Skip to content

Pull requests: state-spaces/mamba

New pull request New
63 Open 148 Closed
63 Open 148 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

Fix Mamba-3 decode path: forward() to step() shape handling
#978 opened Jun 20, 2026 by sanowl Loading…
Make CUDA selective scan install explicit opt-in
#977 opened Jun 15, 2026 by berlinchen7 Collaborator Loading…
Add support for input states to Mamba-3 forward pass
#976 opened Jun 15, 2026 by aakashlahoti Collaborator Loading…
Installation and Build Failure on Modern Cloud Environments (Python 3…
#975 opened Jun 14, 2026 by Turusore08 Loading…
Add state_batch_indices to mamba3_step_fn (paged-inference state pools, parity with selective_state_update)
#971 opened Jun 11, 2026 by jgcb00 Loading…
Remove host syncs from compute_dacs_segsum_triton_varlen
#969 opened Jun 11, 2026 by jgcb00 Loading…
fix: skip duplicate B/C RMSNorm in checkpoint_lvl=1 backward
#964 opened Jun 2, 2026 by Chessing234 Contributor Loading…
1
fix: fall back to source build when version string is not PEP 440
#961 opened May 30, 2026 by Chessing234 Contributor Loading…
Fix Mamba-3 MIMO TileLang kernel cache churn for dynamic batch/seq/head dims
#946 opened May 10, 2026 by Copilot AI • Draft
@aakashlahoti
1
feat: ANE for Mac/iPhone support
#943 opened May 7, 2026 by Dorianhgn • Draft
Update angle normalization using remainder function
#922 opened Apr 16, 2026 by sukjunhwang Loading…
1
Skip selective_scan_cuda build on ROCm/HIP
#915 opened Apr 12, 2026 by ChrisLundquist Loading…
Fix Mamba-3 Triton kernels for AMD ROCm/RDNA4
#914 opened Apr 12, 2026 by ChrisLundquist Loading…
4 of 6 tasks
Cache ctx.saved_tensors in Mamba3 backward to fix FSDP activation checkpointing
#909 opened Apr 9, 2026 by anasiri Loading…
Fix Mamba2 step() D handling when D_has_hdim=True
#893 opened Apr 2, 2026 by ccs1112 Loading…
3 tasks done
2
Allow overriding BASE_WHEEL_URL via an environment variable
#892 opened Apr 2, 2026 by songyz2019 Loading…
selective scan cuda kernel x buffer overflow fix
#883 opened Mar 26, 2026 by darxradi3nt Contributor Loading…
docs: add uv as alternative installation option
#874 opened Mar 23, 2026 by MasahiroOgawa Loading…
2 tasks done
feat: add speed benchmark, interactive example, and RTX 4080 test
#873 opened Mar 23, 2026 by MasahiroOgawa Loading…
4 tasks done
refactor: rename benchmark for clarity, add benchmark_README.md
#872 opened Mar 23, 2026 by MasahiroOgawa Loading…
1 task done
fix: add sm_89 (Ada Lovelace) gencode and fix Python/dep versions
#871 opened Mar 23, 2026 by MasahiroOgawa Loading…
3 tasks done
Respect TORCH_CUDA_ARCH_LIST to speed up builds
#856 opened Mar 13, 2026 by d3banjan Loading…
4 tasks
1
feat: Add Mamba-3 implementation (ICLR 2026)
#850 opened Mar 8, 2026 by payk24 Loading…
4 tasks done
X
#849 opened Mar 4, 2026 by mitchelfalcon Loading…
fix: Add MSVC support for Windows
#846 opened Feb 12, 2026 by LAWArthur Loading…
ProTip! Type g i on any issue or pull request to go back to the issue listing page.