66 Commits
Author SHA1 Message Date
4d58fa254a [MoE] Unify DeepGEMM MegaMoE experts behind one feature-matrix layer; add NVFP4 (#61538)
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
2026-10-15 19:02:01 +00:00
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
984c242ede [Quantization] Support compressed-tensors MXFP4 MoE with FP8 activations via Mxfp4MoEMethod (#59815)
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
2026-10-14 13:33:35 -07:00
Ronaldandaoshen02 19aba3e4d5 [Bugfix][RL] Load attention sinks through weight_loader for layerwise reload (#57797)
Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
2026-10-14 10:17:58 +00:00
c5bda658d2 [Model][MoE] Load MegaMoE experts directly into the DeepGEMM layout (#60096)
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Signed-off-by: Aaron Hao <hao-aaron@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Aaron Hao <hao-aaron@users.noreply.github.com>
2026-10-13 11:52:08 -07:00
Ilya Markov a281ad32f0 [EPLB] Add EPLB support to dspark draft models (DeepSeekV4) (#57549)
Signed-off-by: Markov Ilya <markovilya197@gmail.com>
2026-10-12 10:03:09 -04:00
81198e97ba [Refactor] Abstract DeepGemm MegaMOE backend (#58053)
Signed-off-by: Hui Shi <kalinshi@tencent.com>
Co-authored-by: LuneExplorer <liyihan23@mails.tsinghua.edu.cn>
Co-authored-by: ArmageddonKnight <bojian@cs.toronto.edu>
2026-10-08 05:05:39 +00:00
aoshen02andClaude Opus 5.5 4a30c4cad0 [Model][DeepSeek-V4] Make MegaMoE shared-expert finalize independent of linear post-load order (#59927)
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 08:39:49 +00:00
Julien Debache 111f71a6db [feat] FlashInfer CuteDSL MegaMoE integration (#54049)
Signed-off-by: jdebache <jdebache@nvidia.com>
2026-10-02 15:43:33 -04:00
Woosuk KwonandCodex 9a70c233cd [Perf][DSV4.1] Fuse TP all-reduce with mHC input preparation (#57643)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: Codex <noreply@openai.com>
2026-09-20 20:28:06 -07:00
Itay AlroyandTyler Michael Smith 017dced6a6 [DeepSeek V4] Fix fused MoE expert distribution (#57465)
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2026-09-18 15:17:07 +00:00
41f9104fa6 [DSv4.1] Integrate Mega-Gate from DeepGEMM (#56266)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: zjy0516 <riverclouds.zhu@qq.com>
2026-09-17 12:06:36 -07:00
Woosuk Kwon 30b847c1fa [Perf][DSV4.1] Remove MegaMoE padding and shared padding workaround (#57204)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-09-16 19:09:30 +00:00
Harry MellorandClaude Opus 5 651a88c09a [CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 10:57:24 -07:00
Harry MellorandClaude Opus 5 8c1557a79c Add pydocstyle to the ruff rules (#52136)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Harry Mellor <hej.mellor@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 16:44:29 +00:00
Canlin GuoandCodex 03f67b3ad1 [Perf][DSV4.1] Pad shared experts for native MegaMoE fusion (#56568)
Signed-off-by: Canlin <canlinguosdu@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
2026-09-16 17:08:42 +08:00
Roberto L. CastroandCodex e52be1a62d [9/N][warmup][DSv4] Migrate MHC TileLang kernels (#50178)
Signed-off-by: LopezCastroRoberto <rocastro@redhat.com>
Signed-off-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
Co-authored-by: Codex <codex@openai.com>
2026-09-12 17:25:23 -07:00
Andreas KaratzasandOpenAI Codex 9dd969da09 [Model][ROCm] Enable DeepSeek V4 Vision (#55107)
Signed-off-by: Andreas Karatzas <Andreas.Karatzas@amd.com>
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
2026-09-11 09:17:30 -07:00
9b959b8657 [Model] DeepSeek-V4.1-Flash Model Definitions (#56228)
Signed-off-by: Roger Wang <hey@rogerw.io>
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Kimi Code <noreply@moonshot.cn>
Co-authored-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Codex <noreply@openai.com>
2026-09-10 20:16:26 +08:00
d87a440f88 [Core][MRV2] Support eagle3 spec decode with pipeline parallel (#50514)
Signed-off-by: Yongqin Wang <yongqinwang@roblox.com>
Co-authored-by: Cursor <noreply@cursor.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
2026-09-05 11:45:34 +08:00
Isotr0py 1356635d83 [New model][Multimodal] Add DeepSeek-V4-Flash-Vision-Exp support (#54566)
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
2026-09-02 11:09:09 -07:00
8fe9317f2e [Model][MoE] DeepSeek-V4: add opt-in FlashInfer moe_ep expert backend (#49636)
Signed-off-by: Md Saidul Hoque Anik <mhoqueanik@nvidia.com>
Signed-off-by: Anerudhan Gopal <agopal@nvidia.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Anerudhan Gopal <agopal@nvidia.com>
2026-08-25 03:35:01 +00:00
Giancarlo Delfin d154d90d6c [Model Runner V2] batch-sharded sample (#50465)
Signed-off-by: Giancarlo Delfin <gdelfin@inferact.ai>
2026-08-24 15:49:57 -07:00
ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟andJee Jee Li 702e1d7186 [LoRA] feat: Support LoRA for DeepSeek V4 (#53361)
Signed-off-by: Hollow Man <hollowman@opensuse.org>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
2026-08-24 03:57:57 +00:00
Canlin Guo 4f6885fffc [DSV4][Kernel] Fuse shared experts into MegaMoE (#53040)
Signed-off-by: Canlin Guo <canlinguosdu@gmail.com>
2026-08-20 12:07:02 -07:00
Harry Mellor 1fe3a1571a Reduce AutoWeightsLoader kwargs (#53106)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2026-08-20 15:40:11 +00:00
Woosuk KwonandOpenAI Codex f1178f3a06 Revert DSv4 eager workspace reuse (#52836)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: OpenAI Codex <codex@openai.com>
2026-08-18 17:02:59 -07:00
ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟 a9f4afb66f [Bugfix] Fix DeepSeek V4 mHC broadcast buffer for dummy load (#51368)
Signed-off-by: Hollow Man <hollowman@opensuse.org>
2026-08-18 17:02:18 -07:00
ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟 d5f5de7a7d [Bugfix] Fix DeepSeek V4 mHC broadcast buffer for weight sync (#52626)
Signed-off-by: Hollow Man <hollowman@opensuse.org>
2026-08-18 10:55:37 +08:00
aoshen02 4eef91c03c [Model] Support R3 capture with DeepGEMM MegaMoE (#51831)
Signed-off-by: aoshen02 <aoshen@inferact.ai>
2026-08-12 01:07:07 -07:00
Wentao Ye fd04edef3e [Refactor] Delete dead code in models (#51838)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2026-08-11 15:44:26 -04:00
Woosuk KwonandOpenAI Codex 79c865b838 [Perf] Narrow DeepSeek V4 eager CUDA graph region (#51430)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: OpenAI Codex <codex@openai.com>
2026-08-10 10:52:10 -07:00
Woosuk Kwon 38a466e7b6 [DSV4] Implement Sequence Parallelism (#46789)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-08-01 10:22:33 -07:00
Wentao Yeandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> df71917cf1 [DSv4 Perf] Optimize workspace reuse for eager break, 3.9% E2E TTFT improvement. (#49236)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-07-31 08:45:58 -07:00
bnellnmandHarry Mellor 6e311c6e20 [MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)
Signed-off-by: Bill Nell <bnell@redhat.com>
Signed-off-by: bnellnm <49004751+bnellnm@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2026-07-31 11:28:18 +01:00
Wentao Ye 904fae8be1 [DSv4 Perf] Fix redundant memory allocation and copy for dsv4 pp buffer, 448 MiB GPU memory saved (#50312)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2026-07-30 11:38:51 -04:00
Mike G 76bf55240c [Bugfix] Fix DeepSeek-V4 DSpark draft shared-expert padding for TP > 8 (#49415)
Signed-off-by: Mike G <180722391+mikekg@users.noreply.github.com>
2026-07-23 05:06:21 +00:00
Wentao Yeandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 442c421e79 [Perf] Remove redundant repeat and copy for dsv4, 1.8% E2E TPOT improvement. (#48137)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-07-15 00:48:10 +00:00
f5a8d73377 [Spec Decode] DSpark (#46995)
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
Signed-off-by: Giancarlo Delfin <gdelfin@inferact.ai>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Giancarlo Delfin <gdelfin@inferact.ai>
Co-authored-by: mgoin <mgoin64@gmail.com>
2026-07-01 08:30:24 -07:00
Harry Mellor 1ab9522935 Remove more unnecessary load_weights methods (#47058)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2026-06-30 15:22:16 +01:00
Ilya MarkovandMarkov Ilya a2abce646f [EPLB] Mask padding in EPLB load recording (#38128)
Signed-off-by: ilmarkov <markovilya197@gmail.com>
Signed-off-by: Markov Ilya <markovilya19@gmail.com>
Co-authored-by: Markov Ilya <markovilya19@gmail.com>
2026-06-28 19:43:58 -07:00
Woosuk Kwonandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 0a3e2dbc09 [Optimization] Skip DP padding tokens in MoE (#46428)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-23 14:54:46 -07:00
44d95069e9 Enable DeepSeek V4 and GLM-5.1 on SM120 (#43477)
Signed-off-by: Zihua Wu <13583761+lucifer1004@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>
2026-06-22 11:54:14 -07:00
Jeff (Junze) Maandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2a6c6b9429 [DeepSeek-V4] Support TEP=16 for the block-FP8 shared expert (#46001)
Signed-off-by: Jeff Ma <jeffjma@umich.edu>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-18 20:10:12 -07:00
Tyler Michael SmithandClaude Opus 4.6 e2f993dc41 [WideEP] Integrate DeepEP v2 (#41183)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-06-08 18:07:29 -07:00
bnellnm dc68bd8c41 [MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)
Signed-off-by: Bill Nell <bnell@redhat.com>
2026-06-08 10:42:58 -04:00
Woosuk Kwon 4efd6ffde0 [DSV4] Refactor DeepseekV4Attention (#44569)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-06-04 20:23:07 -07:00
Wei Zhao 969aec4bc8 [Bugfix] Fix Deepseek v4 non-mega-moe model init error (#44356)
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
2026-06-02 18:26:30 -07:00
Woosuk Kwon b254e0456c [DSV4] Minor cleanup for DeepseekV4MegaMoEExperts (#44367)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-06-02 17:54:27 -07:00
Wei ZhaoandWei Zhao 2427094152 [Feature] Support EPLB for DeepSeek v4 Mega Moe (#43339)
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
Co-authored-by: Wei Zhao (Engrg-Hardware 1) <weizha@login-lyris01.lyris.clusters.nvidia.com>
2026-06-02 10:56:44 -07:00