Open Steviey opened 1 year ago
Plot Animated Learning Curves!
There is a 'ready to go Python-code' to plot learning curves. How could we intigrate this into the 'reticulate version'?
https://github.com/awslabs/gluonts/issues/2220
I have done this with some other R-models already. H2o has a convenient API for this too.
It looks very cool with ggplot's...
myPlot<-myPlot + transition_reveal(Training_Rounds)
...and can offer so much insights into cv/training, possible under- and overfitting of a model.
I would do it myself- but need some guidance where and how to place and trigger the python function (callback) and collect the data into R.
callback_list<-list(modeltime.gluonts::TrainingHistory())
'TrainingHistory' is not an exported object from 'namespace:modeltime.gluonts'
Plot Animated Learning Curves!
There is a 'ready to go Python-code' to plot learning curves. How could we intigrate this into the 'reticulate version'?
https://github.com/awslabs/gluonts/issues/2220
I have done this with some other R-models already. H2o has a convenient API for this too.
It looks very cool with ggplot's...
myPlot<-myPlot + transition_reveal(Training_Rounds)...and can offer so much insights into cv/training, possible under- and overfitting of a model.
I would do it myself- but need some guidance where and how to place and trigger the python function (callback) and collect the data into R.
callback_list<-list(modeltime.gluonts::TrainingHistory())'TrainingHistory' is not an exported object from 'namespace:modeltime.gluonts'