Commits

Branch selector

 main

User selector

WoosukKwon

Datepicker

All time

Commit History

Commits on Nov 1, 2023

• Remove `MPTConfig` (#1529)

  WoosukKwon

  authored
  1fe0990

  Copy full SHA for 1fe0990

Commits on Oct 31, 2023

• Fix integer overflows in attention & cache ops (#1514)

  WoosukKwon

  authored
  0ce8647

  Copy full SHA for 0ce8647

Commits on Oct 29, 2023

• Fix bias in InternLM (#1501)

  WoosukKwon

  authored
  aa9af07

  Copy full SHA for aa9af07

Commits on Oct 17, 2023

• [BugFix] Define `__eq__` in SequenceGroupOutputs (#1389)

  WoosukKwon

  authored
  f8a1e39

  Copy full SHA for f8a1e39

• Add Mistral 7B to `test_models` (#1366)

  WoosukKwon

  authored
  9524867

  Copy full SHA for 9524867

• Change scheduler & input tensor shape (#1381)

  WoosukKwon

  authored
  c1376e0

  Copy full SHA for c1376e0

Commits on Oct 16, 2023

• Fix sampler test (#1379)

  WoosukKwon

  authored
  d3a5bd9

  Copy full SHA for d3a5bd9

• Fix PyTorch index URL in workflow (#1378)

  WoosukKwon

  authored
  e8ef4c0

  Copy full SHA for e8ef4c0

• Fix PyTorch version to 2.0.1 in workflow (#1377)

  WoosukKwon

  authored
  348897a

  Copy full SHA for 348897a

• Implement PagedAttention V2 (#1348)

  WoosukKwon

  authored
  928de46

  Copy full SHA for 928de46

• Minor fix on AWQ kernel launch (#1356)

  WoosukKwon

  authored
  29678cd

  Copy full SHA for 29678cd

Commits on Oct 14, 2023

• Fix error message on `TORCH_CUDA_ARCH_LIST` (#1239)

  Show description for d0740df
  

  WoosukKwon

  and

  yunfeng-scale

  authored
  d0740df

  Copy full SHA for d0740df

Commits on Oct 13, 2023

• Bump up transformers version & Remove MistralConfig (#1254)

  WoosukKwon

  authored
  e7c8555

  Copy full SHA for e7c8555

Commits on Oct 12, 2023

• Add blacklist in model checkpoint (#1325)

  WoosukKwon

  authored
  875afe3

  Copy full SHA for 875afe3

Commits on Oct 2, 2023

• [Minor] Fix type annotations (#1238)

  WoosukKwon

  authored
  84e4e37

  Copy full SHA for 84e4e37

Commits on Sep 28, 2023

• Bump up the version to v0.2.0 (#1212)

  WoosukKwon

  authored
  e2fb71e

  Copy full SHA for e2fb71e

• Provide default max model length (#1224)

  WoosukKwon

  authored
  f936657

  Copy full SHA for f936657

• Fix OOM in attention kernel test (#1223)

  WoosukKwon

  authored
  6f88f76

  Copy full SHA for 6f88f76

• Add Mistral to supported model list (#1221)

  WoosukKwon

  authored
  202351d

  Copy full SHA for 202351d

• [Fix] Remove false assertion (#1222)

  WoosukKwon

  authored
  2e8e49f

  Copy full SHA for 2e8e49f

• Fix Mistral model (#1220)

  WoosukKwon

  authored
  a8e98ae

  Copy full SHA for a8e98ae

Commits on Sep 27, 2023

• Automatically configure `max_num_batched_tokens` (#1198)

  WoosukKwon

  authored
  a19bc5c

  Copy full SHA for a19bc5c

Commits on Sep 26, 2023

• Add comments on RoPE initialization (#1176)

  WoosukKwon

  authored
  03ffd0a

  Copy full SHA for 03ffd0a

• [Setup] Enable `TORCH_CUDA_ARCH_LIST` for selecting target GPUs (#1074)

  WoosukKwon

  authored
  a425bd9

  Copy full SHA for a425bd9

Commits on Sep 24, 2023

• Fix config for Falcon (#1164)

  WoosukKwon

  authored
  9f6be86

  Copy full SHA for 9f6be86

Commits on Sep 22, 2023

• Announce the First vLLM Meetup (#1148)

  WoosukKwon

  authored
  8d926e9

  Copy full SHA for 8d926e9

• Use `--ipc=host` in docker run for distributed inference (#1125)

  WoosukKwon

  authored
  7d7e3b7

  Copy full SHA for 7d7e3b7

Commits on Sep 21, 2023

• Add float16 and float32 (#1115)

  WoosukKwon

  authored
  1ac4ccf

  Copy full SHA for 1ac4ccf

• Replace DtypeTensor (#1123)

  WoosukKwon

  authored
  2ac4d5e

  Copy full SHA for 2ac4d5e

Commits on Sep 20, 2023

• Add gpu_memory_utilization and swap_space to LLM (#1090)

  WoosukKwon

  authored
  bc06445

  Copy full SHA for bc06445

Commits on Sep 19, 2023

• Add pyarrow to dependencies & Print warning on Ray import error (#1094)

  WoosukKwon

  authored
  400b828

  Copy full SHA for 400b828

Commits on Sep 18, 2023

• Add minimum capability requirement for AWQ (#1064)

  WoosukKwon

  authored
  2b1c116

  Copy full SHA for 2b1c116

• Convert before transpose (#1073)

  WoosukKwon

  authored
  cc796b1

  Copy full SHA for cc796b1

Commits on Sep 16, 2023

• Implement AWQ quantization support for LLaMA (#1032)

  Show description for e3e79e9
  

  5 peopleauthored
  e3e79e9

  Copy full SHA for e3e79e9

Commits on Sep 15, 2023

• Fix warning message on LLaMA FastTokenizer (#1037)

  WoosukKwon

  authored
  64ca424

  Copy full SHA for 64ca424

Pagination

Previous

Next