Some code refactoring and clean-up to remove dependency on Keras and TensorFlow
from studioml code proper. Examples and unit tests still depend on these modules of course.
Also clean up and fixes for unit tests, focusing on setups we are actually using -
namely, local runner mode and sending workloads to external executor through RMQ queues.
Some code refactoring and clean-up to remove dependency on Keras and TensorFlow from studioml code proper. Examples and unit tests still depend on these modules of course. Also clean up and fixes for unit tests, focusing on setups we are actually using - namely, local runner mode and sending workloads to external executor through RMQ queues.