Various Experiments, possibly ML related
Prototyping to tested code
How can pytest be used in Jupyter notebooks? And why does it make sense? This talk discusses how Jupyter notebooks form an effective environment for prototyping and how code can be refactored code into modules. A particular emphasis is placed on testing and the use of ipytest.
Causality and function approximations
How do causal analysis and function approximations interact? This blog post demonstrates how results for tabular methods break down for models with finite capacity.
Keyword detection from scratch
Detecting keywords in speech from data-collection to modelling, includes code to listen continuously for commands.
- chmp: support code as a python package
This project uses pipenv to organize dependencies and common scripts. To setup a virtual environment with all requirements use:
pipenv sync --dev
After that the following tasks can be performed:
# run all pre-commit tasks (docs, formatting, tests) pipenv run precommit # run pre-commit tasks and integration tests pipenv run precommit-full # run notebook integration tests pipenv run integration # run tests pipenv run test # update the documentation pipenv run docs