Konstantin T
Konstantin T
It is worth considering the automatic generation of tests for created services. This will improve the quality of services, perhaps launching tests can be made a separate stage of validation....
Checklist: - [x] Sklearn - [x] XGBoost - [x] CatBoost - [x] LightGBM - [ ] LightAutoML - [ ] Torch - [ ] Tensorflow - [ ] Fedot -...
Make basic visualization, first step: check LightAutoML reports. CC: @pacifikus
To improve the quality of the generated services, you can perform fuzzing testing of the API for the generated services. Perhaps this should be done as a separate step for...
In some cases it will be convenient to use python `wheels`/`tar.gz` as a distribution package, it is necessary to support this build capability. CC: @pacifikus
Template validation with mypy allows you to statically check the correctness of types in the template for backends, but also creates complications. For example, it creates additional dependencies for validation....
Research for speedup test-suite running possibilities. (maybe local and/or CI) You can check: https://github.com/cleder/awesome-python-testing CC: @pacifikus
Mac os runners doesn't have `docker` from the box. We can use https://github.com/docker-practice/actions-setup-docker, but i see very slow docker-setup job for mac runners, for example in job: https://github.com/qnbhd/deployme/actions/runs/3059827694/jobs/4937780059 we can...