Skip to content

Pull requests: vllm-project/vllm-gaudi

New pull request New
60 Open 1,575 Closed
60 Open 1,575 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

Enable HPU residual fix for qwen3_next (Qwen3-Coder-Next)
#1676 opened Aug 3, 2026 by libinta Collaborator Loading…
NIXL examples: port disagg accuracy test script fixes to v0.26.0
#1675 opened Jul 31, 2026 by skaulintel Contributor Loading…
4
Port of: Add pure-Python MiniMax-M3 tool-call parser for HPU- #1655
#1674 opened Jul 31, 2026 by iboiko-habana Collaborator Loading…
5
test: add security validation unit tests
#1669 opened Jul 30, 2026 by adobrzyn Collaborator Loading…
5
[DO NOT MERGE] Enable Nemotron-H FP8 on HPU: quant-aware Mamba in_proj + non-gated FP8 MoE
#1667 opened Jul 30, 2026 by rsmyrek Contributor Loading…
13
Enable Nemotron-H FP8 on HPU: quant-aware Mamba in_proj + non-gated FP8 MoE
#1666 opened Jul 30, 2026 by rsmyrek Contributor • Draft
19
Warm up native-resolution vision towers at explicit resolutions and item counts
#1661 opened Jul 28, 2026 by libinta Collaborator Loading…
1
Fix #1612: don't wrap INT8 W8A8 loader with FP8-only gaudi_weight_wrapper
#1659 opened Jul 28, 2026 by rsmyrek Contributor Loading…
6
Parallelize some full_tests e2e groups per-card
#1658 opened Jul 28, 2026 by iboiko-habana Collaborator Loading…
6
feat(model) NemotronH support
#1645 opened Jul 27, 2026 by narain1 • Draft
4
Tests optimization
#1643 opened Jul 27, 2026 by iboiko-habana Collaborator Loading…
5
Improve MiniMax-M3 MoE decode performance after upstream merge
#1640 opened Jul 27, 2026 by mkrze Contributor Loading…
3
Add pure-Python MiniMax-M3 tool-call parser for HPU
#1639 opened Jul 27, 2026 by mkrze Contributor Loading…
4
[Do not merge] Enable Nemotron-H (TP8xPP2) on HPU
#1634 opened Jul 24, 2026 by ldurejko Loading…
8
Run Qwen3-30B MoE FP8 load/generate CI tests in parallel
#1633 opened Jul 24, 2026 by iboiko-habana Collaborator Loading…
8
Fix model banner and runtime logging issues - multi models runs
#1608 opened Jul 8, 2026 by 12010486 Contributor Loading…
11
Fix runtime errors on Qwen single process model swapping
#1604 opened Jul 7, 2026 by 12010486 Contributor Loading…
13
[ATEA] [google/gemma-4-31B] Gaudi3 /hpu support
#1596 opened Jul 5, 2026 by sureshnam Collaborator Loading…
9
Gptoss mxfp4 bias v0.23.0
#1577 opened Jun 30, 2026 by SKRohit Contributor Loading…
[kimi2.6] enable model text only initial
#1574 opened Jun 28, 2026 by sureshnam Collaborator Loading…
3
Pin vLLM core version to VLLM_GAUDI_COMMIT in RHEL UBI Dockerfile
#1571 opened Jun 26, 2026 by ghandoura Contributor Loading…
2
MXFP4 GPT-OSS SwiGLU-OAI + bias serving
#1567 opened Jun 25, 2026 by osavchenkox Contributor Loading…
2 of 3 tasks
14
granite-4.0-h-small made compatible with single process models swap
#1565 opened Jun 25, 2026 by 12010486 Contributor Loading…
24
[DRAFT][GAUDISW-248216] Flag compact-GDN per-layer states as independent for the Synapse bridge
#1563 opened Jun 25, 2026 by moshehoori • Draft
4
Port of #1558 Adapt multi_model_api_server to vLLM ServingTokenizatio…
#1559 opened Jun 24, 2026 by iboiko-habana Collaborator Loading…
2
ProTip! Add no:assignee to see everything that’s not assigned.