Discussions

• 1 You must be logged in to vote

  💬

  Dataloader reinsertion for recursive predictionsrs

  leonardcaquot94 started Oct 11, 2024 in General

  

  

  0
• 8 You must be logged in to vote

  💬

  Conflict between Lightning and Huggingface Transformers (device_map).

  richarddwang started Jun 20, 2023 in General

  

  

  

  

  4
• 1 You must be logged in to vote

  ⚡

  I want to apply custom learning rate scheduler.

  sooftware asked Jun 19, 2021 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Answered

  

  

  

  

  11
• 11 You must be logged in to vote

  ⚡

  Multi-GPU Inference

  distributed Generic distributed-related topic callback: prediction writer trainer: predict

  ricardorei asked Sep 1, 2021 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  

  

  27
• 1 You must be logged in to vote

  ⚡

  Custom train loop to perform partial batch updates

  leonardcaquot94 asked Oct 4, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  0
• 1 You must be logged in to vote

  ⚡

  Alternate "prediction" loops

  kaboroevich asked Oct 4, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Closed · Answered

  

  

  2
• 1 You must be logged in to vote

  ⚡

  val loss and accuracy not logging

  gkrampah asked Oct 2, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  0
• 19 You must be logged in to vote

  ⚡

  All time taken by {method 'enable' of '_lsprof.Profiler' objects} in Advanced profiler output

  profiler

  SeguinBe asked Feb 7, 2022 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  

  

  9
• 3 You must be logged in to vote

  ⚡

  How to log hyperparameters in Tensorboard HPARAMS tab?

  00sapo asked May 16, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  

  

  3
• 2 You must be logged in to vote

  ⚡

  LightningCLI access DataModule methods in model constructor

  lightningcli pl.cli.LightningCLI pl Generic label for PyTorch Lightning package

  bstee615 asked Jun 24, 2022 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Answered

  

  

  

  

  11
• 1 You must be logged in to vote

  💬

  How to train two modules with two different loss in one train_step alternately?

  Kamichanw started Sep 21, 2024 in General

  

  

  0
• 1 You must be logged in to vote

  🤖

  Mitigating program hang from on_train_epoch_end() with self.all_gather() call

  isaacgerg asked Sep 20, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🤖

  Low GPU utilization

  jbuckman asked Sep 19, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 1 You must be logged in to vote

  💭

  Saving conformer checkpoint and resuming train progress from checkpoint

  LuluW8071 asked Sep 18, 2024 in code help: NLP / ASR / TTS · Closed · Answered

  

  

  1
• 2 You must be logged in to vote

  🤖

  How to use different strategies between training and validation loops?

  Meehai asked Sep 18, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 1 You must be logged in to vote

  🤖

  triage for handling uneven inputs

  DucoG asked Sep 14, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 6 You must be logged in to vote

  💬

  I have runned: pip install pytorch-lightning, but get error: No module named pytorch_lightning

  muyuuuu started Jan 15, 2022 in General

  

  

  

  

  7
• 1 You must be logged in to vote

  🤖

  How to auto-resume torchrun multi-node training after hitting SLURM time limit?

  amorehead asked Sep 8, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 1 You must be logged in to vote

  ⚡

  RuntimeError: Input type (c10::Half) and bias type (float) should be the same

  morestart asked Sep 7, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  0
• 1 You must be logged in to vote

  💬

  How to log yaml config when using the WandbLogger?

  celsofranssa started Sep 7, 2024 in General

  

  

  0
• 1 You must be logged in to vote

  🤖

  It takes a lot of time to resume training after running evaluation [FSDP]. What can be the reason behind this?

  psr-ai asked Sep 6, 2024 in DDP / multi-GPU / multi-node · Unanswered

  

  

  0
• 1 You must be logged in to vote

  ⚡

  When ensembling a kfold trained set, does it lose the ability to set other models?

  VDFaller asked Sep 6, 2024 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  0
• 1 You must be logged in to vote

  😎

  With FSDP, when using model's forward in on_fit_end() hook weights are misshappen

  QuentinAndre11 asked Sep 6, 2024 in code help: CV · Unanswered

  

  

  0
• 1 You must be logged in to vote

  ⚡

  load_from_checkpoint has different validation results

  lanlanlan3 asked Jul 1, 2022 in Lightning Trainer API: Trainer, LightningModule, LightningDataModule · Unanswered

  

  

  

  

  7
• 2 You must be logged in to vote

  🤖

  DeepSpeedPlugin with activation checkpoint fails

  nachshonc asked Aug 26, 2021 in DDP / multi-GPU / multi-node · Answered

  

  

  

  

  8