193 Commits
Author SHA1 Message Date
MOVZXandHinata 5aa2ac64f8 [Mini-VLLM] Remove LoRA, plugins, speech-to-text and scale-out surfaces
Production runs one static model behind the OpenAI-compatible API. Every
adapter path and non-OpenAI HTTP surface is unreachable there.

- Delete the LoRA stack: vllm/lora, LoRARequest on the engine wire, the
  scheduler, worker, cudagraph and MoE plumbing, LoRA metrics, the lora
  CLI group, and the LoRA tests.
- Drop the plugin loader in vllm/plugins. Platform resolution is now
  CUDA-only and the general/stat-logger plugin hooks are gone.
- Delete the speech-to-text and realtime entrypoints, the scale-out
  render routes, and vllm/config/speech_to_text.py. GenerationTask is
  generate-only.
- Cut DecodeBenchConnector and the disagg-prefill doc and image. The
  Multi, Offloading and Example connectors stay; production uses the
  CPU KV offload tier.
- Delete the ROCm-only fused KDA chunk kernel (2346 lines) with its CMake
  arch gate, declarations and op registrations. The CUDA build never
  defined the guard and no Python caller existed.
- Delete the orphan Mistral and Pixtral chat fixtures and examples/.
- Fix _check_model: removing the LoRA early return made every request
  fail with "The model ... does not exist. Valid aliases: ...".

Verified on an RTX 3090: import sweep reports 1242 modules with 7 known
optional-dependency failures; a live server answers chat, completions,
responses, tokenize and detokenize with zero log errors and 404 for the
removed surfaces; collection yields 8558 tests with the 3 pre-existing
errors; ruff adds no new finding; the incremental _C build succeeds.

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-19 21:06:19 +07:00
MOVZXandHinata 5f3c7bd8ab [Mini-VLLM] Remove the ROCm device path, debug routes and arch-gated tests
- Delete csrc/rocm, hipify, the _rocm_C extension and the setup.py rocm
  target path; drop the rocm requirement files
- Delete the aiter, rdna3, skinny-gemm and ROCm fused-MoE kernel modules
  and retarget the linear kernel registries away from them
- Reduce platforms/rocm.py to the GCN architecture predicates that live
  code still imports, and delete the RocmPlatform class and plugin
- Drop the VLLM_SERVER_DEV_MODE debug surface: reset, collective_rpc,
  server_info and sleep routes plus the log_error_stack flag
- Delete lm-format-enforcer, lark and otel-semantic-conventions-ai pins;
  the live server process maps none of them
- Delete tests that only build architectures this fork no longer
  registers, and test directories left without any test file
- Restore the tests/v1/kv_connector mock helpers that scheduler tests
  import at module level

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-19 19:31:07 +07:00
MOVZXandHinata 54a396586e [Mini-VLLM] Repair test suite after SM86 cut
- Restore example connector, test helpers, and dropped patch tests
- Guard MLA metadata import in the example connector
- Point the PLE test at common.ops.ple after upstream #60021 move
- Cut tests for removed features: tools, DRY, epd proxy, STT batch,
  diffusion scheduler, INC/Quark PLE cases, mamba-2 mixer

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-19 01:06:30 +07:00
MOVZXandHinata e61dcb9143 [Mini-vLLM] SM86-only cut: models, backends, Rust, tests
- Keep only Qwen3.5, Qwen3.5-MTP and Qwen4Exp model paths; trim the model registry to 41 entries
- Delete idle attention backends (MLA, FlashInfer, ROCm-only, quant backends) and dead GDN CuTeDSL kernels
- Remove inc/quark quant loaders; keep compressed-tensors and Marlin
- Drop the Rust frontend, pooling entrypoints, and idle test suites
- Rebuild kernels for sm_80;sm_86 and strip the extensions
- Verified: full import sweep clean (optional deps only) and a dummy-load Qwen3.5 engine run on RTX 3090 via the Triton GDN path

Co-authored-by: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-18 19:19:56 +07:00
MOVZX be95edb7f6 Merge branch 'main' into patches
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-16 17:35:42 +07:00
Andreas KaratzasandClaude 29345f42a6 [ROCm][Qwen4Exp] Support HC and MoE sequence parallelism on ROCm (#61471)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-10-15 20:11:17 -05:00
d92a4f25f2 [watermarking] Add DFlash support to dual-key gumbel-max watermarking (#59554)
Signed-off-by: Shern Shiou Tan <shernshiou@gmail.com>
Signed-off-by: Simon Veitner <sveitner@redhat.com>
Signed-off-by: Raphael Rialland <raphael.rialland@mistral.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Simon Veitner <sveitner@redhat.com>
Co-authored-by: Raphael Rialland <raphael.rialland@mistral.ai>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
2026-10-15 22:06:06 +00:00
qizixiClaude Opus 5.5mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
7cd7982dfd [Spec Decode] Don't reserve DFlash/DSpark draft slots in the scheduler token budget (#59468)
Signed-off-by: zixi-qi <zixi@inferact.ai>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-16 01:23:58 +08:00
MOVZX 76df77c04e Merge branch 'main' into patches
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-15 21:16:39 +07:00
Harry MellorClaude Opus 5.5mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>Kevin H. Luu
b965ee564d [Chore] Use upstream Transformers DeepseekV4Config (#61392)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
2026-10-15 01:08:08 -07:00
48f17b77bb [Bugfix] Reject mismatched float16/bfloat16 KV cache dtype (#60885)
Signed-off-by: cipheraxat <cipheraxat@users.noreply.github.com>
Co-authored-by: cipheraxat <cipheraxat@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 08:06:59 +00:00
Harry Huangandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> c75cb63cd2 [Qwen4Exp] support Sequence Parallelism for Qwen3.8-Flash-Next (#56322)
Signed-off-by: huanghaoyan.hhy <huanghaoyan.hhy@alibaba-inc.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-12 20:56:44 +08:00
Nick Hill 7d0b4e57aa [ROCm] Remove MRV1 model exceptions (#56360)
Signed-off-by: Nick Hill <nickhill123@gmail.com>
2026-10-10 05:09:19 +00:00
04462732a4 [Bugfix] Reject max_num_scheduled_tokens=0 at config time instead of hanging (#60714)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Hui Li <40770106+untrall@users.noreply.github.com>
2026-10-09 19:15:10 +00:00
MOVZX ac305c44cb Merge branch 'main' into patches
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-09 17:41:13 +00:00
Woosuk Kwon 194bd72be7 [Compile] Remove torch.compile-based sequence parallelism and AsyncTP (#60517) 2026-10-08 08:28:23 -07:00
819852df6b [ModelRunner V2] Speculative Decoding NGram GPU Implementations (#40704)
Signed-off-by: PatchouliTaisa <patchychen@tencent.com>
Signed-off-by: PatchouliTaisa <pyramkar@gmail.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Signed-off-by: Xuanan Chen <xuananchenc@nvidia.com>
Co-authored-by: PatchouliTaisa <patchychen@tencent.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Xuanan Chen <xuananchenc@nvidia.com>
Co-authored-by: Giancarlo Delfin <32987265+TheEpicDolphin@users.noreply.github.com>
2026-10-06 13:32:59 -07:00
Nishant RevurandClaude ad0f67a7ec [Bugfix][Spec Decode] Keep heterogeneous-vocab draft models on Model Runner V1 (#59541)
Signed-off-by: Nishant Revur <kumrfni@amazon.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-10-06 18:03:44 +00:00
Raphaël RiallandandSimon Veitner c32495d58f [watermark] compatibility validation (#56801)
Signed-off-by: Raphael Rialland <raphael.rialland@mistral.ai>
Signed-off-by: Simon Veitner <sveitner@redhat.com>
Co-authored-by: Simon Veitner <sveitner@redhat.com>
2026-10-05 14:37:13 -04:00
Matthew BonanniandClaude Opus 5.5 54d93af9fb [Bugfix][HiSparse] Reject cudagraph_mode=FULL at startup (#59688)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 12:42:27 -04:00
MOVZX 2877121b2d Merge remote-tracking branch 'upstream/main' into patches
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-10-03 21:59:03 +07:00
Harry MellorandClaude Opus 5 e3cae8d2ac [Model] Use upstream GLM-5.3 and Qwen4-Exp configs and processor (#57387)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-10-02 02:25:18 +00:00
cb6458c5ce [Core] Remove deprecated mamba_cache_mode "all" (#58997)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Thomas Parnell <tpa@zurich.ibm.com>
2026-10-01 02:59:37 -07:00
70be6aae19 [Feature][Spec Decode] Support sampling mask replay for MRV2 MTP (#59359)
Signed-off-by: Cuiping Cheng <chengcuipingswu@163.com>
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: Cuiping Cheng <chengcuipingswu@163.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 13:51:39 +08:00
Andrii SkliarandAndrii Skliar deb148e730 [Mamba] Add FlashInfer ReplaySSM support for MTP (#52928)
Signed-off-by: Andrii Skliar <askliar@nvidia.com>
Co-authored-by: Andrii Skliar <askliar@nvidia.com>
2026-09-30 21:41:17 +00:00
0f8b398158 [watermarking] add context deduplication support to speculative decoding (#56807)
Signed-off-by: Raphael Rialland <raphael.rialland@mistral.ai>
Signed-off-by: Simon Veitner <sveitner@redhat.com>
Co-authored-by: Simon Veitner <sveitner@redhat.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
2026-09-30 12:22:10 -04:00
90e13fc757 [Bugfix][HiSparse] Never allocate GPU pages without host backing (#59036)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-30 04:13:29 +00:00
jackandQwertyJack 79a0307d74 [Config] Remove Engram CUDA-alike device restrictions (#59171)
Signed-off-by: QwertyJack <7554089+QwertyJack@users.noreply.github.com>
Co-authored-by: QwertyJack <7554089+QwertyJack@users.noreply.github.com>
2026-09-29 12:42:02 +00:00
MOVZXandHinata ae0350c0de [Spec Decode][MRV2] RecoverSSM for Qwen GDN and PLE short conv (PR #58863)
Intake of open upstream PR vllm-project/vllm#58863. Active only with
VLLM_USE_V2_MODEL_RUNNER=1; inert otherwise.

Speculative decoding on Gated DeltaNet snapshotted the full recurrent
state after every draft position and reserved one Mamba block per draft
position. On a 24 GB card that is reserved VRAM we never use, plus
per-step copy traffic.

RecoverSSM, the protocol Kimi-K3 KDA already uses, replaces this:
verify runs from a read-only checkpoint and stores a small per-token
replay record [correction (V), normalized key (K), gate (1)] in fp32;
commit replays only the accepted tokens and writes the recurrent state
once, plus the block-boundary state in align mode. The PLE dilated
short conv compacts its accepted window with the same plan.

Conflict in vllm/models/qwen4_exp/nvidia/mtp.py: our get_top_tokens()
work put LocalArgmaxMixin on Qwen4ExpMTP, the PR adds SupportsReplaySSM
to the same base list. Kept both.

Co-Authored-By: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-09-29 13:39:09 +07:00
MOVZX a63dca4afb Merge remote-tracking branch 'upstream/main' into patches
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-09-29 12:54:18 +07:00
Ren YuzhouandClaude ec5e0c352f [Bugfix][Engram] Keep THP tables private when resolving shared memory (#59068)
Signed-off-by: Ren Yuzhou <54501155+yuzhouo7@users.noreply.github.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-09-29 05:19:49 +00:00
Tyler Michael SmithandCodex 4406874e7c [Bugfix][FT] Pass stateless process-group timeouts explicitly (#58611)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
2026-09-29 03:14:02 +00:00
MOVZXandHinata 43fbd6510b [Qwen4Exp] Serve PLE rows from the loader's mapped checkpoint shards (PR #58815)
Adds an opt-in backend, --engram-config '{"host_file_gather": true}', that
serves Qwen4Exp PLE rows from the checkpoint files through the page cache
instead of allocating the 47.7 GiB table on GPU or in pinned host memory.

- Qwen4ExpPLEFileGatherEmbedding allocates no table rows; load_weights keeps
  the default loader's zero-copy safetensors views of each shard. At bind
  (post-load hook) every view must be file-backed per /proc/self/maps (HF
  cache symlinks work; heap/pinned/anonymous memory is rejected) and the
  shards must cover the whole table.
- Each step, Qwen4ExpModelState.prepare_inputs computes the PLE n-gram ids,
  copies them to pinned host memory, syncs the stream once, issues
  POSIX_FADV_WILLNEED over the coalesced page ranges of the step's rows,
  gathers the rows on the host, and copies them into a fixed device staging
  buffer. The forward reads only the staging buffer, so FULL CUDA graph
  capture is unaffected.
- Rows stay in the checkpoint dtype (FP8 table stays FP8, dequantized on the
  GPU); dummy loads serve zeros; weight reload is rejected; takes precedence
  over cpu_offload; rejected on ROCm and with elastic EP.

Co-Authored-By: Hinata <hinata@idihore.id>
Signed-off-by: MOVZX <movzx@yahoo.com>
2026-09-27 09:54:42 +07:00
Jürgen SchmiedandClaude a80e41d690 [Mamba] Share RecoverSSM commit kernels and helpers across models
RecoverSSM's commit-plan and conv-window compaction kernels are model independent.
Move them from models/kimi_k3 into model_executor/layers/mamba/ops/recoverssm.py
(Kimi-K3 keeps its private names as import aliases; kernel source unchanged), add
mamba/recoverssm_utils.py (activation predicate, argument validation) and the
generic spec-row/request mapping helpers to v1/attention/backends/recoverssm_metadata.py.

CacheConfig.use_kda_recoverssm becomes use_recoverssm: the switch now selects RecoverSSM
for any model whose recurrent layers implement it, not only KDA.

No functional change for Kimi-K3.

Signed-off-by: Jürgen Schmied <juergenschmied70@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-09-27 00:07:04 +02:00
a4eb3f25d6 [Spec Decode] Enable Gemma4 DSpark adaptive verification with FlashInfer (#57263)
Signed-off-by: zixi-qi <zixi@inferact.ai>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
2026-09-26 17:01:17 +08:00
73a78e6f1f [SpecDecode] Add LiLiCorr drafter (#57934)
Signed-off-by: Andrii Skliar <askliar@nvidia.com>
Signed-off-by: Andrii Skliar <andreyws96@gmail.com>
Co-authored-by: Andrii Skliar <askliar@nvidia.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Benjamin Chislett <chislett.ben@gmail.com>
2026-09-25 10:08:28 -04:00
267eee54bf [Qwen4Exp][ROCm] PLE n-gram table CPU offload (#57497)
Signed-off-by: Mathew Odden <modden@redhat.com>
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: opencode+deepseek-v4-flash+vllm <opencode+deepseek-v4-flash+vllm@example.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
2026-09-24 23:28:20 -05:00
stefankoncarevicandAndreas Karatzas e33de821c0 [ROCm][CI] skip the ROCm MRV1 default where MRV1 cannot serve the config (#58535)
Signed-off-by: Stefan Koncarevic <Stefan.Koncarevic@amd.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
2026-09-24 15:36:36 -05:00
LucainClaude Opus 5Harry Mellormergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
344fcc252a [Bugfix] Resolve the Hub revision once per repo (#56092)
Signed-off-by: Wauplin <lucainp@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-09-24 08:48:24 -07:00
Harry MellorandClaude Opus 5.5 d881812500 [Chore] Use Transformers v5 names and drop redundant processor use_fast (#58550)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-24 14:14:10 +00:00
Lio Einaudi 711fc55c10 [Perf] Use breakable CUDA graphs (no torch.compile) by default under VLLM_BATCH_INVARIANT so the tuned matmul configs see the runtime M (#57586)
Signed-off-by: LioEinaudi <zhao3024667639@gmail.com>
2026-09-23 14:39:38 -04:00
826e300cde [Bugfix] Skip VllmConfig re-validation for with_hf_config submodel views (#58212)
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Kimi Code <noreply@moonshot.ai>
Co-authored-by: Roger Wang <rogerw@inferact.ai>
2026-09-23 00:50:17 +00:00
Artem PerevedentsevandCNE Pierre FICHEPOIL 1ea7c63f4a [Bugfix][Structured Output] Disallow MRV1 + PP>1 + async sched + structured output (#56250)
Signed-off-by: Artem Perevedentsev <aperevedents@nvidia.com>
Co-authored-by: CNE Pierre FICHEPOIL <pierre-1.fichepoil@gendarmerie.interieur.gouv.fr>
2026-09-22 18:00:05 +04:00
Jeff (Junze) Ma 0961bbae28 [Spec Decode] Enable async scheduling for DFlash (#58065) 2026-09-21 22:05:44 -07:00
Nicolò Lucchesi 54020c3c3e [Engram] Drop redundant VLLM_PLE_CPU_OFFLOAD env var (#57937)
Signed-off-by: NickLucche <nicolo.lucchesi@mistral.ai>
2026-09-21 20:16:02 +02:00
Tony Lin dc479a2d83 [Bugfix] Fix external LB DP rank handling when replicas share nodes (#53743)
Signed-off-by: Tony Lin <tony.lin@intel.com>
2026-09-21 17:19:14 +02:00
76ffb0d374 [Model][Engram] Share host tables across co-located DP replicas by default (#57651)
Signed-off-by: Juntian Liu <Juntianl777@gmail.com>
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: zjy0516 <riverclouds.zhu@qq.com>
2026-09-21 08:26:56 +00:00
JohnQinAMD 6dfd87f596 [ROCm][DSv4.1] Keep the Engram tables in host memory on ROCm (#57491) 2026-09-21 10:17:11 +08:00
5302d1fe40 [Model Runner V2] Support custom logits processors (#56497)
Signed-off-by: David Cheung <d7cheung@gmail.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
2026-09-19 17:46:49 +00:00
Laura Lopez-RealLaura López Realmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
a85dd55177 Fix Laguna patch mutating flat RoPE parameters (#57189)
Signed-off-by: Laura López Real <laulopezreal@users.noreply.github.com>
Co-authored-by: Laura López Real <laulopezreal@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-09-17 12:22:57 +00:00