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

622 Open 1,803 Closed
622 Open 1,803 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

Build PyTorch/XLA with the same version of CUDA as PyTorch CI CI related change xla:gpu
#8700 opened Feb 11, 2025 by tengyifei
Gradient Accumulation API aliasing runtime
#8697 opened Feb 10, 2025 by rpsilva-aws
@rpsilva-aws
4
Enable 2D shardign with minibatch=True for SPMD enhancement New feature or request SPMD / Distributed
#8696 opened Feb 10, 2025 by miladm
@lsy323
1
In the case of xla amp, turning on gradient checkpoint, cause OOM gradient_checkpointing xla:gpu
#8695 opened Feb 10, 2025 by mars1248
1
[torch_xla] scan only captures aten operations dynamo enhancement New feature or request SPMD / Distributed
#8691 opened Feb 9, 2025 by tengyifei
2
Check if the current Pallas kernels have enough test coverage. pallas testing Testing and coverage related issues.
#8687 opened Feb 8, 2025 by vanbasten23
@vanbasten23
1
Segment fault on Llama3 and Mixtral model using PyTorch/XLA nightly libtpu-update
#8683 opened Feb 6, 2025 by zpcore
2
Introduce a mark_sharding that also shards the backward enhancement New feature or request SPMD / Distributed
#8678 opened Feb 4, 2025 by tengyifei
@tengyifei
1
Unify all pallas kernel code in one directory pallas
#8676 opened Feb 4, 2025 by miladm
@vanbasten23
1
cummax: returned indices are not consistent with PyTorch. bug Something isn't working op lowering pytorch api XLA behavior doesn't match Pytorch eager frontend
#8675 opened Feb 4, 2025 by ysiraichi
@ysiraichi
1
torch.nan_to_num doesn't work with -inf/inf bug Something isn't working pytorch api XLA behavior doesn't match Pytorch eager frontend
#8674 opened Feb 4, 2025 by Akshat-Tripathi
@ysiraichi
1
Autocast Policy for new xla backend. enhancement New feature or request
#8672 opened Feb 4, 2025 by avizon-aws
1
Improve testing coverage across all device hw types enhancement New feature or request
#8669 opened Feb 3, 2025 by rpsilva-aws
6
Missing sharding specs when annotating sharding over views bug Something isn't working SPMD / Distributed
#8662 opened Feb 1, 2025 by rpsilva-aws
@rpsilva-aws
6
TPU test failed at head: ERROR: test_scan_with_rematerialization (__main__.ScanTest) bug Something isn't working CI CI related change xla:tpu TPU specific issues and PRs
#8661 opened Jan 31, 2025 by tengyifei
@tengyifei
3
Torch XLA Model all_gather does not work with tensors of different sizes along dimension 0 enhancement New feature or request SPMD / Distributed usability Bugs/features related to improving the usability of PyTorch/XLA
#8660 opened Jan 31, 2025 by ajayvohra2005
@zpcore
5
RNN / GRU / LSTM implementation for torch_xla enhancement New feature or request
#8655 opened Jan 30, 2025 by qihqi
@tengyifei
2
how to use device_map in diffusers.stablediffusionpipeline.from_pretrained usability Bugs/features related to improving the usability of PyTorch/XLA xla:tpu TPU specific issues and PRs
#8646 opened Jan 29, 2025 by chaowenguo
3
Make Mixtral pallas kernels Dynamo/AOTAutograd traceable pallas
#8642 opened Jan 28, 2025 by tengyifei
1
split on second dimension of 2D array not working functionalization-disabled Issues specifically for when functionalization is disabled. pytorch api XLA behavior doesn't match Pytorch eager frontend
#8640 opened Jan 28, 2025 by jeffhataws
@ysiraichi
4
[torchax][RFC] Anchor on the device API everywhere RFC torchxla2
#8638 opened Jan 28, 2025 by tengyifei
@tengyifei
3
[torchax] RNG handling in a jitted graph is unsound bug Something isn't working torchxla2
#8636 opened Jan 28, 2025 by tengyifei
[torchax] jit compile the model constructor enhancement New feature or request torchxla2
#8635 opened Jan 28, 2025 by tengyifei
[scan] Avoid re-tracing the combine function on every call enhancement New feature or request
#8632 opened Jan 27, 2025 by tengyifei
Add type annotation to PyTorch/XLA code and tests & expand tests for various types as needed TECHNICAL_DEBT Technical Debt Is Evil usability Bugs/features related to improving the usability of PyTorch/XLA
#8627 opened Jan 26, 2025 by miladm
@ysiraichi @zpcore
1
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.