Search results

SequentialLR scheduler, bug with last_epoch = N

module: LrScheduler

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#162591 In pytorch/pytorch;

· Jostarndt opened on Sep 10, 2025

Support optimisation of sparse parameters via LBFGS

module: optimizerRelated to torch.optimRelated to torch.optim

module: sparseRelated to torch.sparseRelated to torch.sparse

needs designWe want to add this feature but we need to figure out how firstWe want to add this feature but we need to figure out how first

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#160637 In pytorch/pytorch;

· Thomas-Christie opened on Aug 14, 2025

Allowing <code>optimizer.step(..., *, set_grad_to_none=True)</code> (explicit kwarg opt-in, seems should not break existing code)

actionable

enhancementNot as big of a feature, but technically not a bug. Should be easy to fixNot as big of a feature, but technically not a bug. Should be easy to fix

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#158638 In pytorch/pytorch;

· vadimkantorov opened on Jul 18, 2025

Transition Adam and maybe other optimizers to fused code path by default to avoid <code>foreach=True</code>-specific VRAM peak due to temp TensorList for bias-corrected moments

module: optimizerRelated to torch.optimRelated to torch.optim

module: performanceIssues related to performance, either of kernel code or framework glueIssues related to performance, either of kernel code or framework glue

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#158371 In pytorch/pytorch;

· vadimkantorov opened on Jul 15, 2025

Can't alter the learning rate though loading optimizer states

has workaround

module: LrScheduler

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#156846 In pytorch/pytorch;

· Alex17Li opened on Jun 25, 2025

get_ema_multi_avg_fn() equation is a little confused

easylabel for easy docathon taskslabel for easy docathon tasks

good first issue

module: docsRelated to our documentation, both in docs/ and docblocksRelated to our documentation, both in docs/ and docblocks

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#155551 In pytorch/pytorch;

· nstalways opened on Jun 10, 2025

Use opmath_t and not double compute in fused optimizers

module: cudaRelated to torch.cuda, and CUDA support in generalRelated to torch.cuda, and CUDA support in general

module: optimizerRelated to torch.optimRelated to torch.optim

module: performanceIssues related to performance, either of kernel code or framework glueIssues related to performance, either of kernel code or framework glue

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#153649 In pytorch/pytorch;

· janeyx99 opened on May 15, 2025

Adam optimizer ValueError: beta1 as a Tensor

module: optimizerRelated to torch.optimRelated to torch.optim

needs reproductionEnsure you have actionable steps to reproduce the issue. Someone else needs to confirm the repro.Ensure you have actionable steps to reproduce the issue. Someone else needs to confirm the repro.

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#149508 In pytorch/pytorch;

· Vetti420 opened on Mar 19, 2025

Add Optimizer: Sharpness Aware Minimization

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#149269 In pytorch/pytorch;

· divyansha opened on Mar 16, 2025

addition of muon optimizer to torch.optim

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#148819 In pytorch/pytorch;

· SwamiKannan opened on Mar 8, 2025

Adam doesn't work with nonzero-dim Tensor betas

module: optimizerRelated to torch.optimRelated to torch.optim

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#147921 In pytorch/pytorch;

· Tony-Y opened on Feb 26, 2025

To avoid default eps ineffective in some precision(float16), dynamically set the default value of eps in optimizers such as Adam based on precision.

module: optimizerRelated to torch.optimRelated to torch.optim

needs designWe want to add this feature but we need to figure out how firstWe want to add this feature but we need to figure out how first

triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Status: Open.

#147437 In pytorch/pytorch;

· 1994cxy opened on Feb 19, 2025