Repository for the deep learning project.
The aim of this project is to find answers and solutions to the following questions and tasks related to Hyper Parameter Optimization (HPO) and Neural Architecture Search (NAS).
- How well can traditional HPO optimizers like SMAC perform for NAS. Can we identify under what conditions and properties it can outperform other methods provided by NASLib?
- Main Task: Identify where HPO based NAS is strongest, using SMAC on the NASBench201 search space for CIFAR10. Use DeepCave to analyse and Highlight these strengths.
Other questions raised during the project will also be discussed.
Install python 3.9 inside of a virtual environment (Conda recommended)
Clone NASLib inside this directory with the following command:
git clone https://github.com/automl/NASLib.git
Switch to the NASLib directory that you just cloned and follow the setup guide of NASLib:
https://github.com/automl/NASLib
Preferably switch to the Develop tree inside of NASLib