-
Notifications
You must be signed in to change notification settings - Fork 115
Pull requests: vllm-project/compressed-tensors
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix
get_head_dim crash on transformers>=5 heterogeneous (per-layer) configs
#844
opened Aug 19, 2026 by
yashb98
Loading…
Fix dequantize strategy misidentification for block-quantized weights
ready
When a PR is ready for full CI testing before merge
#843
opened Aug 19, 2026 by
kylesayrs
Collaborator
Loading…
1 of 2 tasks
Add ReadTheDocs build for compressed-tensors
documentation
Improvements or additions to documentation
#842
opened Aug 18, 2026 by
dsikka
Collaborator
Loading…
Remove GPTQ Group/Dynamic Activation Ordering
ready
When a PR is ready for full CI testing before merge
#840
opened Aug 18, 2026 by
Roderick-Wu
Collaborator
Loading…
[Quantization] Fix decompression dtype
ready
When a PR is ready for full CI testing before merge
#839
opened Aug 18, 2026 by
kylesayrs
Collaborator
Loading…
[Converter] magnitude-expert-pruner
documentation
Improvements or additions to documentation
#834
opened Aug 14, 2026 by
soyr-redhat
Contributor
Loading…
6 tasks done
[Docs][3/n] Update int8 and mixed examples
documentation
Improvements or additions to documentation
ready
When a PR is ready for full CI testing before merge
#833
opened Aug 14, 2026 by
dsikka
Collaborator
Loading…
[Docs][2/n] Update bit packing examples
documentation
Improvements or additions to documentation
ready
When a PR is ready for full CI testing before merge
#831
opened Aug 13, 2026 by
dsikka
Collaborator
Loading…
Enable Triton kernel for cast_to_fp4 only
ready
When a PR is ready for full CI testing before merge
#827
opened Aug 11, 2026 by
HDCharles
Collaborator
Loading…
Fix potentially dangerous triton pointers and make sure we run on the right GPU
needs-rebase
#818
opened Aug 6, 2026 by
ElizaWszola
Contributor
•
Draft
Fuse quantize and FP4 pack triton kernels
needs-rebase
#815
opened Aug 5, 2026 by
ElizaWszola
Contributor
Loading…
Fused triton quantize-dequantize in forward helpers
#812
opened Aug 3, 2026 by
ElizaWszola
Contributor
•
Draft
feat: layerwise decompression/compression support
#811
opened Aug 2, 2026 by
kylesayrs
Collaborator
Loading…
4 tasks
feat: skip NVFP4 compression on meta-device tensors
#810
opened Aug 2, 2026 by
kylesayrs
Collaborator
Loading…
3 tasks
save_mtp_tensors_to_checkpoint lazy download
#791
opened Jul 24, 2026 by
kylesayrs
Collaborator
Loading…
add imatrix calibration data resolution to quant config
#787
opened Jul 20, 2026 by
Roderick-Wu
Collaborator
Loading…
[Offload] Support views
documentation
Improvements or additions to documentation
#786
opened Jul 20, 2026 by
kylesayrs
Collaborator
Loading…
Skip default-valued fields during QuantizationArgs serialization
needs-rebase
quality-failed
#784
opened Jul 15, 2026 by
kylesayrs
Collaborator
Loading…
Quant buffer pool for triton kernels
needs-rebase
#782
opened Jul 15, 2026 by
ElizaWszola
Contributor
•
Draft
[Perf] Add pluggable backend dispatch for quantization lifecycle ops
needs-rebase
#773
opened Jul 7, 2026 by
ishrith-gowda
Contributor
Loading…
[New] [Offload] [1/2] Disambiguate synchronous
__setitem__/offload from update_offload
#768
opened Jul 6, 2026 by
kylesayrs
Collaborator
Loading…
Previous Next
ProTip!
Adding no:label will show everything without a label.