Skip to content

Pull requests: vllm-project/speculators

New pull request New
29 Open 335 Closed
29 Open 335 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

clean up docs documentation Improvements or additions to documentation
#467 opened Apr 23, 2026 by shanjiaz Collaborator • Draft
4 tasks done
3
(WIP) Add multi-gpu marker quality-failed
#466 opened Apr 23, 2026 by shanjiaz Collaborator • Draft
4 tasks
2
Add FP8 quantization for hidden state data generation needs-rebase two-reviews
#461 opened Apr 22, 2026 by shubhra Collaborator Loading…
6
feat(mtp): add MTP model architecture, converter, and stitcher two-reviews
#452 opened Apr 21, 2026 by rahul-tuli Collaborator Loading…
3 of 4 tasks
@rahul-tuli
13
Add guidellm-based performance benchmarking utilities documentation Improvements or additions to documentation
#434 opened Apr 17, 2026 by anmarques Collaborator Loading…
2 of 4 tasks
10
fix: resolve Pydantic v2 + transformers MRO conflicts two-reviews
#431 opened Apr 16, 2026 by ianliuy Loading…
1
7
docs: add FastMTP documentation, benchmarks, and integration tests documentation Improvements or additions to documentation
#430 opened Apr 16, 2026 by ianliuy Loading…
1
8
Fix local_rank usage in distributed batch sampler for multi-node training
#427 opened Apr 15, 2026 by ianliuy Loading…
7
Add simple timer metrics for trainer profiling quality-failed
#425 opened Apr 14, 2026 by mgoin Member • Draft
2
Add support for async hidden states connector
#424 opened Apr 14, 2026 by fynnsu Collaborator Loading…
3 of 4 tasks
5
Update pytest requirement from ~=8.2.2 to ~=9.0.3 dependencies Pull requests that update a dependency file python Pull requests that update python code quality-failed
#421 opened Apr 14, 2026 by dependabot Bot Loading…
2
Update setuptools-git-versioning requirement from <3,>=2.0 to >=3.0.1,<4 dependencies Pull requests that update a dependency file python Pull requests that update python code
#415 opened Apr 14, 2026 by dependabot Bot Loading…
1
[WIP] P-eagle Training Phase 1 implementation documentation Improvements or additions to documentation
#393 opened Apr 11, 2026 by shanjiaz Collaborator • Draft
3 of 4 tasks
9
feat(fast-mtp): FastMTP speculative head finetuning for Qwen3-Next documentation Improvements or additions to documentation training
#366 opened Mar 26, 2026 by rahul-tuli Collaborator • Draft
@rahul-tuli
4
fix(datagen): fix VllmHiddenStatesGenerator for hybrid KV cache models and large datasets
#365 opened Mar 26, 2026 by rahul-tuli Collaborator • Draft
@rahul-tuli
1
[Doc] Add a choosing algorithms section documentation Improvements or additions to documentation
#355 opened Mar 22, 2026 by DonaghBr Loading…
4 tasks done
1
[WIP]update generator to support vllm latest main
#348 opened Mar 16, 2026 by shanjiaz Collaborator • Draft
3 of 4 tasks
2
add multimodal preprocessing support
#344 opened Mar 13, 2026 by shx2005 Loading…
3 of 4 tasks
6
add speculators finetune examples documentation Improvements or additions to documentation needs-rebase
#339 opened Mar 10, 2026 by Annarine Loading…
4 tasks
4
feat: P-EAGLE Phase 1 - COD sampling, model, config, and loss utilities needs-rebase
#327 opened Mar 4, 2026 by NJX-njx Loading…
1
11
docs: add algorithm documentation for EAGLE-3 and FastMTP documentation Improvements or additions to documentation
#326 opened Mar 4, 2026 by NJX-njx Loading…
7
Avoid D2H copies in VllmHiddenStatesGenerator to speed up offline data generation by x5
#322 opened Mar 2, 2026 by qGentry Loading…
5
Materialize non-layer modules after FSDP sharding
#314 opened Feb 27, 2026 by guan404ming Contributor Loading…
7
Feature/tool use support
#300 opened Feb 24, 2026 by olisicky Loading…
10
add multinode train support
#299 opened Feb 24, 2026 by Liccol Loading…
15
ProTip! Filter pull requests by the default branch with base:main.