6.887 Machine Learning for Systems

Lab 2: Smart Caching using Bayesian Optimization

Due: Friday, Oct 8 at 12:00 PM

Instructions:

Step 1

Clone this repository with this command:

git clone https://github.com/jialinding/6.887Lab2.git
cd 6.887Lab2

Step 2

If your Python version is at least 3.8, then you can directly use pip to install the required packages:

python3 -m pip install jupyter numpy scikit-optimize cloudpickle

However, if your Python version is less than 3.8, or if you prefer to use an environment, use the following commands to create a conda environment and install required packages

conda create --name lab2_mlforsys python=3.8
conda activate lab2_mlforsys
conda install jupyter numpy cloudpickle
conda install -c conda-forge scikit-optimize

Step 3

Launch jupyter notebook:

jupyter notebook

If you're having trouble with importing skopt in the first cell, make sure your scikit-learn version is scikit-learn==0.24.2. Apparently scikit-learn v1.0 is causing issues with skopt.

Step 4

Once completed, upload your notebook through Canvas. If you do not have access to Canvas, email your completed notebook directly to Jialin ([email protected]) and Pouya ([email protected]).