Optimizer Type
Namespace: DiffSharp.Optim
Assembly: DiffSharp.Core.dll
Base Type: obj
Constructors
| Constructor | Description |
Optimizer(model)
Full Usage:
Optimizer(model) Parameter Types:
Return Type: Optimizer Signature: (model: Model) -> Optimizer |
Parameter Types:
Optimizer
|
Instance members
| Instance member | Description |
this.model
Full Usage:
this.model Return Type: Model Signature: Model |
Return Type:
Model
|
this.step()
Full Usage:
this.step() Signature: unit -> unit |
|
this.updateRule arg1 arg2
Full Usage:
this.updateRule arg1 arg2 Parameter Types:
Return Type: Tensor Signature: string -> Tensor -> Tensor Modifiers: abstract |
Parameter Types:
Tensor
|
Static members
| Static member | Description |
Optimizer.adam(model, dataloader, loss, ?lr, ?beta1, ?beta2, ?eps, ?weightDecay, ?reversible, ?iters, ?epochs, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine)
Full Usage:
Optimizer.adam(model, dataloader, loss, ?lr, ?beta1, ?beta2, ?eps, ?weightDecay, ?reversible, ?iters, ?epochs, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine) Parameter Types:
Signature: (model: Model * dataloader: DataLoader * loss: (Tensor -> Tensor -> Tensor) * lr: Tensor * beta1: Tensor * beta2: Tensor * eps: Tensor * weightDecay: Tensor * reversible: bool * iters: int * epochs: int * threshold: double * print: bool * printEvery: int * printPrefix: string * printPostfix: string * printNewLine: bool) -> unit |
Parameter Types:
|
Optimizer.adam(f, x0, ?lr, ?beta1, ?beta2, ?eps, ?iters, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine)
Full Usage:
Optimizer.adam(f, x0, ?lr, ?beta1, ?beta2, ?eps, ?iters, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine) Parameter Types:
Return Type: Tensor * Tensor Signature: (f: (Tensor -> Tensor) * x0: Tensor * lr: Tensor * beta1: Tensor * beta2: Tensor * eps: Tensor * iters: int * threshold: double * print: bool * printEvery: int * printPrefix: string * printPostfix: string * printNewLine: bool) -> Tensor * Tensor |
Parameter Types:
Tensor * Tensor
|
Optimizer.sgd(model, dataloader, loss, ?lr, ?momentum, ?nesterov, ?weightDecay, ?reversible, ?iters, ?epochs, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine)
Full Usage:
Optimizer.sgd(model, dataloader, loss, ?lr, ?momentum, ?nesterov, ?weightDecay, ?reversible, ?iters, ?epochs, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine) Parameter Types:
Signature: (model: Model * dataloader: DataLoader * loss: (Tensor -> Tensor -> Tensor) * lr: Tensor * momentum: Tensor * nesterov: bool * weightDecay: Tensor * reversible: bool * iters: int * epochs: int * threshold: double * print: bool * printEvery: int * printPrefix: string * printPostfix: string * printNewLine: bool) -> unit |
Parameter Types:
|
Optimizer.sgd(f, x0, ?lr, ?momentum, ?nesterov, ?iters, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine)
Full Usage:
Optimizer.sgd(f, x0, ?lr, ?momentum, ?nesterov, ?iters, ?threshold, ?print, ?printEvery, ?printPrefix, ?printPostfix, ?printNewLine) Parameter Types:
Return Type: Tensor * Tensor Signature: (f: (Tensor -> Tensor) * x0: Tensor * lr: Tensor * momentum: Tensor * nesterov: bool * iters: int * threshold: double * print: bool * printEvery: int * printPrefix: string * printPostfix: string * printNewLine: bool) -> Tensor * Tensor |
Parameter Types:
Tensor * Tensor
|
