Skip to content

Pull requests: pytorch/xla

New pull request New
145 Open 6,006 Closed
145 Open 6,006 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
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[Computation Hash] Introduce deterministic hash for user computations tpuci
#8539 opened Jan 7, 2025 by rpsilva-aws Loading…
1
7
Implement torch.linspace and torch.logspace
#8533 opened Jan 6, 2025 by vyom1611 Loading…
6
Lowering Aten op to composite op instead of small ops tpuci
#8502 opened Dec 18, 2024 by Zantares Loading…
13
training UX: automatic generating make_train_step
#8495 opened Dec 17, 2024 by qihqi Loading…
10
Lower cummax op op lowering
#8491 opened Dec 13, 2024 by zyy-martin Loading…
1
Support S32 index operations for copy and fill
#8463 opened Dec 6, 2024 by rpsilva-aws Loading…
3
Multi-queries Paged attn fails with continuator halted
#8459 opened Dec 5, 2024 by vanbasten23 • Draft
update test_replace_xla_tensor
#8458 opened Dec 5, 2024 by laithsakka Loading…
Remove the constraint of pages_per_sequence%num_kv_pages_per_compute_… tpuci
#8457 opened Dec 5, 2024 by vanbasten23 • Draft
[WIP] Update test_ops.py
#8449 opened Dec 4, 2024 by ManfeiBai • Draft
[torch_xla2] Create jtorchvision_roi_align.py
#8394 opened Nov 19, 2024 by ManfeiBai • Draft
norm and pca_lowrank op info test from issue #7528
#8384 opened Nov 14, 2024 by vyom1611 Loading…
WIP implement lowering for torchvision::roi_align
#8362 opened Nov 7, 2024 by barney-s Loading…
1
custom_kernel: fix shape mismatch by sharding segment_ids in flash attn. pallas
#8333 opened Oct 29, 2024 by dudulightricks • Draft
4
[release] add permission of iam.auto.tfvars
#8326 opened Oct 25, 2024 by ManfeiBai Loading…
2
print payload DO_NOT_MERGE Not for merging.
#8312 opened Oct 24, 2024 by zpcore • Draft
debug DO_NOT_MERGE Not for merging.
#8293 opened Oct 21, 2024 by zpcore • Draft
[DO NOT REVIEW] Experiment the pl.debug_print in Pallas.
#8284 opened Oct 18, 2024 by vanbasten23 • Draft
[DO NOT REVIEW YET] Extend paged attention
#8237 opened Oct 8, 2024 by vanbasten23 • Draft
Pin update
#8079 opened Sep 26, 2024 by qihqi Loading…
Check warning msg in second captured warning.
#8069 opened Sep 25, 2024 by zhuhaozhe Loading…
Support FSDPv2 compute dtype
#8056 opened Sep 24, 2024 by lausannel Loading…
4
Print the training update after the device execution DO_NOT_REVIEW_YET
#8054 opened Sep 24, 2024 by JackCaoG • Draft
Part 2. Introduce multi-node SPMD initialization for Neuron
#8046 opened Sep 19, 2024 by rpsilva-aws • Draft
2
Forward XLATensorImpl::is_contiguous_custom to TensorImpl.
#8032 opened Sep 17, 2024 by ysiraichi Loading…
14
ProTip! no:milestone will show everything without a milestone.