Gemini321 / AutoML

A simple implementation of ENAS

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

AutoML

Deep Reinforcement Learning for Efficient Neural Architecture Search (ENAS) in PyTorch, i.e., AutoML. Code based on the paper https://arxiv.org/abs/1802.03268 and repo https://github.com/RualPerez/AutoML

How to run the pipeline:

  1. Clone the whole repository git clone https://github.com/Gemini321/AutoML.git

  2. Create virtualenv virtualenv -p /usr/bin/python3.7 virtualenv/AutoML

  3. Activate virtualenv source virtualenv/AutoML/bin/activate

  4. Install libraries pip3 install -r requirements.txt

  5. Run the main script, for instance: python3 main.py --num_episodes 5 --batch 5

Note that you can get a help of how to run the main script by: python3 main.py -h

Once the whole steps have run successfully, the next times you only need to run the last step 5.

Output: The main script saves the trained policy/controller net as policy.pt

File description

File / Folder Description
main.py Main script with runs the AutoML experiment
*.py An auxiliar script necessary to run main.py, its detailed description has been written as python documentation
Policy_Gradient_AutoML.ipynb Jupyter Notebook designed to help users to understand how this project has been developed
article.pdf Basic article that describes the principles of this project (theory-related). Here it can be found the results.
requirements.txt Version of the python libraries necessary to run the main script
images/ Images used for the article

About

A simple implementation of ENAS


Languages

Language:Jupyter Notebook 92.8%Language:Python 7.2%