Deep.Net


TrainingLog

Namespace: Models

Training history module.

Nested types and modules

TypeDescription
Entry
Log<'P>

Functions and values

Function or valueDescription
best log
Signature: log:Log<'?175866> -> (Entry * Tensor<'?175866>) option
Type parameters: '?175866
bestIter log
Signature: log:Log<'?175864> -> int
Type parameters: '?175864
create minImprovement bestOn
Signature: minImprovement:float -> bestOn:Partition -> Log<'?175857>
Type parameters: '?175857
itersWithoutImprovement log
Signature: log:Log<'?175868> -> int
Type parameters: '?175868
lastIter log
Signature: log:Log<'?175862> -> int
Type parameters: '?175862
record entry parVals log
Signature: entry:Entry -> parVals:Tensor<'?175860> -> log:Log<'?175860> -> Log<'?175860>
Type parameters: '?175860
relevantLoss bestOn entry
Signature: bestOn:Partition -> entry:Entry -> float
removeToIter iter log
Signature: iter:int -> log:Log<'?175870> -> Log<'?175870>
Type parameters: '?175870
Fork me on GitHub