To do:
- Specify CNN model (David)
- Create hyperoptimization script, export best model as TFModel (David)
- Run hyperopt script on Hops (?)
- Look into Hops dataset file reader error (Tobias)
- Reading the GET data from JSON file of url's into suitable format for Spark (Simon)
- See how TensorFlowOnSpark uses a trained model for inferencing, TFCluster.inference(dataRDD) (Tobias)
- Spark vs Spark Streaming vs Spark Structured Streaming for this project (Simon)