Lightning-AI/lightning issues and pull requests

#18690 - Enable passing `load_state_dict(..., assign=True|False)` in FabricModule

Pull Request - State: closed - Opened by awaelchli 12 months ago - 1 comment
Labels: feature, ready, fabric

#18689 - Add recommendation for matmul precision in docs

Pull Request - State: closed - Opened by awaelchli 12 months ago - 1 comment
Labels: ready, pl

#18688 - Pickle error when saving the lr scheduler defined after fabric.setup or by LambdaLR in Fabric

Issue - State: closed - Opened by hiyyg almost 1 year ago - 10 comments
Labels: bug, fabric, ver: 2.0.x, repro needed

#18687 - ModelCheckpoint Doesn't Delete Old Best Checkpoints When Resuming Training

Issue - State: open - Opened by danielzeng-gt almost 1 year ago - 2 comments
Labels: bug, callback: model checkpoint, ver: 1.9.x

#18686 - Maintain float32 precision at minimum in ResultMetric

Pull Request - State: closed - Opened by awaelchli almost 1 year ago - 2 comments
Labels: bug, ready, logging, pl

#18685 - Bump coverage from 7.3.0 to 7.3.1 in /requirements

Pull Request - State: closed - Opened by dependabot[bot] almost 1 year ago - 2 comments
Labels: ready, fabric, app, pl

#18684 - Bump httpx from 0.24.1 to 0.25.0 in /requirements

Pull Request - State: closed - Opened by dependabot[bot] almost 1 year ago - 2 comments
Labels: ready, app

#18681 - Bitsandbytes docs improvements

Pull Request - State: closed - Opened by carmocca about 1 year ago - 1 comment
Labels: ready, docs, fabric, pl

#18681 - Bitsandbytes docs improvements

Pull Request - State: closed - Opened by carmocca about 1 year ago - 1 comment
Labels: ready, docs, fabric, pl

#18680 - Avoid importing bitsandbytes unless requested

Pull Request - State: closed - Opened by carmocca about 1 year ago - 3 comments
Labels: ready, refactor, fabric, plugin

#18679 - Support combinations of precision plugins

Issue - State: open - Opened by carmocca about 1 year ago - 7 comments
Labels: feature, design, fabric, plugin, pl

#18677 - Set an upper limit on CPU threads in distributed training

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 5 comments
Labels: ready, fabric, strategy: ddp, performance, pl

#18677 - Set num threads

Pull Request - State: open - Opened by awaelchli about 1 year ago
Labels: fabric, pl

#18676 - fix `pydantic` compatibility for 2.0+ & allow new `fastAPI`

Pull Request - State: closed - Opened by Borda about 1 year ago - 2 comments
Labels: ready, priority: 1, app, pl

#18675 - intro granularity for docs build

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, docs, ci, fabric, app, pl

#18675 - intro granularity for docs build

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, docs, ci, fabric, app, pl

#18674 - App: Fix dispatch return value

Pull Request - State: closed - Opened by ethanwharris about 1 year ago - 1 comment
Labels: ready, app

#18674 - App: Fix dispatch return value

Pull Request - State: closed - Opened by ethanwharris about 1 year ago - 1 comment
Labels: ready, app

#18672 - Add a warning for problematic dataloader settings when `reload_dataloaders_every_n_epochs>0`

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: feature, ready, data handling, pl

#18671 - Forbid passing precision and a precision plugin

Pull Request - State: closed - Opened by carmocca about 1 year ago - 3 comments
Labels: ready, breaking change, fabric, plugin, pl

#18670 - `save_last: True` saves 2 checkpoints every time

Issue - State: closed - Opened by stas00 about 1 year ago - 6 comments
Labels: feature, help wanted, docs, callback: model checkpoint, repro needed

#18669 - LightningCLI + --trainer.gradient_clip_val XX doesn't work?

Issue - State: open - Opened by profPlum about 1 year ago
Labels: bug, needs triage, ver: 1.9.x

#18669 - LightningCLI + --trainer.gradient_clip_val XX doesn't work?

Issue - State: open - Opened by profPlum about 1 year ago
Labels: bug, needs triage, ver: 1.9.x

#18669 - LightningCLI + --trainer.gradient_clip_val XX doesn't work?

Issue - State: open - Opened by profPlum about 1 year ago - 1 comment
Labels: question, ver: 1.9.x

#18668 - PyTorch Profiler Stats Only Showing for "Records"

Issue - State: open - Opened by alexander-zhang about 1 year ago
Labels: bug, profiler, ver: 1.9.x

#18668 - PyTorch Profiler Stats Only Showing for "Records"

Issue - State: open - Opened by alexander-zhang about 1 year ago
Labels: bug, profiler, ver: 1.9.x

#18667 - Adding test for legacy checkpoint created with 2.0.9.post0

Pull Request - State: closed - Opened by pl-ghost about 1 year ago - 2 comments
Labels: checkpointing, tests, pl

#18667 - Adding test for legacy checkpoint created with 2.0.9.post0

Pull Request - State: closed - Opened by pl-ghost about 1 year ago - 2 comments
Labels: checkpointing, tests, pl

#18666 - docs: switch HTTPerror to custom runtime error

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, docs, ci, priority: 1, pl

#18666 - docs: switch HTTPerror to custom runtime error

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, docs, ci, priority: 1, pl

#18666 - docs: switch HTTPerror to custom runtime error

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, docs, ci, priority: 1, pl

#18665 - `torch.backends.cudnn.allow_tf32` is no longer on by default

Issue - State: closed - Opened by stas00 about 1 year ago - 4 comments
Labels: help wanted, docs, accelerator: cuda, performance, ver: 2.1.x

#18664 - Forbid non-FSDP precision plugins with FSDP

Pull Request - State: closed - Opened by carmocca about 1 year ago - 1 comment
Labels: ready, fabric, strategy: fsdp, pl

#18664 - Forbid non-FSDP precision plugins with FSDP

Pull Request - State: open - Opened by carmocca about 1 year ago - 1 comment
Labels: fabric, strategy: fsdp, pl

#18663 - CLI `--lr_scheduler` doesn't work?

Issue - State: open - Opened by profPlum about 1 year ago - 3 comments
Labels: bug, help wanted, docs, lightningcli, ver: 1.9.x

#18662 - releasing hotfix for Conda users [rebase & merge]

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci, priority: 1, release

#18662 - releasing hotfix for Conda users [rebase & merge]

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci, priority: 1, release

#18660 - App: Drop actions

Pull Request - State: closed - Opened by ethanwharris about 1 year ago - 2 comments
Labels: ready, app

#18660 - App: Drop actions

Pull Request - State: closed - Opened by ethanwharris about 1 year ago - 2 comments
Labels: ready, app

#18660 - App: Drop actions

Pull Request - State: closed - Opened by ethanwharris about 1 year ago - 2 comments
Labels: ready, app

#18659 - Call `wandb.finish` in the `WandbLogger.finalize`

Pull Request - State: closed - Opened by kunaltyagi about 1 year ago - 6 comments
Labels: logger: wandb, community, pl

#18659 - Call `wandb.finish` in the `WandbLogger.finalize`

Pull Request - State: closed - Opened by kunaltyagi about 1 year ago - 6 comments
Labels: logger: wandb, community, pl

#18657 - Updated check on model step output types from dict to Mapping

Pull Request - State: closed - Opened by nik777 about 1 year ago - 1 comment
Labels: ready, lightningmodule, community, pl

#18657 - Updated check on model step output types from dict to Mapping

Pull Request - State: open - Opened by nik777 about 1 year ago - 1 comment
Labels: ready, lightningmodule, community, pl

#18656 - Changed check on model step output types from dict to Mapping

Pull Request - State: closed - Opened by nik777 about 1 year ago
Labels: pl

#18655 - Bitsandbytes precision plugin

Pull Request - State: closed - Opened by carmocca about 1 year ago - 3 comments
Labels: feature, ready, fabric, experimental, plugin, pl

#18655 - Bitsandbytes precision plugin

Pull Request - State: open - Opened by carmocca about 1 year ago - 1 comment
Labels: feature, fabric, experimental, plugin, pl

#18654 - ci: timeouts for docs build

Pull Request - State: closed - Opened by Borda about 1 year ago - 2 comments
Labels: ready, docs, ci, priority: 1, dependencies

#18653 - [WIP] Downgrade the panel dependency

Pull Request - State: open - Opened by awaelchli about 1 year ago - 1 comment
Labels: docs, app

#18653 - [WIP] Downgrade the panel dependency

Pull Request - State: open - Opened by awaelchli about 1 year ago - 1 comment
Labels: docs, app

#18653 - [WIP] Downgrade the panel dependency

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: docs, app

#18652 - Provide a non-expert mode for the DeviceStatsMonitor

Issue - State: open - Opened by awaelchli about 1 year ago
Labels: feature, help wanted, callback: device stats

#18651 - Set the generator seed in `random_split` usages in the examples

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: ready, docs, data handling, fabric, lightningdatamodule, pl

#18651 - Set the generator seed in `random_split` usages in the examples

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: ready, docs, data handling, fabric, lightningdatamodule, pl

#18650 - Why does running Lightning on SLURM with python perform worse than with srun?

Issue - State: open - Opened by Muennighoff about 1 year ago - 6 comments
Labels: help wanted, question, environment: slurm, ver: 2.0.x

#18649 - Update `persistent_workers` recommendation when using spawn launcher

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 2 comments
Labels: ready, docs, data handling, performance, pl

#18649 - Update `persistent_workers` recommendation when using spawn launcher

Pull Request - State: open - Opened by awaelchli about 1 year ago - 2 comments
Labels: docs, data handling, performance, pl

#18648 - Model trained with Deepspeed stage 3 shape not match when loading

Issue - State: open - Opened by yinweisu about 1 year ago - 1 comment
Labels: bug, strategy: deepspeed, ver: 2.0.x, repro needed

#18648 - Model trained with Deepspeed stage 3 shape not match when loading

Issue - State: open - Opened by yinweisu about 1 year ago
Labels: bug, needs triage, ver: 2.0.x

#18647 - Fix KeyError when calling `Fabric.load_raw` before setting up an FSDP model

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: bug, ready, fabric, strategy: fsdp

#18646 - LightningModule.log_dict: fix type hints for MetricCollection

Pull Request - State: closed - Opened by adamjstewart about 1 year ago - 5 comments
Labels: ready, code quality, community, pl

#18646 - LightningModule.log_dict: fix type hints for MetricCollection

Pull Request - State: open - Opened by adamjstewart about 1 year ago - 5 comments
Labels: ready, code quality, community, pl

#18645 - Spurious validation step when restarting with a checkpoint when `max_steps` is set in the trainer

Issue - State: open - Opened by arnaudstiegler about 1 year ago
Labels: bug, help wanted, loops, ver: 2.0.x

#18645 - Spurious validation step when restarting with a checkpoint when max_steps is set in the trainer

Issue - State: open - Opened by arnaudstiegler about 1 year ago
Labels: bug, needs triage, ver: 2.0.x

#18645 - Spurious validation step when restarting with a checkpoint when max_steps is set in the trainer

Issue - State: open - Opened by arnaudstiegler about 1 year ago
Labels: bug, needs triage, ver: 2.0.x

#18644 - Module variable is initialized twice

Issue - State: open - Opened by mfoglio about 1 year ago
Labels: bug, needs triage, ver: 1.7.x

#18644 - Module variable is initialized twice

Issue - State: open - Opened by mfoglio about 1 year ago
Labels: bug, needs triage, ver: 1.7.x

#18644 - Module variable is initialized twice

Issue - State: closed - Opened by mfoglio about 1 year ago - 1 comment
Labels: question, ver: 1.7.x

#18643 - switch download UI from GCP to S3

Pull Request - State: open - Opened by Borda about 1 year ago - 1 comment
Labels: ci

#18643 - switch download UI from GCP to S3

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci

#18642 - Introduce Cache 1/n

Pull Request - State: closed - Opened by tchaton about 1 year ago - 3 comments
Labels: ready, ci, dependencies

#18641 - Incorrect type hints for `log_dict(MetricCollection())`?

Issue - State: closed - Opened by adamjstewart about 1 year ago - 3 comments
Labels: bug, help wanted, good first issue, code quality, ver: 2.0.x

#18641 - Incorrect type hints for `log_dict(MetricCollection())`?

Issue - State: closed - Opened by adamjstewart about 1 year ago - 3 comments
Labels: bug, help wanted, good first issue, code quality, ver: 2.0.x

#18640 - Investigate Resident Memory Increase during Inference

Issue - State: open - Opened by ZekunZh about 1 year ago - 3 comments
Labels: bug, help wanted, performance, ver: 2.0.x

#18640 - Investigate Resident Memory Increase during Inference

Issue - State: open - Opened by ZekunZh about 1 year ago - 2 comments
Labels: bug, help wanted, performance, ver: 2.0.x

#18638 - Tutorial DataModule leads to Validation data leakage in Multiprocessing

Issue - State: closed - Opened by profPlum about 1 year ago - 12 comments
Labels: docs, lightningdatamodule

#18637 - ci: prune install AWS CLI

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci

#18637 - ci: prune install AWS CLI

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci

#18637 - ci: prune install AWS CLI

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci

#18635 - use `-m build` instead of particular cmds

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ci

#18634 - Remove outdated workaround for PyTorch autocast bug

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 1 comment
Labels: ready, refactor, fabric, pl

#18633 - Truncating Batches is Dangerous

Issue - State: open - Opened by profPlum about 1 year ago - 1 comment
Labels: question, ver: 1.9.x

#18633 - Truncating Batches is Dangerous

Issue - State: closed - Opened by profPlum about 1 year ago - 3 comments
Labels: question, ver: 1.9.x

#18632 - Relax constraints on jsonargparse dependency

Pull Request - State: closed - Opened by adamjstewart about 1 year ago - 1 comment
Labels: ready, lightningcli, community, pl, dependencies

#18632 - Relax constraints on jsonargparse dependency

Pull Request - State: closed - Opened by adamjstewart about 1 year ago - 1 comment
Labels: ready, lightningcli, community, pl, dependencies

#18631 - Fsspec fix find checkpoints

Pull Request - State: closed - Opened by mraxcel about 1 year ago
Labels: pl

#18630 - Avoid redundant input-type casting in FSDP precision

Pull Request - State: closed - Opened by awaelchli about 1 year ago - 2 comments
Labels: bug, ready, fabric, strategy: fsdp, pl

#18629 - enable codespell for docs & fixing +TPU

Pull Request - State: open - Opened by Borda about 1 year ago - 2 comments
Labels: ci, fabric, app, pl

#18629 - enable codespell for docs & fixing +TPU

Pull Request - State: closed - Opened by Borda about 1 year ago - 2 comments
Labels: ready, ci, fabric, app, pl

#18628 - ci/groupcheck: fix TPU parameters [fabric]

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci, priority: 1

#18628 - ci/groupcheck: fix TPU parameters [fabric]

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci, priority: 1

#18628 - ci/groupcheck: fix TPU parameters [fabric]

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ready, ci, priority: 1

#18627 - ci: fix installing env. for Flag apps

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ci

#18627 - ci: fix installing env. for Flag apps

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ci

#18627 - ci: fix installing env. for Flag apps

Pull Request - State: closed - Opened by Borda about 1 year ago - 1 comment
Labels: ci

#18626 - DataLoader Limited to 2 CPUs leading to bottleneck in training

Issue - State: open - Opened by Andres-Carranza about 1 year ago - 2 comments
Labels: bug, performance, ver: 2.0.x

#18626 - DataLoader Limited to 2 CPUs leading to bottleneck in training

Issue - State: open - Opened by Andres-Carranza about 1 year ago - 1 comment
Labels: bug, performance, ver: 2.0.x

#18626 - DataLoader Limited to 2 CPUs leading to bottleneck in training

Issue - State: open - Opened by Andres-Carranza about 1 year ago - 4 comments
Labels: bug, performance, ver: 2.0.x

#18625 - Bump playwright from 1.36.0 to 1.38.0 in /requirements

Pull Request - State: open - Opened by dependabot[bot] about 1 year ago - 1 comment
Labels: ci, app

#18625 - Bump playwright from 1.36.0 to 1.38.0 in /requirements

Pull Request - State: closed - Opened by dependabot[bot] about 1 year ago - 1 comment
Labels: ready, ci, app

GitHub / Lightning-AI/lightning issues and pull requests