library("rplatypus")
history <- training_history(fit)
head(history)
# For a detector the loss is broken into four terms on purpose: a run whose
# coordinate loss falls while its objectness loss does not is finding the right
# places and refusing to commit, and the reverse is confident nonsense.The epoch-by-epoch record
Description
The epoch-by-epoch record
Usage
training_history(object)
Arguments
object
|
A platypus_fit().
|
Value
A data frame with one row per model per epoch.