Python Environments

Simple, repeatable environment management for AI projects.

Consistent Python environments prevent silent dependency drift and reduce debug time.

Naming Convention

Use one environment per project or per major experiment track.

Example:

triwei-llm-core
triwei-diffusion-lab
triwei-rag-stack

Base Workflow

1. Create the environment.
2. Install only required dependencies.
3. Freeze versions after successful runs.
4. Record upgrades in a short changelog section.

Reset Rule

If dependency conflicts stack up, create a fresh environment instead of patching repeatedly.