Commits

Branch selector

 fix_vllm_config

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 17, 2025

• fix bug

  Show description for 52770f4
  heheda12345

  committed
  52770f4

  Copy full SHA for 52770f4

Commits on Jan 16, 2025

• [Model] Add support for deepseek-vl2-tiny model (#12068)

  Show description for 62b06ba
  Isotr0py

  authored
  62b06ba

  Copy full SHA for 62b06ba

• [misc] Add LoRA kernel micro benchmarks (#11579)

  varun-sundar-rabindranath

  authored
  5fd24ec

  Copy full SHA for 5fd24ec

• [Bugfix] Fix max image feature size for Llava-one-vision (#12104)

  Show description for 874f7c2
  ywang96

  authored
  874f7c2

  Copy full SHA for 874f7c2

• [core] LLM.collective_rpc interface and RLHF example (#12084)

  Show description for 92e793d
  youkaichao

  authored
  92e793d

  Copy full SHA for 92e793d

• Support torchrun and SPMD-style offline inference (#12071)

  Show description for bf53e0c
  youkaichao

  authored
  bf53e0c

  Copy full SHA for bf53e0c

• [Bugfix] Remove hardcoded `head_size=256` for Deepseek v2 and v3 (#12067)

  Show description for dd7c9ad
  Isotr0py

  authored
  dd7c9ad

  Copy full SHA for dd7c9ad

• Various cosmetic/comment fixes (#12089)

  Show description for 9aa1519
  mgoin

  authored
  9aa1519

  Copy full SHA for 9aa1519

• [Doc] Add documentation for specifying model architecture (#12105)

  DarkLight1337

  authored
  f8ef146

  Copy full SHA for f8ef146

• [Core] Default to using per_token quantization for fp8 when cutlass is supported. (#8651)

  Show description for fa0050d
  

  elfiegg

  and

  mgoin

  authored
  fa0050d

  Copy full SHA for fa0050d

Commits on Jan 15, 2025

• Allow hip sources to be directly included when compiling for rocm. (#12087)

  tvirolai-amd

  authored
  cd9d06f

  Copy full SHA for cd9d06f

• [Bugfix] Fix _get_lora_device for HQQ marlin (#12090)

  Show description for ebd8c66
  

  varun-sundar-rabindranath

  and

  Varun Sundar Rabindranath

  authored
  ebd8c66

  Copy full SHA for ebd8c66

• [V1][Core] Autotune encoder cache budget (#11895)

  Show description for 70755e8
  ywang96

  authored
  70755e8

  Copy full SHA for 70755e8

• [Bugfix] use right truncation for non-generative tasks (#12050)

  Show description for edce722
  joerunde

  authored
  edce722

  Copy full SHA for edce722

• [Doc]: Update `OpenAI-Compatible Server` documents (#12082)

  maang-h

  authored
  57e729e

  Copy full SHA for 57e729e

• [Misc][Quark] Upstream Quark format to VLLM (#10765)

  Show description for de0526f
  

  3 peopleauthored
  de0526f

  Copy full SHA for de0526f

• Misc: allow to use proxy in `HTTPConnection` (#12042)

  Show description for 5ecf3e0
  zhouyuan

  authored
  5ecf3e0

  Copy full SHA for 5ecf3e0

• [Model]: Support internlm3 (#12037)

  RunningLeon

  authored
  97eb97b

  Copy full SHA for 97eb97b

• [Platform] Do not raise error if _Backend is not found (#12023)

  Show description for 3adf0ff
  

  wangxiyuan

  and

  MengqingCao

  authored
  3adf0ff

  Copy full SHA for 3adf0ff

• Type-fix: make execute_model output type optional (#12020)

  youngkent

  authored
  ad388d2

  Copy full SHA for ad388d2

• Fix: cases with empty sparsity config (#12057)

  Show description for cbe9439
  rahul-tuli

  authored
  cbe9439

  Copy full SHA for cbe9439

• [V1][Prefix Cache] Move the logic of num_computed_tokens into KVCacheManager (#12003)

  heheda12345

  authored
  994fc65

  Copy full SHA for 994fc65

• [Doc] Update examples to remove SparseAutoModelForCausalLM (#12062)

  Show description for 3f9b7ab
  kylesayrs

  authored
  3f9b7ab

  Copy full SHA for 3f9b7ab

• [core] platform agnostic executor via collective_rpc (#11256)

  Show description for ad34c0d
  youkaichao

  authored
  ad34c0d

  Copy full SHA for ad34c0d

• [core] Turn off GPU communication overlap for Ray executor (#12051)

  Show description for f218f9c
  ruisearch42

  authored
  f218f9c

  Copy full SHA for f218f9c

• [Misc] Add multipstep chunked-prefill support for FlashInfer (#10467)

  elfiegg

  authored
  0794e74

  Copy full SHA for 0794e74

• [V1][BugFix] Fix edge case in VLM scheduling (#12065)

  Show description for b7ee940
  WoosukKwon

  authored
  b7ee940

  Copy full SHA for b7ee940

• [Platform] move current_memory_usage() into platform (#11369)

  Show description for 9ddac56
  shen-shanshan

  authored
  9ddac56

  Copy full SHA for 9ddac56

• [HPU][Bugfix] Don't use /dev/accel/accel0 for HPU autodetection in setup.py (#12046)

  Show description for 1a51b9f
  kzawora-intel

  authored
  1a51b9f

  Copy full SHA for 1a51b9f

• [Kernel] Support MulAndSilu (#11624)

  Show description for 42f5e7c
  jeejeelee

  authored
  42f5e7c

  Copy full SHA for 42f5e7c

Commits on Jan 14, 2025

• [Misc] Merge bitsandbytes_stacked_params_mapping and packed_modules_mapping (#11924)

  Show description for a3a3ee4
  jeejeelee

  authored
  a3a3ee4

  Copy full SHA for a3a3ee4

• [Doc]: Update the Json Example of the `Engine Arguments` document (#12045)

  maang-h

  authored
  87054a5

  Copy full SHA for 87054a5

• Explain where the engine args go when using Docker (#12041)

  Show description for c9d6ff5
  hmellor

  authored
  c9d6ff5

  Copy full SHA for c9d6ff5

• [Bugfix][Kernel] Give unique name to BlockSparseFlashAttention (#12040)

  Show description for a2d2acb
  heheda12345

  authored
  a2d2acb

  Copy full SHA for a2d2acb

• [Platform] Add output for Attention Backend (#11981)

  Show description for 2e0e017
  wangxiyuan

  authored
  2e0e017

  Copy full SHA for 2e0e017

Pagination

Previous

Next