List view

Add logger state dumping and restoring

checkpointingRelated to checkpointingRelated to checkpointing

communityThis PR is from the communityThis PR is from the community

featureIs an improvement or enhancementIs an improvement or enhancement

has conflicts

loggerRelated to the LoggersRelated to the Loggers

plGeneric label for PyTorch Lightning packageGeneric label for PyTorch Lightning package

Status: Draft (not ready).

Lightning-AI/pytorch-lightningnumber 13069

#13069 In Lightning-AI/pytorch-lightning;

· ashleve opened on May 13, 2022

Requirement files do not correspond to extra dependencies

ciContinuous IntegrationContinuous Integration

help wantedOpen to be worked onOpen to be worked on

Status: Open.

#15210 In Lightning-AI/pytorch-lightning;

· otaj opened on Oct 20, 2022

LightningModule self.log add_dataloader_idx doesn't reduce properly the metric across dataloaders

bugSomething isn't workingSomething isn't working

priority: 1Medium priority taskMedium priority task

Status: Open.

#11126 In Lightning-AI/pytorch-lightning;

· tchaton opened on Dec 17, 2021

check_val_every_n_epoch bug with list of dataloaders

bugSomething isn't workingSomething isn't working

loopsRelated to the Loop APIRelated to the Loop API

Status: Open.

#12145 In Lightning-AI/pytorch-lightning;

· TrentBrick opened on Feb 28, 2022

DeepSpeed breaks learning rate scheduler checkpointing

bugSomething isn't workingSomething isn't working

lr scheduler

repro neededThe issue is missing a reproducible exampleThe issue is missing a reproducible example

strategy: deepspeed

Status: Open.

#12576 In Lightning-AI/pytorch-lightning;

· gahdritz opened on Apr 1, 2022

SimpleProfiler not working on windows

bugSomething isn't workingSomething isn't working

profiler

Status: Open.

#12998 In Lightning-AI/pytorch-lightning;

· rohitgr7 opened on May 6, 2022

lr_find crashes with Deepspeed

bugSomething isn't workingSomething isn't working

strategy: deepspeed

tuner

Status: Open.

#13370 In Lightning-AI/pytorch-lightning;

· ant-louis opened on Jun 22, 2022

Multiple GPU per node could fail silently with KubeflowEnvironment

bugSomething isn't workingSomething isn't working

environment: kubeflow

help wantedOpen to be worked onOpen to be worked on

Status: Open.

#13374 In Lightning-AI/pytorch-lightning;

· RamtinRassoli opened on Jun 22, 2022

test rule in Makefile not working

bugSomething isn't workingSomething isn't working

ciContinuous IntegrationContinuous Integration

Status: Open.

#13468 In Lightning-AI/pytorch-lightning;

· CyprienRicque opened on Jun 30, 2022

<code>RichProgressBar</code> in v1.6 is slower than v1.5

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

performance

plGeneric label for PyTorch Lightning packageGeneric label for PyTorch Lightning package

priority: 1Medium priority taskMedium priority task

progress bar: rich

Status: Open.

#13937 In Lightning-AI/pytorch-lightning;

· akihironitta opened on Jul 29, 2022

Store logger experiment id in checkpoint to enable correct resuming of experiments

checkpointingRelated to checkpointingRelated to checkpointing

featureIs an improvement or enhancementIs an improvement or enhancement

help wantedOpen to be worked onOpen to be worked on

loggerRelated to the LoggersRelated to the Loggers

Status: Open.

#5342 In Lightning-AI/pytorch-lightning;

· ashleve opened on Jan 3, 2021

Datapipe handling with more than 1 GPU gives torch.multiprocessing.spawn.ProcessExitedException: process 1 terminated with exit code 1

bugSomething isn't workingSomething isn't working

data handlingGeneric data-related topicGeneric data-related topic

strategy: ddpDistributedDataParallelDistributedDataParallel

Status: Open.

#13039 In Lightning-AI/pytorch-lightning;

· franchesoni opened on May 11, 2022

ModelPruning callback logs wrong sparsity

bugSomething isn't workingSomething isn't working

callback: pruning

help wantedOpen to be worked onOpen to be worked on

priority: 2Low priority taskLow priority task

Status: Open.

#13595 In Lightning-AI/pytorch-lightning;

· SungFeng-Huang opened on Jul 11, 2022

TorchScript export incompatible with DeviceDtypeModuleMixin properties

bugSomething isn't workingSomething isn't working

lightningmodulepl.LightningModulepl.LightningModule

priority: 2Low priority taskLow priority task

Status: Open.

#13887 In Lightning-AI/pytorch-lightning;

· mattm458 opened on Jul 27, 2022

Async checkpointing race conditions

bugSomething isn't workingSomething isn't working

checkpointingRelated to checkpointingRelated to checkpointing

tests

Status: Open.

#14035 In Lightning-AI/pytorch-lightning;

· awaelchli opened on Aug 5, 2022

<code>self.log</code> raised error when number of dataloader is not consistent

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

loggingRelated to the `LoggerConnector` and `log()`Related to the `LoggerConnector` and `log()`

Status: Open.

#16431 In Lightning-AI/pytorch-lightning;

· ding3820 opened on Jan 19, 2023

DDP PackedSequence

bugSomething isn't workingSomething isn't working

data handlingGeneric data-related topicGeneric data-related topic

Status: Open.

#16397 In Lightning-AI/pytorch-lightning;

· stsouko opened on Jan 17, 2023

Tuner starts from validation even though num_sanity_val_steps is set to 0

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

tuner

Status: Open.

#16488 In Lightning-AI/pytorch-lightning;

· fierval opened on Jan 24, 2023

<code>configure_optimizers</code> in LightningCLI doesn't work in distributed mode

featureIs an improvement or enhancementIs an improvement or enhancement

help wantedOpen to be worked onOpen to be worked on

lightningclipl.cli.LightningCLIpl.cli.LightningCLI

strategy: ddpDistributedDataParallelDistributedDataParallel

Status: Open.

#16489 In Lightning-AI/pytorch-lightning;

· stevenmanton opened on Jan 24, 2023

Validation has an impact on training

bugSomething isn't workingSomething isn't working

reproducibility

Status: Open.

#16827 In Lightning-AI/pytorch-lightning;

· valentinp72 opened on Feb 21, 2023

Incompatibility with MLFlow 1.30

3rd partyRelated to a 3rd-partyRelated to a 3rd-party

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

logger: mlflow

Status: Open.

#17084 In Lightning-AI/pytorch-lightning;

· whrichd opened on Mar 14, 2023

Loss does not decrease on single TPU

accelerator: tpuTensor Processing UnitTensor Processing Unit

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

plGeneric label for PyTorch Lightning packageGeneric label for PyTorch Lightning package

repro neededThe issue is missing a reproducible exampleThe issue is missing a reproducible example

Status: Open.

#14457 In Lightning-AI/pytorch-lightning;

· pjspol opened on Aug 30, 2022

Checkpoint do not save under multinode training mode when using Webdataset

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

questionFurther information is requestedFurther information is requested

repro neededThe issue is missing a reproducible exampleThe issue is missing a reproducible example

Status: Open.

#16893 In Lightning-AI/pytorch-lightning;

· superhero-7 opened on Feb 28, 2023

Schedule in PyTorchProfiler doesn't work

bugSomething isn't workingSomething isn't working

help wantedOpen to be worked onOpen to be worked on

profiler

Status: Open.

#14063 In Lightning-AI/pytorch-lightning;

· fedorovgv opened on Aug 6, 2022