Just to supplement that if the ML model is saved and then loaded within the same execution, calling load() will not cause the mentioned exception. Copying the model directory from UC volume to ephemeral storage attached to the driver node is also a w...
