Skip to content

Issues: pytorch/xla

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

599 Open 1,772 Closed
599 Open 1,772 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

Slow XLA training performance.
#8541 opened Jan 7, 2025 by tzstoyanov
[Computation Hash] User Computation hash is agnostic to debug metadata
#8538 opened Jan 7, 2025 by rpsilva-aws
4
[Computation Hash] User Computation hash disregards protobuf requirements
#8537 opened Jan 6, 2025 by rpsilva-aws
1
1
PyTorch toy model fails to execute on DTensor constructed from "xla" type device mesh
#8534 opened Jan 6, 2025 by awshaichen
Input tensor is not an XLA tensor: torch.FloatTensor
#8532 opened Jan 4, 2025 by chaowenguo
PyTorch DTensor device mesh interface with device type "xla" fails at get_rank()
#8528 opened Jan 3, 2025 by awshaichen
Device-to-host transfers are very slow and asymmetrical
#8525 opened Jan 3, 2025 by neel04
1
SPMD with scalar parameter
#8522 opened Dec 29, 2024 by giladxw
1
how to clear tpu memory in pytorch xla
#8520 opened Dec 24, 2024 by chaowenguo
1
how to release tpu memory after del diffusers pipeline
#8516 opened Dec 22, 2024 by ghost
1
multi_queries_paged_attention_kernel fails with Llama3 70B on a TPU-v4-16 with sequence length of 256
#8515 opened Dec 21, 2024 by OhadRubin
2
Adding XLA as a backend for DeepSpeed
#8514 opened Dec 21, 2024 by radna0
Input tensor is not an XLA tensor on AWS Trainium instance
#8510 opened Dec 20, 2024 by JmeanJmy
3
API guide code snippets don't work
#8497 opened Dec 17, 2024 by richardsliu
5
AMP BF16 issue with batch norm layer
#8496 opened Dec 17, 2024 by lukeliu15
1
How to do multi-machine SPMD/FSDPv2 training with TPU？
#8492 opened Dec 13, 2024 by radna0
6
2 questions for the composite op feature stablehlo StableHLO related work
#8486 opened Dec 12, 2024 by Zantares
@lsy323
3
Program Hang/Stuck after using F.interpolate? VAE Decode Step of HunyuanVideo model.
#8470 opened Dec 9, 2024 by radna0
2
Pallas Kernel Expected Output Shape Error Using Grids
#8469 opened Dec 8, 2024 by dshalem
6
Useful Q8 Kernels For TPUs/XLA Support
#8466 opened Dec 6, 2024 by radna0
2.6 backport PR request list backport_2.6
#8455 opened Dec 5, 2024 by tengyifei
9
how to auto convert back to bfloat16 after conv1 and conv2
#8454 opened Dec 5, 2024 by ghost
1
Request for Wheel with Older GLIBC
#8430 opened Dec 3, 2024 by ScopeX-ASU
2
Pallas custom kernels: vmem error when using larger models.
#8429 opened Dec 2, 2024 by dshalem
2
TPU memory use increased significantly in torch/xla - 2.6.0.dev20241107
#8423 opened Nov 27, 2024 by dudulightricks
4
ProTip! Type g i on any issue or pull request to go back to the issue listing page.