Search results

inconsistent behavior for torch.as_tensor with cupy array due to a driver bug

Status: Open.

#158316 In pytorch/pytorch;

· youkaichao opened on Jul 15, 2025

<code>torch.nn.functional.celu</code> produces inconsistent output on CPU vs CUDA

Status: Open.

#158314 In pytorch/pytorch;

· jiren-the-gray opened on Jul 15, 2025

Add support for training FP8_E4M3

Status: Open.

#158313 In pytorch/pytorch;

· edgeinfinity1 opened on Jul 15, 2025

<code>torch.nn.LogSoftmax</code> produces inf value on CPU but not on GPU

Status: Open.

#158310 In pytorch/pytorch;

· jiren-the-gray opened on Jul 15, 2025

Memory Error for _flash_attention_forward on AMD Radeon RX 7900 XTX

Status: Open.

#158308 In pytorch/pytorch;

· BonsonW opened on Jul 15, 2025

Torch nightly from 20250630 and onwards enables (forced) logging/debugging to INFO on any application.

Status: Open.

#158307 In pytorch/pytorch;

· Panchovix opened on Jul 15, 2025

DISABLED test_triton_relu_fused_gemm (__main__.MultiKernelTest)

module: inductor

module: rocmAMD GPU support for PytorchAMD GPU support for Pytorch

oncall: pt2

skippedDenotes a (flaky) test currently skipped in CI.Denotes a (flaky) test currently skipped in CI.

Status: Open.

#158274 In pytorch/pytorch;

· huydhn opened on Jul 14, 2025

[QST] [DTensor] MoE + EP support

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

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.

#158244 In pytorch/pytorch;

· jeromeku opened on Jul 14, 2025

[Windows][inductor] Windows compilers building was slower 4x than on Linux.

module: inductor

module: windowsWindows support for PyTorchWindows support for PyTorch

oncall: pt2

Status: Open.

#158240 In pytorch/pytorch;

· xuhancn opened on Jul 14, 2025

Incompatibility between torch.no_grad, torch.autocast, and FSDP2

actionable

module: fsdp

oncall: distributedAdd this issue/PR to distributed oncall triage queueAdd this issue/PR to distributed oncall triage queue

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.

#158232 In pytorch/pytorch;

· KimbingNg opened on Jul 14, 2025

Degraded JIT execution performance on 5090 Windows

oncall: jitAdd this issue/PR to JIT oncall triage queueAdd this issue/PR to JIT oncall triage queue

Status: Open.

#158227 In pytorch/pytorch;

· blace-ai-transfer opened on Jul 14, 2025

Title: torch.compile() produces incorrect outputs for transformer model with causal attention

oncall: pt2

Status: Open.

#158226 In pytorch/pytorch;

· zhouxiaoyaozzz opened on Jul 14, 2025