Refactoring for Studios

This release begins the process of migrating to Lightning Studios :)

Update pre-commit configs and UI tools

This release addresses issued with pre-commit that originating in the config file and not the pip installed version.

Additional changes included opting for optuna as the default experiment manager, and streamlit + plotly as the default UI tools.

Implement Clean Template

This release implemented a clean template for users.

The previous version of lightning-pod with an end-to-end example is located at lightning-pod-example.

Implement Pure WANDB Sweep

this release introduces a 'pure' wandb Sweep; meaning Optuna is not used. additionally, the flows module from the previous release was placed in the components modules.

Add lightning_pod.flows

Added basic examples for hydra, wandb, and optuna+wandb in lightning_pod.flows/.

Implement PodTrainer

This release implemented a common PodTrainer in lightning_pod.core.trainer. Example usage is found in examples/.

New Lightning Import Conventions

This release updates the project to remove instances of import pytorch_lightning as pl in favor of import lightning as L and from lightning.pytorch import some_module_name.

0.0.5

add support for lightning-hpo and poetry

What's Changed

• Feature/make new by @JustinGoheen in #1
• Feature/hydra cli interop by @JustinGoheen in #2
• update with new functionality by @JustinGoheen in #3
• Enh/refactor network by @JustinGoheen in #4
• Feature/implement seed by @JustinGoheen in #5
• Improve CLI by @JustinGoheen in #6
• add type hints to core by @JustinGoheen in #7
• update pipeline type hints by @JustinGoheen in #8
• add type hints to cli by @JustinGoheen in #9
• add poetry by @JustinGoheen in #13

Full Changelog: https://github.com/JustinGoheen/lightning-pod/commits/0.0.5