How to load a .rda neuralnetwork model of R in python

Question

I have done a binary classification model using neuralnet() function in R. An example code is here . I have done same network using python keras , which has less prediction accuracy due to its parameters adjustments. So i have saved the R's neural network model using save(model_nnet, file = "/tmp/model_nnet.rda") and loaded using load(file = "/tmp/model_nnet.rda") .

My question is can i load this file in python? and use it for some predicton?

Answer 1

RData files are R's storage format for R objects, and so can contain complex data structures and metadata. This is not easy to translate into data formats for other languages.

You will have to load the RData file into R and write the components of the object you need to text files using other packages - for example you could write it to a JSON file using one of R's JSON packages, and then read it in using a Python JSON module. But you will have to manually extract the parameters and numbers and so o from your model fit and construct an object that can be written to JSON. Its possible someone has already done this for your class of object.