Skip to content

Pull requests: pytorch/helion

New pull request New
Clear current search query, filters, and sorts
126 Open 2,119 Closed
126 Open 2,119 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

[autotune] Feat: Overlapping GPU benchmarking and CPU compilation CLA Signed This label is managed by the Meta Open Source bot.
#1416 opened Feb 10, 2026 by hinriksnaer Collaborator • Draft
10 👍 2
Use persistent+tensor_descriptor defaults for dot kernels on Blackwell (sm100+) CLA Signed This label is managed by the Meta Open Source bot.
#1964 opened Apr 7, 2026 by choijon5 Contributor Loading…
2 👍 1
Feat: Early term during benchmarking CLA Signed This label is managed by the Meta Open Source bot.
#1392 opened Feb 6, 2026 by hinriksnaer Collaborator • Draft
4 👍 1
[Autotuner] Add generation_ctx() context manager for each autotuner generation CLA Signed This label is managed by the Meta Open Source bot.
#1745 opened Mar 19, 2026 by yf225 Contributor Loading…
1 task done
4 👍 1
Validate user provided settings CLA Signed This label is managed by the Meta Open Source bot.
#947 opened Oct 15, 2025 by choijon5 Contributor Loading…
3
Add epilogue subtiling CLA Signed This label is managed by the Meta Open Source bot.
#948 opened Oct 15, 2025 by PaulZhang12 Contributor Loading…
26
Disallow both interpret modes active CLA Signed This label is managed by the Meta Open Source bot.
#1023 opened Oct 25, 2025 by yf225 Contributor • Draft
1
Add static_shapes="none" mode to disable 0/1 shape specialization CLA Signed This label is managed by the Meta Open Source bot.
#1053 opened Oct 30, 2025 by Itssshikhar Loading…
1
@yf225
27
Add OpenEvolve-based Autotuner for Helion GPU Kernels CLA Signed This label is managed by the Meta Open Source bot.
#1082 opened Nov 4, 2025 by mycpuorg • Draft
1
[Auto-Recovery] Refactor search algorithms for checkpointable state CLA Signed This label is managed by the Meta Open Source bot.
#1946 opened Apr 4, 2026 by yf225 Contributor Loading…
3
[Auto-Recovery] Add checkpoint save/load/resume CLA Signed This label is managed by the Meta Open Source bot.
#1947 opened Apr 4, 2026 by yf225 Contributor Loading…
2
Raise error when autotune_with_torch_compile_fusion is set without torch_compile_fusion CLA Signed This label is managed by the Meta Open Source bot.
#1951 opened Apr 4, 2026 by yf225 Contributor Loading…
3 tasks
1
[compiler] add basic lambda support CLA Signed This label is managed by the Meta Open Source bot.
#1958 opened Apr 6, 2026 by hinriksnaer Collaborator Loading…
1
only check accuracy of mutated args CLA Signed This label is managed by the Meta Open Source bot.
#1961 opened Apr 6, 2026 by shunting314 Contributor Loading…
3
autotune fp8 matmul reduce scatter in unit test CLA Signed This label is managed by the Meta Open Source bot.
#1962 opened Apr 6, 2026 by shunting314 Contributor Loading…
fix _clone_args when there are non-tensor input CLA Signed This label is managed by the Meta Open Source bot.
#1963 opened Apr 6, 2026 by shunting314 Contributor Loading…
3
warning for process group name not found conditionally CLA Signed This label is managed by the Meta Open Source bot.
#1973 opened Apr 7, 2026 by shunting314 Contributor Loading…
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…
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
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] 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
[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
no tensor descriptor for the 2d paralell gemm example CLA Signed This label is managed by the Meta Open Source bot.
#1848 opened Mar 27, 2026 by shunting314 Contributor Loading…
1
[Pallas] Fix adjust_block_size_constraints to unblock test_bmm and test_attention_block_pointer CLA Signed This label is managed by the Meta Open Source bot.
#1881 opened Mar 30, 2026 by AmesingFlank Contributor Loading…
5
[pallas-tpu] we are capping block sizes after they have been determined - that breaks in non-static shape regimes CLA Signed This label is managed by the Meta Open Source bot.
#1888 opened Mar 30, 2026 by v0i0 Contributor Loading…
1
ProTip! no:milestone will show everything without a milestone.