Optimize

Main optimization module

Static members

Static memberDescription
Minimize(f, w0, par)
Signature: (f:(DV -> D) * w0:DV * par:Params) -> DV * D * DV [] * D []

Minimize vector-to-scalar function f, starting from initial parameter vector w0. Uses the optimization configuration given in par.
Minimize(f, w0)
Signature: (f:(DV -> D) * w0:DV) -> DV * D * DV [] * D []

Minimize vector-to-scalar function f, starting from initial parameter vector w0. Uses the default optimization configuration in Params.Default.
Train(f, w0, d, v, par)
Signature: (f:(DV -> DM -> DM) * w0:DV * d:Dataset * v:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d, and also monitoring the loss for the validation data given in dataset v. Uses the optimization configuration given in par.
Train(f, w0, d, v)
Signature: (f:(DV -> DM -> DM) * w0:DV * d:Dataset * v:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d, and also monitoring the loss for the validation data given in dataset v. Uses the default optimization configuration in Params.Default.
Train(f, w0, d, par)
Signature: (f:(DV -> DM -> DM) * w0:DV * d:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the optimization configuration given in par.
Train(f, w0, d)
Signature: (f:(DV -> DM -> DM) * w0:DV * d:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the default optimization configuration in Params.Default.
Train(f, w0, d, v, par)
Signature: (f:(DV -> DV -> DV) * w0:DV * d:Dataset * v:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d, and also monitoring the loss for the validation data given in dataset v. Uses the optimization configuration given in par.
Train(f, w0, d, v)
Signature: (f:(DV -> DV -> DV) * w0:DV * d:Dataset * v:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d, and also monitoring the loss for the validation data given in dataset v. Uses the default optimization configuration in Params.Default.
Train(f, w0, d, par)
Signature: (f:(DV -> DV -> DV) * w0:DV * d:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the optimization configuration given in par.
Train(f, w0, d)
Signature: (f:(DV -> DV -> DV) * w0:DV * d:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the default optimization configuration in Params.Default.
Train(f, w0, d, v, par)
Signature: (f:(DV -> DV -> D) * w0:DV * d:Dataset * v:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d and also monitoring the loss for the validation data given in dataset v. Uses the optimization configuration given in par.
Train(f, w0, d, v)
Signature: (f:(DV -> DV -> D) * w0:DV * d:Dataset * v:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d and also monitoring the loss for the validation data given in dataset v. Uses the default optimization configuration in Params.Default
Train(f, w0, d, par)
Signature: (f:(DV -> DV -> D) * w0:DV * d:Dataset * par:Params) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the optimization configuration given in par.
Train(f, w0, d)
Signature: (f:(DV -> DV -> D) * w0:DV * d:Dataset) -> DV * D * DV [] * D []

Train model function f, starting from initial parameter vector w0, by computing the loss for the training data given in dataset d. Uses the default optimization configuration in Params.Default

Basics
Examples