Search results

Missing <code>MATCH_CLASS</code> bytecode in dynamo

dynamo-triage-dec2025small/mid-sized dynamo tasks that we would like to be completed in the near futuresmall/mid-sized dynamo tasks that we would like to be completed in the near future

Status: Open.

#171274 In pytorch/pytorch;

· guilhermeleobas opened on Dec 24, 2025

[Flex Flash] CuteDSL 4.3.4 has much higher cpu overhead

module: flex attention

oncall: pt2

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.

#171267 In pytorch/pytorch;

· drisspg opened on Dec 24, 2025

[CUDA Graphs] Capture failure leaves generator state corrupted, causing "Offset increment outside graph capture encountered unexpectedly" on subsequent RNG operations

Status: Open.

#171263 In pytorch/pytorch;

· alpha0422 opened on Dec 24, 2025

TorchInductor returns wrong output shape for aten.isin with scalar input

Status: Open.

#171262 In pytorch/pytorch;

· jikechao opened on Dec 24, 2025

torch.AcceleratorError: CUDA error: device-side assert triggered

Status: Open.

#171259 In pytorch/pytorch;

· Cookiee235 opened on Dec 24, 2025

[Inductor] TorchInductor produces inconsistent results with eager for <code>aten.fractional_max_pool2d</code>

Status: Open.

#171258 In pytorch/pytorch;

· Cookiee235 opened on Dec 24, 2025

[Inductor] Significant numerically inconsistent results for aten.trapezoid

Status: Open.

#171256 In pytorch/pytorch;

· Cookiee235 opened on Dec 24, 2025

[RFC][X86][Inductor] Convert fp32 -&gt; fp8_e4m3 with the new AVX10.2 ISA

Status: Open.

#171254 In pytorch/pytorch;

· Xia-Weiwen opened on Dec 24, 2025

[Bug] softplus returns inf for large beta while other frameworks return finite value

Status: Open.

#171249 In pytorch/pytorch;

· Thrsu opened on Dec 24, 2025

onnx test fail: "NativeOnnxOpsTest.test_attention_export_gqa"

Status: Open.

#171246 In pytorch/pytorch;

· yingluAMD opened on Dec 24, 2025

<code>torch._inductor.config.skip_l1_cache</code> could fail due to the lack of key check in <code>buffer_read_counts</code>

module: inductor

oncall: pt2

Status: Open.

#171244 In pytorch/pytorch;

· crcrpar opened on Dec 24, 2025

DISABLED test_comprehensive_fft_fftn_xpu_float16 (__main__.TestInductorOpInfoXPU)

module: xpuIntel XPU related issuesIntel XPU related issues

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

Status: Open.

#171236 In pytorch/pytorch;

· chuanqi129 opened on Dec 24, 2025