Skip to content

Issues: intel/torch-xpu-ops

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

123 Open 340 Closed
123 Open 340 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[Windows][ARC][UT] test_weight_norm.py::TestNNMethod::test_weight_norm_different_type failed with error AssertionError: Tensor-likes are not close! client hw: Arc os: Windows Windows Platform
#1401 opened Feb 24, 2025 by huaiyuzh
@xytintel
[Windows][ARC][UT] test_rms_norm.py::TestNNMethod::test_rms_norm_bw failed with error AssertionError: Tensor-likes are not close! client hw: Arc
#1400 opened Feb 24, 2025 by huaiyuzh PT2.7
@xytintel @LuFinch
1
[Microbench] loss.soft_margin_loss forward/backward have performance regression
#1399 opened Feb 24, 2025 by huaiyuzh
@xytintel
hf_T5_generate inference accuracy fail_to_run E2E hw: PVC
#1392 opened Feb 21, 2025 by kaileiyx PT2.7
@etaf
moondream got different accuracy results 2 eager runs E2E hw: PVC
#1391 opened Feb 21, 2025 by kaileiyx PT2.7
@retonym
DebertaForQuestionAnswering amp_bf16/amp_fp16 training/inference accuracy/performance got failed E2E hw: PVC
#1390 opened Feb 21, 2025 by kaileiyx PT2.7
@etaf
DebertaForMaskedLM amp_bf16/amp_fp16 training/inference accuracy/performance got failed E2E
#1389 opened Feb 21, 2025 by kaileiyx PT2.7
@etaf
Most of E2E models failed with torch._inductor.exc.InductorError: RuntimeError: Triton Error [ZE]: 0x78000011 on BMG windows. client dependency component: Triton E2E hw: BMG module: dependency bug Problem is not caused by us, but caused by the library we use
#1385 opened Feb 20, 2025 by libohao1201 PT2.7
@Stonepia
2
pad_sequence and gru.input have performance regression in molan
#1381 opened Feb 19, 2025 by huaiyuzh PT2.8
@xytintel
Sort has performance regression in model pointnet-atlas(~5% on single tile and ~10% on scaling up)
#1380 opened Feb 19, 2025 by huaiyuzh PT2.8
@xytintel
[iGPU] The device does not have the ext_intel_free_memory aspect client dependency component: driver hw : LNL hw : MTL MTL platform module: dependency bug Problem is not caused by us, but caused by the library we use os: Windows Windows Platform
#1352 opened Feb 10, 2025 by Stonepia 2.8
@Stonepia @guangyey
4
UT test_roi_align_backward will be hung on Windows with nightly wheel client hw : LNL os: Windows Windows Platform
#1350 opened Feb 10, 2025 by mengfei25 PT2.8
@chunhuanMeng
erfcx_xpu and ndtri_xpu not implemented for 'BFloat16' ut_triaged
#1338 opened Feb 6, 2025 by huaiyuzh PT2.7
@chunhuanMeng
1
fractional_max_pool2d and fractional_max_pool3d cause an IPEX UT fail.
#1337 opened Feb 6, 2025 by huaiyuzh
@xytintel
"index_copy_xpu" not implemented for 'Float8_e4m3fn' module: OP Impl
#1336 opened Feb 6, 2025 by huaiyuzh PT2.8
@xytintel @yucai-intel
2
Torchbench timm_regnet training BF16 accuracy regression Accuracy component: torchbench dtype: bfloat16 E2E hw: PVC os: Ubuntu training
#1334 opened Feb 5, 2025 by mengfei25 PT2.8
internal compiler error: in extract_insn when compiling pytorch with xpu with gcc 12
#1332 opened Feb 3, 2025 by jingxu10 2.8
@xytintel @chunhuanMeng
Not Implemented Error "torch.ao.quantization.quantize_dynamic" module: OP Impl module: quant
#1329 opened Jan 29, 2025 by gurwinderintel PT2.8
@ZhiweiYan-96
1
Recommended building and runtime dependences for MKL. dependency component: oneMKL
#1325 opened Jan 24, 2025 by fengyuan14 2.8
@CuiYifeng
2
[Win] UR Error when OOM and break the tensor context client dependency component: driver module: dependency bug Problem is not caused by us, but caused by the library we use os: Windows Windows Platform
#1324 opened Jan 24, 2025 by Stonepia PT2.7
@guangyey
5
upstream test failures with test_ops.py module: OP Impl
#1315 opened Jan 22, 2025 by daisyden
@ZhiweiYan-96 @CuiYifeng @xytintel
Models got fail accuracy on BMG but passed on PVC Accuracy E2E hw: BMG os: Windows Windows Platform
#1305 opened Jan 20, 2025 by mengfei25 PT2.7
@retonym
1
Torchbench demucs training got fail accuracy Accuracy component: torchbench dtype: amp_bf16 dtype: amp_fp16 dtype: float32 E2E hw: PVC os: Ubuntu training
#1296 opened Jan 17, 2025 by mengfei25 PT2.7
@retonym
[DONT CLOSE] Last Known Workable Pytorch Commits
#1280 opened Jan 11, 2025 by chuanqi129
233
Detectron2 inference accuracy got failed Accuracy component: torchbench E2E inference
#1278 opened Jan 10, 2025 by mengfei25 PT2.7
@retonym
1
ProTip! Add no:assignee to see everything that’s not assigned.