Setup and execution routes
Begin with workspace acquisition, Python preparation, environment choices and troubleshooting.
Use first run for foundational commands. Modules 6, 7, and 13 use core regression dependencies. Modules 8–12 and 14 use the separate CPU framework environment.
Support and evidence
- Core execution: Python 3.12 on Windows/Linux and Python 3.13 on Windows.
- Framework execution: isolated Linux CPU actual training/inference/app checks. Use the Linux CPU reference for framework work.
- Notebooks: curated cells execute in fresh namespaces; interface and assistive-technology behavior need separate manual checks.
- Hosted routes: choose a provider that can access the complete source and supplies a matching runtime. Validate the selected module in that session; provider execution is outside the recorded CI platform scope.
- Reading: static explanations, tables, and actual verified reports; no browser Python execution is implied.
The original-source ZIP includes scripts, notebooks, shared implementation, and the app. A notebook alone lacks imports. External data/model caches are separate and retain provenance. If a download link does not open, use the offline access route.
Full site reproduction
The build requires reports from both notebook groups, retained in runs/verified. It fails rather than inventing missing results. Follow framework setup for the full sequence.
CI assembles complete-reading-site only after core/framework execution succeeds, then checks links, structure, archives, extracted source, and package building. Use the resulting artifact directly or serve it through your chosen hosting process. Review accessibility status before requiring an interface.