LSTM-Hyperparameter-Optimization-Network-Traffic

Dependency

• keras
• pandas
• numpy
• scikit-learn
• scipy
• seaborn
• matplotlib
• yaml

Run Instructions

• To run the the code please run the following command from your project directory

$ python BO.py [link-number] [lstm-architecture]

To predict on link 3 using 1 layer LSTM:

$ python BO.py 3 L1

To predict on link 4 using 4 layer LSTM:

$ python BO.py 4 L4

• To run optimization on all 8 links for an architetcure please use the following command from your project directory

$ sh run.sh [lstm-architecture]

To run optimization for 2 layer LSTM:

$ sh run.sh L2

To run optimization for 4 layer LSTM:

$ sh run.sh L4

Contacts

Shahriar: [email protected]
Bashir: [email protected]

Acknowledgement

This project is a collaboration with Lawrence Berkeley National Laboratory under DOE Contract number for Deep Learning FP00006145 investigating Large-Scale Deep Learning for Intelligent Networks.