.
- # * Python '2.7' is available when `runtime_version` is set to '1.15' or
- # earlier.
- #
- # Read more about the Python versions available for [each runtime
- # version](/ml-engine/docs/runtime-version-list).
- "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. # Optional. *Only* specify this field in a
- # projects.models.versions.patch
- # request. Specifying it in a
- # projects.models.versions.create
- # request has no effect.
- #
- # Configures the request-response pair logging on predictions from this
- # Version.
- # Online prediction requests to a model version and the responses to these
- # requests are converted to raw strings and saved to the specified BigQuery
- # table. Logging is constrained by [BigQuery quotas and
- # limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits,
- # AI Platform Prediction does not log request-response pairs, but it continues
- # to serve predictions.
- #
- # If you are using [continuous
- # evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to
- # specify this configuration manually. Setting up continuous evaluation
- # automatically enables logging of request-response pairs.
- "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1.
- # For example, if you want to log 10% of requests, enter `0.1`. The sampling
- # window is the lifetime of the model version. Defaults to 0.
- "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format:
- # "<var>project_id</var>.<var>dataset_name</var>.<var>table_name</var>"
- #
- # The specified table must already exist, and the "Cloud ML Service Agent"
- # for your project must have permission to write to it. The table must have
- # the following [schema](/bigquery/docs/schemas):
- #
- # <table>
- # <tr><th>Field name</th><th style="display: table-cell">Type</th>
- # <th style="display: table-cell">Mode</th></tr>
- # <tr><td>model</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>model_version</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>time</td><td>TIMESTAMP</td><td>REQUIRED</td></tr>
- # <tr><td>raw_data</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>raw_prediction</td><td>STRING</td><td>NULLABLE</td></tr>
- # <tr><td>groundtruth</td><td>STRING</td><td>NULLABLE</td></tr>
- # </table>
+ "labels": { # Optional. One or more labels that you can add, to organize your model versions. Each label is a key-value pair, where both the key and the value are arbitrary strings that you supply. For more information, see the documentation on using labels.
+ "a_key": "A String",
},
- "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the
- # model. You should generally use `auto_scaling` with an appropriate
- # `min_nodes` instead, but this option is available if you want more
- # predictable billing. Beware that latency and error rates will increase
- # if the traffic exceeds that capability of the system to serve it based
- # on the selected number of nodes.
- "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up,
- # starting from the time the model is deployed, so the cost of operating
- # this model will be proportional to `nodes` * number of hours since
- # last billing cycle plus the cost for each prediction performed.
+ "container": { # Specify a custom container to deploy. Our ContainerSpec is a subset of the Kubernetes Container specification. https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#container-v1-core
+ "ports": [ # Immutable. List of ports to expose from the container. Exposing a port here gives the system additional information about the network connections a container uses, but is primarily informational. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network.
+ { # ContainerPort represents a network port in a single container.
+ "containerPort": 42, # Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
+ },
+ ],
+ "env": [ # Immutable. List of environment variables to set in the container.
+ { # EnvVar represents an environment variable present in a Container.
+ "name": "A String", # Name of the environment variable. Must be a C_IDENTIFIER.
+ "value": "A String", # Variable references $(VAR_NAME) are expanded using the previous defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
+ },
+ ],
+ "command": [ # Immutable. Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
+ "image": "A String", # Docker image name. More info: https://kubernetes.io/docs/concepts/containers/images
+ "args": [ # Immutable. Arguments to the entrypoint. The docker image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
},
- "createTime": "A String", # Output only. The time the version was created.
"lastUseTime": "A String", # Output only. The time the version was last used for prediction.
- "framework": "A String", # Optional. The machine learning framework AI Platform uses to train
- # this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`,
- # `XGBOOST`. If you do not specify a framework, AI Platform
- # will analyze files in the deployment_uri to determine a framework. If you
- # choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version
- # of the model to 1.4 or greater.
- #
- # Do **not** specify a framework if you're deploying a [custom
- # prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines).
- #
- # If you specify a [Compute Engine (N1) machine
- # type](/ml-engine/docs/machine-types-online-prediction) in the
- # `machineType` field, you must specify `TENSORFLOW`
- # for the framework.
- "predictionClass": "A String", # Optional. The fully qualified name
- # (<var>module_name</var>.<var>class_name</var>) of a class that implements
- # the Predictor interface described in this reference field. The module
- # containing this class should be included in a package provided to the
- # [`packageUris` field](#Version.FIELDS.package_uris).
- #
- # Specify this field if and only if you are deploying a [custom prediction
- # routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines).
- # If you specify this field, you must set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and
- # you must set `machineType` to a [legacy (MLS1)
- # machine type](/ml-engine/docs/machine-types-online-prediction).
- #
- # The following code sample provides the Predictor interface:
- #
- # <pre style="max-width: 626px;">
- # class Predictor(object):
- # """Interface for constructing custom predictors."""
- #
- # def predict(self, instances, **kwargs):
- # """Performs custom prediction.
- #
- # Instances are the decoded values from the request. They have already
- # been deserialized from JSON.
- #
- # Args:
- # instances: A list of prediction input instances.
- # **kwargs: A dictionary of keyword args provided as additional
- # fields on the predict request body.
- #
- # Returns:
- # A list of outputs containing the prediction results. This list must
- # be JSON serializable.
- # """
- # raise NotImplementedError()
- #
- # @classmethod
- # def from_path(cls, model_dir):
- # """Creates an instance of Predictor using the given path.
- #
- # Loading of the predictor should be done in this method.
- #
- # Args:
- # model_dir: The local directory that contains the exported model
- # file along with any additional files uploaded when creating the
- # version resource.
- #
- # Returns:
- # An instance implementing this Predictor class.
- # """
- # raise NotImplementedError()
- # </pre>
- #
- # Learn more about [the Predictor interface and custom prediction
- # routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
- "isDefault": True or False, # Output only. If true, this version will be used to handle prediction
- # requests that do not specify a version.
- #
- # You can change the default version by calling
- # projects.methods.versions.setDefault.
- "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help
- # prevent simultaneous updates of a model from overwriting each other.
- # It is strongly suggested that systems make use of the `etag` in the
- # read-modify-write cycle to perform model updates in order to avoid race
- # conditions: An `etag` is returned in the response to `GetVersion`, and
- # systems are expected to put that etag in the request to `UpdateVersion` to
- # ensure that their change will be applied to the model as intended.
"serviceAccount": "A String", # Optional. Specifies the service account for resource access control.
- "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
- "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to
- # create the version. See the
- # [guide to model
- # deployment](/ml-engine/docs/tensorflow/deploying-models) for more
- # information.
- #
- # When passing Version to
- # projects.models.versions.create
- # the model service uses the specified location as the source of the model.
- # Once deployed, the model version is hosted by the prediction service, so
- # this location is useful only as a historical record.
- # The total number of model files can't exceed 1000.
- "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment.
- #
- # For more information, see the
- # [runtime version list](/ml-engine/docs/runtime-version-list) and
- # [how to manage runtime versions](/ml-engine/docs/versioning).
- "description": "A String", # Optional. The description specified for the version when it was created.
+ "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment. For more information, see the [runtime version list](/ml-engine/docs/runtime-version-list) and [how to manage runtime versions](/ml-engine/docs/versioning).
+ "name": "A String", # Required. The name specified for the version when it was created. The version name must be unique within the model it is created in.
+ "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the model. You should generally use `auto_scaling` with an appropriate `min_nodes` instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
+ "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed, so the cost of operating this model will be proportional to `nodes` * number of hours since last billing cycle plus the cost for each prediction performed.
+ },
+ "state": "A String", # Output only. The state of a version.
+ "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines) or [scikit-learn pipelines with custom code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code). For a custom prediction routine, one of these packages must contain your Predictor class (see [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally, include any dependencies used by your Predictor or scikit-learn pipeline uses that are not already included in your selected [runtime version](/ml-engine/docs/tensorflow/runtime-version-list). If you specify this field, you must also set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
+ "A String",
+ ],
+ "pythonVersion": "A String", # Required. The version of Python used in prediction. The following Python versions are available: * Python '3.7' is available when `runtime_version` is set to '1.15' or later. * Python '3.5' is available when `runtime_version` is set to a version from '1.4' to '1.14'. * Python '2.7' is available when `runtime_version` is set to '1.15' or earlier. Read more about the Python versions available for [each runtime version](/ml-engine/docs/runtime-version-list).
}
- updateMask: string, Required. Specifies the path, relative to `Version`, of the field to
-update. Must be present and non-empty.
-
-For example, to change the description of a version to "foo", the
-`update_mask` parameter would be specified as `description`, and the
-`PATCH` request body would specify the new value, as follows:
-
-```
-{
- "description": "foo"
-}
-```
-
-Currently the only supported update mask fields are `description`,
-`requestLoggingConfig`, `autoScaling.minNodes`, and `manualScaling.nodes`.
-However, you can only update `manualScaling.nodes` if the version uses a
-[Compute Engine (N1)
-machine type](/ml-engine/docs/machine-types-online-prediction).
+ updateMask: string, Required. Specifies the path, relative to `Version`, of the field to update. Must be present and non-empty. For example, to change the description of a version to "foo", the `update_mask` parameter would be specified as `description`, and the `PATCH` request body would specify the new value, as follows: ``` { "description": "foo" } ``` Currently the only supported update mask fields are `description`, `requestLoggingConfig`, `autoScaling.minNodes`, and `manualScaling.nodes`. However, you can only update `manualScaling.nodes` if the version uses a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction).
x__xgafv: string, V1 error format.
Allowed values
1 - v1 error format
@@ -1666,66 +539,33 @@
Returns:
An object of the form:
- { # This resource represents a long-running operation that is the result of a
- # network API call.
- "error": { # The `Status` type defines a logical error model that is suitable for # The error result of the operation in case of failure or cancellation.
- # different programming environments, including REST APIs and RPC APIs. It is
- # used by [gRPC](https://github.com/grpc). Each `Status` message contains
- # three pieces of data: error code, error message, and error details.
- #
- # You can find out more about this error model and how to work with it in the
- # [API Design Guide](https://cloud.google.com/apis/design/errors).
- "details": [ # A list of messages that carry the error details. There is a common set of
- # message types for APIs to use.
+ { # This resource represents a long-running operation that is the result of a network API call.
+ "name": "A String", # The server-assigned name, which is only unique within the same service that originally returns it. If you use the default HTTP mapping, the `name` should be a resource name ending with `operations/{unique_id}`.
+ "response": { # The normal response of the operation in case of success. If the original method returns no data on success, such as `Delete`, the response is `google.protobuf.Empty`. If the original method is standard `Get`/`Create`/`Update`, the response should be the resource. For other methods, the response should have the type `XxxResponse`, where `Xxx` is the original method name. For example, if the original method name is `TakeSnapshot()`, the inferred response type is `TakeSnapshotResponse`.
+ "a_key": "", # Properties of the object. Contains field @type with type URL.
+ },
+ "done": True or False, # If the value is `false`, it means the operation is still in progress. If `true`, the operation is completed, and either `error` or `response` is available.
+ "error": { # The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors). # The error result of the operation in case of failure or cancellation.
+ "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "details": [ # A list of messages that carry the error details. There is a common set of message types for APIs to use.
{
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
],
- "message": "A String", # A developer-facing error message, which should be in English. Any
- # user-facing error message should be localized and sent in the
- # google.rpc.Status.details field, or localized by the client.
- "code": 42, # The status code, which should be an enum value of google.rpc.Code.
+ "message": "A String", # A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the google.rpc.Status.details field, or localized by the client.
},
- "done": True or False, # If the value is `false`, it means the operation is still in progress.
- # If `true`, the operation is completed, and either `error` or `response` is
- # available.
- "response": { # The normal response of the operation in case of success. If the original
- # method returns no data on success, such as `Delete`, the response is
- # `google.protobuf.Empty`. If the original method is standard
- # `Get`/`Create`/`Update`, the response should be the resource. For other
- # methods, the response should have the type `XxxResponse`, where `Xxx`
- # is the original method name. For example, if the original method name
- # is `TakeSnapshot()`, the inferred response type is
- # `TakeSnapshotResponse`.
+ "metadata": { # Service-specific metadata associated with the operation. It typically contains progress information and common metadata such as create time. Some services might not provide such metadata. Any method that returns a long-running operation should document the metadata type, if any.
"a_key": "", # Properties of the object. Contains field @type with type URL.
},
- "metadata": { # Service-specific metadata associated with the operation. It typically
- # contains progress information and common metadata such as create time.
- # Some services might not provide such metadata. Any method that returns a
- # long-running operation should document the metadata type, if any.
- "a_key": "", # Properties of the object. Contains field @type with type URL.
- },
- "name": "A String", # The server-assigned name, which is only unique within the same service that
- # originally returns it. If you use the default HTTP mapping, the
- # `name` should be a resource name ending with `operations/{unique_id}`.
}</pre>
</div>
<div class="method">
<code class="details" id="setDefault">setDefault(name, body=None, x__xgafv=None)</code>
- <pre>Designates a version to be the default for the model.
-
-The default version is used for prediction requests made against the model
-that don't specify a version.
-
-The first version to be created for a model is automatically set as the
-default. You must make any subsequent changes to the default version
-setting manually using this method.
+ <pre>Designates a version to be the default for the model. The default version is used for prediction requests made against the model that don't specify a version. The first version to be created for a model is automatically set as the default. You must make any subsequent changes to the default version setting manually using this method.
Args:
- name: string, Required. The name of the version to make the default for the model. You
-can get the names of all the versions of a model by calling
-projects.models.versions.list. (required)
+ name: string, Required. The name of the version to make the default for the model. You can get the names of all the versions of a model by calling projects.models.versions.list. (required)
body: object, The request body.
The object takes the form of:
@@ -1740,335 +580,77 @@
Returns:
An object of the form:
- { # Represents a version of the model.
- #
- # Each version is a trained model deployed in the cloud, ready to handle
- # prediction requests. A model can have multiple versions. You can get
- # information about all of the versions of a given model by calling
- # projects.models.versions.list.
- "labels": { # Optional. One or more labels that you can add, to organize your model
- # versions. Each label is a key-value pair, where both the key and the value
- # are arbitrary strings that you supply.
- # For more information, see the documentation on
- # <a href="/ml-engine/docs/tensorflow/resource-labels">using labels</a>.
- "a_key": "A String",
+ { # Represents a version of the model. Each version is a trained model deployed in the cloud, ready to handle prediction requests. A model can have multiple versions. You can get information about all of the versions of a given model by calling projects.models.versions.list.
+ "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to create the version. See the [guide to model deployment](/ml-engine/docs/tensorflow/deploying-models) for more information. When passing Version to projects.models.versions.create the model service uses the specified location as the source of the model. Once deployed, the model version is hosted by the prediction service, so this location is useful only as a historical record. The total number of model files can't exceed 1000.
+ "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. Online prediction requests to a model version and the responses to these requests are converted to raw strings and saved to the specified BigQuery table. Logging is constrained by [BigQuery quotas and limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits, AI Platform Prediction does not log request-response pairs, but it continues to serve predictions. If you are using [continuous evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to specify this configuration manually. Setting up continuous evaluation automatically enables logging of request-response pairs. # Optional. *Only* specify this field in a projects.models.versions.patch request. Specifying it in a projects.models.versions.create request has no effect. Configures the request-response pair logging on predictions from this Version.
+ "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format: " project_id.dataset_name.table_name" The specified table must already exist, and the "Cloud ML Service Agent" for your project must have permission to write to it. The table must have the following [schema](/bigquery/docs/schemas): Field nameType Mode model STRING REQUIRED model_version STRING REQUIRED time TIMESTAMP REQUIRED raw_data STRING REQUIRED raw_prediction STRING NULLABLE groundtruth STRING NULLABLE
+ "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1. For example, if you want to log 10% of requests, enter `0.1`. The sampling window is the lifetime of the model version. Defaults to 0.
},
- "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only
- # applies to online prediction service. If this field is not specified, it
- # defaults to `mls1-c1-m2`.
- #
- # Online prediction supports the following machine types:
- #
- # * `mls1-c1-m2`
- # * `mls1-c4-m2`
- # * `n1-standard-2`
- # * `n1-standard-4`
- # * `n1-standard-8`
- # * `n1-standard-16`
- # * `n1-standard-32`
- # * `n1-highmem-2`
- # * `n1-highmem-4`
- # * `n1-highmem-8`
- # * `n1-highmem-16`
- # * `n1-highmem-32`
- # * `n1-highcpu-2`
- # * `n1-highcpu-4`
- # * `n1-highcpu-8`
- # * `n1-highcpu-16`
- # * `n1-highcpu-32`
- #
- # `mls1-c1-m2` is generally available. All other machine types are available
- # in beta. Learn more about the [differences between machine
- # types](/ml-engine/docs/machine-types-online-prediction).
- "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom
- # prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines)
- # or [scikit-learn pipelines with custom
- # code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code).
- #
- # For a custom prediction routine, one of these packages must contain your
- # Predictor class (see
- # [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally,
- # include any dependencies used by your Predictor or scikit-learn pipeline
- # uses that are not already included in your selected [runtime
- # version](/ml-engine/docs/tensorflow/runtime-version-list).
- #
- # If you specify this field, you must also set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
- "A String",
- ],
- "acceleratorConfig": { # Represents a hardware accelerator request config. # Optional. Accelerator config for using GPUs for online prediction (beta).
- # Only specify this field if you have specified a Compute Engine (N1) machine
- # type in the `machineType` field. Learn more about [using GPUs for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
- # Note that the AcceleratorConfig can be used in both Jobs and Versions.
- # Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and
- # [accelerators for online
- # prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
+ "predictionClass": "A String", # Optional. The fully qualified name (module_name.class_name) of a class that implements the Predictor interface described in this reference field. The module containing this class should be included in a package provided to the [`packageUris` field](#Version.FIELDS.package_uris). Specify this field if and only if you are deploying a [custom prediction routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify this field, you must set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and you must set `machineType` to a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction). The following code sample provides the Predictor interface: class Predictor(object): """Interface for constructing custom predictors.""" def predict(self, instances, **kwargs): """Performs custom prediction. Instances are the decoded values from the request. They have already been deserialized from JSON. Args: instances: A list of prediction input instances. **kwargs: A dictionary of keyword args provided as additional fields on the predict request body. Returns: A list of outputs containing the prediction results. This list must be JSON serializable. """ raise NotImplementedError() @classmethod def from_path(cls, model_dir): """Creates an instance of Predictor using the given path. Loading of the predictor should be done in this method. Args: model_dir: The local directory that contains the exported model file along with any additional files uploaded when creating the version resource. Returns: An instance implementing this Predictor class. """ raise NotImplementedError() Learn more about [the Predictor interface and custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
+ "framework": "A String", # Optional. The machine learning framework AI Platform uses to train this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`, `XGBOOST`. If you do not specify a framework, AI Platform will analyze files in the deployment_uri to determine a framework. If you choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version of the model to 1.4 or greater. Do **not** specify a framework if you're deploying a [custom prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines). If you specify a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction) in the `machineType` field, you must specify `TENSORFLOW` for the framework.
+ "description": "A String", # Optional. The description specified for the version when it was created.
+ "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in response to increases and decreases in traffic. Care should be taken to ramp up traffic according to the model's ability to scale or you will start seeing increases in latency and 429 response codes. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify `manual_scaling`.
+ "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed. Therefore, the cost of operating this model will be at least `rate` * `min_nodes` * number of hours since last billing cycle, where `rate` is the cost per node-hour as documented in the [pricing guide](/ml-engine/docs/pricing), even if no predictions are performed. There is additional cost for each prediction performed. Unlike manual scaling, if the load gets too heavy for the nodes that are up, the service will automatically add nodes to handle the increased load as well as scale back as traffic drops, always maintaining at least `min_nodes`. You will be charged for the time in which additional nodes are used. If `min_nodes` is not specified and AutoScaling is used with a [legacy (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 0, in which case, when traffic to a model stops (and after a cool-down period), nodes will be shut down and no charges will be incurred until traffic to the model resumes. If `min_nodes` is not specified and AutoScaling is used with a [Compute Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction), `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a Compute Engine machine type. Note that you cannot use AutoScaling if your version uses [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use ManualScaling. You can set `min_nodes` when creating the model version, and you can also update `min_nodes` for an existing version: update_body.json: { 'autoScaling': { 'minNodes': 5 } } HTTP request: PATCH https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes -d @./update_body.json
+ },
+ "isDefault": True or False, # Output only. If true, this version will be used to handle prediction requests that do not specify a version. You can change the default version by calling projects.methods.versions.setDefault.
+ "createTime": "A String", # Output only. The time the version was created.
+ "routes": { # RouteMap is used to override HTTP paths sent to a Custom Container. If specified, the HTTP server implemented in the ContainerSpec must support the route. If unspecified, standard HTTP paths will be used.
+ "predict": "A String", # HTTP path to send prediction requests.
+ "health": "A String", # HTTP path to send health check requests.
+ },
+ "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
+ "machineType": "A String", # Optional. The type of machine on which to serve the model. Currently only applies to online prediction service. If this field is not specified, it defaults to `mls1-c1-m2`. Online prediction supports the following machine types: * `mls1-c1-m2` * `mls1-c4-m2` * `n1-standard-2` * `n1-standard-4` * `n1-standard-8` * `n1-standard-16` * `n1-standard-32` * `n1-highmem-2` * `n1-highmem-4` * `n1-highmem-8` * `n1-highmem-16` * `n1-highmem-32` * `n1-highcpu-2` * `n1-highcpu-4` * `n1-highcpu-8` * `n1-highcpu-16` * `n1-highcpu-32` `mls1-c1-m2` is generally available. All other machine types are available in beta. Learn more about the [differences between machine types](/ml-engine/docs/machine-types-online-prediction).
+ "acceleratorConfig": { # Represents a hardware accelerator request config. Note that the AcceleratorConfig can be used in both Jobs and Versions. Learn more about [accelerators for training](/ml-engine/docs/using-gpus) and [accelerators for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus). # Optional. Accelerator config for using GPUs for online prediction (beta). Only specify this field if you have specified a Compute Engine (N1) machine type in the `machineType` field. Learn more about [using GPUs for online prediction](/ml-engine/docs/machine-types-online-prediction#gpus).
"type": "A String", # The type of accelerator to use.
"count": "A String", # The number of accelerators to attach to each machine running the job.
},
- "state": "A String", # Output only. The state of a version.
- "name": "A String", # Required. The name specified for the version when it was created.
- #
- # The version name must be unique within the model it is created in.
- "autoScaling": { # Options for automatically scaling a model. # Automatically scale the number of nodes used to serve the model in
- # response to increases and decreases in traffic. Care should be
- # taken to ramp up traffic according to the model's ability to scale
- # or you will start seeing increases in latency and 429 response codes.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use specify
- # `manual_scaling`.
- "minNodes": 42, # Optional. The minimum number of nodes to allocate for this model. These
- # nodes are always up, starting from the time the model is deployed.
- # Therefore, the cost of operating this model will be at least
- # `rate` * `min_nodes` * number of hours since last billing cycle,
- # where `rate` is the cost per node-hour as documented in the
- # [pricing guide](/ml-engine/docs/pricing),
- # even if no predictions are performed. There is additional cost for each
- # prediction performed.
- #
- # Unlike manual scaling, if the load gets too heavy for the nodes
- # that are up, the service will automatically add nodes to handle the
- # increased load as well as scale back as traffic drops, always maintaining
- # at least `min_nodes`. You will be charged for the time in which additional
- # nodes are used.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [legacy
- # (MLS1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 0, in which case, when traffic to a model stops
- # (and after a cool-down period), nodes will be shut down and no charges will
- # be incurred until traffic to the model resumes.
- #
- # If `min_nodes` is not specified and AutoScaling is used with a [Compute
- # Engine (N1) machine type](/ml-engine/docs/machine-types-online-prediction),
- # `min_nodes` defaults to 1. `min_nodes` must be at least 1 for use with a
- # Compute Engine machine type.
- #
- # Note that you cannot use AutoScaling if your version uses
- # [GPUs](#Version.FIELDS.accelerator_config). Instead, you must use
- # ManualScaling.
- #
- # You can set `min_nodes` when creating the model version, and you can also
- # update `min_nodes` for an existing version:
- # <pre>
- # update_body.json:
- # {
- # 'autoScaling': {
- # 'minNodes': 5
- # }
- # }
- # </pre>
- # HTTP request:
- # <pre style="max-width: 626px;">
- # PATCH
- # https://ml.googleapis.com/v1/{name=projects/*/models/*/versions/*}?update_mask=autoScaling.minNodes
- # -d @./update_body.json
- # </pre>
- },
- "explanationConfig": { # Message holding configuration options for explaining model predictions. # Optional. Configures explainability features on the model's version.
- # Some explanation features require additional metadata to be loaded
- # as part of the model payload.
- # There are two feature attribution methods supported for TensorFlow models:
- # integrated gradients and sampled Shapley.
- # [Learn more about feature
- # attributions.](/ai-platform/prediction/docs/ai-explanations/overview)
- "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage # Attributes credit by computing the Aumann-Shapley value taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: http://proceedings.mlr.press/v70/sundararajan17a.html
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1703.01365
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help prevent simultaneous updates of a model from overwriting each other. It is strongly suggested that systems make use of the `etag` in the read-modify-write cycle to perform model updates in order to avoid race conditions: An `etag` is returned in the response to `GetVersion`, and systems are expected to put that etag in the request to `UpdateVersion` to ensure that their change will be applied to the model as intended.
+ "explanationConfig": { # Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. [Learn more about feature attributions.](/ai-platform/prediction/docs/ai-explanations/overview) # Optional. Configures explainability features on the model's version. Some explanation features require additional metadata to be loaded as part of the model payload.
+ "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs. # Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "xraiAttribution": { # Attributes credit by computing the XRAI taking advantage # Attributes credit by computing the XRAI taking advantage
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- # of the model's fully differentiable structure. Refer to this paper for
- # more details: https://arxiv.org/abs/1906.02825
- # Currently only implemented for models with natural image inputs.
- "numIntegralSteps": 42, # Number of steps for approximating the path integral.
- # A good value to start is 50 and gradually increase until the
- # sum to diff property is met within the desired error range.
+ "integratedGradientsAttribution": { # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365 # Attributes credit by computing the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365
+ "numIntegralSteps": 42, # Number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is met within the desired error range.
},
- "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that # An attribution method that approximates Shapley values for features that
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- # contribute to the label being predicted. A sampling strategy is used to
- # approximate the value rather than considering all subsets of features.
- "numPaths": 42, # The number of feature permutations to consider when approximating the
- # Shapley values.
+ "sampledShapleyAttribution": { # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features. # An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
+ "numPaths": 42, # The number of feature permutations to consider when approximating the Shapley values.
},
},
- "pythonVersion": "A String", # Required. The version of Python used in prediction.
- #
- # The following Python versions are available:
- #
- # * Python '3.7' is available when `runtime_version` is set to '1.15' or
- # later.
- # * Python '3.5' is available when `runtime_version` is set to a version
- # from '1.4' to '1.14'.
- # * Python '2.7' is available when `runtime_version` is set to '1.15' or
- # earlier.
- #
- # Read more about the Python versions available for [each runtime
- # version](/ml-engine/docs/runtime-version-list).
- "requestLoggingConfig": { # Configuration for logging request-response pairs to a BigQuery table. # Optional. *Only* specify this field in a
- # projects.models.versions.patch
- # request. Specifying it in a
- # projects.models.versions.create
- # request has no effect.
- #
- # Configures the request-response pair logging on predictions from this
- # Version.
- # Online prediction requests to a model version and the responses to these
- # requests are converted to raw strings and saved to the specified BigQuery
- # table. Logging is constrained by [BigQuery quotas and
- # limits](/bigquery/quotas). If your project exceeds BigQuery quotas or limits,
- # AI Platform Prediction does not log request-response pairs, but it continues
- # to serve predictions.
- #
- # If you are using [continuous
- # evaluation](/ml-engine/docs/continuous-evaluation/), you do not need to
- # specify this configuration manually. Setting up continuous evaluation
- # automatically enables logging of request-response pairs.
- "samplingPercentage": 3.14, # Percentage of requests to be logged, expressed as a fraction from 0 to 1.
- # For example, if you want to log 10% of requests, enter `0.1`. The sampling
- # window is the lifetime of the model version. Defaults to 0.
- "bigqueryTableName": "A String", # Required. Fully qualified BigQuery table name in the following format:
- # "<var>project_id</var>.<var>dataset_name</var>.<var>table_name</var>"
- #
- # The specified table must already exist, and the "Cloud ML Service Agent"
- # for your project must have permission to write to it. The table must have
- # the following [schema](/bigquery/docs/schemas):
- #
- # <table>
- # <tr><th>Field name</th><th style="display: table-cell">Type</th>
- # <th style="display: table-cell">Mode</th></tr>
- # <tr><td>model</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>model_version</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>time</td><td>TIMESTAMP</td><td>REQUIRED</td></tr>
- # <tr><td>raw_data</td><td>STRING</td><td>REQUIRED</td></tr>
- # <tr><td>raw_prediction</td><td>STRING</td><td>NULLABLE</td></tr>
- # <tr><td>groundtruth</td><td>STRING</td><td>NULLABLE</td></tr>
- # </table>
+ "labels": { # Optional. One or more labels that you can add, to organize your model versions. Each label is a key-value pair, where both the key and the value are arbitrary strings that you supply. For more information, see the documentation on using labels.
+ "a_key": "A String",
},
- "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the
- # model. You should generally use `auto_scaling` with an appropriate
- # `min_nodes` instead, but this option is available if you want more
- # predictable billing. Beware that latency and error rates will increase
- # if the traffic exceeds that capability of the system to serve it based
- # on the selected number of nodes.
- "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up,
- # starting from the time the model is deployed, so the cost of operating
- # this model will be proportional to `nodes` * number of hours since
- # last billing cycle plus the cost for each prediction performed.
+ "container": { # Specify a custom container to deploy. Our ContainerSpec is a subset of the Kubernetes Container specification. https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#container-v1-core
+ "ports": [ # Immutable. List of ports to expose from the container. Exposing a port here gives the system additional information about the network connections a container uses, but is primarily informational. Not specifying a port here DOES NOT prevent that port from being exposed. Any port which is listening on the default "0.0.0.0" address inside a container will be accessible from the network.
+ { # ContainerPort represents a network port in a single container.
+ "containerPort": 42, # Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.
+ },
+ ],
+ "env": [ # Immutable. List of environment variables to set in the container.
+ { # EnvVar represents an environment variable present in a Container.
+ "name": "A String", # Name of the environment variable. Must be a C_IDENTIFIER.
+ "value": "A String", # Variable references $(VAR_NAME) are expanded using the previous defined environment variables in the container and any service environment variables. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. Defaults to "".
+ },
+ ],
+ "command": [ # Immutable. Entrypoint array. Not executed within a shell. The docker image's ENTRYPOINT is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
+ "image": "A String", # Docker image name. More info: https://kubernetes.io/docs/concepts/containers/images
+ "args": [ # Immutable. Arguments to the entrypoint. The docker image's CMD is used if this is not provided. Variable references $(VAR_NAME) are expanded using the container's environment. If a variable cannot be resolved, the reference in the input string will be unchanged. The $(VAR_NAME) syntax can be escaped with a double $$, ie: $$(VAR_NAME). Escaped references will never be expanded, regardless of whether the variable exists or not. More info: https://kubernetes.io/docs/tasks/inject-data-application/define-command-argument-container/#running-a-command-in-a-shell
+ "A String",
+ ],
},
- "createTime": "A String", # Output only. The time the version was created.
"lastUseTime": "A String", # Output only. The time the version was last used for prediction.
- "framework": "A String", # Optional. The machine learning framework AI Platform uses to train
- # this version of the model. Valid values are `TENSORFLOW`, `SCIKIT_LEARN`,
- # `XGBOOST`. If you do not specify a framework, AI Platform
- # will analyze files in the deployment_uri to determine a framework. If you
- # choose `SCIKIT_LEARN` or `XGBOOST`, you must also set the runtime version
- # of the model to 1.4 or greater.
- #
- # Do **not** specify a framework if you're deploying a [custom
- # prediction routine](/ml-engine/docs/tensorflow/custom-prediction-routines).
- #
- # If you specify a [Compute Engine (N1) machine
- # type](/ml-engine/docs/machine-types-online-prediction) in the
- # `machineType` field, you must specify `TENSORFLOW`
- # for the framework.
- "predictionClass": "A String", # Optional. The fully qualified name
- # (<var>module_name</var>.<var>class_name</var>) of a class that implements
- # the Predictor interface described in this reference field. The module
- # containing this class should be included in a package provided to the
- # [`packageUris` field](#Version.FIELDS.package_uris).
- #
- # Specify this field if and only if you are deploying a [custom prediction
- # routine (beta)](/ml-engine/docs/tensorflow/custom-prediction-routines).
- # If you specify this field, you must set
- # [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater and
- # you must set `machineType` to a [legacy (MLS1)
- # machine type](/ml-engine/docs/machine-types-online-prediction).
- #
- # The following code sample provides the Predictor interface:
- #
- # <pre style="max-width: 626px;">
- # class Predictor(object):
- # """Interface for constructing custom predictors."""
- #
- # def predict(self, instances, **kwargs):
- # """Performs custom prediction.
- #
- # Instances are the decoded values from the request. They have already
- # been deserialized from JSON.
- #
- # Args:
- # instances: A list of prediction input instances.
- # **kwargs: A dictionary of keyword args provided as additional
- # fields on the predict request body.
- #
- # Returns:
- # A list of outputs containing the prediction results. This list must
- # be JSON serializable.
- # """
- # raise NotImplementedError()
- #
- # @classmethod
- # def from_path(cls, model_dir):
- # """Creates an instance of Predictor using the given path.
- #
- # Loading of the predictor should be done in this method.
- #
- # Args:
- # model_dir: The local directory that contains the exported model
- # file along with any additional files uploaded when creating the
- # version resource.
- #
- # Returns:
- # An instance implementing this Predictor class.
- # """
- # raise NotImplementedError()
- # </pre>
- #
- # Learn more about [the Predictor interface and custom prediction
- # routines](/ml-engine/docs/tensorflow/custom-prediction-routines).
- "isDefault": True or False, # Output only. If true, this version will be used to handle prediction
- # requests that do not specify a version.
- #
- # You can change the default version by calling
- # projects.methods.versions.setDefault.
- "etag": "A String", # `etag` is used for optimistic concurrency control as a way to help
- # prevent simultaneous updates of a model from overwriting each other.
- # It is strongly suggested that systems make use of the `etag` in the
- # read-modify-write cycle to perform model updates in order to avoid race
- # conditions: An `etag` is returned in the response to `GetVersion`, and
- # systems are expected to put that etag in the request to `UpdateVersion` to
- # ensure that their change will be applied to the model as intended.
"serviceAccount": "A String", # Optional. Specifies the service account for resource access control.
- "errorMessage": "A String", # Output only. The details of a failure or a cancellation.
- "deploymentUri": "A String", # Required. The Cloud Storage location of the trained model used to
- # create the version. See the
- # [guide to model
- # deployment](/ml-engine/docs/tensorflow/deploying-models) for more
- # information.
- #
- # When passing Version to
- # projects.models.versions.create
- # the model service uses the specified location as the source of the model.
- # Once deployed, the model version is hosted by the prediction service, so
- # this location is useful only as a historical record.
- # The total number of model files can't exceed 1000.
- "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment.
- #
- # For more information, see the
- # [runtime version list](/ml-engine/docs/runtime-version-list) and
- # [how to manage runtime versions](/ml-engine/docs/versioning).
- "description": "A String", # Optional. The description specified for the version when it was created.
+ "runtimeVersion": "A String", # Required. The AI Platform runtime version to use for this deployment. For more information, see the [runtime version list](/ml-engine/docs/runtime-version-list) and [how to manage runtime versions](/ml-engine/docs/versioning).
+ "name": "A String", # Required. The name specified for the version when it was created. The version name must be unique within the model it is created in.
+ "manualScaling": { # Options for manually scaling a model. # Manually select the number of nodes to use for serving the model. You should generally use `auto_scaling` with an appropriate `min_nodes` instead, but this option is available if you want more predictable billing. Beware that latency and error rates will increase if the traffic exceeds that capability of the system to serve it based on the selected number of nodes.
+ "nodes": 42, # The number of nodes to allocate for this model. These nodes are always up, starting from the time the model is deployed, so the cost of operating this model will be proportional to `nodes` * number of hours since last billing cycle plus the cost for each prediction performed.
+ },
+ "state": "A String", # Output only. The state of a version.
+ "packageUris": [ # Optional. Cloud Storage paths (`gs://…`) of packages for [custom prediction routines](/ml-engine/docs/tensorflow/custom-prediction-routines) or [scikit-learn pipelines with custom code](/ml-engine/docs/scikit/exporting-for-prediction#custom-pipeline-code). For a custom prediction routine, one of these packages must contain your Predictor class (see [`predictionClass`](#Version.FIELDS.prediction_class)). Additionally, include any dependencies used by your Predictor or scikit-learn pipeline uses that are not already included in your selected [runtime version](/ml-engine/docs/tensorflow/runtime-version-list). If you specify this field, you must also set [`runtimeVersion`](#Version.FIELDS.runtime_version) to 1.4 or greater.
+ "A String",
+ ],
+ "pythonVersion": "A String", # Required. The version of Python used in prediction. The following Python versions are available: * Python '3.7' is available when `runtime_version` is set to '1.15' or later. * Python '3.5' is available when `runtime_version` is set to a version from '1.4' to '1.14'. * Python '2.7' is available when `runtime_version` is set to '1.15' or earlier. Read more about the Python versions available for [each runtime version](/ml-engine/docs/runtime-version-list).
}</pre>
</div>