Skip to content

Pull requests: pytorch/helion

New pull request New
89 Open 2,027 Closed
89 Open 2,027 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[compiler][autotuner] Compiler seed heuristics CLA Signed This label is managed by the Meta Open Source bot.
#2392 opened May 10, 2026 by ethche Contributor Loading…
6
[autotune] Add observed heuristic seeds CLA Signed This label is managed by the Meta Open Source bot.
#2378 opened May 10, 2026 by choijon5 Contributor • Draft
Attention Perf: Transpose blocked K right before QK instead of pre-transposing before the kernel CLA Signed This label is managed by the Meta Open Source bot.
#2374 opened May 9, 2026 by AmesingFlank Contributor Loading…
Attention Perf: Multiply Q in-loop to avoid memory spillage CLA Signed This label is managed by the Meta Open Source bot.
#2373 opened May 9, 2026 by AmesingFlank Contributor Loading…
Seeding the configs for aymmetric skinny matmuls CLA Signed This label is managed by the Meta Open Source bot.
#2357 opened May 7, 2026 by umechand-amd Collaborator Loading…
1
Avoid dynamic shape recompiles for 0/1 tensor dimensions CLA Signed This label is managed by the Meta Open Source bot.
#2353 opened May 7, 2026 by oulgen Contributor Loading…
1
[runtime:pallas] migrate to torch_tpu's new Pallas buffer donation API CLA Signed This label is managed by the Meta Open Source bot.
#2351 opened May 7, 2026 by cota Collaborator Loading…
2
WIP Pallas atomic all launchers v2 CLA Signed This label is managed by the Meta Open Source bot.
#2346 opened May 7, 2026 by thcmbs Collaborator • Draft
Add multi tile loop support to autodiff CLA Signed This label is managed by the Meta Open Source bot.
#2338 opened May 7, 2026 by karthickai Contributor • Draft
[DO NOT MERGE] [Pallas] Manual full-coverage TPU benchmark trigger CLA Signed This label is managed by the Meta Open Source bot.
#2322 opened May 6, 2026 by norx1991 Contributor • Draft
Add RemoteCacheBackend ABC for pluggable remote autotune caching CLA Signed This label is managed by the Meta Open Source bot.
#2317 opened May 6, 2026 by fulvius31 Collaborator Loading…
3
[Pallas] Skip factory tensor padding for Pallas backend CLA Signed This label is managed by the Meta Open Source bot.
#2279 opened May 5, 2026 by norx1991 Contributor • Draft
[TPU][Pallas] relax tolerances and fix Pallas autotuning OOM in layer_norm CLA Signed This label is managed by the Meta Open Source bot.
#2272 opened May 5, 2026 by yarongmu-google Collaborator • Draft
A jagged_hstu_attention example that works on Pallas TPU CLA Signed This label is managed by the Meta Open Source bot.
#2218 opened May 3, 2026 by AmesingFlank Contributor Loading…
[Pallas] Use LONG_INT_TYPE for jagged offsets in examples and tests CLA Signed This label is managed by the Meta Open Source bot.
#2132 opened Apr 28, 2026 by norx1991 Contributor • Draft
[Autotuner] Long-lived worker pool for parallel precompile CLA Signed This label is managed by the Meta Open Source bot.
#2128 opened Apr 28, 2026 by choijon5 Contributor • Draft
[Autotuner] Raise default min_improvement_delta to 0.003 CLA Signed This label is managed by the Meta Open Source bot.
#2119 opened Apr 27, 2026 by choijon5 Contributor • Draft
autotuner: cap tile size for imbalanced 2D grid dims CLA Signed This label is managed by the Meta Open Source bot.
#2102 opened Apr 24, 2026 by umechand-amd Collaborator Loading…
9
[Pallas] Switch gather to jnp.take_along_axis (for JAX issue filing) CLA Signed This label is managed by the Meta Open Source bot.
#2061 opened Apr 20, 2026 by AmesingFlank Contributor • Draft
[Pallas] Lower aten gather using one_hot + sum for TPU compatibility, unblocking cross_entropy CLA Signed This label is managed by the Meta Open Source bot.
#2060 opened Apr 20, 2026 by AmesingFlank Contributor Loading…
2
Use torch.gather instead of generic int indexing for cross_entropy example CLA Signed This label is managed by the Meta Open Source bot.
#2058 opened Apr 20, 2026 by AmesingFlank Contributor • Draft
2
[Pallas] Implement indirect gather via exact one-hot matmul CLA Signed This label is managed by the Meta Open Source bot.
#2035 opened Apr 17, 2026 by thcmbs Collaborator • Draft
CI: retry flaky CUDA compute check CLA Signed This label is managed by the Meta Open Source bot.
#2034 opened Apr 17, 2026 by choijon5 Contributor • Draft
[TPU][Pallas]Fix example/cross_entropy.py on Pallas TPU CLA Signed This label is managed by the Meta Open Source bot.
#2019 opened Apr 14, 2026 by yarongmu-google Collaborator • Draft
9
WIP: fp8 all gather matmul CLA Signed This label is managed by the Meta Open Source bot.
#1974 opened Apr 7, 2026 by shunting314 Contributor Loading…
ProTip! Adding no:label will show everything without a label.