Skip to content

Optional Local Setup (Conda + AI Help)

Use Binder for the smoothest course experience. Use local setup only if you prefer running notebooks on your own machine.

Binder vs Local

  • Use Binder when you want zero setup and fast onboarding.
  • Use Local Conda when you want persistent files, custom tooling, or offline edits.
  • Use Colab for Lab 4 (neural networks), where GPU support is helpful.

Local Setup (Step-by-Step)

  1. Install Miniconda: https://www.anaconda.com/docs/getting-started/miniconda/main
  2. Clone this repository: 
    git clone https://github.com/wfondrie/ushupo-ml-short-course.git
cd ushupo-ml-short-course
  3. Create/update the environment: 
    make env
  4. Activate the environment: 
    conda activate ./envs
  5. Launch Jupyter Lab: 
    make jupyter

Sanity Check

After Jupyter opens:

  1. Open notebooks/1_introduction.ipynb.
  2. Run all cells from a fresh kernel.
  3. Confirm there are no import errors and plots render correctly.

If a notebook needs project data, run:

make data

Troubleshooting Quick Notes

  • conda: command not found: close/reopen terminal after Miniconda install, then retry.
  • Wrong kernel/interpreter: in Jupyter, switch kernel to the env created from ./envs.
  • Binder is slow to start: wait for image build to finish, then refresh once.
  • Package mismatch after edits: rerun make env to resync dependencies from environment.yml.

Prompt Template for AI Tools

Use this when asking for setup help:

I am using the wfondrie/ushupo-ml-short-course repo. Please guide me step-by-step to run notebooks with Conda using make env, conda activate ./envs, and make jupyter. Assume I am a beginner. Ask me to run one command at a time and verify each output before moving on.