Skip to content

Pull requests: vllm-project/llm-compressor

New pull request New
97 Open 1,587 Closed
97 Open 1,587 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 #2984: prevent DDP race on shared HF cache during model loading bug Something isn't working enhancement New feature or request llama For any PR / issue related to Llama herd support smoothquant For any issue / PR related to SmoothQuant support transforms Related to transforms-based modifiers like SpinQuant and Quip w4a16
#3030 opened Aug 14, 2026 by sonalibiswas242 Loading…
1
10
[Feature] Add layerwise hidden-state QAD for quantized block reconstruction quality-failed
#3029 opened Aug 14, 2026 by BeichenHuang Loading…
11
HIGGS: ILP-based mixed-precision quantization quality-failed
#3028 opened Aug 13, 2026 by HDCharles Collaborator Loading…
4 tasks
9
Add config filter flag to test runner script
#3027 opened Aug 13, 2026 by HDCharles Collaborator Loading…
2 tasks
8
docs: model-size guidance for scheme selection, language alignment for calibration data documentation Improvements or additions to documentation two-reviews When a PR requires two reviews
#3025 opened Aug 12, 2026 by spped2000 Loading…
8
fix(autoround): prevent input_capture_hook from accumulating GPU memory during optimization autoround For any PR / issue related to autoround support ready When a PR is ready for full CI testing before merge two-reviews When a PR requires two reviews
#3024 opened Aug 12, 2026 by xesdiny Loading…
1
14
[Tests] Weekly lm eval tests
#3022 opened Aug 11, 2026 by kylesayrs Collaborator Loading…
9
[error message improvement] include num_calibration_samples/max_seq_length in oom resolution message needs-rebase
#3020 opened Aug 11, 2026 by brian-dellabetta Collaborator Loading…
1
6
[lut-b] data-free round-to-nearest example
#3018 opened Aug 11, 2026 by brian-dellabetta Collaborator • Draft
4 of 5 tasks
7
[WIP] Add explicit LinearExperts2D -> 3D MoE repack for HF-native saves two-reviews When a PR requires two reviews
#3017 opened Aug 11, 2026 by GOavi101 Contributor Loading…
8
Roderick wu/sanity check fouroversix quality-failed
#3015 opened Aug 10, 2026 by Roderick-Wu Collaborator • Draft
7
Reject long untruncated calibration samples when max_seq_length is unset two-reviews When a PR requires two reviews
#3012 opened Aug 9, 2026 by rishabhsinha17 Loading…
1
9
[TEST] No-trigger test: .md file change documentation Improvements or additions to documentation two-reviews When a PR requires two reviews
#3009 opened Aug 7, 2026 by rashmigottipati Collaborator Loading…
11
[TEST] No-trigger test: tests/e2e/ change two-reviews When a PR requires two reviews
#3008 opened Aug 7, 2026 by rashmigottipati Collaborator Loading…
8
[TEST] Trigger test: .buildkite/ change two-reviews When a PR requires two reviews
#3007 opened Aug 7, 2026 by rashmigottipati Collaborator Loading…
6
[TEST] Trigger test: src/ change two-reviews When a PR requires two reviews
#3006 opened Aug 7, 2026 by rashmigottipati Collaborator Loading…
8
[Tests] Nightly lmeval tests
#3004 opened Aug 7, 2026 by kylesayrs Collaborator Loading…
8
Add native FP8 expert support to REAP two-reviews When a PR requires two reviews
#3003 opened Aug 6, 2026 by ieBoytsov Loading…
8
[Performance] Speed up subgraph tracing - add direct dict lookup for O(1) node membership checks enhancement New feature or request Refactor Code cleanup and/or improvements to existing features tracing Issues related to model tracing two-reviews When a PR requires two reviews
#2999 opened Aug 4, 2026 by wanadzhar913 Contributor Loading…
19
Avoid CUDA indexing failures in iMatrix grid search updates two-reviews When a PR requires two reviews
#2997 opened Aug 3, 2026 by yasu-oh Loading…
11
Let AWQ mappings absorb scales into a sliced projection awq For any issue / PR related to AWQ support two-reviews When a PR requires two reviews
#2996 opened Aug 3, 2026 by Ar4ikov Loading…
@brian-dellabetta
9
Add layerwise decompression and compression to sequential pipeline quality-failed
#2995 opened Aug 2, 2026 by kylesayrs Collaborator Loading…
5 tasks
8
feat: add Kimi-K3 model definition and quantization example quality-failed
#2994 opened Aug 2, 2026 by kylesayrs Collaborator Loading…
3 tasks
23
feat: resumable model-free PTQ with meta-tensor skip quality-failed
#2993 opened Aug 2, 2026 by kylesayrs Collaborator Loading…
3 tasks
8
[Performance] Speed up subgraph tracing enhancement New feature or request tracing Issues related to model tracing two-reviews When a PR requires two reviews
#2992 opened Aug 2, 2026 by YingqiDuan Loading…
9
ProTip! Adding no:label will show everything without a label.