Commits · 83e1c76dbe07e30b7f4e6dbe17ba580f4afc98f0 · OpenDAS / vllm_cscc

08 Jan, 2026 1 commit
- [Misc] Tidy up some spec decode logic in GPUModelRunner (#31591) · a3d909ad
  Nick Hill authored Jan 08, 2026
```
Signed-off-by: Nick Hill <nickhill123@gmail.com>
```
  a3d909ad
07 Jan, 2026 2 commits
- [refactor] refactor memory constants usage (#31865) · c907d221
  Ning Xie authored Jan 08, 2026
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
  c907d221
- [Chore] Migrate V0 attention utils (#31891) · b665bbc2
  Cyrus Leung authored Jan 07, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  b665bbc2
02 Jan, 2026 1 commit
- [BugFix] Support online dense model DP without overhead (#30739) · bd877162
  Nick Hill authored Jan 02, 2026
```
Signed-off-by: Nick Hill <nhill@redhat.com>
Signed-off-by: njhill <nickhill123@gmail.com>
```
  bd877162
31 Dec, 2025 1 commit
- [BugFix] Fix async scheduling for pooling models (#31584) · 6c2cfb62
  Nick Hill authored Dec 31, 2025
```
Signed-off-by: njhill <nickhill123@gmail.com>
```
  6c2cfb62
10 Dec, 2025 1 commit

[Attention] Make seq_lens_cpu optional in CommonAttentionMetadata to enable... · abe93bce

Lucas Wilkinson authored Dec 09, 2025


[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>

abe93bce

09 Dec, 2025 2 commits
- [Model Runner V2] Support num NaNs in logits (#30187) · d471b2af
  Woosuk Kwon authored Dec 09, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  d471b2af
- [Model Runner V2] Fix Triton warning on tl.where (#30355) · 9e6562a3
  Woosuk Kwon authored Dec 09, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  9e6562a3
06 Dec, 2025 1 commit
- [Model Runner V2] Support min-p sampling (#30171) · a238cbd8
  Woosuk Kwon authored Dec 05, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  a238cbd8
04 Dec, 2025 1 commit
- [Model Runner V2] Implement get_num_sampled_and_rejected kernel (#30029) · cc050558
  Woosuk Kwon authored Dec 04, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  cc050558
30 Nov, 2025 2 commits
- [Model Runner V2] Use packed mask for prompt bin counts (#29756) · ec38a736
  Woosuk Kwon authored Nov 30, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  ec38a736
- [Core] Enable `inputs_embeds_size` separate from `hidden_size` (#29741) · 64bc09ba
  Cyrus Leung authored Nov 30, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  64bc09ba
29 Nov, 2025 5 commits
- [Model Runner V2] Fuse penalties and temperature into single kernel (#29720) · f223ed41
  Woosuk Kwon authored Nov 29, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  f223ed41
- [Model Runner V2] Add sample/ directory and reorganize files (#29719) · 6afc0ffa
  Woosuk Kwon authored Nov 29, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  6afc0ffa
- [Model Runner V2] Don't use UVA buffer for prefill_len (#29713) · 4a80ad0a
  Woosuk Kwon authored Nov 28, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  4a80ad0a
- [Model Runner V2] Refactor prefill token preparation (#29712) · ca1b1e72
  Woosuk Kwon authored Nov 28, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  ca1b1e72
- [Model Runner V2] Support penalties using bin counts (#29703) · 1dcafb3d
  Woosuk Kwon authored Nov 28, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  1dcafb3d
27 Nov, 2025 7 commits
- [Model Runner V2][BugFix] Keep reference to GPU tensors in AsyncOutput (#29623) · ae0ce1be
  Woosuk Kwon authored Nov 27, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  ae0ce1be
- [Bugfix] Fix pre-commit (#29601) · e6d4f3c2
  Cyrus Leung authored Nov 27, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  e6d4f3c2
- [Frontend][torch.compile] CompilationConfig Overhaul (#20283): Set up -O infrastructure (#26847) · 0838b52e
  Morrison Turnansky authored Nov 27, 2025
```
Signed-off-by: morrison-turnansky <mturnans@redhat.com>
Signed-off-by: adabeyta <aabeyta@redhat.com>
Signed-off-by: Morrison Turnansky <mturnans@redhat.com>
Co-authored-by: adabeyta <aabeyta@redhat.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  0838b52e
- [Model Runner V2] Implement multi-step Eagle with CUDA graph (#29559) · da3222f3
  Woosuk Kwon authored Nov 27, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  da3222f3
- [Model Runner V2] Refactor CudaGraphManager (#29583) · 11ea5ec1
  Woosuk Kwon authored Nov 26, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  11ea5ec1
- [Model Runner V2] Minor cleanup for build_attn_metadata (#29576) · ee80aee1
  Woosuk Kwon authored Nov 26, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  ee80aee1
- [Model Runner V2] Minor code cleanup (#29570) · 0aeb698b
  Woosuk Kwon authored Nov 26, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  0aeb698b
24 Nov, 2025 4 commits
- [Model Runner V2] Simplify Eagle bookkeeping with num_rejected (#29347) · f32c7d6f
  Woosuk Kwon authored Nov 24, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  f32c7d6f
- [Model Runner V2] Add minor clarification comments for Eagle (#29332) · 97588c4d
  Woosuk Kwon authored Nov 24, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  97588c4d
- [Model Runner V2] Change Numba AoT to JIT (#29328) · cec418b5
  Woosuk Kwon authored Nov 24, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  cec418b5
- [Model Runner V2] Implement Single-step Eagle 1 (#29300) · cc313cb7
  Woosuk Kwon authored Nov 24, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  cc313cb7
23 Nov, 2025 5 commits
- [Model Runner V2] Add apply_temperature option to gumbel_sample (#29276) · 3e1ad406
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  3e1ad406
- [Model Runner V2] Optimize CUDA graph capture time (#29275) · 62d54ba4
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  62d54ba4
- [Model Runner V2] Support spec decoding [1/N] (#29274) · b004c004
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  b004c004
- [Model Runner V2] Change bookkeeping logic in preparation for spec decoding (#29194) · 7f12c82f
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  7f12c82f
- [Model Runner V2] Minor fix for cudagraph_utils (#29256) · 20ee418a
  Woosuk Kwon authored Nov 22, 2025
  
  20ee418a
22 Nov, 2025 2 commits
- [Model Runner V2] Limit cudagraph size to max decode batch size (#29221) · e9056056
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  e9056056
- [CI] Bug: Fix triton import issue (#29202) · 1d34eb11
  Wentao Ye authored Nov 21, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  1d34eb11
21 Nov, 2025 3 commits
- [Model Runner V2] Optimize Gumbel Sampling Kernel (#29210) · e9af6ba6
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  e9af6ba6
- [Chore] Fix pre-commit error after #25266 (#29190) · 1bed891f
  Woosuk Kwon authored Nov 21, 2025
  
  1bed891f
- GPU Model Runner V2 (#25266) · 30b44a15
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  30b44a15