Skip to content

Pull requests: ROCm/pytorch

New pull request New
90 Open 2,735 Closed
90 Open 2,735 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

Add portable Docker build workflow using public TheRock wheels
#3149 opened Apr 13, 2026 by ethanwee1 • Draft
6
Update Numba version constraints to support Python 3.14
#3148 opened Apr 10, 2026 by sohbodas Loading…
3
[release/2.10] Cherry-pick: [FSDP2] support dataclass args/kwargs output without memory leakage (#174692)
#3146 opened Apr 10, 2026 by chinmaydk99 • Draft
1
[release/2.11][ROCm][inductor] Additional GEMM, pointwise and reduction configs.
#3145 opened Apr 9, 2026 by naromero77amd Loading…
1
[release/2.11] Update composable_kernel submodule with gfx1033 support
#3144 opened Apr 9, 2026 by harkgill-amd Loading…
1 task done
1
Align Radeon targets for supported and preferred hipBLASLt backend with release/2.10
#3135 opened Apr 7, 2026 by mstankov-amd Loading…
1
[AUTOGENERATED] [release/2.8] ProcessGroupGloo: fix CUDA tensor stream handling with futures (#170812)
#3128 opened Apr 2, 2026 by rocm-repo-management-api-6 bot Loading…
1
[release/2.10] cmake version fix
#3116 opened Apr 1, 2026 by JeniferC99 Loading…
1 task done
1
CI: Add ROCm nightly docker workflow
#3115 opened Mar 31, 2026 by leo-automation Collaborator Loading…
14
Use natively hip_bfloat16 as a member of BFloat16 class to take advan…
#3114 opened Mar 31, 2026 by anatoliylitv • Draft
1 task
1
Anatoliylitv/bloat 16 use hip bfloat 16 as union
#3113 opened Mar 31, 2026 by anatoliylitv • Draft
1 task
1
Anatoliylitv/bloat 16 use hip bfloat 16 as parent
#3112 opened Mar 31, 2026 by anatoliylitv • Draft
1 task
1
[WIP] PyTorch ROCTx Observer Implementation module: rocm WIP
#3108 opened Mar 27, 2026 by EvanPhoukong Loading…
1 task
1
[WIP] Support CU-masked stream for selective backward execution and rccl
#3107 opened Mar 27, 2026 by chien-an-chen • Draft
1 task
2
[triton 3.7] Remove temporary ROCm test skips
#3105 opened Mar 25, 2026 by naromero77amd Loading…
2 of 3 tasks
5
Fix preffered blas backend test
#3085 opened Mar 19, 2026 by mstankov-amd Loading…
11
Update kineto submodule
#3084 opened Mar 18, 2026 by mwootton Loading…
1
Update kineto submodule
#3083 opened Mar 18, 2026 by mwootton Loading…
1
Direct, immediate, and local fix to UpSampleNearest2D fwd to work around the UINT32_MAX limit of HIP runtime
#3082 opened Mar 18, 2026 by glen-amd Loading…
7
[IFU] Add automated issue creation workflow
#3067 opened Mar 13, 2026 by chinmaydk99 Loading…
2
Use Github app secrets
#3065 opened Mar 11, 2026 by jithunnair-amd Collaborator • Draft
1 task
1
Get scalar from VRAM when using ROCm and large bar is enabled
#3048 opened Mar 6, 2026 by pschlan-amd Loading…
1 task
3
Added hipblaslt bias fused kernels into autotune backend for addmm
#3012 opened Feb 28, 2026 by guangzlu Loading…
1 task done
1
Support for AMD GPU address sanitizer.
#3007 opened Feb 26, 2026 by carlobertolli Loading…
1
Atomics on 16 bits: prevent reading 4 bytes for 2-byte locations.
#3005 opened Feb 26, 2026 by carlobertolli • Draft
1 task
1
ProTip! Filter pull requests by the default branch with base:develop.