tensorflow

Documentation

Index

• func CategorizeAttributes(trainStmt *ir.TrainStmt) (trainParams, validateParams, modelParams map[string]interface{})
• func DeriveFeatureColumnCodeAndFieldDescs(trainStmt *ir.TrainStmt) (featureColumnsCode []string, fieldDescs map[string][]*ir.FieldDesc, err error)
• func DocGenInMarkdown() string
• func Evaluate(stmt *ir.EvaluateStmt, session *pb.Session) (string, error)
• func Explain(stmt *ir.ExplainStmt, session *pb.Session) (string, error)
• func InitializeAttributes(trainStmt *ir.TrainStmt) error
• func IsPAI() bool
• func Pred(predStmt *ir.PredictStmt, session *pb.Session) (string, error)
• func Train(trainStmt *ir.TrainStmt, session *pb.Session) (string, error)

Functions

func CategorizeAttributes

func CategorizeAttributes(trainStmt *ir.TrainStmt) (trainParams, validateParams, modelParams map[string]interface{})

CategorizeAttributes returns attributes like train.*, validation.* and model.* to separated maps.

func DeriveFeatureColumnCodeAndFieldDescs

func DeriveFeatureColumnCodeAndFieldDescs(trainStmt *ir.TrainStmt) (featureColumnsCode []string, fieldDescs map[string][]*ir.FieldDesc, err error)

DeriveFeatureColumnCodeAndFieldDescs generates tensorflow feature column code and field descs from IR.

func DocGenInMarkdown

func DocGenInMarkdown() string

DocGenInMarkdown generates the doc of the XGBoost in Markdown format.

func Evaluate

func Evaluate(stmt *ir.EvaluateStmt, session *pb.Session) (string, error)

Evaluate generates a Python program to evaluate a trained model.

func Explain

func Explain(stmt *ir.ExplainStmt, session *pb.Session) (string, error)

Explain generates a Python program to explain a trained model.

func InitializeAttributes

func InitializeAttributes(trainStmt *ir.TrainStmt) error

InitializeAttributes initializes the attributes of TensorFlow and does type checking for them

func IsPAI

func IsPAI() bool

IsPAI tells if we are using PAI platform currently

func Pred

func Pred(predStmt *ir.PredictStmt, session *pb.Session) (string, error)

Pred generates a Python program for predict using a TensorFlow model.

func Train

func Train(trainStmt *ir.TrainStmt, session *pb.Session) (string, error)

Train generates a Python program for train a TensorFlow model.