support pathlib.Path paths in lightning API

[dyollb]

Proposed refactor

It would be nice if the api would support pathlib.Path paths.

methods like

• load_from_checkpoint
• save_checkpoint

and variables like

• default_root_dir (in Trainer)

Motivation

• pathlib.Path paths are concise, e.g. root_dir / "model.pt"
• this could be done via e.g. a Union[str, Path] or similar.

Pitch

Additional context

---

If you enjoy Lightning, check out our other projects! ⚡

• Metrics: Machine learning metrics for distributed, scalable PyTorch applications.

• Lite: enables pure PyTorch users to scale their existing code on any kind of device while retaining full control over their own loops and optimization logic.

• Flash: The fastest way to get a Lightning baseline! A collection of tasks for fast prototyping, baselining, fine-tuning, and solving problems with deep learning.

• Bolts: Pretrained SOTA Deep Learning models, callbacks, and more for research and production with PyTorch Lightning and PyTorch.

• Lightning Transformers: Flexible interface for high-performance research using SOTA Transformers leveraging PyTorch Lightning, Transformers, and Hydra.

cc @Borda

[awaelchli]

@dyollb Thanks for noticing this! We actually do support it 🎉 . load_from_checkpoint just needs a tiny update in the type annotation. See my changes in #15540 .
Trainer.save_checkpoint also supports it today.

Would you be interested in looking into Trainer.default_root_dir? Everything should be working there too, just needs a type update in the Trainer argument.