Gpt Oss Enablement#441

Closed

hlahkar wants to merge 55 commits intovllm-project:mainfrom

hlahkar:gpt_oss_porting

Commits on Oct 24, 2025

Initial Commit GPT-OSS
hlahkar
committed
Add Naive Attention
hlahkar
committed
[Bugfix] Fix bucketing UT (#367)

kzawora-intel
authored and
hlahkar
committed
[GITHUB ACTION] Remove commits comparison so we can rerun (#373)

xuechendi
authored and
hlahkar
committed
[CI] Set seeds for e2e tests (#368)

kzawora-intel
authored and
hlahkar
committed
Fix dp padding after upstream change #25768 (#362)

authored and
hlahkar
committed
Create LICENSE (#379)

kzawora-intel
authored and
hlahkar
committed
Change to starting page and installation (#371)

authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fix upstream crash introduced by #24486 + #24926 + #25103 + #25807 (#366)

authored and
hlahkar
committed
Enable Parallel Compilation feature for compile mode by default (#370)

authored and
hlahkar
committed
[SW-239226] Adjust junit xml filenames for retry mechanism (#382)

tlipinski1337
authored and
hlahkar
committed
Docs installation, quick start and build fixes (#384)

PatrykWo
authored and
hlahkar
committed
Correct htexp._data_ptr utility (#387)

xinyu-intel
authored and
hlahkar
committed
ray: pin ray to <2.49.0 (#386)

xinyu-intel
authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fix #24172, [Refactor]: Use M-RoPE interface directly while defining model class instead of maintaining model specific M-RoPE implementation in mrope.py (#388)

authored and
hlahkar
committed
[Bugfix] Fix min linear decode value (#391)

authored and
hlahkar
committed
[SW-241908] Omit all prompt buckets that exceed max_num_batched_tokens (#331)

authored and
hlahkar
committed
Experimental - fatal errro from 0.12 release (#398)

authored and
hlahkar
committed
Port: [Docs] CI failures chapter (#276) (#389)

authored and
hlahkar
committed
Fix issue with async_scheduling when dealing with chunked input (#360)

authored and
hlahkar
committed
nixl: support mla kvcache transfer (#403)

xinyu-intel
authored and
hlahkar
committed
Unified Attention Accuracy Bugfixes (#393)

authored and
hlahkar
committed
Minor optimizationm for bucketing calc (#395)

michalkuligowski
authored and
hlahkar
committed
Fix linear assert (#401)

kamil-kaczor
authored and
hlahkar
committed
Enviroment logs - disable prefix caching with conti pa + add vllm brnach+commit value to logs (#402)

adobrzyn
authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Upstream vllm fixes for #26355 and #26737 (#407)

authored and
hlahkar
committed
Cherrypick cd docker fixes/commits from v0.10.2 to main v0.11.0 (#341)

authored and
hlahkar
committed
Unit test for prefix caching in Gaudi plugin (#349)

iirzynska
authored and
hlahkar
committed
Add missing prompt bucket to warmup, when max_ctx is 0 (#352)

authored and
hlahkar
committed
Unified attention improvemets (#363)

authored and
hlahkar
committed
[NIXL][BUGFIX][Gaudi2Gaudi accuracy] use 4d kv_cache for nixl_connector KV register and update host_buffer accordingly (#411)

xuechendi
authored and
hlahkar
committed
Multi-image generation CI tests (#377)

authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fix for Separate out vllm.utils.collections #26990 (#413)

authored and
hlahkar
committed
Add fp8 calibration procedure (#309)

afierka-intel
authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fix for #27022 (#418)

adobrzyn
authored and
hlahkar
committed
[CI]unified attn is too easy to fail, add small RTOL (#422)

xuechendi
authored and
hlahkar
committed
Update supported_features.md (#180)

authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fixes for upstream #26908 and #27143 and #27169 (#427)

iboiko-habana
authored and
hlahkar
committed
[NIXL]Enable prefill TP < Decode TP with host_buffer (#421)

xuechendi
authored and
hlahkar
committed
Fix typo in installation.md: correct script name to install_nixl.py (#385)

authored and
hlahkar
committed
[SW-242466] Update not_over_max_model_len filter to fix warmup perf regression (#424)

skavulya
authored and
hlahkar
committed
Docs update post v0.11 (#428)

PatrykWo
authored and
hlahkar
committed
[FIX_FOR_VLLM_LATEST] Fix for #26440 (#442)

iboiko-habana
authored and
hlahkar
committed
[main] Defragmenter warmup accuracy workaround (#436)

authored and
hlahkar
committed
Update docs: Quickstart - Executing inference (#410)

authored and
hlahkar
committed
[Security] Update requirements.txt (#443) (#445)

afierka-intel
authored and
hlahkar
committed
[GITHUB ACTION] Always run same job to same node (#450)

xuechendi
authored and
hlahkar
committed
reuse DP allgather tensor across layers (#415)

authored and
hlahkar
committed
Update test case
hlahkar
committed
Support DP for unified attention (#242)

authored and
hlahkar
committed
[Linear warmup] Default values optimization (#426)

adobrzyn
authored and
hlahkar
committed
Buckets from file - alpha version (#375)

authored and
hlahkar
committed
Fix math log2 exponential bucket error if max_model_len <= block_size (#451)

skavulya
authored and
hlahkar
committed
Fix requirements filtering in HPU Dockerfiles (#419)

jakub-sochacki
authored and
hlahkar
committed
View to be shape agnostic
hlahkar
committed