Repository Issue Activity (beta)

#21809 Add an end-to-end anomaly detection tutorial using PyTorch Lightning

#21808 `accumulate_grad_batches` is silently ignored when epoch length < `accumulate_grad_batches`, because Lightning always steps on the last batch

#21797 Support remote/cloud storage (fsspec) for ModelParallelStrategy checkpointing

#21804 SIGSEGV (Segmentation Fault) when saving checkpoints with FSDP on CPU

#21795 Graceful stop on an extrenal file marker

#21800 Checkpoints silently fail to save due to swallowed PermissionError in `_atomic_save`

#21798 Community eval tip

#21771 Profiler misses dataloader worker initialization (and outdated docs)

#21778 FSDPStrategy fails on CPU with RuntimeError: FSDP needs a non-CPU accelerator device when validation is bypassed

#21787 WandbLogger marks crashed runs as "Finished" instead of "Failed" in W&B sweeps

#21589 Feature Request : Gradient Statistics Monitoring Callback

#21724 EMA validation swap models before first update_parameters

#21779 Support remote/cloud storage (fsspec) for FSDP checkpointing

#21548 Cannot make test case unused parameters for proposed strategy in PR

#18740 manual_backward and .backward() have different behaviour.

#17212 PyTorch Lightning DDP crashes with unused parameters

#9136 AdvancedProfiler: ValueError: Attempting to stop recording an action (run_test_evaluation) which was never started.

#12119 Use :emphasize-lines: in sphinx docs to highlight code.

#21554 Set experiment tags in MLflow

#21433 MLFlow Logger Tests Not Passing

#21437 MLflow model format required for the Register Model

#21617 CometLogger does not write logs in offline mode

#21694 Lightning project has been quarantined on pypi

#21782 Log key prefix controls for Trainer, WandbLogger, and CometLogger

#21491 Add interval: "batch" option for learning rate schedulers