Skip to content

Pull requests: bitsandbytes-foundation/bitsandbytes

New pull request New
42 Open 711 Closed
42 Open 711 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

[ROCm] Preserve float bits in SIMT DPP reduction
#2089 opened Sep 15, 2026 by liminfei-amd Loading…
1
fix(optim): reject bias_correction and adam_w_mode in LAMB
#2088 opened Sep 15, 2026 by ConnorMoss02 Loading…
fix(optim): reject nesterov and dampening in SGD and LARS
#2085 opened Sep 11, 2026 by ConnorMoss02 Loading…
Fix Params4bit.to() leaving CPU-packed weights packed
#2083 opened Sep 10, 2026 by 22elix3r Loading…
1
2
Fix PytorchLARS updates when momentum is zero
#2082 opened Sep 10, 2026 by mottopanikeiku Loading…
1
[WIP] Prototype original-parameter AdamW8bit state for FSDP1
#2081 opened Sep 10, 2026 by gss10282025 Loading…
Warn once when Linear4bit computes in float32 on GPU
#2077 opened Sep 9, 2026 by thangdangjp Loading…
Make AdEMAMix32bit inherit AdEMAMix so it allocates the double state buffer
#2076 opened Sep 8, 2026 by caiotheodoro Loading…
1
chore(deps): bump softprops/action-gh-release from 3.0.2 to 3.0.3 in the actions group Dependencies github_actions Pull requests that update GitHub Actions code
#2074 opened Sep 7, 2026 by dependabot Bot Loading…
1
[ROCm] Allow explicit GPU architecture override
#2073 opened Sep 4, 2026 by LOGO127 Loading…
1
Add AVX2 path for CPU 4-bit (NF4/FP4) blockwise dequantization
#2072 opened Sep 4, 2026 by rachitj27 Loading…
fix(cuda): chunk 4-bit quant/dequant launches above INT32_MAX elements (#1785)
#2071 opened Sep 4, 2026 by Anai-Guo Contributor Loading…
fix(optim): fall back to non-paged state on all devices without managed memory Cross Platform Low Priority (will be worked on after all priority issues) Optimizers Issues or feature requests relating to optimizers
#2069 opened Sep 3, 2026 by egeozkoc Contributor Loading… v0.50.3
3
Fix CPU Hub kernel backend selection x64 CPU
#2065 opened Aug 30, 2026 by SiluPanda Loading… v0.50.3
1
7
fix(optim): mark unused amsgrad/optim_bits as deprecated in 8-bit Adam classes Optimizers Issues or feature requests relating to optimizers
#2063 opened Aug 28, 2026 by kritikagarg Loading…
1
Fix N-D input handling in default backend int8 ops Cross Platform
#2062 opened Aug 27, 2026 by 2sumtech Loading…
docs: correct Adagrad and matrix multiplication wording
#2058 opened Aug 25, 2026 by Excelius-Wang Loading…
Add Apple MPS NF4 benchmark
#2057 opened Aug 25, 2026 by hamedrabah Loading…
1
Tune SM103 fused 4-bit GEMM SIMT dispatch CUDA Issues and PRs related to the CUDA backend, excluding installation/support help.
#2054 opened Aug 22, 2026 by heiheiha798 Loading…
Remove redundant CUB barriers from blockwise dequantization
#2053 opened Aug 22, 2026 by heiheiha798 Loading…
Respect the current CUDA stream in blockwise quantization
#2052 opened Aug 22, 2026 by heiheiha798 Loading…
Fuse nested 4-bit scale reconstruction on SM103
#2051 opened Aug 22, 2026 by heiheiha798 Loading…
Use the current CUDA stream for non-paged optimizer updates Optimizers Issues or feature requests relating to optimizers
#2050 opened Aug 22, 2026 by heiheiha798 Loading…
Fix CPU dequantize_4bit output shape for 1-D inputs aarch64 x64 CPU
#2048 opened Aug 20, 2026 by 2sumtech Loading…
1
3
Fix mixed INT8 FakeTensor output metadata
#2044 opened Aug 19, 2026 by tandede Loading…
1
ProTip! Type g i on any issue or pull request to go back to the issue listing page.