Obtain and navigate the student workspace
Use a repository checkout or the complete source archive supplied for your class. Running the downloaded course locally does not require a GitHub account.
If a download link does not open
Lesson pages link to downloads on the reading site. If one of those links does not open, for example while the site is being updated or when you are offline, nothing is wrong with your computer; use the route below.
- Get the complete course source. Download it from the repository with the download route below, or ask your instructor for the course source archive,
ml_ai_it-source.zip. Extract it and keep its folder structure. - Use the in-source copy of each download.
| Link on a lesson page | Where it is in the source, or how to produce it |
|---|---|
| Complete source archive | The archive from step 1 |
| Module notebook | notebooks/lessons/<module folder>/, for example notebooks/lessons/05-regression/regression.ipynb |
| Practice or solutions notebook | notebooks/practice/moduleNN-practice.ipynb and moduleNN-solutions.ipynb |
Worked report (example/index.html) |
Run the module's command; open index.html in the output folder you named |
| Executed notebook copy | Open the source notebook, restart the kernel and run all cells |
| Sample data and schemas | data/samples/ and data/schemas/ |
| Attributed data cache | The module's download command saves the same files under runs/datasets/ and checks them against the reviewed checksums |
| Capstone notebook or report | projects/phishing-triage/notebooks/complete-study.ipynb; run it to produce the report |
Instructors can build the archive from a checkout with python scripts/build_downloads.py --source-only ml_ai_it-source.zip. It contains only the published source allowlist, with no run outputs, caches or environments.
Download route
No GitHub account is needed.
- Open the repository.
- Choose Code, then Download ZIP. Extract the archive; opening the compressed view is insufficient.
- Open the extracted folder containing README.md, pyproject.toml, examples and data. Its download name may include main; commands work from that folder regardless of its name.
- On Windows, open Terminal/PowerShell in that folder. On Linux/macOS, open a terminal and use cd followed by the folder path. Quote paths containing spaces.
- Run the folder check below, then follow first run.
Windows PowerShell:
Get-Location
Get-ChildItem
py --version
py examples/first-run/scripts/first_run.py --storage-percent 85 --output runs/module01/first-run.txt
Linux/macOS:
pwd
ls
python3 --version
python3 examples/first-run/scripts/first_run.py --storage-percent 85 --output runs/module01/first-run.txt
Expected result: Review storage. Find runs/module01/first-run.txt in the same extracted project. Change 85 to 70, save to another output filename and explain the changed decision.
Git route
If you already use Git, clone https://github.com/ntious/ml_ai_it.git, then change into ml_ai_it. Downloaded archives do not have Git history and do not require Git for learning.
Open a notebook
Install the core environment and notebook interface using first run. Launch the interface with that environment's Python. In the browser file list, open notebooks/lessons and the selected module. Run cells in order with Shift+Enter, save with Ctrl+S or the File menu, then restart the kernel and run all. Download your saved notebook and results before leaving an ephemeral hosted session.
Reading reports from GitHub
The reading-site address hosts the lesson pages, worked reports, executed notebook copies and the complete source archive. Every page is built from checked source by the course's automated workflow.
If the site is unavailable, use the offline access route. Maintainers and reviewers can also download the complete-reading-site artifact from a successful Learning resource checks run under Actions, extract it and open index.html. That fallback may require GitHub authentication and artifacts have limited retention. Running the scripts locally produces the same report format from real execution.
Do not build the entire site just to study one module. Site assembly is a maintainer workflow requiring both notebook execution profiles. See troubleshooting for common errors.