Cheng Wan e703f9e566 [refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

..

deepseek_common

Fix gfx95 bpreshuffle FP8 activation scale layout (#29275)

2026-07-08 10:55:09 -07:00

triton_ops

[codex] Centralize more inline Triton kernels (#27429)

2026-06-11 22:17:26 +08:00

afmoe.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

apertus.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

arcee.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

baichuan.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

bailing_moe_linear.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

bailing_moe_nextn.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

bailing_moe.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

bert.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

chatglm.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

clip.py

VLM: add Conv2dLayer/Conv3dLayer to fix PyTorch 2.9.1 CuDNN Conv3d (#20282)

2026-03-15 19:17:44 +08:00

cohere2_moe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

cohere2_vision.py

model: support Command A plus (#26106)

2026-06-03 11:23:04 +08:00

commandr.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

dbrx.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

deepseek_janus_pro.py

…

deepseek_nextn.py

[AMD] Fix GLM-5.2 MTP Quark excludes (#30265)

2026-07-08 14:57:10 -07:00

deepseek_ocr.py

[CI] Fix lint brought by #27527 (#28988)

2026-06-22 20:40:06 -07:00

deepseek_v2.py

Fix gfx95 bpreshuffle FP8 activation scale layout (#29275)

2026-07-08 10:55:09 -07:00

deepseek_v4_nextn.py

[refactor] Split the DP gathered-buffer state between flags.dp and ctx.forward (#30491)

2026-07-09 02:09:21 -07:00

deepseek_v4.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

deepseek_vl2.py

[VLM] Chunk-aware ViT encoding with per-image cache and lazy device transfer (#22038)

2026-04-04 16:55:17 +08:00

deepseek.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

dflash.py

Add Laguna XS.2.1 DFlash support to SGLang (#29446)

2026-07-01 20:26:23 -07:00

dots_ocr.py

…

dots_vlm_vit.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

dots_vlm.py

…

ernie4_eagle.py

…

ernie4.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

ernie45_moe_vl.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

ernie45_vl.py

…

exaone4.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

exaone_moe_mtp.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

exaone_moe.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

exaone.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

falcon_h1.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

gemma2_reward.py

[Score API] Add return_pooled_hidden_states to Scoring API for SequenceClassification / RewardModel (#22427)

2026-04-15 14:58:56 -07:00

gemma2.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gemma3_causal.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

gemma3_mm.py

feat(model_runner): remove pool/backend refs from ForwardBatch via ForwardContext (#25983)

2026-05-21 14:01:49 -07:00

gemma3n_audio.py

…

gemma3n_causal.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

gemma3n_mm.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

gemma4_audio.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gemma4_causal.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gemma4_mm.py

add dflash gemma4 support (#27471)

2026-06-17 16:39:28 -07:00

gemma4_mtp.py

[model] support encoder-free unified Text/Vision/Audio model (#27167)

2026-06-03 23:58:06 +08:00

gemma4_unified.py

[model] support encoder-free unified Text/Vision/Audio model (#27167)

2026-06-03 23:58:06 +08:00

gemma4_vision.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gemma.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

glm4_moe_lite_nextn.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

glm4_moe_lite.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

glm4_moe_nextn.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

glm4_moe.py

[AMD] Fix GLM-5.2 MTP Quark excludes (#30265)

2026-07-08 14:57:10 -07:00

glm4.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

glm4v_moe.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

glm4v.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

glm_ocr_nextn.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

glm_ocr.py

[Bug Fix] GLM-V / GLM-OCR: field detection for transformers 5.x and MTP omission fix (#21134)

2026-03-23 13:19:48 -07:00

glmasr.py

…

gpt2.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gpt_bigcode.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gpt_j.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

gpt_oss.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

granite.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

granitemoe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

granitemoehybrid.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

grok.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

hrm_text.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

hunyuan_v3_nextn.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

hunyuan_v3.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

hunyuan.py

chore: cleanup garbage code (#29770)

2026-07-02 16:14:01 +08:00

idefics2.py

VLM: add Conv2dLayer/Conv3dLayer to fix PyTorch 2.9.1 CuDNN Conv3d (#20282)

2026-03-15 19:17:44 +08:00

internlm2_reward.py

[Score API] Add return_pooled_hidden_states to Scoring API for SequenceClassification / RewardModel (#22427)

2026-04-15 14:58:56 -07:00

internlm2.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

interns1.py

[Bugfix] Fix Intern-S1 FP8 expert count lookup (#28629)

2026-06-18 17:49:59 +08:00

interns1pro.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

interns2preview.py

Support Intern-S2-Preview (#24875)

2026-05-10 22:17:30 +08:00

internvl.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

iquest_loopcoder.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

jet_nemotron.py

feat(model_runner): remove pool/backend refs from ForwardBatch via ForwardContext (#25983)

2026-05-21 14:01:49 -07:00

jet_vlm.py

…

kimi_k25_eagle3.py

(perf): Shard Kimi-K2.5 Eagle3 draft fc + symm-mem AG (#29223)

2026-06-27 17:40:32 -07:00

kimi_k25.py

Remove transformers 5.12.1 dead-code workarounds (#29758)

2026-07-03 00:03:06 +08:00

kimi_linear.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

kimi_vl_moonvit.py

[XPU] Fix Device Assignment (#26257)

2026-05-29 09:38:11 +08:00

kimi_vl.py

[EPD][VLM] Support Kimi VL EPD (#22490)

2026-04-16 12:40:02 +08:00

laguna.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

lfm2_moe.py

[LFM2-MoE] Support Transformers v5 packed MoE expert weights (#29659)

2026-07-02 17:13:02 +08:00

lfm2_vl.py

[VLM] fix LFM2-VL offline inference and GPU JPEG decode (#22448)

2026-04-15 09:13:25 +08:00

lfm2.py

feat(short-conv): shared ShortConvAttnBackend for ZAYA1 CCA + LFM2 short conv (#29867)

2026-07-01 20:08:00 -07:00

lightonocr.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

llada2.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

llama4.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

llama_classification.py

[Score API] Add SequenceClassification Model support (#22118)

2026-04-08 01:30:58 -07:00

llama_eagle3.py

[Spec] Remove deprecated EAGLE v1 DRAFT_EXTEND forward mode (#28129)

2026-06-13 13:00:21 -07:00

llama_eagle.py

[Refactor] Pass PP start_layer via model constructor instead of forward_batch.token_to_kv_pool (#25825)

2026-05-19 22:16:07 -07:00

llama_embedding.py

…

llama_reward.py

[Score API] Add return_pooled_hidden_states to Scoring API for SequenceClassification / RewardModel (#22427)

2026-04-15 14:58:56 -07:00

llama.py

Support speculative decoding on CPU (#27862)

2026-07-09 10:27:09 +08:00

llava.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

llavavid.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

locate_anything.py

model: support nvidia/LocateAnything-3B (#28958)

2026-06-30 00:16:42 +08:00

longcat_flash_nextn.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

longcat_flash.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

midashenglm.py

Upgrade transformers==5.3.0 (#17784)

2026-03-18 13:50:43 -07:00

mimo_audio.py

Support MiMo v2 ASR (#26278)

2026-06-11 13:43:13 -07:00

mimo_mtp.py

Support speculative decoding on CPU (#27862)

2026-07-09 10:27:09 +08:00

mimo_v2_asr.py

Support MiMo v2 ASR (#26278)

2026-06-11 13:43:13 -07:00

mimo_v2_nextn.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

mimo_v2.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

mimo_vl.py

fix(mimo-vl): pass padded_context_dim to Qwen2_5_VisionPatchMerger (#29994)

2026-07-04 08:08:25 +08:00

mimo.py

…

mindspore.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

minicpm3.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

minicpm.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

minicpmo.py

[Deps] Bump transformers to 5.12.1 (#29393)

2026-06-30 00:54:01 -07:00

minicpmv_vit.py

[Model] Add MiniCPM-V 4.6 support (#24855)

2026-05-10 12:24:17 -04:00

minicpmv.py

[XPU] Fix Device Assignment (#26257)

2026-05-29 09:38:11 +08:00

minimax_m2.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

ministral3.py

[Bugfix] Fix Ministral3 init argument forwarding (#29111)

2026-06-25 18:26:45 +08:00

mistral_eagle.py

debug followup (#24058)

2026-04-29 23:03:27 +08:00

mistral_large_3_eagle.py

[CP] 1/N: Support MLA Prefill Context Parallel (#23292)

2026-05-23 03:07:09 -07:00

mistral_large_3.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

mistral.py

Bugfix (#24027)

2026-04-29 21:13:51 +08:00

mixtral_quant.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

mixtral.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

mllama4.py

[CPU] Padding for dim divisibility in TP3/6 cases (#20072)

2026-07-02 13:14:39 +08:00

mllama.py

[CPU] Padding for dim divisibility in TP3/6 cases (#20072)

2026-07-02 13:14:39 +08:00

moss_vl.py

[Deps] Bump transformers to 5.12.1 (#29393)

2026-06-30 00:54:01 -07:00

nano_nemotron_vl.py

[NemotronH] V3 Omni wrapper: WeightsMapper + config round-trip (#25023)

2026-05-26 20:34:05 +00:00

nemotron_h_mtp.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

nemotron_h_utils.py

Use Flashinfer allreduce fusion for MNNVL allreduce for Nemotron (#28346)

2026-06-18 17:30:12 -07:00

nemotron_h.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

nemotron_nas.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

nvila_lite.py

[Cherry pick to release/v0.5.15] Fix NVILA weight loading (#30400)

2026-07-07 13:21:55 -07:00

nvila.py

[Cherry pick to release/v0.5.15] Fix NVILA weight loading (#30400)

2026-07-07 13:21:55 -07:00

olmo2.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

olmo.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

olmoe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

opt.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

orion.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

paddleocr_vl.py

VLM: add Conv2dLayer/Conv3dLayer to fix PyTorch 2.9.1 CuDNN Conv3d (#20282)

2026-03-15 19:17:44 +08:00

parakeet.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

persimmon.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

phi3_small.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

phi4mm_audio.py

…

phi4mm_utils.py

…

phi4mm.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

phi.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

phimoe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

pixtral.py

Bugfix (#24027)

2026-04-29 21:13:51 +08:00

points_v15_chat.py

…

qwen2_5_vl.py

[CPU] Padding for dim divisibility in TP3/6 cases (#20072)

2026-07-02 13:14:39 +08:00

qwen2_audio.py

…

qwen2_classification.py

[Score API] Add SequenceClassification Model support (#22118)

2026-04-08 01:30:58 -07:00

qwen2_eagle.py

Support speculative decoding on CPU (#27862)

2026-07-09 10:27:09 +08:00

qwen2_moe.py

[refactor] Adopt get_parallel() everywhere and close out the parallel wrapper surface (#30492)

2026-07-09 02:09:39 -07:00

qwen2_rm.py

[Score API] Add return_pooled_hidden_states to Scoring API for SequenceClassification / RewardModel (#22427)

2026-04-15 14:58:56 -07:00

qwen2_vl.py

VLM: add Conv2dLayer/Conv3dLayer to fix PyTorch 2.9.1 CuDNN Conv3d (#20282)

2026-03-15 19:17:44 +08:00

qwen2.py

Support speculative decoding on CPU (#27862)

2026-07-09 10:27:09 +08:00

qwen3_5_mtp.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

qwen3_5.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

qwen3_asr.py

model: support qwen3-asr (#22073)

2026-04-07 13:27:05 +08:00

qwen3_classification.py

[Model] Qwen3ForPooledOutput: forward get_input_embeddings to inner model (#23434)

2026-04-29 12:25:06 -07:00

qwen3_moe_mtp.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

qwen3_moe.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

qwen3_next_mtp.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

qwen3_next.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

qwen3_omni_moe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

qwen3_rm.py

…

qwen3_vl_moe.py

[BugFix][EPD]Fix Qwen3VLMoe encoder-only AttributeError (#25759)

2026-05-21 11:07:21 +08:00

qwen3_vl.py

[refactor] Read resolved config from server_args fields; retire the flags mirror tier (#30346)

2026-07-07 21:28:34 -07:00

qwen3.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

qwen.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

radio.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

registry.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

roberta.py

…

sarashina2_vision.py

…

sarvam_moe.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

sdar_moe.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

sdar.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

siglip2.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

siglip.py

[CPU] Add frontend support for Gemma (#12590)

2026-03-18 23:02:26 -07:00

solar.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

stablelm.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

starcoder2.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

step3_vl_10b.py

[FIX]Fix Step3-VL multi-image embedding and local patch splitting (#24970)

2026-06-17 10:31:32 -07:00

step3_vl.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

step3p5_mtp.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

step3p5.py

[refactor] ctx.resources: named slots, stream leases, and workspace buffer leases (#30348)

2026-07-07 21:30:10 -07:00

step3p7.py

model: support Step-3.7-Flash (#26565)

2026-05-29 08:00:54 +08:00

teleflm.py

…

torch_native_llama.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

transformers.py

[Fix] model init / XPU / transformers-v5 / bench-image fixes (#28292)

2026-06-23 12:58:13 +08:00

unlimited_ocr.py

model: support baidu unlimited-ocr (#29186)

2026-06-27 23:36:19 +08:00

utils.py

[AMD] Support triton backend decode context parallel for Qwen3.5 (#25090)

2026-06-23 17:35:56 -07:00

voxtral.py

chore: add vLLM SPDX copyright headers to ported files (#25182)

2026-05-13 15:17:30 -07:00

whisper.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

xverse_moe.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

xverse.py

Add get_parallel(): a structured accessor for parallel-topology state (#28567)

2026-06-17 20:23:43 -07:00

yivl.py

Upgrade transformers from 5.5.4 to 5.6.0 (#23525)

2026-04-26 22:33:54 -07:00

zaya.py

feat(short-conv): shared ShortConvAttnBackend for ZAYA1 CCA + LFM2 short conv (#29867)

2026-07-01 20:08:00 -07:00