Lucas Wilkinson
|
abe93bce59
|
[Attention] Make seq_lens_cpu optional in CommonAttentionMetadata to enable true async spec-decode (#29624)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Benjamin Chislett <chislett.ben@gmail.com>
|
2025-12-09 17:18:10 -08:00 |
|
Matthew Bonanni
|
1d93f11675
|
[Attention][CUDAGraph] Remove CG padding from attention backends (#29352)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
|
2025-12-02 13:48:08 -05:00 |
|
Benjamin Chislett
|
304419576a
|
[Perf] Refactor cudagraph_support to enable full CUDA graphs for spec decoding with FlashInfer (#28479)
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
|
2025-11-13 01:56:40 +09:00 |
|
fhl2000
|
284cc92275
|
[MISC] cudagraph_capture_sizes related improvements (#26016)
Signed-off-by: fhl <2410591650@qq.com>
Signed-off-by: fhl2000 <63384265+fhl2000@users.noreply.github.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
|
2025-10-24 05:11:05 -07:00 |
|
Vadim Gimpelson
|
785d8b6410
|
[PERF] Qwen3-next MTP speedup (change bool mask indexing to index_select / index_copy to reduce d2h) (#26437)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
|
2025-10-16 12:18:31 +08:00 |
|
Harry Mellor
|
8fcaaf6a16
|
Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
|
2025-10-12 09:51:31 -07:00 |
|
Roger Wang
|
43c146ca42
|
[Misc] Clean up unnecessary E501 ignore (#26274)
Signed-off-by: Roger Wang <hey@rogerw.io>
|
2025-10-06 07:29:18 +00:00 |
|
Harry Mellor
|
d6953beb91
|
Convert formatting to use ruff instead of yapf + isort (#26247)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
|
2025-10-05 07:06:22 -07:00 |
|
Tao He
|
99b3a504c5
|
[Qwen3-Next][GDN] fixes cuda graph capturing bug in GDN metadata and a stride bug in causal_conv_1d. (#25743)
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
|
2025-09-26 01:18:58 -07:00 |
|
Benjamin Chislett
|
c30b405b8f
|
[Spec Decode] Enable FlashInfer Spec Decoding (#25196)
Signed-off-by: Benjamin Chislett <benjamin.chislett@centml.ai>
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
Co-authored-by: lhsjohn <huashuoli@tencent.com>
|
2025-09-23 22:29:58 -04:00 |
|
Thomas Parnell
|
a903669e10
|
[V1] Remove V0 code paths for Hybrid models (#25400)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
|
2025-09-23 08:26:13 -07:00 |
|
Vadim Gimpelson
|
072d7e53e5
|
[PERF] Add conv1d metadata to GDN attn (#25105)
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
|
2025-09-18 14:27:49 +00:00 |
|
Tao He
|
dd6a910aac
|
[Bugfix][Qwen3-Next] fixes the varlen issue in qwen3-next's MTP implementation. (#24957)
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
|
2025-09-17 21:59:09 +08:00 |
|
Tao He
|
8226dd56bf
|
[Qwen3Next] Fixes the cuda graph capture conditions under large batch sizes (#24660) (#24667)
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
|
2025-09-12 22:31:32 +00:00 |
|
Tao He
|
e93f4cc9e3
|
Add the support for the qwen3 next model (a hybrid attention model). (#24526)
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
|
2025-09-11 15:32:09 +08:00 |
|