This repository was archived by the owner on Nov 1, 2024. It is now read-only.
Update dependency pytorch_lightning to >=1.9.5,<1.10 - autoclosed #3
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![renovate[bot]](https://avatars.githubusercontent.com/in/2740?s=60&v=4){kind=small}
renovate
bot
changed the title
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR contains the following updates:
>=1.5.4,<1.7->>=1.9.5,<1.10Release Notes
Lightning-AI/lightning (pytorch_lightning)
v1.9.5: Minor patch releaseCompare Source
App
Changed
healthzendpoint to plugin server (#16882)Fabric
Changed
TorchCollectiveworks on thetorch.distributedWORLD process group by default (#16995)Fixed
_cuda_clearCublasWorkspaceson teardown (#16907)PyTorch
Changed
NeptuneLogger(#16761):log()method withappend()andextend().Handleras an alternative toRunfor therunargument. This means that you can call it likeNeptuneLogger(run=run["some/namespace"])to log everything to thesome/namespace/location of the run.sys.argvand args inLightningCLI(#16808)Depercated
ShardedTensorstate dict hooks inLightningModule.__init__withtorch>=2.1(#16892)lightning.pytorch.core.saving.ModelIOclass interface (#16974)Fixed
num_nodesnot being set forDDPFullyShardedNativeStrategy(#17160)DeepSpeedStrategy(#16973)richthat prevented Lightning to be imported in Google Colab (#17156)_cuda_clearCublasWorkspaceson teardown (#16907)psutilpackage is now required for CPU monitoring (#17010)Contributors
@awaelchli, @belerico, @carmocca, @colehawkins, @dmitsf, @Erotemic, @ethanwharris, @kshitij12345, @borda
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.9.4: Weekly patch releaseCompare Source
App
Removed
testing.run_app_in_cloudin favor of headless login and app selection (#16741)Fabric
Added
Fabric(strategy="auto")support (#16916)Fixed
find_usable_cuda_devices(num_devices=-1)(#16866)PyTorch
Added
Fabric(strategy="auto")support. It will choose DDP over DDP-spawn, contrary tostrategy=None(default) (#16916)Fixed
lightning.pytorch.utilities.parsing.get_init_args(#16851)Contributors
@ethanwharris, @carmocca, @awaelchli, @justusschock , @dtuit, @Liyang90
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.9.3: Weekly patch releaseCompare Source
App
Fixed
lightning opencommand and improved redirects (#16794)Fabric
Fixed
accelerator=tpuanddevices > 1(#16806)--acceleratorand--precisionin Fabric CLI whenacceleratorandprecisionare set to non-default values in the code (#16818)PyTorch
Fixed
accelerator=tpuanddevices > 1(#16806)Contributors
@ethanwharris, @carmocca, @awaelchli, @borda, @tchaton, @yurijmikhalevich
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.9.2: Weekly patch releaseCompare Source
App
Added
rm: Delete files from your Cloud Platform Filesystemlightning connect datato register data connection to private s3 buckets (#16738)Fabric
Fixed
PyTorch
Changed
Fixed
min_epochsormin_steps(#16719)Contributors
@akihironitta, @awaelchli, @borda, @tchaton
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.9.1: Weekly patch releaseCompare Source
App
Added
lightning opencommand (#16482)ls: List files from your Cloud Platform Filesystemcd: Change the current directory within your Cloud Platform filesystem (terminal session based)pwd: Return the current folder in your Cloud Platform Filesystemcp: Copy files between your Cloud Platform Filesystem and local filesystemcdinto non-existent folders (#16645)cp(upload) at project level (#16631)lsandcp(download) at project level (#16622)lightning connect datato register data connection to s3 buckets (#16670)Changed
LightningClient(retry=False)toretry=True(#16382)lightning.app.components.LiteMultiNodetolightning.app.components.FabricMultiNode(#16505)lightning connecttolightning connect appfor consistency (#16670)Fixed
lightning cp(#16626)Fabric
Fixed
accelerator="mps"andddpstrategy pairing (#16455)torch_xlarequirement (#16476)torch.distributedis not available (#16658)Pytorch
Fixed
save_hyperparameterson mixin classes that don't subclassLightningModule/LightningDataModule(#16369)MLFlowLoggerlogging the wrong keys with.log_hyperparams()(#16418)MLFlowLoggerand long values are truncated (#16451)torch_xlarequirement (#16476)torch.distributedis not available (#16658)Contributors
@akihironitta, @awaelchli, @borda, @BrianPulfer, @ethanwharris, @hhsecond, @justusschock, @Liyang90, @RuRo, @senarvi, @shenoynikhil, @tchaton
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.9.0: Stability and additional improvementsCompare Source
App
Added
Changed
DeviceStatsMonitor(#16002)
lightning_app.components.serve.gradiotolightning_app.components.serve.gradio_server(#16201)Fixed
relpathbug on Windows (#16164)LooseVersion(#16162)lightning loginwith env variables would not correctly save the credentials (#16339)Fabric
Added
Fabric.launch()to programmatically launch processes (e.g. in Jupyter notebook) (#14992)runmethod (#14992)Fabric.setup_module()andFabric.setup_optimizers()to support strategies that need to set up the model before an optimizer can be created (#15185)lightning_fabric.accelerators.find_usable_cuda_devicesutility function (#16147)Fabric(callbacks=...)and emitting events throughFabric.call()(#16074)Fabric(loggers=...)to support different Logger frameworks in FabricFabric.logfor logging scalars using multiple loggersFabric.log_dictfor logging a dictionary of multiple metrics at onceFabric.loggersandFabric.loggerattributes to access the individual logger instancesself.logandself.log_dictin a LightningModule when using Fabricself.loggerandself.loggersin a LightningModule when using Fabriclightning_fabric.loggers.TensorBoardLogger(#16121)lightning_fabric.loggers.CSVLogger(#16346).zero_grad(set_to_none=...)on the wrapped optimizer regardless of which strategy is used (#16275)Changed
LightningLitetoFabric(#15932, #15938)Fabric.run()method is no longer abstract (#14992)XLAStrategynow inherits fromParallelStrategyinstead ofDDPSpawnStrategy(#15838)DDPSpawnStrategyintoDDPStrategyand removedDDPSpawnStrategy(#14952).setup_dataloaders()now calls.set_epoch()on the distributed sampler if one is used (#16101)Strategy.reducetoStrategy.all_reducein all strategies (#16370)Removed
strategy='ddp_sharded'|'ddp_sharded_spawn'). Use Fully-Sharded Data Parallel instead (strategy='fsdp') (#16329)Fixed
DistributedSampler(#16101)PyTorch
Added
MetricCollectionwith enabled compute groups (#15580)pl.loggers.WandbLogger(#16173)LRFinder(#15304)pl.utilities.upgrade_checkpointscript (#15333)axto the.lr_find().plot()to enable writing to a user-defined axes in a matplotlib figure (#15652)log_modelparameter toMLFlowLogger(#9187)self.log(..., logger=True)is called without a configured logger (#15814)LightningCLIsupport for optimizer and learning schedulers via callable type dependency injection (#15869)DDPFullyShardedNativeStrategystrategy (#15826)DDPFullyShardedNativeStrategy(cpu_offload=True|False)via bool instead of needing to pass a configuration object (#15832)LightningModule.configure_optimizers(#16189)Changed
tensorboardtotensorboardxinTensorBoardLogger(#15728)LightningModule.load_from_checkpointautomatically upgrade the loaded checkpoint if it was produced in an old version of Lightning (#15237)Trainer.{validate,test,predict}(ckpt_path=...)no longer restores theTrainer.global_stepandtrainer.current_epochvalue from the checkpoints - From now on, onlyTrainer.fitwill restore this value (#15532)ModelCheckpoint.save_on_train_epoch_endattribute is now computed dynamically every epoch, accounting for changes to the validation dataloaders (#15300)MLFlowLoggernow logs hyperparameters and metrics in batched API calls (#15915)on_train_batch_{start,end}hooks in conjunction with taking adataloader_iterin thetraining_stepno longer errors out and instead shows a warning (#16062)tensorboardXto extra dependencies. Use theCSVLoggerby default (#16349)Deprecated
description,env_prefixandenv_parseparameters inLightningCLI.__init__in favour of giving them throughparser_kwargs(#15651)pytorch_lightning.profilerin favor ofpytorch_lightning.profilers(#16059)Trainer(auto_select_gpus=...)in favor ofpytorch_lightning.accelerators.find_usable_cuda_devices(#16147)pytorch_lightning.tuner.auto_gpu_select.{pick_single_gpu,pick_multiple_gpus}in favor ofpytorch_lightning.accelerators.find_usable_cuda_devices(#16147)nvidia/apexdeprecation (#16039)pytorch_lightning.plugins.NativeMixedPrecisionPluginin favor ofpytorch_lightning.plugins.MixedPrecisionPluginLightningModule.optimizer_step(using_native_amp=...)argumentTrainer(amp_backend=...)argumentTrainer.amp_backendpropertyTrainer(amp_level=...)argumentpytorch_lightning.plugins.ApexMixedPrecisionPluginclasspytorch_lightning.utilities.enums.AMPTypeenumDeepSpeedPrecisionPlugin(amp_type=..., amp_level=...)argumentshorovoddeprecation (#16141)Trainer(strategy="horovod")HorovodStrategyclasspytorch_lightning.lite.LightningLitein favor oflightning.fabric.Fabric(#16314)FairScaledeprecation (in favor of PyTorch's FSDP implementation) (#16353)pytorch_lightning.overrides.fairscale.LightningShardedDataParallelclasspytorch_lightning.plugins.precision.fully_sharded_native_amp.FullyShardedNativeMixedPrecisionPluginclasspytorch_lightning.plugins.precision.sharded_native_amp.ShardedNativeMixedPrecisionPluginclasspytorch_lightning.strategies.fully_sharded.DDPFullyShardedStrategyclasspytorch_lightning.strategies.sharded.DDPShardedStrategyclasspytorch_lightning.strategies.sharded_spawn.DDPSpawnShardedStrategyclassRemoved
pytorch_lightning.utilities.memory.get_gpu_memory_mapin favor ofpytorch_lightning.accelerators.cuda.get_nvidia_gpu_stats(#15617)pytorch_lightning.profiler.base.AbstractProfilerin favor ofpytorch_lightning.profilers.profiler.Profiler(#15637)pytorch_lightning.profiler.base.BaseProfilerin favor ofpytorch_lightning.profilers.profiler.Profiler(#15637)pytorch_lightning.utilities.meta(#16038)LightningDeepSpeedModule(#16041)pytorch_lightning.accelerators.GPUAcceleratorin favor ofpytorch_lightning.accelerators.CUDAAccelerator(#16050)pytorch_lightning.profiler.*classes in favor ofpytorch_lightning.profilers(#16059)pytorch_lightning.utilities.climodule in favor ofpytorch_lightning.cli(#16116)pytorch_lightning.loggers.basemodule in favor ofpytorch_lightning.loggers.logger(#16120)pytorch_lightning.loops.basemodule in favor ofpytorch_lightning.loops.loop(#16142)pytorch_lightning.core.lightningmodule in favor ofpytorch_lightning.core.module(#16318)pytorch_lightning.callbacks.basemodule in favor ofpytorch_lightning.callbacks.callback(#16319)Trainer.reset_train_val_dataloaders()in favor ofTrainer.reset_{train,val}_dataloader(#16131)LightningCLI(seed_everything_default=None)(#16131)strategy='ddp_sharded'|'ddp_sharded_spawn'). Use Fully-Sharded Data Parallel instead (strategy='fsdp') (#16329)Fixed
reduce_boolean_decisionto accommodateany-analogous semantics expected by theEarlyStoppingcallback (#15253)intervalkey of the scheduler would be ignored during manual optimization, making the LearningRateMonitor callback fail to log the learning rate (#16308)MLFlowLoggernot finalizing correctly when status code 'finished' was passed (#16340)Contributors
@1SAA, @akihironitta, @AlessioQuercia, @awaelchli, @bipinKrishnan, @Borda, @carmocca, @dmitsf, @erhoo82, @ethanwharris, @Forbu, @hhsecond, @justusschock, @lantiga, @lightningforever, @Liyang90, @manangoel99, @mauvilsa, @nicolai86, @nohalon, @rohitgr7, @schmidt-jake, @speediedan, @yMayanand
If we forgot someone due to not matching commit email with GitHub account, let us know :]
v1.8.6: Weekly patch releaseCompare Source
App
Added
Requestannotation inconfigure_apihandlers (#16047)work.deletemethod to delete the work (#16103)display_nameproperty to LightningWork for the cloud (#16095)ColdStartProxyto the AutoScaler (#16094)ready(#16075)readyfor components (#16129)Changed
start_methodfor creating Work processes locally on macOS is now 'spawn' (previously 'fork') (#16089)lightning.app.utilities.cloud.is_running_in_cloudnow returnsTrueduring the loading of the app locally when running with--cloud(#16045)True(#16009)Fixed
PythonServermessaging "Your app has started" (#15989)AutoScalerwould fail with min_replica=0 (#16092AutoScalerUI (#16128)streamlit(#16139)Full Changelog: Lightning-AI/pytorch-lightning@1.8.5.post0...1.8.6
v1.8.5.post0: Minor patch releaseCompare Source
App
self.lightningignore(#16080)Pytorch
Full Changelog: Lightning-AI/pytorch-lightning@1.8.5...1.8.5.post0
v1.8.5: Weekly patch releaseCompare Source
App
Added
Lightning{Flow,Work}.lightningignoresattributes to programmatically ignore files before uploading to the cloud (#15818).lightningignorethat ignoresvenv(#16056)Changed
Fixed
DDPStrategyimport in app framework (#16029)AutoScalerraising an exception when non-default cloud compute is specified (#15991)Pytorch
Full Changelog: Lightning-AI/pytorch-lightning@1.8.4.post0...1.8.5
v1.8.4.post0: Minor patch releaseCompare Source
App
L.app.structures(#15964)Pytorch
XLAProfilernot recording anything due to mismatching of action names (#15885)Full Changelog: Lightning-AI/pytorch-lightning@1.8.4...1.8.4.post0
v1.8.4: Weekly patch releaseCompare Source
App
Added
code_dirargument to tracer run (#15771)lightning run modelto launch aLightningLiteaccelerated script (#15506)lightning delete appto delete a lightning app on the cloud (#15783)AutoScalercomponent (#15769)readyof the LightningFlow to inform when theOpen Appshould be visible (#15921)_start_methodto customize how to start the works (#15923)configure_layoutmethod to theLightningWorkwhich can be used to control how the work is handled in the layout of a parent flow (#15926)lightning run app organization/name(#15941)Changed
MultiNodecomponents now warn the user when running withnum_nodes > 1locally (#15806)BuildConfig(requirements=[...])is passed but arequirements.txtfile is already present in the Work (#15799)BuildConfig(dockerfile="...")is passed but aDockerfilefile is already present in the Work (#15799)Removed
SingleProcessRuntime(#15933)Fixed
enable_spawnmethod of theWorkRunExecutor(#15812)L.app.structureswould cause multiple apps to be opened and fail with an error in the cloud (#15911)ImportErroron Multinode if package not present (#15963)Lite
shuffle=Falsehaving no effect when using DDP/DistributedSampler (#15931)Pytorch
Changed
Fixed
fit_loop.restartingto beFalsefor lr finder (#15620)torch.jit.script-ing a LightningModule causing an unintended error message about deprecateduse_ampproperty (#15947)Full Changelog: Lightning-AI/pytorch-lightning@1.8.3...1.8.4
v1.8.3.post2: Dependency hotfixCompare Source
🤖
v1.8.3.post1: Hotfix for Python ServerCompare Source
App
Changed
Full Changelog: Lightning-AI/pytorch-lightning@1.8.3...1.8.3
v1.8.3.post0: Hotfix for requirementsCompare Source
v1.8.3: Weekly patch releaseCompare Source
App
Changed
lightning add ssh-keyCLI command has been transitioned tolightning create ssh-keylightning remove ssh-keyCLI command has been transitioned tolightning delete ssh-keyLightningTrainerScriptstart-up time (#15751)StreamlitFrontendto support upload in localhost (#15684)Fixed
LightningFlow(#15750)Lite
Changed
Pytorch
Changed
tensorboardtotensorboardxinTensorBoardLogger(#15728)Full Changelog: Lightning-AI/pytorch-lightning@1.8.2...1.8.3
v1.8.2: Weekly patch releaseCompare Source
App
Added
Changed
.lightningfile (#15654)Fixed
Lite
Fixed
LightningLite(strategy="ddp_spawn", ...)toLightningLite(strategy="ddp", ...)when on an LSF cluster (#15103)Pytorch
Fixed
Trainer(strategy="ddp_spawn", ...)toTrainer(strategy="ddp", ...)when on an LSF cluster (#15103][https://github.com/PyTorchLightning/pytorch-lightning/issues/15103](https://redirect.github.com/PyTorchLightning/pytorch-lightning/issues/15103)3))Full Changelog: Lightning-AI/pytorch-lightning@1.8.1...1.8.2
v1.8.1: Weekly patch releaseCompare Source
App
Added
startmethod to the work (#15523)MultiNodeComponent to run with distributed computation with any frameworks (#15524)RunWorkExecutorto the work and provides default ones for theMultiNodeComponent (#15561)start_with_flowflag to theLightningWorkwhich can be disabled to prevent the work from starting at the same time as the flow (#15591)bi-directionaldelta updates between the flow and the works (#15582)--setupflag tolightning run appCLI command allowing for dependency installation via app comments (#15577)Changed
flow.flowsto be recursive wont to align the behavior with theflow.works(#15466)paramsargument inTracerPythonScript.runno longer prepends--automatically to parameters (#15518)Fixed
Configuration
📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about this update again.
This PR was generated by Mend Renovate. View the repository job log.