Commits

Branch selector

 main

User selector

comaniac

Datepicker

All time

Commit History

Commits on Sep 24, 2024

• [MISC] Skip dumping inputs when unpicklable (vllm-project#8744)

  comaniac

  authored
  b8747e8

  Copy full SHA for b8747e8

Commits on Sep 12, 2024

• [MISC] Dump model runner inputs when crashing (vllm-project#8305)

  comaniac

  authored
  a65cb16

  Copy full SHA for a65cb16

Commits on Sep 10, 2024

• [MISC] Keep chunked prefill enabled by default with long context when prefix caching is enabled (vllm-project#8342)

  comaniac

  authored
  b1f3e18

  Copy full SHA for b1f3e18

Commits on Sep 4, 2024

• [MISC] Replace input token throughput with total token throughput (vllm-project#8164)

  Show description for 77d9e51
  

  comaniac

  and

  mgoin

  authored
  77d9e51

  Copy full SHA for 77d9e51

• [MISC] Consolidate FP8 kv-cache tests (vllm-project#8131)

  comaniac

  authored
  2ad2e56

  Copy full SHA for 2ad2e56

Commits on Sep 3, 2024

• [Performance] Enable chunked prefill and prefix caching together (vllm-project#8120)

  Show description for bd852f2
  

  3 peopleauthored
  bd852f2

  Copy full SHA for bd852f2

Commits on Aug 28, 2024

• [Performance] Enable chunked prefill and prefix caching together (vllm-project#7753)

  comaniac

  authored
  e358053

  Copy full SHA for e358053

Commits on Aug 27, 2024

• Revert vllm-project#7509 (vllm-project#7887)

  comaniac

  authored
  9606c71

  Copy full SHA for 9606c71

Commits on Aug 26, 2024

• [Performance][BlockManagerV2] Mark prefix cache block as computed after schedule (vllm-project#7822)

  comaniac

  authored
  2deb029

  Copy full SHA for 2deb029

Commits on Aug 19, 2024

• [MISC] Add prefix cache hit rate to metrics (vllm-project#7606)

  comaniac

  authored
  3ac50b4

  Copy full SHA for 3ac50b4

Commits on Aug 5, 2024

• [MISC] Use non-blocking transfer in prepare_input (vllm-project#7172)

  comaniac

  authored
  ef527be

  Copy full SHA for ef527be

Commits on Jul 31, 2024

• [MISC] Introduce pipeline parallelism partition strategies (vllm-project#6920)

  Show description for bd70013
  

  comaniac

  and

  youkaichao

  authored
  bd70013

  Copy full SHA for bd70013

Commits on Jul 25, 2024

• [Bugfix] Fix decode tokens w. CUDA graph (vllm-project#6757)

  comaniac

  authored
  309aaef

  Copy full SHA for 309aaef

Commits on Jul 23, 2024

• [Core] Modulize prepare input and attention metadata builder (vllm-project#6596)

  comaniac

  authored
  e0c1575

  Copy full SHA for e0c1575

Commits on Jul 18, 2024

• [Misc] Minor patch for draft model runner (vllm-project#6523)

  comaniac

  authored
  8a74c68

  Copy full SHA for 8a74c68

• [Model] Pipeline parallel support for Mixtral (vllm-project#6516)

  comaniac

  authored
  b5af8c2

  Copy full SHA for b5af8c2

Commits on Jul 17, 2024

• [Core] Refactor _prepare_model_input_tensors - take 2 (vllm-project#6164)

  comaniac

  authored
  2fa4623

  Copy full SHA for 2fa4623

Commits on Jul 16, 2024

• [Misc] Log spec decode metrics (vllm-project#6454)

  comaniac

  authored
  160e1d8

  Copy full SHA for 160e1d8

Commits on Jul 12, 2024

• [Bugfix] Fix illegal memory access in FP8 MoE kernel (vllm-project#6382)

  comaniac

  authored
  75f64d8

  Copy full SHA for 75f64d8

• [MISC] Upgrade dependency to PyTorch 2.3.1 (vllm-project#5327)

  comaniac

  authored
  aa48e50

  Copy full SHA for aa48e50

Commits on Jun 29, 2024

• [Kernel] Raise an exception in MoE kernel if the batch size is larger then 65k (vllm-project#5939)

  comaniac

  authored
  f7dac83

  Copy full SHA for f7dac83

Commits on Jun 28, 2024

• [Spec Decode] Introduce DraftModelRunner (vllm-project#5799)

  comaniac

  authored
  b2c6202

  Copy full SHA for b2c6202

Commits on Jun 17, 2024

• Fix w8a8 benchmark and add Llama-3-8B (vllm-project#5562)

  comaniac

  authored
  e2b85cf

  Copy full SHA for e2b85cf

Commits on Jun 13, 2024

• [MISC] Remove FP8 warning (vllm-project#5472)

  Show description for 30299a4
  

  comaniac

  and

  pcmoritz

  authored
  30299a4

  Copy full SHA for 30299a4

Commits on Jun 12, 2024

• [Kernel] Vectorized FP8 quantize kernel (vllm-project#5396)

  Show description for 5985e34
  comaniac

  authored
  5985e34

  Copy full SHA for 5985e34

Commits on Jun 5, 2024

• [Model] Correct Mixtral FP8 checkpoint loading (vllm-project#5231)

  comaniac

  authored
  5563a4d

  Copy full SHA for 5563a4d

Commits on May 31, 2024

• [Model] Enable FP8 QKV in MoE and refine kernel tuning script (vllm-project#5039)

  comaniac

  authored
  e9899fb

  Copy full SHA for e9899fb

Commits on May 22, 2024

• [Misc] Take user preference in attention selector (vllm-project#4960)

  comaniac

  authored
  ee3eea0

  Copy full SHA for ee3eea0

• [Misc] Load FP8 kv-cache scaling factors from checkpoints (vllm-project#4893)

  Show description for a3a73ab
  comaniac

  authored
  a3a73ab

  Copy full SHA for a3a73ab

Commits on May 16, 2024

• [Speculative decoding][Re-take] Enable TP>1 speculative decoding (vllm-project#4840)

  Show description for 973617a
  

  comaniac

  and

  cadedaniel

  authored
  973617a

  Copy full SHA for 973617a

Commits on May 13, 2024

• [Speculative decoding] Improve n-gram efficiency (vllm-project#4724)

  comaniac

  authored
  ce532ff

  Copy full SHA for ce532ff

Commits on May 10, 2024

• [Kernel] Refactor FP8 kv-cache with NVIDIA float8_e4m3 support (vllm-project#4535)

  comaniac

  authored
  c833101

  Copy full SHA for c833101

Commits on May 8, 2024

• [Dynamic Spec Decoding] Auto-disable by the running queue size (vllm-project#4592)

  Show description for f942efb
  

  comaniac

  and

  cadedaniel

  authored
  f942efb

  Copy full SHA for f942efb

Commits on May 4, 2024

• [Misc][Refactor] Introduce ExecuteModelData (vllm-project#4540)

  comaniac

  authored
  bc8ad68

  Copy full SHA for bc8ad68

Commits on Apr 26, 2024

• [Misc][Refactor] Generalize linear_method to be quant_method (vllm-project#4373)

  comaniac

  authored
  a62aaf1

  Copy full SHA for a62aaf1

Pagination

Previous

Next