Commits

Branch selector

rope-fp32

User selector

All users

All time

Commit history

Commits on Sep 18, 2023

• Minor

  WoosukKwon

  committed
  4124386View commit details

  Copy full SHA for 4124386

  Browse repository at this point

• Reduce #tests

  WoosukKwon

  committed
  1c5848aView commit details

  Copy full SHA for 1c5848a

  Browse repository at this point

• Minor

  WoosukKwon

  committed
  90157a3View commit details

  Copy full SHA for 90157a3

  Browse repository at this point

• format

  WoosukKwon

  committed
  f1291e7View commit details

  Copy full SHA for f1291e7

  Browse repository at this point

• Fix test

  WoosukKwon

  committed
  8dda602View commit details

  Copy full SHA for 8dda602

  Browse repository at this point

Commits on Sep 17, 2023

• Merge branch 'main' into rope-fp32

  WoosukKwon

  committed
  c3082d2View commit details

  Copy full SHA for c3082d2

  Browse repository at this point

Commits on Sep 16, 2023

• Implement AWQ quantization support for LLaMA (#1032)

  Show description for e3e79e9
  

  5 peopleauthored
  e3e79e9View commit details

  Copy full SHA for e3e79e9

  Browse repository at this point

Commits on Sep 15, 2023

• Abort when coroutine is cancelled (#1020)

  rucyang

  authored
  b9fe461View commit details

  Copy full SHA for b9fe461

  Browse repository at this point

• Fix warning message on LLaMA FastTokenizer (#1037)

  WoosukKwon

  authored
  64ca424View commit details

  Copy full SHA for 64ca424

  Browse repository at this point

• Only fail if logit_bias has actual values (#1045)

  LLukas22

  authored
  b5f93d0View commit details

  Copy full SHA for b5f93d0

  Browse repository at this point

• Add pandas to requirements.txt (#1047)

  Show description for a589369
  WoosukKwon

  authored
  a589369View commit details

  Copy full SHA for a589369

  Browse repository at this point

Commits on Sep 14, 2023

• Fix detokenization leaving special tokens (#1044)

  Show description for dd54a4b
  Yard1

  authored
  dd54a4bView commit details

  Copy full SHA for dd54a4b

  Browse repository at this point

• Announce paper release (#1036)

  WoosukKwon

  authored
  eda1a7cView commit details

  Copy full SHA for eda1a7c

  Browse repository at this point

Commits on Sep 13, 2023

• [FIX] Minor bug fixes (#1035)

  Show description for f04908c
  zhuohan123

  authored
  f04908cView commit details

  Copy full SHA for f04908c

  Browse repository at this point

• Add Model Revision Support (#1014)

  Show description for ab019ee
  

  3 peopleauthored
  ab019eeView commit details

  Copy full SHA for ab019ee

  Browse repository at this point

• Use TGI-like incremental detokenization (#984)

  Yard1

  authored
  9841d48View commit details

  Copy full SHA for 9841d48

  Browse repository at this point

• Fix typo in README.md (#1033)

  eltociear

  authored
  3272d7aView commit details

  Copy full SHA for 3272d7a

  Browse repository at this point

Commits on Sep 12, 2023

• Make `max_model_len` configurable (#972)

  Yard1

  authored
  0bb1e88View commit details

  Copy full SHA for 0bb1e88

  Browse repository at this point

• add option to shorten prompt print in log (#991)

  Show description for d6545ad
  

  3 peopleauthored
  d6545adView commit details

  Copy full SHA for d6545ad

  Browse repository at this point

Commits on Sep 11, 2023

• Bump up the version to v0.1.7 (#1013)

  WoosukKwon

  authored
  90eb3f4View commit details

  Copy full SHA for 90eb3f4

  Browse repository at this point

• Use FP32 in RoPE initialization (#1004)

  Show description for e67b4f2
  

  WoosukKwon

  and

  imoneoi

  authored
  e67b4f2View commit details

  Copy full SHA for e67b4f2

  Browse repository at this point

• Update setup.py (#1006)

  WoosukKwon

  authored
  d6770d1View commit details

  Copy full SHA for d6770d1

  Browse repository at this point

Commits on Sep 10, 2023

• [Docs] Update installation page (#1005)

  WoosukKwon

  authored
  b9cecc2View commit details

  Copy full SHA for b9cecc2

  Browse repository at this point

• fix: CUDA error when inferencing with Falcon-40B base model (#992)

  kyujin-cho

  authored
  898285cView commit details

  Copy full SHA for 898285c

  Browse repository at this point

• Fix op dtype

  WoosukKwon

  committed
  af2b057View commit details

  Copy full SHA for af2b057

  Browse repository at this point

• Use FP32 for cos and sin chache

  WoosukKwon

  committed
  b2c8752View commit details

  Copy full SHA for b2c8752

  Browse repository at this point

Commits on Sep 9, 2023

• Merge branch 'main' into rope-fp32

  WoosukKwon

  committed
  66e8024View commit details

  Copy full SHA for 66e8024

  Browse repository at this point

• Fix wrong dtype in PagedAttentionWithALiBi bias (#996)

  Show description for a62de9e
  Yard1

  authored
  a62de9eView commit details

  Copy full SHA for a62de9e

  Browse repository at this point

• fix "tansformers_module" ModuleNotFoundError when load model with `trust_remote_code=True` (#871)

  Jingru

  authored
  4042d19View commit details

  Copy full SHA for 4042d19

  Browse repository at this point

Commits on Sep 8, 2023

• Bump up the version to v0.1.6 (#989)

  zhuohan123

  authored
  1117aa1View commit details

  Copy full SHA for 1117aa1

  Browse repository at this point

• Start background task in `AsyncLLMEngine.generate` (#988)

  Show description for 0804384
  

  Yard1

  and

  zhuohan123

  authored
  0804384View commit details

  Copy full SHA for 0804384

  Browse repository at this point

• faster startup of vLLM (#982)

  Show description for 4b5bcf8
  

  ri938

  and

  robirv938

  authored
  4b5bcf8View commit details

  Copy full SHA for 4b5bcf8

  Browse repository at this point

Commits on Sep 7, 2023

• Bump up the version to v0.1.5 (#944)

  WoosukKwon

  authored
  852ef5bView commit details

  Copy full SHA for 852ef5b

  Browse repository at this point

• [FIX] Fix Alibi implementation in PagedAttention kernel (#945)

  Show description for db09d4a
  

  3 peopleauthored
  db09d4aView commit details

  Copy full SHA for db09d4a

  Browse repository at this point

• Enable safetensors loading for all models (#974)

  zhuohan123

  authored
  c957c74View commit details

  Copy full SHA for c957c74

  Browse repository at this point