The structure of the code is as follows and everything is in CS 6140:
- conv1d contains the CNN experiment code and logs
- gru contains the GRU experiment code and logs
- lstm contains the vanilla LSTM experiment code and logs
- mini and ml contain the rest experiment code and logs
To reproduce the results and run the code, you might need to pull the original dataset from the github repo
NSDI 19: Is advance knowledge of flow sizes a plausible assumption?