Movatterモバイル変換

MySQL 9.4 Reference Manual

27.3.10.8 Convenience Methods

The GenAI API provides the convenience methods described in this section under theml namespace. These methods act as wrappers for theLLM methods; rather than being invoked asLLM instance methods, they take the model ID as one of the options passed to them.ml.generate() andml.rag() return only the text portions of the values returned by their LLM counterparts.

ml.generate()

This method is a wrapper forLLM.generate(). It loads the model specified by themodel_id specified as one of theoptions, generates a response based on the prompt using this model, and returns the response. The defaultmodel_id ("llama3.2-3b-instruct-v1") is used if one is not specified. Like theLLM method,ml.generate() supports two variants, one for a single invocation, and one for batch processing.

Signature (single job)

String ml.generate(  Stringprompt,  Objectoptions)

Arguments

prompt (String): The desired prompt
options (Object) (default{}): The options employed for generation; these follow the same rules as the options used withLLM.generate()

Return type

String: The text of the response

Usage

//  Both invocations use "llama3.2-3b-instruct-v1" as the model_idlet response = ml.generate("What is Mysql?", {max_tokens: 10})let response = ml.generate("What is Mysql?", {model_id: "llama3.2-3b-instruct-v1", max_tokens: 10})

Signature (batch processing)

undefined ml.generate(  TableinputTable,   StringinputColumn,   StringoutputColumn,   Objectoptions)

Arguments

inputTable (Table): Table to use for operations
inputColumn (String): Name of column frominputTable to be embedded
outputColumn (String): Name of column in which to store embeddings; this can be either a fully-qualified name of a column or the name of the column only; in the latter case, the input table and its schema are used to construct the fully-qualified name
options (Object) (optional; default{}): An object containing the options used for embedding; see the description ofML_EMBED_ROW for available options

Return type

undefined

Usage

let schema = session.getSchema("mlcorpus")let table = schema.getTable("genai_table")ml.generate(table, "input", "mlcorpus.predictions.response", {max_tokens: 10})

ml.embed()

This method is a wrapper forLLM.embed(). Like theLLM method, it supports two variants, one for a single invocation, and one for batch processing.

Signature (single job)

Float32Array ml.embed(  Stringquery,  Objectoptions)

Arguments

query (String): Text of a natural-language query
options (Object) (default{}): The options employed for generation; these follow the same rules as the options used withLLM.embed(); the defaultmodel_id is"all_minilm_l12_v2"

Return type

Float32Array (MySQLVECTOR): The embedding

Usage

//  These produce the same resultlet embedding = ml.embed("What is Mysql?", {model_id: "all_minilm_l12_v2"})let embedding = ml.embed("What is Mysql?", {})

Signature (batch processing)

undefined ml.embed(  TableinputTable,  StringinputColumn,  StringoutputColumn,  Objectoptions)

Arguments

inputTable (Table): Table to use for operations
inputColumn (String): Name of column frominputTable to be embedded
outputColumn (String): Name of column in which to store embeddings; this can be either a fully-qualified name of a column or the name of the column only; in the latter case, the input table and its schema are used to construct the fully-qualified name
options (Object) (optional; default{}): An object containing the options used for embedding; see the description ofML_EMBED_ROW for available options

Return type

undefined

Usage

let schema = session.getSchema("mlcorpus")let table = schema.getTable("genai_table")ml.embed(table, "input", "mlcorpus.predictions.response",         {model_id: "all_minilm_l12_v2"})

ml.load()

This static method loads an existing (and already trained) MySQL HeatWave AutoML model having the name specified. An error is thrown if model with the given name does not exist.

Signature

```
Object ml.load(  Stringname)
```

Arguments

name (String): The name of the model.

Return type

Object: Any ofClassifier,Regressor,Forecaster,AnomalyDetector, orRecommender, depending on the type of model loaded.

For more information, seeML_MODEL_LOAD.

ml.rag()

This is a wrapper forLLM.rag(). Like theLLM method, it supports two variants, one for a single invocation, and one for batch processing.

Signature (single job)

String ml.rag(  Stringquery,  Objectoptions)

Arguments

query (String): Text of a natural-language query
options (Object) (default{}): The options employed for generation; these follow the same rules as the options used withLLM.rag(); the defaultmodel_id is"llama3.2-3b-instruct-v1"

Return type

String: Response text

Usage

//  These produce the same resultlet result = ml.rag("What is MySql?", {schema: ["vector_store"], n_citations: 1, model_options: {model_id: "llama3.2-3b-instruct-v1"}})let result = ml.rag("What is MySql?", {schema: ["vector_store"], n_citations: 1})

Signature (batch processing)

undefined ml.rag(  TableinputTable,  StringinputColumn,  StringoutputColumn,  Objectoptions)

Arguments

inputTable (Table): Table to use for operations
inputColumn (String): Name of column frominputTable to be embedded
outputColumn (String): Name of column in which to store embeddings; this can be either a fully-qualified name of a column or the name of the column only; in the latter case, the input table and its schema are used to construct the fully-qualified name
options (Object) (optional; default{}): An object containing the options used for embedding; see the description ofML_EMBED_ROW for available options

Return type

undefined

Usage

let schema = session.getSchema("mlcorpus")let table = schema.getTable("genai_table")ml.rag(table, "input", "mlcorpus.predictions.response",       {schema: ["vector_store"], n_citations: 1, model_options:       {model_id: "llama3.2-3b-instruct-v1"}});

PREV HOME UP NEXT