Closed semperparatusalpha closed 3 years ago
So far this is not possible as the application was designed to work with Keras models. The technique could, however, be extended to other platforms as well, I just currently do not have the time to do so.
If you would like to create a PR for R based models, here are some pointers:
To render the visualization, I get a model graph from the backend: https://github.com/viscom-ulm/Net2Vis/blob/305979027ef4a4eb1268e90012aaa7ef19e52b15/backend/server.py#L152-L153 This model graph has the following definition: https://github.com/viscom-ulm/Net2Vis/blob/master/backend/translate/graph.py and consists of multiple layers (https://github.com/viscom-ulm/Net2Vis/blob/master/backend/translate/layer.py), which the frontend is able to render.
Basically what you would need to do to make this possible is write a translate function, such as this one for Keras: https://github.com/viscom-ulm/Net2Vis/blob/305979027ef4a4eb1268e90012aaa7ef19e52b15/backend/translate/translate_keras.py#L9-L28 to translate your network into my Graph representation.
If you have any more questions feel free to ask.
Closing this as there has been not activity for a while, and I dont plan to integrate R support myself. Feel free to reopen.
Greetings,
Could the same visualization be run with R code?
Take care