lightning
https://github.com/lightning-ai/lightning
Python
Deep learning framework to train, deploy, and ship AI products Lightning fast.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported2 Subscribers
Add a CodeTriage badge to lightning
Help out
- Issues
- `dirpath` in `OnExceptionCheckpoint` should default to the same directory as `ModelCheckpoint`
- Address FSDP + manual optimization
- neptune.ai logger produces lots of errors when logging "training/epoch"
- `self.all_gather` used in `on_training_epoch_end` reports `RuntimeError`
- LightningCLI cannot use strategy and accelerator class instances.
- CUDA memory increase (caused CUDA OOM) when saving checkpoint at the `train_epoch_end`
- LightningModule.train_dataloader not being called
- Introduce sharded checkpointing for NO_SHARD FSDP.
- Ensure `torch.nn.Modules` in `DataModule` get moved to proper device
- Support `assign` argument in `LightningModule.load_from_checkpoint` method
- Docs
- Python not yet supported