-
-
Notifications
You must be signed in to change notification settings - Fork 860
Pull requests: bitsandbytes-foundation/bitsandbytes
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(nn/Linear4bit): consume QuantState keys in _load_from_state_dict (#1946)
#1958
opened May 28, 2026 by
Anai-Guo
Loading…
docs(cuda): use Windows-native BNB_CUDA_VERSION guidance
#1957
opened May 27, 2026 by
peter941221
Loading…
refactor: Rewrite
assert statements as exceptions
#1931
opened Apr 23, 2026 by
rapsealk
Contributor
Loading…
5 of 7 tasks
ci: add nightly integration tests for transformers, accelerate, peft
CI/CD
#1923
opened Apr 13, 2026 by
Titus-von-Koeller
Collaborator
Loading…
[ROCm] Optimize kgemm_4bit_inference_naive for ROCm, use it for batch sizes other than 1
CUDA
Issues and PRs related to the CUDA backend, excluding installation/support help.
ROCm
#1920
opened Apr 11, 2026 by
sstamenk
Contributor
Loading…
docs: add quantization and energy efficiency guide
Documentation
Improvements or additions to documentation
#1882
opened Feb 24, 2026 by
hongping-zh
Loading…
Fix: Linear8bitLt crash with tied weights (skip_modules)
#1865
opened Feb 14, 2026 by
TimDettmers
Collaborator
Loading…
3 tasks
fix: Stop quantize_blockwise and quantize_4bit from mutating user-provided absmax
#1863
opened Feb 14, 2026 by
TimDettmers
Collaborator
Loading…
3 tasks done
Add k-bit blockwise quantization (K=2-5) with warp-level CUDA kernels
CUDA
Issues and PRs related to the CUDA backend, excluding installation/support help.
#1858
opened Feb 14, 2026 by
TimDettmers
Collaborator
Loading…
Enable Ascend NPU Backend with Custom Ops Integration for NF4 Support
Ascend NPU
Related to Ascend NPU backend
#1695
opened Jul 2, 2025 by
SlightwindSec
Loading…
ProTip!
Adding no:label will show everything without a label.