Commit Graph
558 Commits
Author SHA1 Message Date
Yifan QiaoandClaude Opus 5.5 240785b82c [Bugfix] Keep the GLM-5.3 kpool tail and Qwen4 QSA ring out of the null block (#59528)
Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 20:44:04 +00:00
Shijin Zhang 2274e86842 [Spec Decode][Model] Support DFlash2 draft models with GLM-5.3-Flash (#56983)
Signed-off-by: Shijin Zhang <75300765+Dovis01@users.noreply.github.com>
2026-10-08 14:41:10 -04:00
kyleliang-nvKyle LiangOpenAI Codexmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
1664c4f444 [Bugfix][MiniMax-M3] Guard sparse decode sentinel indices (#54597)
Signed-off-by: Kyle Liang <kylliang@nvidia.com>
Signed-off-by: Kyle Liang <kyleliang-nv@users.noreply.github.com>
Co-authored-by: Kyle Liang <kyleliang-nv@users.noreply.github.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-08 18:27:57 +00:00
Elvir CrnčevićandMisha Goin 8674a794c7 [Perf] Enable Kimi K3 shared expert sharding with DeepEP v2 (#56875)
Co-authored-by: Misha Goin <mgoin64@gmail.com>
2026-10-08 16:42:15 +00:00
9f3d70e081 [Quantization] Enable shared expert fusion compatibility with online shared_expert quantization (showcase: along Quark MXFP4 routed experts) (#55686)
Signed-off-by: Felix Marty <Felix.Marty@amd.com>
Signed-off-by: fxmarty-amd <felmarty@amd.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: Misha Goin <mgoin64@gmail.com>
2026-10-08 12:04:46 -04:00
Mikko TukiainenandCursor e2414ec1db [ROCm][Qwen4Exp] Use fused PLE Triton kernels on the AMD backend (#60021)
Signed-off-by: Mikko Tukiainen <Mikko.Tukiainen@amd.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-08 15:47:56 +00:00
kliuae d80ddaac73 [ROCm][Perf] Kimi-K3 Store only the current rank's shards in latent MoE up-proj (#59591)
Signed-off-by: kliuae <kuanfu.liu@embeddedllm.com>
2026-10-08 14:55:20 +00:00
458ba2edf8 [ROCm][Kimi-K3] Fuse AttnRes output with per-token FP8 input quantization (#59069)
Signed-off-by: whx-sjtu <xiaowang990929@gmail.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com>
2026-10-08 12:01:47 +00:00
Harry MellorandClaude Opus 5.5 21d9758459 Remove Python 3.10 support now that it is EOL (#60402)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 02:30:52 -07:00
Nicolò Lucchesi 70bdce6133 [Model][GLM-5.3-Flash] Warm up the DCP top-k merge kernels for the kpool indexer (#60032)
Signed-off-by: NickLucche <nicolo.lucchesi@mistral.ai>
2026-10-08 11:05:52 +02:00
JohnQinAMDClaude Opus 5Ren Yuzhoumergify[bot] <37929162+mergify[bot]@users.noreply.github.com>Douglas Lehr
fb2ac82458 [ROCm][DSv4.1][Perf] Replace the Torch prefill combine with the shared combine_topk_swa_indices kernel (#56638)
Signed-off-by: JohnQinAMD <yanyuan.qin@amd.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Ren Yuzhou <54501155+yuzhouo7@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Douglas Lehr <91553416+dllehr-amd@users.noreply.github.com>
2026-10-08 14:15:18 +08:00
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
Hyukjoon LeeClaude Opus 5.5mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>Summer Yang
1fbbfa1657 [Bugfix][Kimi-K3] Declare max_tp_shards on the DSpark MLA KV cache spec (#59733)
Signed-off-by: hyukjlee <hyukjlee@amd.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Summer Yang <girasoleyang@gmail.com>
2026-10-08 00:33:32 -04:00
Isotr0pyandShanshan Shen 8352b2427f [Multimodal] Optimize encoder cudagraph padding logic (#58925)
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
Co-authored-by: Shanshan Shen <467638484@qq.com>
2026-10-08 03:38:09 +00:00
8faac3dd2a [Bugfix] [Spec Decode] DeepSeek V4/V4.1 NVFP4: keep ignored MTP/DSpark experts on MXFP4 (#59655)
Signed-off-by: Shobhit Verma <shobhitv@nvidia.com>
Signed-off-by: v-shobhit <161510941+v-shobhit@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Benjamin Chislett <chislett.ben@gmail.com>
2026-10-08 00:10:32 +00:00
f38c6778ce [quantization] Add per-token NVFP4 quantization MoE support for ReLU2 (#56740)
Signed-off-by: Xuanteng Huang <xuantengh@nvidia.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
2026-10-07 22:01:09 +00:00
e572e02d21 [UX] Skip checkpoint shards an MTP head does not load (#60152)
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-07 15:16:19 -04:00
d64a18832b [ROCm][Triton] Migrating MiniMax-M3 kernels from make_block_ptr to tensor descriptors (#59285)
Signed-off-by: jpvillam <juan.villamizar@amd.com>
Signed-off-by: Daniel <danichan@amd.com>
Co-authored-by: jpvillam <juan.villamizar@amd.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: Rohan Potdar <rohan.potdar@amd.com>
Co-authored-by: Daniel <danichan@amd.com>
2026-10-07 11:49:59 -05:00
Micah WilliamsonandFangzhou Ai f6e3944e31 [ROCm][DSv4] AITER MegaMoEV2 Integration For DeepSeek V4 (#59685)
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
Co-authored-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com>
2026-10-07 09:35:53 -05:00
2a54f6b625 [ROCm][Perf][GLM-5.3-Flash] Enable AITER fused shared experts on gfx950 (#59221)
Signed-off-by: Jin Tao <jin.tao@amd.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Simon Danielsson <70206058+simondanielsson@users.noreply.github.com>
2026-10-07 20:53:40 +08:00
d9503dc503 [Bugfix][Qwen4Exp] Derive the QSA head_dim from the unsharded head count (#59945)
Signed-off-by: lesj0610 <lesj0610@godoiksan.org>
Co-authored-by: lesj0610 <lesj0610@godoiksan.org>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-07 12:24:40 +00:00
fe51194e2d [Bugfix][MiniMax-M3] Fence decode indexer shared-memory stage reuse (#60194)
Signed-off-by: Vladimir Anisimov <vanisimov@nvidia.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-07 01:35:12 +00:00
PatrykSafferandMisha Goin e43db1f5e2 Pad to 64 not 128 Q heads in flashMLA sparse (#60029)
Signed-off-by: Patryk Saffer <patryk.saffer@mistral.ai>
Co-authored-by: Misha Goin <mgoin64@gmail.com>
2026-10-06 16:03:05 -04:00
aoshen02andClaude Opus 5.5 9a1f6fd4cb [Model][DeepSeek-V4.1] Permute mega-attention wq_b / wo_a while loading (#60064)
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 05:20:09 +00:00
Juntian LiuClaude Opus 5.5mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
eb2901b735 [Perf][DSv4.1] Run decoder replay layers in CUDA graphs and trim in PIECEWISE graphs (#59532)
Signed-off-by: Juntian Liu <Juntianl777@gmail.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-05 22:11:45 -07:00
b0b23a48d4 [Bugfix] Fix Qwen4Exp FP8 PLE pinned lookup kernel failing to compile below SM89 (#60142)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-06 01:56:44 +00:00
JohnQinAMDClaude Opus 5mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
97bd8c74eb [ROCm][DSv4.1][Perf] One-pass dequantize and a gather-sized grid for the compressed K cache (#56720)
Signed-off-by: JohnQinAMD <yanyuan.qin@amd.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-05 17:57:46 -07:00
b2863c0de9 [Kimi-K3] Use the fused K/V pack kernel on the decode-context-parallel prefill path (#56881)
Signed-off-by: Brandon Zhang <31413216+brandonfzhang@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Wei Zhao <51183510+wzhao18@users.noreply.github.com>
2026-10-06 00:13:55 +00:00
Vadim GimpelsonCursor Agentmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
0eac152707 [Bugfix] Route every speculation-capable row through the speculative path so recurrent state stays consistent (#56531)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-06 03:03:44 +04:00
Mikko TukiainenandCursor d23af12cb1 [ROCm][Perf] Reach the fused QSA pre-indexer from the AMD path (#57947)
Signed-off-by: mjkvaak-amd <mjkvaak-amd@users.noreply.github.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-05 16:22:29 +00:00
Nicolò Lucchesi 30d4032363 [Model][GLM-5.3-Flash] Support DCP for the kpool sparse indexer (#59211)
Signed-off-by: NickLucche <nicolo.lucchesi@mistral.ai>
2026-10-05 16:44:14 +02:00
cjackal 0e468adb43 [Model] Extend device-side mm normalization to Kimi K2.5 / K3 (#59278)
Signed-off-by: cjackal <44624812+cjackal@users.noreply.github.com>
2026-10-05 21:28:13 +08:00
Stefano CastagnettaandThien Tran 4ff028d77e [Bugfix][Qwen4Exp] Honor --kv-cache-dtype-skip-layers in QSA attention (#60023)
Signed-off-by: Stefano Castagnetta <scastagnetta@nvidia.com>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-05 12:13:30 +00:00
Stefano CastagnettaandThien Tran 55b80221ed [Perf][Qwen4Exp] Keep the HC up projection on the skinny GEMM path (#60027)
Signed-off-by: Stefano Castagnetta <scastagnetta@nvidia.com>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-05 20:09:36 +08:00
Mikko TukiainenCursormergify[bot] <37929162+mergify[bot]@users.noreply.github.com>TJian
edca360f13 [Bugfix][Qwen4Exp] Load PLE tables unquantized under Quark checkpoints (#59443)
Signed-off-by: Mikko Tukiainen <Mikko.Tukiainen@amd.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
2026-10-05 10:40:02 +00:00
c04c79e50d [Performance] Add SM121 TP=2 skinny-GEMM plans (#59632)
Signed-off-by: sudhanshu shukla <sudhanshu112233shukla@users.noreply.github.com>
Co-authored-by: sudhanshu shukla <sudhanshu112233shukla@users.noreply.github.com>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-10-05 18:27:51 +08:00
bcsdhjewmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>Codex
ae53b06898 [Bugfix][GLM-5.3-Flash] Keep batch x heads out of gridDim.z in the GLM-5.3-Flash fused recurrent KDA kernel (#56974)
Signed-off-by: NolenLiang <nliang@nvidia.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Codex <noreply@openai.com>
2026-10-05 09:31:45 +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
042ab0305c [Perf][Qwen4Exp] Merge QSA QKVG and indexer QK projections (#59533)
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: namgyu-youn <namgyu.dev@gmail.com>
2026-10-05 07:03:29 +00:00
Vadim GimpelsonandClaude 4f52fa35ef [Bugfix] Fix Qwen4Exp PLE embedding rejecting INC (AutoRound) checkpoints (#59990)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-10-05 01:52:06 +00:00
49e0f47978 [Bugfix][DSv4.1] Keep the compressor ring out of the null block (#58560)
Signed-off-by: bluemelov <hannes.b2411@eva.space>
Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: bluemelov <hannes.b2411@eva.space>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 01:49:41 +00:00
Yongqi WangandKunshang Ji 155d23cb00 [Bugfix][XPU] Make DeepSeek V4 FP8 sparse decode graph-capturable (#59159)
Signed-off-by: Yongqi Wang <yongqi.wang@intel.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
2026-10-04 18:31:15 +08:00
Harry Mellor 9e0b6fcf77 [CI] Bump Transformers version to 5.18.0 (#59621)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2026-10-04 00:42:35 +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
Stefano CastagnettaandBenjamin Chislett f590eb2448 [Perf][Qwen4Exp] Add SM121 TP=1 skinny-GEMM plans (#59753)
Signed-off-by: Stefano Castagnetta <scastagnetta@nvidia.com>
Co-authored-by: Benjamin Chislett <bchislett@nvidia.com>
2026-10-02 17:17:45 +00:00
rmhaskarandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> e42d35d4f3 [Minimax-M3] Keep the native FP8 MMA in the Triton indexer scorers (#59481)
Signed-off-by: Rajas Mhaskar <rmhaskar@nvidia.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-02 12:30:39 -04:00
0979892992 [Bugfix][Multimodal] Fix GLM-5.3-Flash vision tower crashes on image input (#59126)
Signed-off-by: Juan Calderon-Perez <835733+gaby@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Isotr0py <Isotr0py@outlook.com>
2026-10-02 15:25:37 +00:00
Simon Danielssonandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 98b29aa99a [ROCm][Perf][GLM-5.3-Flash] Add AITER topk backend for decodes (#58167)
Signed-off-by: simondanielsson <simon.danielsson99@hotmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-02 11:47:41 +01:00
5688a4dd4a [Bugfix][GLM-5.3] Size the image encoder cache from the exact token ceiling (#59565)
Signed-off-by: Yufeng He <40085740+he-yufeng@users.noreply.github.com>
Co-authored-by: Kimi Code <kimi-code@moonshot.cn>
Co-authored-by: Isotr0py <Isotr0py@outlook.com>
2026-10-02 10:43:29 +00:00
kliuae 0154979656 [ROCm][Perf] Kimi-K3 enable prefill checkpoints on ROCm (#58344)
Signed-off-by: kliuae <kuanfu.liu@embeddedllm.com>
2026-10-02 06:58:04 +00:00