Commits · d263bd9df7b2f5586910e5d006a11ff11ba7c310 · OpenDAS / vllm_cscc

14 Dec, 2024 5 commits
- Enable mypy checking on V1 code (#11105) · 6d917d0e
  Mark McLoughlin authored Dec 14, 2024
```
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
```
  6d917d0e
- [V1][Bugfix] Fix V1 TP trust-remote-code (#11182) · ea7bd68d
  Tyler Michael Smith authored Dec 14, 2024
```
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  ea7bd68d
- [Bugfix][V1] Re-compute an entire block when fully cache hit (#11186) · 9855aea2
  Cody Yu authored Dec 13, 2024
```
Signed-off-by: Cody Yu <hao.yu.cody@gmail.com>
```
  9855aea2
- [V1][Bugfix] Fix EngineCoreProc profile (#11185) · 4b5b8a6a
  Tyler Michael Smith authored Dec 13, 2024
```
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  4b5b8a6a
- [Core] V1: Use multiprocessing by default (#11074) · 4863e5fb
  Russell Bryant authored Dec 13, 2024
```
Signed-off-by: Russell Bryant <rbryant@redhat.com>
```
  4863e5fb
13 Dec, 2024 3 commits
- [core] clean up cudagraph batchsize padding logic (#10996) · be39e3cd
  youkaichao authored Dec 12, 2024
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  be39e3cd
- [Bugfix][V1] Fix 'NoneType' object has no attribute 'hash_value' (#11157) · 34f1a806
  Cody Yu authored Dec 12, 2024
```
Signed-off-by: Cody Yu <hao.yu.cody@gmail.com>
```
  34f1a806
- [Misc][V1] Fix type in v1 prefix caching (#11151) · 78ed8f57
  Cody Yu authored Dec 12, 2024
  
  78ed8f57
12 Dec, 2024 4 commits
- [V1] Fix torch profiling for offline inference (#11125) · 4816d20a
  Roger Wang authored Dec 12, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
```
  4816d20a
- [V1] Use more persistent buffers to optimize input preparation overheads (#11111) · f092153f
  Woosuk Kwon authored Dec 11, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  f092153f
- [Core] cleanup zmq ipc sockets on exit (#11115) · ccede2b2
  Russell Bryant authored Dec 11, 2024
```
Signed-off-by: Russell Bryant <rbryant@redhat.com>
```
  ccede2b2
- [V1] VLM preprocessor hashing (#11020) · 4e116833
  Alexander Matveev authored Dec 11, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
Signed-off-by: Alexander Matveev <alexm@neuralmagic.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
Co-authored-by: Roger Wang <ywang@roblox.com>
```
  4e116833
11 Dec, 2024 6 commits
- [V1][Core] Remove should_shutdown to simplify core process termination (#11113) · 452a723b
  Tyler Michael Smith authored Dec 11, 2024
```
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  452a723b
- [V1] Use input_ids as input for text-only models (#11032) · d643c2ab
  Woosuk Kwon authored Dec 11, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  d643c2ab
- [Doc] Update docs to refer to pooling models (#11093) · cad5c0a6
  Cyrus Leung authored Dec 11, 2024
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  cad5c0a6
- [Misc] Split up pooling tasks (#10820) · 8f10d5e3
  Cyrus Leung authored Dec 11, 2024
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  8f10d5e3
- [Core] v1: Use atexit to handle engine core client shutdown (#11076) · 61b1d2f6
  Russell Bryant authored Dec 11, 2024
```
Signed-off-by: Russell Bryant <rbryant@redhat.com>
```
  61b1d2f6
- [Model] PP support for Mamba-like models (#10992) · ffa48c91
  Mor Zusman authored Dec 11, 2024
```
Signed-off-by: mzusman <mor.zusmann@gmail.com>
```
  ffa48c91
10 Dec, 2024 3 commits
- [torch.compile] add a flag to track batchsize statistics (#11059) · 75f89dc4
  youkaichao authored Dec 10, 2024
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  75f89dc4
- [Bugfix] Backport request id validation to v0 (#11036) · 9b9cef31
  Joe Runde authored Dec 10, 2024
```
Signed-off-by: Joe Runde <Joseph.Runde@ibm.com>
```
  9b9cef31
- [V1] Multiprocessing Tensor Parallel Support for v1 (#9856) · 28b3a1c7
  Tyler Michael Smith authored Dec 10, 2024
```
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  28b3a1c7
09 Dec, 2024 5 commits
- [V1] Do not store `None` in self.generators (#11038) · 6faec545
  Woosuk Kwon authored Dec 09, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  6faec545
- [v1] fix use compile sizes (#11000) · 1a2f8fb8
  youkaichao authored Dec 09, 2024
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  1a2f8fb8
- [V1] Further reduce CPU overheads in flash-attn (#10989) · 3b61cb45
  Woosuk Kwon authored Dec 09, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  3b61cb45
- [V1] Input Batch Relocation (#10962) · 25b79d9f
  Varun Sundar Rabindranath authored Dec 09, 2024
```
Signed-off-by: Varun Sundar Rabindranath <varun@neuralmagic.com>
Co-authored-by: Varun Sundar Rabindranath <varun@neuralmagic.com>
```
  25b79d9f
- [V1] Fix Detokenizer loading in `AsyncLLM` (#10997) · c6903579
  Roger Wang authored Dec 09, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
```
  c6903579
08 Dec, 2024 1 commit
- [V1] Initial support of multimodal models for V1 re-arch (#10699) · a11f3265
  Roger Wang authored Dec 08, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
```
  a11f3265
07 Dec, 2024 1 commit
- [3/N] Support and implement merged input processor for LLaVA model (#10676) · 955fa953
  Cyrus Leung authored Dec 07, 2024
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: Roger Wang <ywang@roblox.com>
```
  955fa953
06 Dec, 2024 1 commit

[torch.compile] add logging for compilation time (#10941) · b031a455

youkaichao authored Dec 06, 2024


Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

b031a455

05 Dec, 2024 1 commit
- [V1] Fix when max_model_len is not divisible by block_size (#10903) · 2a56e126
  Woosuk Kwon authored Dec 04, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  2a56e126
04 Dec, 2024 1 commit
- [LoRA] Change lora_tokenizers capacity (#10796) · 01d079fd
  Xin Yang authored Dec 04, 2024
```
Signed-off-by: Xin Yang <xyang19@gmail.com>
```
  01d079fd
03 Dec, 2024 2 commits
- [V1] VLM - Run the mm_mapper preprocessor in the frontend process (#10640) · 3bc94cab
  Alexander Matveev authored Dec 03, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
Co-authored-by: Roger Wang <ywang@roblox.com>
```
  3bc94cab
- [torch.compile] remove compilation_context and simplify code (#10838) · dc5ce861
  youkaichao authored Dec 02, 2024
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  dc5ce861
02 Dec, 2024 2 commits
- [misc] use out argument for flash attention (#10822) · a4c4daf3
  youkaichao authored Dec 02, 2024
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  a4c4daf3
- [Kernel] Use `out` arg in flash_attn_varlen_func (#10811) · 073a4bd1
  Woosuk Kwon authored Dec 01, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  073a4bd1
01 Dec, 2024 1 commit
- [Misc] Rename embedding classes to pooling (#10801) · d2f058e7
  Cyrus Leung authored Dec 01, 2024
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  d2f058e7
28 Nov, 2024 2 commits
- [V1] Optimize the CPU overheads in FlashAttention custom op (#10733) · 98f47f2a
  Woosuk Kwon authored Nov 28, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  98f47f2a
- [V1] Do not allocate beyond the max_model_len (#10730) · a79b1224
  Woosuk Kwon authored Nov 28, 2024
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  a79b1224
26 Nov, 2024 2 commits
- [V1] Refactor model executable interface for multimodal models (#10570) · 2f0a0a17
  Roger Wang authored Nov 26, 2024
```
Signed-off-by: Roger Wang <ywang@roblox.com>
```
  2f0a0a17
- [V1] Enable profile for LLMEngine (#10665) · 1f6584ee
  Kunshang Ji authored Nov 26, 2024
  
  1f6584ee