Hello all, I would like to use the tensorflow model analysis library (love this tool!). As I understand, tfma evaluation graph is generated at training time, what will be the right way to write the model when using transfer learning? How would I use the Apache beam distributed computing for that?
I would like to start with extending an available google model - transfer learning. Is there an example of how to use transfer learning, maybe inception model as a start, with Apache beam tft / tfma library / other TFX library to generate the model analysis Eival graph? Many thanks for any code snippets. Best, ___________ Eila www.orielresearch.org https://www.meetup.com/Deep-Learning-In-Production
