[sync] vLLM 2026-07-21 — 867ff6973..8def3cdde#1715
Open
AlpinDale wants to merge 81 commits into
Open
Conversation
Upstream-vLLM: d5b1ec268431f394efd6bad0f1b2adec657c31af Co-authored-by: liuzhenwei <zhenweiliu@habana.ai>
… head dims (#48855) Upstream-vLLM: ce4bdcbda4f3c2da67b2fe241e80937ccdcc1fc6 Co-authored-by: Julien Denize <40604584+juliendenize@users.noreply.github.com>
Upstream-vLLM: 850295881041754184717804104fcaadd2b2129e Co-authored-by: liuzhenwei <zhenweiliu@habana.ai>
Upstream-vLLM: 4c6e2e4b308c15fc2bcdf10e278f2591c9cec0dc Co-authored-by: Yejing Lai <yejing.lai@intel.com>
…eter' object has no attribute 'weight_loader' (#48699) Upstream-vLLM: 7b3192523e194e20077017ed104caff37c1ad691 Co-authored-by: Yejing Lai <yejing.lai@intel.com>
Upstream-vLLM: d4b45629172dfb792b44bd142982563a8c8bba9d Co-authored-by: Artur Fierka <artur.fierka@intel.com>
…nd prevent reuse race (#48596) Upstream-vLLM: f38f3d11fb7b180c4d66f4af82cf770f335391cc Co-authored-by: AlexHuang <alex.tech.lab@outlook.com>
…#48417) Upstream-vLLM: ce2aecc4dc42151d63f65d51c675edf28d40d671 Co-authored-by: JooHo Lee <96564470+BWAAEEEK@users.noreply.github.com>
…moe (#48849) Upstream-vLLM: c4dd6d78fd6485e103d71ba815b0a530e5c49439 Co-authored-by: passtoor-agi <305788622+passtoor-agi@users.noreply.github.com>
Upstream-vLLM: 877dae9c684cd8cb2f66eed33f1c3261c6792c10 Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
…(MiniMax M2, Qwen3, MiniCPM5 XML) (#48846) Upstream-vLLM: 11d291511a35bfa2a9ecb8fd21ee1c48b3a78d6b Co-authored-by: mosya415 <abuassiyasin@gmail.com>
…n gfx950 (#48788) Upstream-vLLM: efed8a1e8345de2b4398a48243301d6a5b1a361e Co-authored-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com>
…8729) Upstream-vLLM: bf578e1abdffc2d25232783ff59a3132279e6bdd Co-authored-by: labAxiaoming <34019940+labAxiaoming@users.noreply.github.com>
…KV-cache group (#48012) Upstream-vLLM: 5784507da458656ef8248119590822445d4c67f2 Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Upstream-vLLM: cc25f028b76844024fddd0b4ca18d4ca169b32be Co-authored-by: Michael Goin <mgoin64@gmail.com>
…2E TPOT improvement (#48660) Upstream-vLLM: b5433b6f5079feb32f9f278cf4ae23bd87375148 Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
…e and annotation details (#37524) Upstream-vLLM: fcd2255d16bd3c62493bae5dee769ce998098f21 Co-authored-by: devalshahamd <deval.shah@amd.com>
…#47680) Upstream-vLLM: 41ea2dd44a3a20c46ebeb985de0022c7673fb953 Co-authored-by: aoshen02 <aoshen@inferact.ai>
…ad of sum (#48829) Upstream-vLLM: fae543015cd49dfb4db3afb3a2e96d2eccb3e1db Co-authored-by: Wang Xingda <wangxingda1993@126.com>
Upstream-vLLM: 02c01f442b8bfd0affa1f2ceedda01f2e1384303 Co-authored-by: Michael Goin <mgoin64@gmail.com>
…ms (#48641) Upstream-vLLM: 425c4eafb064f3804538a93584ec07c196d14cf9 Co-authored-by: Michael Goin <mgoin64@gmail.com>
Upstream-vLLM: da64db78b989d820effd090d8ac4b65ade2cbe4f Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
… buffer not block-aligned (#43979) Upstream-vLLM: f12b80c6efa116c7c4020ee611d622ab7030ca40 Co-authored-by: xuebwang-amd <xuebwang@amd.com>
… temporal (#48110) Upstream-vLLM: c71a583aa9f81400528e67e3d818f66b804e8340 Co-authored-by: Francesco Fusco <ffu@zurich.ibm.com>
…_loading (fix replicated / disable_tp weight reload) (#48025) Upstream-vLLM: d96aee09518a57fdfe1e83369d2f4ab515f78c30 Co-authored-by: alexxu-roblox <alexxu@roblox.com>
Upstream-vLLM: c233d90aa826df072872df47b201450059be8e71 Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
…48988) Upstream-vLLM: c7ce03bcbd380d0e94490abb111fe48861c16343 Co-authored-by: Michael Goin <mgoin64@gmail.com>
Upstream-vLLM: 29c0ec4d63d1869f54a9fbcdf082f77534d9211a Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
…rk-per-split heuristic (#46832) Upstream-vLLM: e94243893dd30256f58644ad4ecf779be757dff8 Co-authored-by: frida-andersson <fanderss@amd.com>
…48535) Upstream-vLLM: a287eb163fb6f8f007a4a78411fb54c8dde64cc7 Co-authored-by: yzong-rh <yzong@redhat.com>
Upstream-vLLM: 818cf61e9149c3fdff302cddbb2d026090160aea Co-authored-by: Reid <61492567+reidliu41@users.noreply.github.com>
…-token prompts in chat and raw generate (#49111) Upstream-vLLM: d835ad572ca76d8ed6308e81750813c814fb3343 Co-authored-by: Salt Sato <feathbow@gmail.com>
…00 (#49144) Upstream-vLLM: 530ee36a0d99869804c0af68b41007072b609480 Co-authored-by: hcl <chenglunhu@gmail.com>
…rid caches (#47574) Upstream-vLLM: 8ce53a616ebfc337edb73f847fcefe8af8ba769b Co-authored-by: EdalatiAli <aliedalati@cohere.com>
…ng (#48563) Upstream-vLLM: 0a5069e4e30996866d9d3cc5551f784b85408bb0 Co-authored-by: wangqian <601731555@qq.com>
…loadingSpec (#48679) Upstream-vLLM: f007cceb422599330495b202632d78c7b3b89e61 Co-authored-by: Chang Guo <changg@nvidia.com>
…9159) Upstream-vLLM: 15cb8e140dffd9d2a33737c083fcf1d696e1d1f6 Co-authored-by: Isotr0py <Isotr0py@outlook.com>
…l extension (#49161) Upstream-vLLM: e2d7adeb642bd6e6bdb4f3305005e8608765d7cf Co-authored-by: Bugen Zhao <i@bugenzhao.com>
Upstream-vLLM: b23bd73f540175f9e117eaee5029cd7d8df63964 Co-authored-by: Liangqiusong <xiaolong.guo@intel.com>
Upstream-vLLM: bd091079cba0800d8b8ee8ed22feab5864d1b101 Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
…#49193) Upstream-vLLM: f87836789800ab29d66c5ae7d034e86eff730eaf Co-authored-by: tc-mb <157115220+tc-mb@users.noreply.github.com>
Upstream-vLLM: 9dd62d80abad5d2ebeb0a4a60ad005a23c23d0e9 Co-authored-by: wkutak <wkutak@nvidia.com>
…roups (#48911) Upstream-vLLM: fbfe58133d9fa7c09ad94b21073b43c3accb402b Co-authored-by: coltonottley <coltonottley@gmail.com>
Upstream-vLLM: 7ca017778fc04cfb1a96080e7c484b1b44fc3a64 Co-authored-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
…for spec-decode attention metadata (#44492) Upstream-vLLM: 4ec199b66a791070348f3baf847b3a873c48cdd9 Co-authored-by: Oxana Korzh <okorzh@amd.com>
Upstream-vLLM: 5feb3950e529278b1f9dbe0a4071470d3d382a3d Co-authored-by: Charlie Fu <charlifu@amd.com>
…g (#48639) Upstream-vLLM: 642076d26c98aab899a6cc3dc948856d38c7551b Co-authored-by: Fynn Schmitt-Ulms <fschmitt@redhat.com>
…nal (#48938) Upstream-vLLM: b7c20d0cfa822c011e8d24ce63a2e09bb1cef431 Co-authored-by: Ning Xie <andy.xning@gmail.com>
Upstream-vLLM: 58b2012aa26f2b85560ecd6988c8fa2a773804c1 Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
…nfer monolithic MoE kernel (#44214) Upstream-vLLM: 97a668152b9033caf602b37b88289f03e8faffd8 Co-authored-by: Xuanyu Zhang <xuanyu.zhang@mistral.ai>
…ek-V2/R1 + Kimi-K2.5) (#45964) Upstream-vLLM: 2396a611085d7bc78c0f7c41ab45eccbbc7ec0a7 Co-authored-by: sungsoo ha <sungsooh@nvidia.com>
…#49235) Upstream-vLLM: af91f4b3e4efe921957e12e6dbf05d557845b29a Co-authored-by: Nick Hill <nickhill123@gmail.com>
…cache_config (#48317) Upstream-vLLM: 2e2e626b40e8321731cfa42f0402eb42f229e00a Co-authored-by: ormandj <ormandj@corenode.com>
…ernel_awq (#47268) Upstream-vLLM: 0d9210a50259105d4ce889a78e8142433c273e28 Co-authored-by: flutist-alibaba <30485581+flutist@users.noreply.github.com>
…(#48444) Upstream-vLLM: 0a684ab0c0156fc0c4a8ebae33eba35439a9ef13 Co-authored-by: Alejandro Paredes La Torre <99832715+AlejandroParedesLT@users.noreply.github.com>
Upstream-vLLM: 97a98006b0895bc3d488ec19e898f56c777150b6 Co-authored-by: Chris Leonard <chleonar@redhat.com>
Upstream-vLLM: 72d16aee1576283444352583f36ae65b8ff58455 Co-authored-by: Simon Mo <simon.mo@hey.com>
Upstream-vLLM: 1940c8441eb39d7db88764b8e700d37e50573a84 Co-authored-by: Connor Carpenter <connorcarpenter15@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automated upstream sync: 867ff69733dd..8def3cdde2a0