mirror of
https://github.com/MLSysBook/TinyTorch.git
synced 2026-05-02 17:28:44 -05:00
a7fb897eed
- Enhanced tensor module documentation with mathematical foundations - Improved explanations for scalars, vectors, and matrices - Added NBGrader workflow documentation to activations module - Cleaned up .cursor/rules/ directory structure - Updated user preferences for better development workflow These changes improve the educational content and developer experience while maintaining the core functionality of all modules.
87 lines
4.4 KiB
Plaintext
87 lines
4.4 KiB
Plaintext
---
|
|
alwaysApply: false
|
|
---
|
|
|
|
|
|
|
|
# User Preferences and Conventions
|
|
|
|
## Development Preferences
|
|
|
|
- **Module naming**: Use `{module_name}_dev.py` for development files (e.g., `tensor_dev.py`)
|
|
- **Commit frequency**: Check code into repo at good milestone stages
|
|
- **README focus**: Keep main project README focused on overarching goals, not module details
|
|
- **Component terminology**: Call code components 'modules' instead of 'parts'
|
|
- **Testing framework**: Always use pytest for all tests - no manual testing or other frameworks
|
|
- **Brand formatting**: In headers and titles, use "Tiny🔥Torch" (with flame emoji after "Tiny"), not "🔥TinyTorch". In regular text, "TinyTorch" is acceptable
|
|
|
|
## Educational Workflow
|
|
|
|
- **Self-contained modules**: Each module under `modules/` should be self-contained
|
|
- **Local testing first**: Students run and pass all tests locally within the module before exporting
|
|
- **Progressive learning**: Start simple, build complexity gradually
|
|
- **Notebook conversion**: Prefer writing Python files with markdown syntax, convert to Jupyter afterwards
|
|
|
|
## File Organization
|
|
|
|
- **Source-first approach**: Start with Python source files, generate notebooks from them
|
|
- **Clean structure**: Remove experimental/temporary files before committing
|
|
- **Consistent patterns**: Follow established patterns across all modules
|
|
|
|
## NBDev Integration
|
|
|
|
- **Export workflow**: Use `#| export` directives to mark code for package export
|
|
- **Educational directives**: Use `#| hide` for instructor solutions, visible sections for student work
|
|
- **Package structure**: Code exports from `modules/` to `tinytorch/` package
|
|
|
|
## Testing Standards
|
|
|
|
- **pytest only**: All tests must use pytest framework with proper test classes
|
|
- **Comprehensive coverage**: Tests should cover functionality, edge cases, and integration
|
|
- **Professional structure**: Use test classes, fixtures, and descriptive test names
|
|
- **CLI integration**: Tests must be compatible with `tito module test --module {name}` (requires virtual environment activation)
|
|
- **Environment setup**: Always activate virtual environment with `source bin/activate-tinytorch.sh` before running commands
|
|
|
|
These preferences ensure consistency across the TinyTorch educational framework and smooth development workflow.
|
|
|
|
|
|
# User Preferences and Conventions
|
|
|
|
## Development Preferences
|
|
|
|
- **Module naming**: Use `{module_name}_dev.py` for development files (e.g., `tensor_dev.py`)
|
|
- **Commit frequency**: Check code into repo at good milestone stages
|
|
- **README focus**: Keep main project README focused on overarching goals, not module details
|
|
- **Component terminology**: Call code components 'modules' instead of 'parts'
|
|
- **Testing framework**: Always use pytest for all tests - no manual testing or other frameworks
|
|
- **Brand formatting**: In headers and titles, use "Tiny🔥Torch" (with flame emoji after "Tiny"), not "🔥TinyTorch". In regular text, "TinyTorch" is acceptable
|
|
|
|
## Educational Workflow
|
|
|
|
- **Self-contained modules**: Each module under `modules/` should be self-contained
|
|
- **Local testing first**: Students run and pass all tests locally within the module before exporting
|
|
- **Progressive learning**: Start simple, build complexity gradually
|
|
- **Notebook conversion**: Prefer writing Python files with markdown syntax, convert to Jupyter afterwards
|
|
|
|
## File Organization
|
|
|
|
- **Source-first approach**: Start with Python source files, generate notebooks from them
|
|
- **Clean structure**: Remove experimental/temporary files before committing
|
|
- **Consistent patterns**: Follow established patterns across all modules
|
|
|
|
## NBDev Integration
|
|
|
|
- **Export workflow**: Use `#| export` directives to mark code for package export
|
|
- **Educational directives**: Use `#| hide` for instructor solutions, visible sections for student work
|
|
- **Package structure**: Code exports from `modules/` to `tinytorch/` package
|
|
|
|
## Testing Standards
|
|
|
|
- **pytest only**: All tests must use pytest framework with proper test classes
|
|
- **Comprehensive coverage**: Tests should cover functionality, edge cases, and integration
|
|
- **Professional structure**: Use test classes, fixtures, and descriptive test names
|
|
- **CLI integration**: Tests must be compatible with `tito module test --module {name}` (requires virtual environment activation)
|
|
- **Environment setup**: Always activate virtual environment with `source bin/activate-tinytorch.sh` before running commands
|
|
|
|
These preferences ensure consistency across the TinyTorch educational framework and smooth development workflow.
|
|
|